Conservation and Social Ecology of Freshwater and Coastal Irrawaddy Dolphins in Indonesia

Home , Delphinida, Kumai River, Orcininae, Sesayap River

FACULTATIVE RIVER DOLPHINS:

CONSERVATION AND SOCIAL ECOLOGY OF FRESHWATER AND COASTAL IRRAWADDY DOLPHINS IN INDONESIA

ISBN: 90-76894-51-5 This research was carried out at the Institute for Biodiversity and Ecosystem Dynamics (IBED)/ Zoölogisch Museum Amsterdam (ZMA) Cover photo : Daniëlle Kreb Cover design : Jan van Arkel Printed by : Febodruk B.V., Enschede Financial support for printing received from: J.E. Jurriaanse Stichting and IBED Copyright © D. Kreb 2004

FACULTATIVE RIVER DOLPHINS:

CONSERVATION AND SOCIAL ECOLOGY OF FRESHWATER AND COASTAL IRRAWADDY DOLPHINS IN INDONESIA

ACADEMISCH PROEFSCHRIFT

Ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. mr. P.F. van der Heijden ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit op dinsdag 9 november 2004, te 14.00 uur

door

Daniëlle Kreb

geboren te Emmeloord, Noordoostpolder

Promotiecommissie:

Promotor: Prof. dr. F.R. Schram

Commissieleden: Prof. dr. S.B.J. Menken Prof. dr. W. Admiraal Prof. dr. P.H. van Tienderen Prof. dr. H.H.T. Prins Prof. dr. J.F. Borsani Dr. C. Smeenk

Faculteit: Natuurwetenschappen, Wiskunde en Informatica

Instituut: Instituut voor Biodiversiteit en Ecosysteem Dynamica

Untuk mas Budi dan Jannah

By Hari Moelyono TABLE OF CONTENTS

ACKNOWLEDGEMENTS……………………………………………………….i

SECTION I. GENERAL BACKGROUND

Chapter 1. A general introduction into the phenomenon of facultative river dolphins and the species Orcaella brevirostris………………………….1

Chapter 2. Observations on the occurrence of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia…….11 Zeitschrift für Säugetierkunde 64: 54-58, 1999

Chapter 3. Cetacean diversity and habitat preferences in tropical waters of East Kalimantan, Indonesia……………………………………….19 With Budiono, submitted manuscript

SECTION II. SURVEY TECHNIQUES FOR ABUNDANCE ESTIMATION

Chapter 4. Density and abundance of the Irrawady dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques………………………………..35 The Raffles Bulletin of Zoology, Supplement 10: 85-95, 2002

Chapter 5. Abundance of freshwater Irrawaddy dolphins in the Mahakam River in East Kalimantan, Indonesia, based on mark-recapture analysis of photo-identified individuals……………………………59 In press: Journal of Cetacean Research and Management, 2004

SECTION III. SOCIAL ECOLOGY AND CONSERVATION OF IRRAWADDY RIVER DOLPHINS

Chapter 6. Conservation management of small core areas: Key to survival of a critically endangered population of Irrawaddy River dolphins in Borneo………………………………………………………....81 With Budiono, in press: Oryx, 2004

Chapter 7. Living under an aquatic freeway: Effects of boats on Irrawaddy dolphins (Orcaella brevirostris) in a coastal and riverine environment in Indonesia….…………….……………………………………..105 With Karen D. Rahadi, in press: Aquatic Mammals, 2004

Chapter 8. Marked declines in populations of Irrawaddy dolphins…………...127 With Brian D. Smith and Isabel Beasley, Oryx 37: 401

Chapter 9. Social dynamics of facultative Irrawaddy River dolphins (Orcaella brevirostris) in Borneo: Impacts of habitat…………………………133 Submitted manuscript

Chapter 10. Impacts of habitat on the acoustics of coastal and freshwater Irrawaddy dolphins, Orcaella brevirostris in East Kalimantan, Indonesia…………………………………….…………………...161 With Junio F. Borsani

SECTION IV. GENERAL DISCUSSION- POPULATIONS AND LONG-TERM PROGNOSIS FOR SURVIVAL

Chapter 11. Freshwater distribution of Irrawaddy dolphins based on river “vagrancy” or allopatric “speciation”?……………………………185

Chapter 12. Predicting long-term survival of riverine Irrawaddy dolphins (Orcaella brevirostris) in East Kalimantan using Population Viability Analysis ………………………………………………...199

APPENDIX I Irrawaddy dolphins in the Mahakam River, Indonesia…………213

APPENDIX II Monitor and evaluate ongoing threats to the Irrawaddy Dolphins in the Mahakam River of Indonesia…………………215 Appendix 1, pp. 88-89; Appendix 2, p. 56 in: R.R. Reeves, B.D. Smith, E.A. Crespo and G.N. di Sciara (eds.) (2003). 2002-2010 Conservation Action Plan for the World’s Cetaecans. Dolphins, whales and porpoises. IUCN, Gland, Switzerland

SUMMARY Facultative river dolphins: Conservation and social ecology of freshwater and coastal populations of Irrawaddy dolphins in Indonesia …….………………………………………………….217

SAMENVATTING Facultatieve rivierdolfijnen: Bescherming en sociale ecologie van zoet- en zoutwater populaties van Irrawaddy dolfijnen, Orcaella brevirostris in Oost Kalimantan, Indonesië………….221

RINGKASAN Lumba-lumba sungai: Konservasi dan sosial ekologi dari populasi lumba-lumba Irrawaddy, Orcaella brevirostris pada air tawar dan laut di Kalimantan Timur, Indonesia………………………………….225

CURRICULUM VITAE………………………………………………………...229 Preface and acknowledgements

PREFACE AND ACKNOWLEDGEMENTS

This research began with a simple telephone call from Indonesia to Scotland, where at that time I was assisting a radio-tracking study on wildcats, by my good friend and colleague Vincent Nijman who asked me if I knew that there were river dolphins in the Mahakam River in East Kalimantan (for which tip I owe him). Since I so far had only heard about the obligate river dolphins in the Amazon, Ganges, Indus and Yangtze Rivers, which had already captured my interest and imagination, I was surprised and interested to find out more about it. From the sides of the Provincial Wildlife Conservation Department of East Kalimantan (BKSDA Kaltim) and WWF Indonesia (thanks to former staff member Ron Lilley), there was an interest to conduct a preliminary survey in the freshwater dolphins in the Mahakam, which were locally referred to as the pesut. Thanks to the help of Dr Peter J.H. van Bree, Curator Emeritus of the Zoological Museum of Amsterdam, who helped me prepare a proposal to join the survey and find a sponsor through Marc Argeloo and Jikkie Jonkman from WNF Nederland, I soon flew off on my way to meet my first river dolphin in real. I should say that my first observation of the dolphins thrilled me with admiration and I felt that this survey was not to be my last one especially after the numbers we encountered during the survey were rather low and visible threats were numerous. The research really had to be started from the scratch as no previous systematic studies on Irrawaddy dolphins in East Kalimantan had been done upon which to build. The difficulties in studying cetaceans in general is that it requires a great deal of organisation and preparation in order to work as efficiently as possible because of the use of boats, which sometimes is an unpredictable and costly factor. Some creativity and patience was required at times when the working schedule needed to be adjusted when dealing with engine problems or during bad weather conditions, especially at sea where lack of freshwater also was problematic at times. During this research I have not only learned a lot about dolphins, boats, rivers and seas, but also about local human cultures, of which I found the mutual respect and hospitality that I encountered heart warming. My impressions, which turned out to be realistic based on interviews, were that many fishermen in the Mahakam actually had an appreciation for the dolphins and did not wish the dolphins to disappear from the river This encouraged me in my attempts to set up a conservation program. A range of activities focusing on increasing local awareness of the younger generation, fishermen, politics

i Preface and acknowledements

and society in general, have been conducted since late 2000 until now by the local NGO Yayasan Konservasi RASI (Conservation Foundation for Rare Aquatic Species of Indonesia). I am most grateful to my promoter Professor Frederick R. Schram for taking this project on his shoulders and for his enduring support and scientific guidance. All my manuscripts have been commented and corrected by him. In this regard, a special note of thanks should go to Dr. Arne Mooers, who brought Fred and me together. I also owe a great deal of gratitude to Dr Peter J.H. van Bree, who helped me throughout the study with literature, good advice, and assistance with locating other support. I also thank Dr Jan Wattel for his help in finding financial support. Harm van der Geest is also thanked for his good tip. My special thanks go to my first counterpart Ir Ade M. Rachmat (M.Sc.), former head of the BKSDA Kaltim, who has so unfortunately deceased some months ago. He invited me as a guest to participate in the first preliminary survey, covered all costs and had me stay in his house in Samarinda in between the surveys. My next counterparts Ir Padmo Wiyoso, former head of BKSDA Kaltim and Prof A. Arrifien Bratawinata of the University of Mulawarman in Samarinda (UNMUL) are also thanked for their support of the study. I would like to thank the Indonesian Institute for Sciences (LIPI), the Provincial Wildlife Conservation Department (BKSDA Kaltim) and local governments of Central- (KUKER) and West Kutai (KUBAR) for granting permission to conduct field research. At the Division of Inter-Institutional Cooperation of LIPI, Ibu Ina Syarief and Ibu Krisbiwatti have been very helpful in arranging necessary permits and letters and are thanked for this. I thank all my field assistants gratefully: Hardy, Syafrudin, Chaironi, Zainudin, Syoim, Rudiansyah, Bambang, Sonaji, Marzuki, Iwiet, Munadianto, Hendra, Deni, Ramon, Audrie and particularly Ahank, Arman, Budiono, Yusri, Karen and Syachrani, who assisted during most surveys and with full dedication. The boatsmen Pak Sairapi, Pak Muis, Pak Mahyuni, Pak Iwan, Pak Johan, Pak Kasino and Pak Anto are thanked gratefully. I also enjoyed the company of Djupri, Fleur Butcher and Pak Sega during the first preliminary survey in 1997. Funding for fieldwork was provided by Ocean Park Conservation Foundation, Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Gibbon Foundation; Netherlands Ministry of Agriculture, Nature Management and Fisheries (PIN/ KNIP); Van Tienhoven Stichting; World Wildlife Fund for Nature (Netherlands); Stichting Doctor Catharina van Tussenbroek Fonds, Coastal Resource

ii Preface and acknowledgements

Management Program/ Proyek Pesisir and Amsterdamse Universiteits Vereniging. Funds for printing costs were provided by the Jurriaanse Stichting and the Institute for Biodiversity and Ecosystem Dynamics. All sponsors mentioned are acknowledged gratefully. During workshops and congresses it has been very pleasant to meet with a number of cetacean colleagues of which some have been particularly helpful during the research; First of all I owe a great deal of gratitude to Thomas A. Jefferson, who allowed me to assist in his research on Indo-Pacific humpback dolphins in Hong Kong for 2 months in 1998; he taught me many useful survey techniques and has always been helpful in all my research questions and requests for literature. I also thank the “Ocean Park crew” for their support, interest in my project and their good company. Randall Reeves, William Perrin, Bernd Würsig, Brian Smith, Fabrizio Borsani, Tony Martin, Jeanette Thomas, Tamara Mcguire, Ian Baird, Randall Wells, Vincent Nijman, Miquel Vences, Isabel Beasley, Guido Parra, Chris Smeenk, Martjan Lammertink, Gabriella Fredriksson, Willie Smits, Arne Mooers, Kees Hazevoet, Martin Genner, Resit Sözer, Matthijs Couwelaar, Bert Hoeksema and Annelies Pierrot-Bults are thanked for sending me literature, exchanging ideas, reviewing my manuscripts, and their general support. I also thank my colleagues at the Zoological Museum, in particular Tineke Prins, Martjan Lammertink, Vincent Nijman, Mansour Aliabadi, Miquel Vences, Kees Roselaar, Tonnie Dunselman, Andre Walgreen, Adri Rol, Mohamed El Moussaoui, Wouter Kraandijk, Thomas van Wissen, Hans van Brandwijk, Tatjana Das, Wouter Los and colleagues at ETI for providing a pleasant working environment and support. I am grateful to all librarians of the Plantage Library and the Zoological Museum for use of their facilities. I owe a great deal to Fabrizio Borsani for providing me with a good hydrophone to make acoustic recordings. Kelly Robertson is thanked for analysing genetic skin samples of the pesut (this study is still in progress). All the co-authors in manuscripts of this thesis are thanked for their cooperation. I would like to thank my husband Budiono and Gustinah for translation of the summary and all chapter abstracts in Bahasa Indonesian. Tineke Prins and Kees Hazevoet are acknowledged for correcting the Dutch summary and Jan van Arkel for preparing the cover design and some figures for this thesis. I also thank Hari Moelyono for his nice pen-drawings of the pesut and for handling the main editing process of the small Video CD on the pesut and its conservation. I thank Syachrani and Erwin van Faassen for solving all my computer problems and program installations.

iii Preface and acknowledements

Paulien de Bruijn is thanked for providing a Dutch breeze for a while during my long stay in Indonesia. I also thank her and Linda Zwiggelaar for their help as paranymphs. My daughter was born when I was still in the middle of my analysis and thanks to Gaguk and the Titaantjes, especially Martine, Melanie, and Alette, who took over a great deal of her daily care, I was able to continue my analysis both in Indonesia and in the Netherlands. I am also especially grateful for the participation of Prof. dr. W. Admiraal, Prof. dr. P.H. van Tienderen, Prof. dr. S.B.J. Menken, Prof. dr. H.H.T. Prins, Prof. dr. J.F. Borsani, and Dr. C. Smeenk in the Doctorate Commission. I would like to express my gratitude to all fishermen and residents mostly along the Mahakam River, Balikpapan Bay and Berau Islands who participated in interviews or provided us with information on the pesut and/ or cetacean species or on fisheries. I am also grateful for the hospitality with which our team was always received in every village to which we came. I thank my foster family of Pak Usman in Long Bagun, the family of Pak Mahyuni in Data Bilang, the family of nenek and Masman, the family of Nina losmen in Muara Pahu, and the family of Pak Yan in Muara Kaman for their true friendship. Naturally, I would like to thank my best friends and family, especially Joleen, Linda, and my sisters Conny and Monique in the Netherlands and family in Indonesia, who have been very encouraging to me to pursue my scientific research, but in whose company I could also relax and have fun and enjoy the other non-scientific side of life. I thank Ronald for his encouragements and for sharing mountain walks through which I developed determination and which set off my exploration of fauna and flora. I am very grateful to my parents, who have always helped me with everything I needed so that I could concentrate on my work and have no other worries. Finally, I would like to thank my daughter Jannah and my husband mas Budi, for being the lights of my life, and for giving me the necessary distraction and faith in my work - Damai Selamanya.

iv General introduction in facultative river dolphins and Orcaella brevirostris

CHAPTER 1

A general introduction into the phenomenon of facultative river dolphins and the species Orcaella brevirostris

Behaviours displayed by coastal Irrawaddy dolphins in captivity ( Laem Sing, Thailand) such as this spy-hopping behaviour, has also been observed in wild Irrawaddy dolphins in the Mahakam River.

1 Chapter 1 (Facultative) river dolphins and river wanderers

The order of Cetacea is composed of a variety of 85 recognized species and 41 subspecies of baleen whales and toothed whales and dolphins (Perrin et al., 2002; Reeves et al., 2003). Cetaceans have originally successfully spread out over vast areas of the worlds‟ oceans and inner seas. The freshwater habitat has been “conquered” at first by four “older” river dolphin (sub)species, the Amazon dolphin or boto Inia geoffrensis, the Yangtze dolphin or baiji Lipotes vexilifer, the Ganges dolphin or shushuk Platanista gangetica gangetica and the Indus dolphin or bhulan P. gangetica minor where they adapted even further to “microhabitats”; lakes, confluence areas, rapid stream areas (Best & da Silva, 1993) and flooded forests (Layne, 1958). The boto has been suggested to have entered the Amazon basin from the Pacific some 15 million years ago (Grabert, 1983) or more recently (1.8-5 million years ago) from the Atlantic Ocean (Brooks et al.; Gaskin, 1982). Occasional river wanderers include representatives of several families of toothed whales: Within the Delphinidae family, the Indo-Pacific humpbacked dolphin Sousa chinensis has been recorded in the Fuchow River (now: Fuchung Jiang) and rivers flowing to Canton (Guangzhou) and 750 miles up the Yangtze, at least as far as Hankow (now: Hankou, near Wuhan) (True 1889). In Indonesia, they are reported to occur about 30 km upstream the Kapuas River in western Kalimantan (information of local fishermen) and in the Dali River in north-eastern Sumatra (Suwelo 1988). In Australia, they are found in the Brisbane River in Queensland (Klinowska, 1991). The Atlantic hump-backed dolphin Sousa teuszii is known to enter the Niger River and the Baniala River in Nigeria (Klinowska, 1991).This species is also known to occur in the Rio Gêba in Guiné Bissau (Spaans, 1990). The common dolphin Delphinus delphis has been observed in the Hudson River, north-eastern USA, as far as 230 km (Stoner, 1938). Bottlenose dolphins Tursiops truncatus have been reported in the Casamance River in Senegal and in the Rio Gêba in Guinea-Bissau (Spaans, 1990). They are expected to occur on other rivers in western Africa as well (Hazevoet, pers. comm. 1997). Two representatives of the Cephalorhynchinae, the Chilean or black dolphin Cephalorhynchus eutropia and the New-Zealand dolphin Cephalorhynchus hectori, occur in rivers. The first one moves at least 5 km up the Valdavia River (Goodall et al., 1988). The latter often enters and travels some distance upstream in several turbid rivers in flood during their northwards summer „migrations‟ (Watson 1981). Within the family of Phocoenidae, the harbour porpoise Phocoena phocoena, can also be found in tidal rivers (Klinowska, 1991). One individual was described to have reached Paris after entering the Seine River and in the 17th century harbour porpoises could be found in the canals of Amsterdam (Delsman, 1922; van Bree, pers. comm. 1997). Even two species of baleen whales might occasionally wander upstream rivers. The minke whale Balaenoptera acutorostrata and the humpback whale Megaptera novaeangliae, have been recorded respectively 16 km upstream the Snohomish River in Washington State (Scheffer & Slipp 1948) and 16 km up the Sacramento River in northern California (Warhol, 1986).

2 General introduction in facultative river dolphins and Orcaella brevirostris

Although the cetaceans above may move very far upstream and even remain there for weeks, they are most likely temporary visitors. Their usual range includes river mouths, bays, lagoons, estuarine complexes and virtually any shallow water marine region. However, the most conspicuous river „wanderer‟ is the white whale or beluga Delphinapterus leucas. This large, white dolphin moves regularly and sometimes in groups, very far upstream rivers. Below follows a list of those rivers where Belugas have been recorded very far upstream, or where they have been reported more often. In Alaska, one individual was found at 1500 km from the Bering Sea upstream the Yukon river near the Kuskokwim River and Nulato (Nelson & True, 1887). According to Lee (1878) the Beluga occurs, during the summer months, in all the mouths and in nearly all bigger rivers at the west coast of the Hudson bay as well as the Greenland coast. The St. Lawrence Beluga population in Quebec is regularly found at Ile de Coudres, about 600 km upstream the mouth. In almost all big river mouths and rivers in Russian-Siberian waters, groups of Belugas were regularly seen some hundred up to two thousand km upstream. Kleinenberg et al. (1969) provided an overview of their Russian-Siberian distribution. The most extreme wanderings included a record of 2000 km upstream the Amur River in eastern Siberia until the Argun River in China . In Europe, their river wanderings are very occasional and therefore caused much excitement and publicity. The occurrence of a Beluga in the Schelde River in Belgium until Dendermonde (c. 100 km from the mouth) in 1711 caused so much excitement that a statue of 2.5 meter length was made and is carried like a trophy around the town at each 25th anniversary. (Gewalt 2001 The one-month wandering of a Beluga in 1966 up the river Rhein until Bad Honnef (c. 400 km from the mouth) made world wide news in press (Gewalt 2001). Underlying factors of these riverine migrations are explained in terms of the riverine migration of prey species. For example, salmons are one of their prey species, which move upstream to lay their eggs and they become an easier prey in the shallow waters where they cannot swim so fast. Other reasons were proposed by Mohr (1952 in Gewalt 1976) in terms of dolphins having lost their direction and in terms of their active curiosity to explore. The fact that they also occur as groups in the river and that this happens frequently might also indicate that their riverine occurrence is probably based on more than an error. Three species of cetaceans, which have established separate populations in rivers and in near-shore, marine waters include the species Irrawaddy dolphin Orcaella brevirostris, the tucuxi Sotalia fluviatilis and the finless porpoise Neophocaena phocaenaenoides (Smith & Jefferson, 2002). These represent more recent colonizers of freshwater habitats, in comparison to the obligate river dolphins, and they have been described as “facultative” river cetaceans, due to their species‟ flexibility to inhabit marine and freshwater environments (Leatherwood & Reeves, 1994). Nevertheless, the freshwater populations may actually represent obligate freshwater populations. The time period of invasion or process of adaptation of these relative newcomers in rivers are unknown and some hypothesis are offered in Chapter 11. The tucuxi is sympatrically ocurring throughout much of its range with the boto and inhabits rivers and lake systems of Amazonia, the lower Orinoco River, and coastal marine waters from the

3 Chapter 1 Florianópolis region of Brazil, north to at least Nicaragua (Carr & Bonde, 2000; IWC 2001). The finless porpoise occurs sympatrically with the baiji in the Yangtze River and lakes system. Additionally, they inhabit shallow nearshore marine waters along the coasts of southern and eastern Asia, from the Persian Gulf east to Sendai Bay, Japan, and south to Java (Reeves et al., 1997, 2000; Parsons & Wang, 1998; Kasuya, 1999). Just like the finless porpoise, Irrawaddy dolphins have a wide, but patchily distribution occurring in shallow, near-shore tropical and subtropical marine waters of the Indo- Pacific, from north-eastern India in the west, northeast to the Philippines and south to northern Australia, including most of the Indonesian archipel (Dolar et al., 2002; Rudolph et al., 1997; Stacey & Leatherwood, 1997; Stacey & Arnold, 1999). Their coastal distribution is mostly concentrated in estuaries and mangrove bay areas (Chapter 3). Their freshwater distribution is limited to three major river systems: the Mahakam in Indonesia, the Ayeyarwady in Myanmar, and the Mekong in Laos, Cambodia and Vietnam. Besides, they also occur in two completely or partially isolated brackish water bodies, including Chilka Lake in India and Songkhla Lake in Thailand (Beasley et al., 2002; Smith & Jefferson, 2002). Although the concept of stocks was already commonly applied in conservation and management of whales by the International Whaling Commission, in the 1988- 1992 Action Plan of the IUCN/SSC Cetacean Specialist Group a rationale was provided by Perrin (1988), for also including endangered populations besides species in conservation action plans. In the next action plan of the IUCN/SSC Cetacean Specialist Group (Reeves and Leatherwood, 1994) two projects were proposed involving the investigation of the riverine status of facultative river dolphins, namely: “Investigation of status and conservation of Irrawaddy dolphins in southern Asia” and “Investigation of status and establishing protected areas for pesut in Indonesia”. Following the latter recommendation a preliminary survey was initiated in 1997 to assess threats, distribution range and densities of the Irrawaddy dolphin in the Mahakam, locally referred to as pesut (Chapter 2), after which a more intensive research was carried out in the form of this Ph.D study. Whether cetacean species are estuarine and/ or occasional river visitors, or are obligate riverine, it is clear that many species depend on the river or the river run-off in estuaries and are very vulnerable to the effects of habitat degradation. Therefore, a holistic approach of protection of the entire riverine ecosystem is of utmost importance within the conservation of (facultative) river dolphins. However, the key will lie in effective conservation of manageable sites, in which positive results for the local community may serve as exemplary for other sites so that gradually a large proportion of the river may be effectively protected (Chapter 6).

Orcaella brevirostris (Gray, 1866)

Type species: Orca (Orcaella) brevirostris Gray, 1866: 285. Type locality: “East coast of India in the harbour of Vizagapatam (presently named Vishakhapatnam)”.

4 General introduction in facultative river dolphins and Orcaella brevirostris

Orcaella fluminalis Anderson, 1871:80. Type locality: “1500 km from the mouth in the fomerly named Irrawaddy River in Burmah (presently named Ayeyarwady River, Myanmar)”. General concensus: One species, Orcaella brevirostris (Loze, 1973; Pilleri & Gihr, 1974; Rice, 1998). Common name: Irrawaddy dolphin; local name in the Mahakam River: pesut.

The most recent systematic placement of the species is within the Order of Cetacea, Suborder Odontoceti, Superfamily Delphinoidea, Family Delphinidae, Subfamily Orcininae. Although, Orcaella has been placed in other families, i.e., Delphinapteridae together with the beluga Delphinapterus leucas (Kasuya, 1973); Monodontidae together with the beluga and narwhal Monodon monoceros (Barnes, 1984; Gaskin, 1982; Pilleri et al., 1989); Orcaellidae (Nishiwaki, 1973), the most recent morphological and molecular data suggest that Irrawaddy dolphins belong to the family of Delphinidae (Arnold & Heinsohn, 1996; Le Duc et al., 1999). They have been placed in the following subfamilies based on morphological data: Orcininae (Fraser & Purves, 1960); Globicephalinae (de Muizon, 1988); the monotypic Orcaellinae (Perrin, 1989). Most recent research involved the use of molecular data, which placed the Irrawaddy dolphin closest to the killer whale Orcinus orca (Arnason & Gullberg, 1996) and into the Orcininae (LeDuc et al., 2002), although the relationship was relatively distant and bootstrap support was low.The taxonomic status at the intrapsecific level remains unclear (Stacey & Arnold, 1999). Hower, recent studies of skull morphology suggest possible specific differences between Australia/ New Guinea and Southasian forms (Beasley et al., 2002) Earlier a short account of the marine and freshwater distribution of Irrawaddy dolphins was given. Figure 1 shows locations of actual records, which are mostly based on Mörzer Bruyns, 1966, Stacey and Leatherwood, 1998, and some derived from various other sources. In Indonesian waters they were found some 16 km upstream the Belawan Deli River (north-eastern Sumatra); Surabaya (northeast coast Java); Cilacap, Segara Anakan (south coast of Central Java); Makassar (southwest coast Sulawesi); between Pulo Superiori and Pulo Biak; mouths of muddy waters (south coast West Papua), Mahakam River, Belitung Island (Mörzer Bruyns, 1966); coastal area of Kumai River (Central Kalimantan) (Kartasana & Suwelo, 1994); Seribu Islands (Java Sea); delta Kendawangan River(south coast West Kalimantan) (Rudolph et al., 1997), c. 380 km upstream the Barito River below Puruk Cahu (South Kalimantan); Kajan River (north East Kalimantan) (Delsman, 1922); Balikpapan and Sangkulirang Bay and coastal areas in between (coast East Kalimantan); Mahakam Delta (Kreb, this thesis, chapter 3); confluence of Sekonjer River and Kumai River (Central Kalimantan); delta Sesayap River (north East Kalimantan) (Erik Meijaard, in litt., 1997).

5 Chapter 1

Figure 1. Map with Irrawaddy dolphin distribution in South East Asia. Black dots representing actual records from literature and own observations.

6 General introduction in facultative river dolphins and Orcaella brevirostris Aims of the study

The general aims of this study were to investigate the conservation biology, social organization, social communication, and marine and freshwater distribution patterns (stock identification) of the freshwater and coastal Irrawaddy dolphin populations in south East Asia with special reference to the Mahakam River in East Kalimantan and adjacent coastal areas. The study‟s ultimate goal is to contribute to the conservation of Indonesia‟s only freshwater dolphin population that inhabits the Mahakam River and lakes in East Kalimantan, Indonesia, to fill in the gap in literature on social systems within (facultative) river dolphins and to an appropriate action plan to ensure the survival of the pesut. Detailed objectives involved: Conducting a preliminary survey prior to the Ph.D study to assess total dolphin distribution range and dolphin densities in different river areas, as well as to obtain an indication of threats to the population (Chapter 2); to assess cetacean diversity and distribution of coastal Irrawaddy dolphins along the coast of East Kalimantan (Chapter 3); to assess total population abundance through different seasons and by aid of several survey techniques, i.e., direct counts, density sampling, and mark-recapture analysis through photo-identification (Chapters 4 & 5); to study habitat use and preferences, site fidelity, short-and long-term movement patterns, threat analysis and developing a conservation program to protect the pesut population and its habitat in the Mahakam River (Chapter 6); to study specifically the effects of boats on dolphins‟ (surfacing) behaviour (Chapter 7); to provide an overview of status and threats of Irrawaddy dolphins throughout South East Asia (Chapter 8); to compare the social structures and breeding strategies of coastal and freshwater populations of Irrawaddy dolphins and additionally, study association patterns of individual dolphins in the Mahakam (Chapter 9); to compare vocalizations of coastal and freshwater populations of Irrawaddy dolphins and among different sites within the river; to investigate whether whistle shapes and frequencies are more determined by ecological, genetic or social factors; and to compare the vocalizations of all populations in East Kalimantan with those from Irrawaddy dolphins in Australian coastal waters and in the Mekong River to investigate whether the acoustics of the Irrawaddy dolphin follow an ecological (freshwater/ coastal) and/or geographical separation (Asia/ Indonesia/ Australia) (Chapter 10); to study the process of isolation of distinctive river and coastal Irrawaddy dolphins through their distribution patterns, (social) biology, historical biogeography, comparisons with other facultative riverine dolphin species. Although genetic material of both riverine and coastal populations were collected, unfortunately, only the skin cell material of the riverine populations yielded enough DNA to be used in the genetic analysis (Chapter 11); finally, to conduct a population viability analysis for a long-term prognosis of survival of the pesut population and assess the direction where conservation effort is mostly required and which events determine the viability of the population (Chapter 12).

7 Chapter 1 References

Arnason, U. & Gulberg, A. 1996. Cytochrome b nucleotide sequences and identification of five primary lineages of extant cetaceans. Molecular Biology and Evolution 13: 407-417. Anderson, J. 1871. Description of a new cetacean from the Irrawaddy River, Burma Orcaella fluminalis Anderson. Proceedings of the Zoological Society of London 39: 142-144. Arnold, P.W. & Heinsohn. 1996. Phylogenetic status of the Irrawaddy dolphin Orcaella brevirostris (Owen in Gray): a cladistic analysis. Memoirs of the Queensland Museum 39: 141-204. Barnes, L.G. 1984. Fossil odontocetes (mammalia: cetacea) from the Almejas Formation, Isla Cedros, Mexico. Paleo. Bios. 42. Museum of Paleontology, University of California Beasley, I., Arnold, P. & Heinsohn, G. Geographical variation in skull morphology of the Irrawaddy dolphin, Orcaella brevirostris (Owen in Gray, 1866). The Raffles Bulletin of Zoology, Supplement 10: 15-34. Brooks, D.R., Thorson, T.B. & Mayer, M.A. 1981. Freshwater stingrays (Potamotrygonidae) and their helminth parasites: testing hypotheses of evolution and coevolution. In: V.A. Funk & D.R. Brooks (eds.), Advances in cladistics. Pp. 147-175. New York Botanical Gardens, New York. Carr, T. & Bonde, R.K. 2000. Tucuxi (Sotalia fluviatilis) occurs in Nicaragua, 800 km north of its previously known range. Marine Mammal Science 16: 447-452. Delsman, H.C.D. 1922. Bruinvisschen en haaien in Borneo‟s groote rivieren. De Tropische Natuur. Orgaan van de Ned.-Indische Natuur-Historische Vereniging 11 (1): 155-157. De Muizon, C. 1988. Les relations phylogénétiques des Delphinida (Cetacea, Mammalia). Annales de Paléontologie : Vertébrés-Invertébrés 74 : 159-227. Dolar, M.L.L., Perrin, W.F., Gaudiano, J.P., Yaptinchay, A.A.S.P., Tan, J.M.L. 2002. Preliminary report on a small estuarine population of Irrawaddy dolphins Orcaella brevirostris in the Philippines. Raffles Bulletin of Zoology, Suppl. 10:155-160. Fraser, F.C. & Purves, P.E. 1960. Hearing in cetaceans. Evolution of the accessory air sacs and the structure and function of the outer and middle ear in recent cetaceans. Bulletin of the British Museum of Natural History (Zoology) 7: 1-140. Gaskin, D.E. 1982. The ecology of whales and dolphins. Heinemann, Toronto. Gewalt, W. 2001. Der Weisswal: Delphinapterus leucas. Hohenwarsleben, Westarp-Wiss., Die Neue Brehm-Bücherei. [In German]. Goodall, R.N.P., Norris, K.S., Galeazzi, A.R., Oporto, J.A. & Cameron, I.S. (1988). On the Chilean dolphin, Cephalorhynchus eutropia (Gray, 1846). Rep. Int. Whal. Commn, Spec. Issue 9: 197-257. Grabert, H. 1983. Migration and speciation of the South American Iniidae (Cetacea: Mammalia). Zeitschrift für Säugetierkunde 49: 334-341. Gray, J. E. 1866. Catalogue of the seals and whales in the British Museum, 2nd edition. British Museum, London, 402 pp. IWC. 2001a. Report of the standing sub-committee on small cetaceans. Journal of Cetacean Research and Management (Special Issue) 2: 1-60. Kartasana, G.F. & Suwelo, I.S. 1994- . The existence of Irrawaddy dolphins at Kumai Bay, Central Kalimantan, Indonesia. Unpublished Manuscript.

8 General introduction in facultative river dolphins and Orcaella brevirostris

Kasuya, T. 1973. Systematic consideration of recent toothed whales based on the morphology of the tympano-periotic bone. Sci. Rep. Whales Res. Inst. 25: 1-103. Kasuya, 1999b: IUCNYablokov, A.V., Bel‟kovich, B.M. & Tarasevich, M.N. 1969. Beluga (Delphinapterus leucas). Investigation of the species. Translated from Russian. IPST Press, Jerusalem, 376 pp. Kleinenberg, S.E., Yablokov, A.V., Bel‟kovich, B.M. & Tarasevich, M.N. 1969. Beluga (Delphinapterus leucas). Investigation of the species. Wiener Bindery, ltd., Jerusalem. Pp. 376. Translated from Russian. Klinowska, M. 1991. Dolphins, Porpoises and Whales of the world. The IUCN Red Data Book. IUCN, Gland and Cambridge. Layne, J.N. 1958. Observations on freshwater dolphins in the upper Amazon. Journal of Mammology 39: 1-22. Leatherwood, S. & Reeves, R. 1994. River dolphins: a review of activities and plans of the Cetacean Specialist Group. Aquatic Mammals 20 (3):137-154. LeDuc, R.G., Perrin, W.F. & Dizon, A.E. 1999. Phylogenetic relationships among the delphinid cetaceans based on full cytochrome b sequences. Marine Mammal Science 15 (3): 619-648. Lee, H. 1878. The White Whale. London. Lloze, R. 1973. Contributions a l’étude anatomique, histologique et biologique de l’Orcaella brevirostris (Gray -1866) (Cetacea-Delphinidae) du Mekong. Thèse présenté a l‟Université Paul Sabatier de Toulouse pour l'obtention du grade de docteur des- sciences naturelles, 598 p. Lowry, L.F., Burns, J.J. & Nelson, R.R. 1987. Polar bear, Ursus maritimus, predation of belugas, Delphinapterus leucas, in the Bering and Chukehi Seas. Can. Field-Nat. 101: 141-146. Mörzer Bruyns, W.F.J. 1966. Some notes on the Irrawaddy dolphin, Orcaella brevirostris (Owen, 1866). Zeitschrift für Säugetierkunde 31: 367-370. Nelson, E.W., True, F.W. 1887. Mammals of Northern Alaska. Report upon Hist. Natur. Collect. made in Alaska. Arctic Ser. of Publ. 3, US Army. Nishiwaki, M. 1972 (year wrongly put in text). Cetacea: general biology. In: Ridgway, S.H. (ed.), Mammals of the Sea. Pp. 6-136. Charles C. Thomas, Springfield, Illinois. Parsons, E.C.M. & Wang, J.Y. 1998. A review of finless porpoises (Neophocaena phocaenoides) from the South China Sea. Proceedings of the Third International Conference on the Marine Biology of the South China Sea, Hong Kong, 28 October-1 November 1996 (ed. B. Morton). Hong Kong University Press. Perrin, W.F. 1989. Dolphins, porpoises and whales. An action plan for conservation of biological diversity: 1988-1992. 2nd ed. IUCN, Gland, Switzerland. Perrin, W.F., Wursig, B. & Thewissen, J.G.M. (eds). 2002. Encyclopedia of marine mammals. Academic Press, San Diego. Pilleri, G. & Gihr, M. 1974. Contribution to the knowledge of the cetaceans of Southwest and Monsoon Asia (Persian Gulf, Indus Delta, Malabar, Andaman Sea and Gulf of Siam). In: G. Pilleri (ed.), Investigations on Cetacea Vol. 4. Pp. 95-149. Berne, Switzerland. Pilleri, G., Gihr, M. & Kraus, C. 1989. The organ of hearing in Cetacea 2. Paleobiological evolution. Investigations on Cetacea 22: 5-185.

9 Chapter 1 Reeves, R.R., Wang, J.Y., & Leatherwood, S. 1997. The finless porpoise, Neophocaena phocaenoides (G. Cuvier 1829): a summary of current knowledge and recommendations for conservation action. Asian Marine Biology 14: 111-143. Reeves, R. R. & Leatherwood, S. 1994. Dolphins, porpoises, and whales: 1994-1998 action plan for the conservation of cetaceans. IUCN Species Survival Commission, Gland, Switzerland. Reeves, R.R., Jefferson, T.A., Kasuya, T., Smith, B.D., Wang Ding, Wang, P., Wells, R.S., Würsig, B. & Zhou, K. 2000. Report of the workshop to develop a conservation Action Plan for the Yangtze River finless porpoise, Ocean Park, Hong Kong, 16-18 September 1997. Pp. 67-80 in: Biology and Conservation of Freshwater Cetaceans in Asia ( eds. R.R. Reeves, B.D. Smith, & T. Kasuya). IUCN/SSC Occasional Paper 23, Gland Switzerland and Cambridge, UK. Reeves, R.R., Smith, B.D., Crespo, E.A. & di Sciara, G.N. 2003. Dolphins, whales and porpoises. 2002-2010 Conservation Action Plan for the World’s Cetaceans. IUCN, Gland, Switzerland and Cambridge, UK. Rice, D.W. 1998. Marine mammals of the world: systematics and distribution. Special publication No. 4 of the Society for Marine Mammology. Allen Press, Lawrence, Kansas, 231 pp. Rudolph, P., C. Smeenk and S. Leatherwood, 1997. Preliminary checklist of cetacea in the Indonesian Archipelago and adjacent waters. Zoologische Verhandelingen. Leiden, Nationaal naturhistorisch Museum. Scheffer, V.B. & Slipp, J.W. 1948. The whales and dolphins of Washington State with a key to the cetaceans of the west coast of North America. Amer. Midl. Na. 39: 257-337. Smith, T.G. 1985. Polar bears, Ursus maritimus, as predators of belugas, Delphinapterus leucas. Can. Field-Nat. 99: 71-75. Smith, B.D. & Jefferson, T.A. 2002. Status and conservation of facultative freshwater cetaceans in Asia. The Raffles Bulletin of Zoology, Supplement 10: 173-187. Spaans, B. (1990) Dolphins in the coastal area of Guiné Bissau. Lutra 33: 126-133. Stacey, P.J. & Leatherwood, S. 1997. The Irrawaddy dolphin, Orcaella brevirostris : A summary of current knowledge and recommendations for conservation action. Asian Marine Biology 14 : 195-214. Stacey, P.J. & Arnold, P.W. 1999. Orcaella brevirostris. Mammalian Species 616: 1-8. Stoner, D. 1938. New York State records for the common dolphin, Delphinus delphis. N.Y. State Mus. Circ. 21: 1-16. Suwelo, I.S. 1988. Whales and whaling in Indonesia. Paper submitted to the Indian Ocean Sanctuary Administrative Meeting of the International Whaling Commission, Canberra, 18-20 May. True, F.W. 1889. Contributions to the natural history of the cetaceans: A review of the family Delphinidae. Bull. U.S. Nat. Mus. 36, 192 pp. Vlaykov, V.D. (1944) Chasse, biology et valeur économique du Marsouin Blanc ou Béluga (Delphinapterus leucas) du fleuve et du golfe Saint-Laurent. Ph.D. Département des pêcheries, Quebec, contribution No. 14: 194 p.[ In French]. Warhol, P. 1986. Humprey. Whalewatcher 20: 13-15. Watson, L. 1981. Whales of the world. Hutchinson & Co. Ltd, London.

10 Preliminary observations of the Irrawaddy Dolphin in the Mahakam River

CHAPTER 2

Observations on the occurrence of the Irrawaddy Dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia

Zeitschrift für Säugetierkunde 64, pp. 54-58, 1999 (with additions)

In order to collect information of the dolphin (seasonal) distribution range, fish abundance and threats, interviews with fishermen and residents throughout the range were held.

11 Chapter 2

ABSTRACT

A two-months’ preliminary survey by boat was conducted on a population of Irrawaddy dolphins Orcaella brevirostris in the Mahakam River in East Kalimantan, Indonesia in 1997 in order to assess total dolphin distribution range and encounter rates. Dolphins were not homogeneously distributed over the entire river length and over different habitats. Significantly higher encounter rates were found in the middle river section between Muara Kaman (c. 200 km from the mouth) and Long Iram (c. 490 km from the mouth) compared to the lower and upper river sections. This section presumably forms the primary habitat for the dolphins, at least during medium to low water levels when this survey was conducted. Encounter rates are just as low to those reported for the critically endangered Yangtze River dolphin, Lipotes vexilifer and indicate the critical situation of this population. Various factors seem to degrade their habitat and more intensive research to monitor total population abundance is required in order to reassess their status.

RINGKASAN

Dua bulan survei awal dengan menggunakan kapal untuk meneliti populasi lumba- lumba Irrawaddy Orcaella brevirostris di Sungai Mahakam, Kalimantan Timur, Indonesia pada tahun 1997 untuk memperkirakan jangkauan penyebaran dan mencari rata-rata. Lumba-lumba penyebarannya tidak merata sepanjang sungai dan pada habitat yang berbeda. Rataan penemuan yang lebih tinggi ditemukan di bagian tengah sungai antara Muara Kaman (c. 200 km ke hulu) dan Long Iram (c. 490 km ke hulu) dibandingkan dengan bagian hilir dan hulu sungai. Bagian ini mungkin merupakan habitat utama untuk lumba-lumba, setidaknya selama permukaan air sedang sampai permukaan air rendah pada saat survei ini dilakukan. Rata-rata penemuan sama kecilnya dengan yang dilaporkan pada lumba-lumba yang terancam kepunahan di Sungai Yangtze, Lipotes vexilifer dan menunjukkan kritisnya keadaan populasi ini. Banyak faktor-faktor yang nampaknya semakin mengurangi habitat mereka dan diperlukan penelitian yang terus- menerus untuk mengawasi keseluruhan jumlah populasi dengan tujuan untuk mengetahui kembali status mereka.

12 Preliminary observations of the Irrawaddy Dolphin in the Mahakam River

The Irrawaddy dolphin, Orcaella brevirostris (Gray, 1866), is considered a ‘facultative’ river dolphin of which distinct riverine and coastal, marine populations exist. The species is mainly found in shallow coastal waters of the tropical Indo-Pacific, but also in major river systems, in particular: Irrawaddy, Mekong, Mahakam, and the estuaries of the Ganges and Brahmaputra (Thomas, 1892; Lloze 1973; Leatherwood et al., 1984; Marsh et al., 1989). Relatively few published studies exist pertaining specifically to the population of Irrawaddy dolphins, in the local vernacular referred to as Pesut, in the Mahakam River, East Kalimantan, Indonesia. Studies so far have focused on the distribution and daily movement pattern of the species in Semayang-Melintang Lakes and connecting Pela and Melintang tributaries (Priyono, 1994) and on bioacoustics (Kamminga et al., 1983). Although no systematic surveys on their abundance have been conducted so far, the Indonesian Directorate General of Forest Protection and Nature Conservation reported the existence of a population of 100-150 individuals for Semayang Lake, Pela River, and adjacent Mahakam River (Tas’an & Leatherwood, 1984) while an unpublished estimate of 68 individuals in the Mahakam River was reported by Priyono in 1993. In this study, I present results of a preliminary survey, which was conducted on the Mahakam River, its tributaries and adjacent lakes in East Kalimantan, Indonesia. Two surveys were conducted, both at medium to low waterlevels, the first from 27 February till 9 March 1997 and the second from 21 March till 6 April 1997. The river was surveyed using a small motor boat, occasionally by large public boat and by large and small motorized canoes, from Muara Kaman (ca. 200 km upstream) to Burit Halau, at the rapids past Long Bagun (ca. 600 km upstream). In addition, the Semayang, Melintang and Jempang Lakes were surveyed as well as the Pela, Melintang, and Kedang Pahu tributaries. The total survey length was 1085 km. For analysis of sighting frequencies, the river was divided into a lower (from Samarinda, ca. 100 km upstream, until Muara Kaman, middle (from Muara Kaman until Long Iram, ca. 490 km upstream), and upper section (from Long Iram until the rapids after Long Bagun). Tributaries and lakes surveyed were also analysed separately. Encounter rates were calculated for each section by dividing the number of observed dolphins by the number of kilometers searched. For testing whether the sighting frequencies are homogeneously distributed over all sections, and whether significant differences exist between different sections, G-tests of goodness of fit for single classification frequency distributions were used. To obtain a better 2 approximation to x , Williams’ correction to G was applied (Gadj; Sokal & Rohlf, 1981). G values were compared with critical values of the chi-square distribution (table C in Siegel & Castellan, 1988). Because multiple tests were performed, a corrected alpha of 0.01 was used in place of the nominal alpha of 0.05 (Rice, 1989). Dolphins were spotted by eye and by means of binoculars. Group composition, location, diving times, respiration rates and behaviors were recorded and photos taken. Additional data on the occurrence and status of Pesut were collected by interviewing local inhabitants, mainly fishermen.

13 Chapter 2

14 Preliminary observations of the Irrawaddy Dolphin in the Mahakam River

During the present study, a total of 32 dolphins were observed, of which four were juveniles. During the first survey 29 individuals were encountered while during the second only 3 were observed, presumably because more time was spent in the upper section of the Mahakam, where no dolphins were observed. Group size varied from 3 to 7 animals with a median group size of 4 individuals. No minimum estimate of abundance could be made as only three dolphins were identifiable individually on the basis of their dorsal fin (no systematic photos of their dorsal fin were made). Also, there is the possibility that the dolphins might have been encountered more than once during each survey, in case they were heading in the same direction during the night as we were heading during the day. Irrawaddy dolphins were found to be rather inconspicuous; they do not leap high out of the water and may stay submerged for up to 12 minutes, surfacing only briefly. Except for some noises produced with their blow holes, which could be heard over 100 m distance, no audible whistles or pure tones were heard. Pesut appeared to be very social, continuously staying in close contact with one another, regardless of whether they were milling (feeding), travelling, or resting. Table 1 shows the encounter rates, i.e. the number of dolphins per km of river searched, for different sections of the Mahakam River sytem. The dolphins are not homogenously distributed over the whole length of different river sections, tributaries and lakes (Gadj=47.8, df=4, p<0.01). The encounter rates of the middle river section are significantly higher than those of the upper section (Gadj=39.2, df=1, p<0.01). Significantly higher encounter rates were also found for the tributaries when compared to the combined main river sections (Gadj=8.3, df=1, p<0.01). However, all tributary observations of Pesut were made in the relatively short Pela tributary (only 8 km search effort), a connecting tributary to Semayang Lake and the Mahakam River. No sightings were made in the longer tributary Kedang Pahu of which 65 km in total was searched. No significant differences were found between encounter rates of middle river section and tributaries. As all tributary observations were made in the Pela tributary connecting to the middle section of the main river, and observations in the middle section of the Mahakam were significant higher than in the upper section (with a higher search effort), this section presumably forms the primary habitat for the dolphins, when water levels are medium to low.

Table 1. Encounter rates - dolphins observed per km of river searched.

Section Search effort No. of Encounter rate (km) individuals No. of individuals/ km Lower River Section 20 0 0 Middle River Section 432 25 0.06 Upper River Section 505 0 0 Tributaries 78 7 0.09 Lakes 50 0 0

15 Chapter 2

Encounter rates for the Semayang and Melintang Lakes, though lower, were not significantly so, when compared to the combined rates of the river and tributaries (Gadj=3.9, G0.01=6.6). The significant difference in encounter rates between these sections is probably a result of treating dolphins sightings in the Pela tributary as tributary observations. However, the dolphins’ presence in either the Pela tributary or in Semayang Lake might depend on time of the day, as the dolphins are reported to migrate daily between these areas (Priyono, 1994). The absence of observations of dolphins in the lakes most certainly is due to the fact that only 50 km were surveyed of the 10.300 hectares and 8.900 hectares large Semayang and Melintang Lakes, respectively. No significant differences in encounter rates were found between lower and other river sections, possibly due to the low search effort in this section. The encounter rates found for Orcaella brevirostris in the Mahakam River are in the same order of magnitude as that reported for Lipotes vexilifer in the Yangtze River (0.09 dolphins/ km), a population considered to have a high exctinction risk (Hua & Chen, 1992). However, the encounter rate of 0.06 dolphins/ km in the mainstem Mahakam River, is considerably lower than those recorded, at similar medium-low water level conditions, for Inia geoffrensis and Sotalia fluviatilis in the mainstems of the Amazon-Marañon-Ucayali (0.18 and 0.27 dolphins/ km, respectively; Leatherwood, 1996). In the present study, Pesuts were observed up till Tering, 400 km upstream (Fig. 1), but they are said to occur up till the waterfalls after Long Bagun. Although no sightings were made in any of the lakes visited, Pesut has frequently been recorded in Semayang and Melintang Lakes (Tas'an & Leatherwood, 1984), but the dolphins are said to be absent from JempangLake. Whether the Pesut occurs between Samarinda (near the mouth of the river) and the open sea, and in which of the river's tributaries, remains unclear. When water levels are high, dolphins are often observed by local inhabitants high up the Kedang Pahu tributary, past the village of Damai. Although the dolphins at always moved away from our research vessel, they were observed twice near two villages (Muara Pahu and Tering) with high levels of boat traffic. According to local fishermen, they were said to frequent these places almost on a daily basis, presumably because of the higher availability of fish. In conclusion, the results from this preliminary survey seem to indicate that encounter rates of the Irrawaddy dolphin in the Mahakam River are relatively low and fall in the same class of those recorded for the seriously threatened Lipotes vexilifer. Furthermore, middle sections of the river seems to be the primary habitat of Pesut, at least at medium to low water levels. Given the many factors contributing to possible deterioration of dolphin habitat (e.g. pollution from mining, forest fires, logging and siltation), these observations of low encounter rates merits further study.

16 Preliminary observations of the Irrawaddy Dolphin in the Mahakam River

ACKNOWLEDGEMENTS

I wish to thank the East Kalimantan nature conservation authorities, sub Balai Konservasi Sumber Daya Alam, WWF-Indonesia and WWF-Netherlands for their support and cooperation. My special thanks to Ir. A.M. Rachmat, D. Suprijono, F. Butcher and boatsman Pak Sega, which all participated in the survey. I would like to thank Dr P.J.H. van Bree for all his help and support throughout the study, as well for his comments on the manuscript. Dr. C.J. Hazevoet, Dr A.Ø. Mooers, V. Nijman and an anonymous reviewer are also thanked for their comments on this manuscript. I. Lysenko and the World Centre for Monitoring Cambridge are thanked for drawning of the map.

REFERENCES

Christensen, M.S. 1992. Investigations on the ecology and fish fauna of the Mahakam River in East Kalimantan (Borneo), Indonesia. Int. Revue gesamt. Hydrobiol. 77: 593- 608. Hua, Y., Chen, P. 1992. Investigation for impacts of changes of the lower reach of Gezhou Dam between Yichang and Chenglingji on the Baiji, Lipotes vexilifer after its key water control project founded. J. Fish. China 16: 322-329. Leatherwood, J.S. 1996. Distributional ecology and conservation status of river dolphins (Inia geoffrensis and Sotalia fluviatilis) in portions of the Periuvian Amazon. Diss. thesis, Texas University, Texas. Leatherwood, S., Peters, C.B. Santerre, R.. Clarke, J.T. 1984. Observations of cetaceans in the northern Indian Ocean Sanctuary, November 1980-May 1983. Rep. Int. Whal. Commn. 34 : 509-520. Lloze, R.. 1973. Contributions a l’étude anatomique, histologique et biologique de l’Orcaella brevirostris (Gray -1866) (Cetacea-Delphinidae) du Mekong. Diss. thesis Toulouse, France. MacKinnon, K.; Hatta, G.; Halim, H.; Mangalik, A. 1996. The ecology of Kalimantan. Indonesian Borneo. Ecol. Indonesia series 3: 152. Marsh, H., Lloze, R., Heinsohn, G.E., Kasuya, T. 1989. Irrawaddy Dolphin Orcaella brevirostris (Gray, 1866). In: H. Ridgeway and R.J. Harrison (eds.), Handbook of marine mammals. River dolphins and the larger toothed whales 4. Pp. 101-118. Priyono, A. 1994. A study on the habitat of Pesut (Orcaella brevirostris Gray, 1866) in Semayang-Melintang Lakes. Media Konservasi 4: 53-60. Rice, W.R. 1989. Analyzing tables of statistical tests. Evolution. Int. J. Org. Evol. 43: 223-225. Siegel, S., Castellan, N.J., Jr. 1988. Nonparametric statistics for the behavioral sciences. Second edition, McGraw-Hill , Inc. Sokal, R.R.; Rolf, F.J. 1981. Biometry. The principles and practice of statistics in biological research. Second edition, W.H. Freeman and company, New York.

17 Chapter 2

Thomas, O. 1892. Viaggio di L. Fea in Birmania e regioni vicine. XLI. On the Mammalia collected by Signor Leonardo Fea in Burma and Tenasserim. Annali del Museo Civico di Storia Naturale di Genova 1892: 913-949. Tasa’n & Leatherwood, S. 1984. Cetaceans live-captured for Jaya Ancol Oceanarium, Djakarta, 1974-1982. Rep. Int. Whal. Commn. 34: 485-489.

18 Coastal cetacean diversity and habitat preferences in East Kalimantan

CHAPTER 3

Cetacean diversity and habitat preferences in tropical waters of East Kalimantan, Indonesia

Daniëlle Kreb and Budiono

Submitted manuscript

Two Gray’s (pantropical) spinner dolphins, Stenella longirostris with distinctive tripartite color pattern, photographed in the Berau Archipelago, October 2003. Photo: Budiono.

19 Chapter 3

ABSTRACT

East Kalimantan was chosen as a site to investigate cetacean diversity because of its probability as a migratory pathway for cetaceans from the Pacific to the Indian Ocean through the Sulu-Sulawesi Seas and Makassar Straits. The Berau Archipelago in the northeast of East Kalimantan Province provided the highest species richness and cetacean abundance (0.64 individuals/ km searched) compared to two other coastal areas of equal coastline length and nearly similar area size in East Kalimantan. A total of 10 species and subspecies were found along the entire coastline (total study area is 8.538 km2) of which 8 were found in the Berau Archipelago (minimum area size is 170 km2). High cetacean diversity in this area is due to the abundant islands and reefs, in which habitat 60% of all taxa were encountered and which had the highest relative cetacean abundance (0.82 individuals/ km searched) of all habitat types, i.e., offshore and near shore waters, bay and delta. Most sightings were made within 5 km of islands and reefs, so a 5-km-radius protection zone off islands and major reefs may be one conservation recommendation. First sighting records (4) for Indonesia of Stenella l. roseiventris were made.

RINGKASAN

Kalimantan Timur telah dipilih sebagai tempat untuk penelitian keanekaragaman cetacean sebab kemungkinan besar daerah ini digunakan sebagai jalur berpindahnya cetacean dari Laut Pasifik ke Laut Hindia melalui Laut Sulu-Sulawesi dan Selat Makasar. Kepulauan Berau di timur laut Provinsi Kalimantan Timur menyimpan banyak kekayaan dari segi jenis dan jumlah (0,64 ekor/km yang diteliti) dibandingkan dengan dua perairan laut lainnya dengan panjang garis pantai dan ukurannya yang hampir sama di Kalimantan Timur. Sejumlah 10 jenis dan sub jenis ditemukan disepanjang garis pantai (Total daerah studi 8.538 km2), dimana 8 diantaranya ditemukan di Kepulauan Berau (ukuran daerah minimum 170 km2). Tingginya keanekaragaman cetacea di daerah dikarenakan oleh banyaknya pulau dan karang, dimana 60 % habitat dari seluruh jenis ditemukan dan relatif memiliki jumlah cetacea tertinggi (0,82 ekor/ km penelitian) dari seluruh tipe habitat, seperti lepas pantai dan perairan dekat pantai, teluk dan delta. Kebanyakan penampakan terjadi dalam jarak 5 km dari pulau and karang, jadi dalam area radius perlindungan 5 km dari pulau dan karang besar dapat menjadi satu rekomendasi untuk konservasi. Catatan pengamatan yang pertama kali (4) untuk Stenella I. roseiventris di Indonesia dibuat.

20 Coastal cetacean diversity and habitat preferences in East Kalimantan

INTRODUCTION

The coastal waters of East Kalimantan form the western part of the Indo-West Pacific centre of maximum marine biodiversity (Voris, 2000). Historical and ecological perspectives support this hypothesis. During the last ice age (17,000 yrs ago), sea level was situated 120 m lower than now (MacKinnon, 1997). Shelf seas, e.g., the Java Sea, had disappeared and Kalimantan was part of the South East Asian continental mainland. The Indonesian through-flow (Gordon & Fine, 1996) continued to pass east of Kalimantan, through the Sulu-Sulawesi Seas and Makassar Strait carrying larvae and plankton from the Pacific to the Indian Ocean. Similarly, these seas most likely represent a migratory pathway for whales and dolphins. East Kalimantan has a wide range of habitats such as major rivers, deltas, mangroves, island/ reefs and deepwater offshore habitat, which are all inhabited by cetaceans. The Indonesian Archipelago contains some 5 million km2 of territory (including water and land), of which 62% consists of seas within the 12-mile coastal limit (Polunin, 1983). At least 29 species of cetaceans are reported to occur in the seas of the Indonesian Archipelago (Rudolph et al., 1997). However, only a few dedicated studies have been conducted on the abundance, distribution and conservation of cetaceans in Indonesia. Cetaceans are threatened with local extinction in many parts of the world, but nowhere more obviously than in Asia (Reeves et al., 1997). Growing human populations are putting an increasing pressure on natural resources including the stocks of wild fish and crustaceans, supplies of freshwater, and even coastal landscapes themselves, e.g., through ‘reclamation’ projects, harbor construction, mariculture and oil spills. Rivers, estuaries and coastal marine waters are becoming increasingly unhealthy ecosystems for wildlife. Modifications and degradations of the habitats of dolphins and porpoises have often resulted in dramatic declines in their abundance and range (Reeves et al., 1997). The present survey involves a preliminary assessment of cetacean diversity in the waters off the East Kalimantan coast and provides the basis for future conservation- orientated research on cetaceans in this area. The objectives of the preliminary survey were to assess the diversity and occurrence of cetaceans and identify important cetacean areas in terms of species richness and abundance.

METHODS

Survey area

Near-shore and (island) offshore waters were surveyed along a total strip of 700 km of coastline. This coastline was divided into three survey areas of equal length, ca. 230 km (Fig. 1). Survey area 1 in the south included Balikpapan Bay (mangrove), near- shore waters, and the inner and outer Mahakam Delta area (mangrove). Total survey

21 Chapter 3

Berau Delta Survey Area 3

Kaniungan Islands

EAST KALIMANTAN o Mangkaliat 1 Sangkulirang Peninsula Bay

Survey Area 2

0 23 46 0o

Mahakam Delta

Survey Area 1

o Balikpapan 1 Bay o o o 117 118 119

Figure 1. Map of survey areas along the East Kalimantan coastline, Indonesia.

22 Coastal cetacean diversity and habitat preferences in East Kalimantan

area was 2467 km2. The shallow, near-shore strip (< 20 m depth) is quite wide (>5 km <10 km). Survey area 2 had an area of 2732 km2 and included the near-shore waters north of the Mahakam Delta, small delta areas of minor rivers, Sangkulirang Bay (mangrove) and offshore island reefs as far as the Mangkaliat Peninsula. The shallow coastal strip was very narrow (on average < 1 km) in the area north of the Mahakam Delta until Sangkulirang Bay and even narrower along the coast farther eastwards to the Mangkaliat Peninsula (< 100 m). Survey area 3 included the Berau Archipelago with an area of 3339 km2, which contains a high density of islands and reefs, the Berau Delta (mangrove), near- shore waters (>2 km < 4 km north of Kaniungan Islands and < 100m from Mangkaliat Peninsula until Kaniungan Islands) and offshore deepwater habitat (< 900 m deep). The southern Mangkalihat Peninsula narrows the passage between Sulawesi Island and Borneo Island and a shallow shelf is absent.

Field methods

Cetaceans were visually searched for along a strip of 700 km of coastline during vessel-based surveys in six different survey periods, each lasting two weeks on average between May 2000 and October 2003. Total search effort by boat was 4481 km (362 h) during 80 days. Area 1 was surveyed during all seasons (governed by winds from all directions), whereas area 2 was surveyed during eastern wind (calm sea) conditions and area 3 during a transition period from south-western to northern wind conditions with days of mirror-like sea surface alternated with days of beaufort 5 sea state. Only sightings made during days with an average beaufort sea-state of 3 or less were used for relative abundance analysis. Pre-determined survey transects were designed to provide representative survey coverage of various habitats. Searches were conducted alternatively from 2 wooden boats of different lengths, i.e., 16 m and 12 m, and horsepower 16 hp and 26 hp respectively, depending on sea conditions and habitat. When surveying deep, offshore waters and remote survey areas, the latter boat was used, which had an additional outboard engine and was used only off-effort for a fast return to shore. The 3-person- observer team followed a routine survey protocol for observation and data recording, in which the first observer scanned continuously with 7x50 binoculars, the second observer searched for dolphins unaided, and recorded all sighting effort data and environmental and geographical conditions using a GPS every 30 minutes, and the third observer searched at the rear by unaided eye and occasionally used binoculars. Positions changed every 30 minutes. One transect was surveyed in one day, and double sightings on the same transect were avoided. Upon making a sighting, radial distance between boat and dolphins was estimated, and compass bearing of the boat and of the dolphins and coordinates of the sighting location were recorded. Species were then identified. If more than one species was observed, it was recorded whether the multiple species mixed. If the

23 Chapter 3

species did not mix, the mean distance between the single-species groups was recorded. Minimum, maximum and best estimates were made of group size and of the number of calves and juveniles. We attempted to photograph each sighting for confirmation of species identification. Depth at sighting location was determined from an official sea map of the area for study area 3. For the other two study areas, a fish finder was used for depth measurement.

Table 1. Encounter rates of individual cetacean species by habitat type and habitats combined in decreasing order of abundance.

Encounter Mean depth Search rateb (r) Sighting (m) of Mean effort (dolphins/ Mean (Sub)species habitat sightings na G (km)2 km) r Tursiops truncatus offshore 172 (50-400) 6 18 261 0.291 island/ reefs 103 (5-300) 6 13 537 0.204 0.248 Stenella attenuata island/ reefsc 280 1 55 261 0.210 0.210 Stenella longirostris offshored 365 (300-400) 2 45 537 0.168 0.168 Stenella longirostris, sp. offshore 50 1 45 537 0.083 (with short beak)e islands 75 (35-115) 2 28 261 0.107 0.095 Orcaella brevirostris near shore 6.9 (2-23) 18 3 1616 0.029 delta 5.6 (3-10) 5 4.8 1010 0.019 0.085 bay 14.3 (2.5-30) 67 3.4 1057 0.220

Stenella l. roseiventris near shore 23 1 2 1616 0.002 offshore 260 (50-400) 3 10.7 537 0.060 0.030 island 35 1 8 261 0.030 Tursiops sp. nearshore6 12.5 (11-14) 4 9 1616 0.028 0.028 Pseudorca crassidens island 400 1 7 261 0.027 0.027 Peponocephala electra island 400 1 4 261 0.015 0.015 Globicephala macrorhynchus island 280 1 4 261 0.015 0.015 Tursiops aduncus offshore 350 1 7 537 0.013 0.013 Neophocaena phocaenoides near shore 6.3 (2-10) 3 4.7 1616 0.009 0.009 a = number of groups sighted b = habitat specific search effort c = > 20 m depth coastal contour line, > 5 km distance off islands and reefs d = < 20 m depth coastal contour line, > 5 km distance off islands and reefs e = tentative identification of possible sub-species of Stenella longirostris f = < 5 km distance of islands and reefs

24 Coastal cetacean diversity and habitat preferences in East Kalimantan

RESULTS

Species identification

A total of 112 independent sightings were made in the 700-km long- survey strip (20 20’ N, 1190 E – 10 50’ S, 116050 E) in a total survey area of 8.538 km2 (Fig. 1). A total of 868 individual cetaceans of 9 different species, one sub-species and one additional tentatively identified sub-species were encountered (Table 1). Five sightings of the dwarf spinner dolphin sub-species, Stenella l. roseiventris, represent the first records for Indonesia and first record of occurrence for the Sundai region1 (Fig. 2). The dwarf spinner dolphins were estimated to be ¾ the size of the more pelagic Gray’s dolphin, Stenella l. longirostris. Their colour pattern (consisting of two elements) was as dark-gray as for bottlenose dolphins, Tursiops truncatus. Near the abdomen, a not very distinct layer of lighter dark-gray was visible. They lacked the tripartite base pattern and distinct pectoral stripes of the larger pelagic spinner dolphins that we observed. Juvenile dwarf spinner dolphins were also observed. The dwarf spinner dolphins usually moved in small groups (mean n = 8 individuals) and were observed in mixed aggregations (within 30 m distance) in three out of five sightings. In the sightings with Gray’s spinner dolphins, their group formation remained in tact. During the other two sightings the dwarf spinner dolphins were observed in close proximity with other species but did not mix, i.e. the distance between different species was more than 30 m. Three sightings were made in deep water (50-400 m) but in relatively close proximity to islands (< 10 km). Three sightings were made of a variant form of larger pelagic spinner dolphins; these had a shorter beak and may represent an un-described sub-species. These were identified during one single-species sighting, one mixed aggregation with dwarf spinner dolphins and spotted dolphins, Stenella attenuata, and one sighting in close proximity (ca. 100 m) to dwarf spinner dolphins and bottlenose dolphins. Their mean group size was 34 individuals. One sighting was made of a small group of dolphins tentatively identified as Indo- Pacific bottlenose dolphins, Tursiops aduncus (n = 7 individuals), which could be distinguished from common bottlenose dolphins by having a more slender body, longer beak and slightly smaller body size. This small group occurred in area 3 in a mixed species group with an average distance of ca. 50 m from common bottlenose dolphins and ca. 50 m distance from spinner dolphins, which occasionally approached. General bottlenose behaviors included slow travel, milling and feeding, and there were many small tuna in the area. The remaining Tursiops sightings in area 2 and 3 were made near islands and reefs, and offshore waters, and appear to have been of T. truncatus. In area 1, Tursiops sightings were made near-shore, but no positive

1 This dwarf form of spinner dolphins was first described by Wagner 1846 as Delphinus roseiventris based on a specimen from the Arafura Sea in Indonesia. Later specimens were collected from the Mollucas.

25 Chapter 3

Figure 2. Three dwarf spinner dolphins, Stenella l. roseiventris with obscure, lateral color pattern, photographed in the Berau Archipelago, October 2003. Photo: Budiono.

species identification could be made, so all the bottlenose dolphin sightings in this area are referred to as Tursiops sp. No unidentified sightings were made.

Relative species abundance and habitat occurrence

The most abundant species observed was the common bottlenose dolphin (0.25 dolphins per habitat specific search effort in km). Other commonly observed species were the pantropical spotted dolphin (0.21 dolphins/ km) and the spinner dolphin (0.17 dolphins/ km). However, the spotted dolphin was only sighted once, but in a large group of 55 individuals. The species least frequently observed were the finless porpoise, Neophocaena phocaenoides, and the Indo-Pacific bottlenose dolphin (0.009 & 0.013 dolphins/ km, respectively). The cetaceans occurred in five different habitat types: near-shore (< 20 m depth coastal contour line, > 5 km off islands and reefs), offshore (> 20 m depth

26 Coastal cetacean diversity and habitat preferences in East Kalimantan

coastal contour line, > 5 km off islands and reefs), bay, delta, and islands/ reefs (< 5 km from islands and reefs). The dwarf spinner dolphin, common bottlenose dolphin, and the Irrawaddy dolphin, Orcaella brevirostris, had the most variable habitat occurrence as each species occurred in three marine habitat types. The latter species actually occurred in 4 habitat types if one includes the freshwater habitat (Mahakam River). Depths at sighting locations varied between a minimum of 2 m, recorded for finless porpoises and Irrawaddy dolphins in near-shore habitat, and a maximum of 350-400 m, recorded for spinner dolphins, dwarf spinner dolphins, bottlenose dolphins, Indo-Pacific bottlenose dolphins, false killer whales (Pseudorca crassidens) and melon-headed whales (Peponocephala electra) in island and offshore habitat.

Relative cetacean abundance by habitat and survey area

The habitats with highest relative abundance of cetaceans were island and reef (0.82 dolphins/ km searched), followed by, offshore (0.529 dolphins/ km) and bay (0.219 dolphins/ km) (Table 2). Delta and near-shore areas were rather poor by comparison (0.023 & 0.071 dolphins/ km, respectively). Near-shore areas and offshore habitats were moderately rich in species occurrence (both 40% of total number species encountered). Island/ reef habitats had the highest species richness (60% of total no. species). The bays and delta habitats were only frequented by one species, the Irrawaddy dolphin. Coastal Irrawaddy dolphins in and near the Mahakam delta were sighted offshore of the delta at low tide, whereas one inshore sighting at 10 km upstream of the mouth was made at high tide. The mean salinity of 12 ppt (SD = 10; range = 4.6-19.3 ppt) measured at dolphin positions in the delta is associated with brackish waters.

Table 2. Number of individuals and cetacean species encountered in different habitats.

Habitat Survey Total no. Encounter No. of % of total no. of effort (km) individual rate (sub)species (sub)species cetaceans (Dolphins/ (n = 10)a km) Bay 1057 231 0.219 1 10 Delta 1010 24 0.023 1 10 Near shore 1616 115 0.071 4 40 Offshore 537 284 0.529 4 40 Islands/ reefs 261 214 0.820 6 60 Total 4481 868 a = tentative identification of the variant form of Gray’s spinner dolphin with short beaks is excluded.

Relative cetacean abundance also varied by survey area: survey area 3, the Berau Archipelago, scored both the highest encounter rate (0.64 individuals/ km

27 Chapter 3

searched in area 3) as well as highest species richness, i.e., 8 species, which was 2.7 times higher than species richness in the other two areas, whereas the area surveyed was only 1.2 and 1.3 times larger than the other areas (Table 3). The minimum area size within which all 8 species of this area were found was ca.170 km2.

Table 3. Diversity of cetacean (sub-)species and relative individual- and species abundance per survey area.

Survey (Sub)species Surveyed Search Encounter Survey area Species areas habitats effort (km) rate (km2) richnessa (dolphins /km) Area 1 Neophocaena phocaenoides Near shore; 3216 0.12 2467 3 Orcaella brevirostris bay; large delta Tursiops sp. (outer & inner)

Area 2 Orcaella brevirostris Near shore; 549 0.21 2732 3 Stenella l. roseiventris bay; small delta; Tursiops truncatus offshore; islands Area 3 Globicephala Near shore; 714 0.64 3339 8 macrorhynchus large delta Pseudorca crassidens (outer); Peponocephala electra offshore; Stenella attenuata islands Stenella longiristris Stenella longirostris sp Stenella l. roseiventris Tursiops aduncus Tursiops truncates

_ (underline) = habitat in which dolphins were sighted a = including the sub-species Stenella l. roseiventris

Table 4. Mixed species sightings n Mixed species sightings (+ dependent sightings) Groups mixing or not?a

1 Neophocaena phocaenoides; Orcaella brevirostris mixing 2 Orcaella brevirostris; Stenella l. roseiventris not mixing; moving in other directions 3 Stenella longirostris; Stenella l. roseiventris; Tursiops truncatus; all species mixing Tursiops aduncus 4 Stenella longirostris, sp.; Stenella l. roseiventris; Tursiops truncatus not mixing; > 100 m distance among each species 5 Stenella longirostris, sp.; Stenella l. roseiventris; Stenella attenuata all species mixing 6 Pseudorca crassidens; Peponocephala electra mixing 7 Globicephala macrorhynchus; Stenella attenuata not mixing; > 30 m distance among each species 8 Stenella longirostris; Stenella l. roseiventris mixing n = independent sightings during which more than one species was encountered a = groups were considered to mix if the distance between different species was less than 30 m

28 Coastal cetacean diversity and habitat preferences in East Kalimantan

Species composition of sightings

Sightings of mixed species involved 20% (n = 8) of all sightings in habitats where more than one species were observed (n = 40) (Table 4). However, the percentage of sightings of groups which actually mixed was 12.5% (n = 5). The remaining 7.5% (n = 3) involved dependent sightings of groups, which did not mix (minimum distance range = 30 m & 100 m). All identified species mixed at least once with other groups, except for the short-finned pilot whale (n = 10 = 91% of all species). Dwarf spinner dolphins were most often sighted in mixed-species aggregations (n = 3), followed by spinner dolphins (n = 2), whereas all other species were seen to mix only once. Although Indo-Pacific bottlenose dolphins were observed at a close distance (15-20 m) from common bottlenose dolphins, they remained in group formation. In all sightings of mixed groups the different species of dolphins were within close range of each other, but they maintained their own group formation.

DISCUSSION

Species identification

Although dwarf spinner dolphins are usually associated with shallow in-shore waters (Perrin et al., 1999), the observation in deep waters in the Berau Archipelago is not so unusual since the area is very rich in islands and reefs, and deeper waters are interspersed with shallow reefs. Also, all deep-water sightings in this study were made within 10 km of islands and reefs. The dwarf spinner dolphins observed in this area share some characteristics with small spinner dolphins occurring in the Arabian Sea of Oman (Waerebeek et al., 1999) and the Aden Gulf of the Republic of Djibouti (Robineau & Rose, 1983), mainly in size with both forms is the smaller body size in comparison to pantropical (Gray’s) spinner dolphins. Also, one of the two forms of spinner dolphins described for Oman has a dark dorsal overlay, obscuring the tripartite base pattern as seen in pantropical spinner dolphins, and such a dark, not well-distinguished, pattern was also seen in the dwarf spinner dolphins in East Kalimantan. The belly in both had a lighter colour, although the form of Oman described above were pink-bellied, whereas the bellies of dwarf spinner dolphins were still gray-coloured. Pink colour in cetaceans occurring in very warm waters may be an ephemeral feature caused by physiological heat management. The dwarf spinners in East Kalimantan share the two-coloured pattern, a dark cape and slightly lighter pectoral and ventral side with the spinner dolphins in the Aden Gulf. The Indo-Pacific bottlenose dolphins appeared well distinguishable from common bottlenose dolphins, especially when they were encountered during the same sighting. The short-beaked form of spinner dolphins needs further study. However, the fact that these were identified during sightings when all individuals shared this trait, may indicate that this possibly represent a different form and perhaps a new sub-species. Also, the short

29 Chapter 3

beak form of spinner dolphins were never observed in mixed sightings with the pantropical-spinner dolphins. Photographic material may aid in the future identification of dwarf spinner dolphins, Indo-Pacific bottlenose dolphins and the short-beaked form of spinner dolphins in these waters.

Species richness

In spite of the fact that survey effort (km) in area 1 was 5 times as high as in the other areas and covered all seasons, species richness was 2.6 times lower than that found in area 3 and similar to that in area 2. Survey effort in areas 2 and 3 was only made during one season, so the species richness there is likely to be higher than recorded. Based on the relatively high species richness and presence of species with a restricted range and a globally conservation dependent status, the waters near the Berau Islands have both a local and global biodiversity importance. In comparison, 14 species of cetaceans were identified in Komodo (identified as one of the richest marine diversity sites in the Indo-Pacific) National Park waters (1,214 km2 surface waters) (Kahn et al., 2000), whereas in the Berau study area alone, 8 species were encountered in an area of only ca. 170 km2. Although there are undoubtedly other areas of high cetacean diversity in Indonesia, such as reported for Solor and Lembata Island in eastern Indonesia (Weber, 1923; Barnes, 1980; Hembree, 1980), there are no comparative data on local species richness available. Most likely only a proportion of the actual numbers of species that occur in the Berau Archipelago seasonally or year round were observed in this preliminary survey, so the species richness may be even higher.

Conservation recommendations

We found that most sightings and species occurred within 5 km of islands and reefs, so a 5-km-radius protection zone off islands and major reefs may be one conservation recommendation. Otherwise, the restricted range of 170 km2 within which 8 identified cetacean species in the Berau Archipelago were observed has a good conservation potential to become a marine vertebrate sanctuary. The area also hosts a number of shark and turtle species, and during the present survey a sighting of a large group of manta rays, Manta birostris, was made (65 individuals). Also, one dugong, Dugong dugon, was observed. The area also includes four islands that are frequently visited by tourists, so the area has a high potential for eco-tourism. However, any intended dolphin/ whale watching should be controlled and guided by instructed and responsible boat operators. The second area in East Kalimantan coastal waters that is recommended as a conservation site is Balikpapan Bay. A high density of Irrawaddy dolphins (0.22 dolphins/ km search effort) was observed in the bay, as well as occasional sightings of individual dugongs. Just 1 to 10 km outside the bay bottlenose dolphins and finless

30 Coastal cetacean diversity and habitat preferences in East Kalimantan

porpoises were observed in shallow waters. Since in this study area four surveys were carried out in different seasons (northwestern wind; northern wind; southeastern wind; southern wind) and Irrawaddy dolphins occurred during all surveys in the bay, this area has a year-round importance. Besides the extensive sedimentation due to mangrove conversion, which has caused a decrease in sea grass fields and fish resources, and pollution (oil and mining exploitation, local city sewages), no other major threats have been detected for this area.

Research recommendations

Red List designation of three species, i.e., pantropical spotted dolphin, spinner dolphin, and short-finned pilot whale is Lower Risk (conservation dependent) (Reeves et al., 2003). Conservation status for all other species is Data Deficient. The status of the dwarf spinner dolphin has not been evaluated, but it has the most restricted range, being confined to shallow inner waters of Southeast Asia (Perrin et al., 1989; Rudolph & Smeenk 2002) although in this study the species also occurred in deepwater habitat near shore. The lack of data on the status of the species in this study indicates the need for more research to assess each species’ abundance, habitat quality, and fisheries interactions. The freshwater population of Irrawaddy dolphins in the Mahakam River is listed as Critically Endangered (Reeves et al., 2003). Freshwater Irrawaddy dolphins were sighted between 180 km and 480 km upstream of the mouth (Kreb, 2002), whereas the most inshore occurrence of coastal Irrawaddy dolphins is about 20 km upstream of the mouth at high tide according to interviews with fishermen. Since the coastal dolphins have not been sighted or reported to move further upstream than 20 km from the mouth and only enter the delta at high tide, they are considered to belong to a different, coastal stock than the true Mahakam River population, which is considered an isolated population. Future research is needed, focusing on the collection of biopsy samples and DNA analysis of coastal and freshwater Irrawaddy dolphins in order to clarify their status. Future survey effort should focus particularly on the Berau Archipelago and involve investigating which areas have a year-round or seasonal importance for all target species and relating this to ecological and bio-geographical factors. More extensive data than those yielded by the present rapid assessment survey (only two weeks) should be collected in this area during at least one year, covering all seasons. These data are needed to prepare a conservation action plan for all threatened target species and their habitats if degraded, possibly through establishment of protected marine parks and local education/ awareness campaigns and a long-term cetacean monitoring program.

31 Chapter 3

ACKNOWLEDGEMENTS

We thank the Indonesian Institute of Sciences (LIPI), East Kalimantan Nature Conservation Authorities (BKSDA Kaltim), Mulawarman University Samarinda (UNMUL) for providing permits and a counterpart during the survey period. We thank the fieldobservers Achmad Chaironi, Ahank, Arman, Audrie J. Siahainenia, Bambang Yanupuspita, Karen D. Rahadi, Muhamed Syafrudin, M. Syoim, Matthijs Couwelaar, Ramon, Rudiansyah, Sonaji, Syahrani and boat drivers Pak Iwan, Pak Johan, Pak Muis, Pak Kasino and Pak Anto. We would like to thank the following persons for their hospitality: Pak Djamhari and villagers at Derawan Island; conservation staff of Turtle Foundation at Sangalaki Island; villagers at Kaniungan Island. We are grateful for the support received during the study period from NWO/ WOTRO (Foundation for the advancement of tropical research), Ocean Park Conservation Foundation, Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Gibbon Foundation; Netherlands Program International Nature Management (PIN/ KNIP) of Ministry of Agriculture, Nature Management and Fisheries; Van Tienhoven Stichting; World Wildlife Fund for Nature (Netherlands); Amsterdamse Universiteits Vereniging; Coastal Resource Management Program/ Proyek Pesisir. We thank the counterpart A. Arrifien Bratawinata, Frederick R. Schram, Peter J.H. van Bree, Chris Smeenk, Thomas A. Jefferson, Bert W. Hoeksema, Annelies Pierrot-Bults and the Plantage Library for their guidance, support and literature. William F. Perrin, Vincent Nijman and two anonymous reviewers are thanked for their corrections of the manuscript.

REFERENCES

Barnes, R. H. 1980. Cetaceans and cetacean hunting: Lamalera, Indonesia. Report on World Wildlife Fund Project 1428: 1-82. Gordon, A. L. & Fine, R. A. 1996. Pathways of water between the Pacific and Indian oceans in the Indonesian seas. Nature 379: 146-149. Hembree, E.D., 1980. Biological aspects of the cetacean fishery at Lamalera, Lembata. Report on World Wildlife Fund Project 1428: 1-55. Kahn, B., James-Kahn, Y. & Pet, J. 2000. Komodo National Park Cetacean surveys - A rapid ecological assessment of cetacean diversity, distribution and abundance. Indonesian Journal of Coastal and Marine Resources 3: 41-59. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press. Perrin, W. F., Dolar, M. L. L. & Robineau, D. 1999. Spinner dolphins (Stenella longirostris) of the western Pacific and South East Asia: pelagic and shallow-water forms. Marine Mammal Science 15: 1029-1053. Reeves, R. R., Wang, Y. J. & Leatherwood, S. 1997. The Finless Porpoise, Neophocaena

32 Coastal cetacean diversity and habitat preferences in East Kalimantan

Phocaenoides (G. Cuvier, 1829): A summary of current knowledge and recommendations for conservation action. Asian Marine Biology 14: 111-143. Reeves, R. R., Smith, B. D., Crespo, E. A. & Notarbartolo di Sciara, G. 2003. Dolphins, whales and porpoises: 2002-2010 conservation action plan for the world’s cetaceans. IUCN/SCC Cetacean Specialist Group. IUCN, Gland, Switzerland and Cambridge, UK. Polunin, N. V. C. 1983. The marine resources of Indonesia. Oceanography and Marine Biology, an annual review 21: 455-531. Robineau, D., & J. Rose, 1983. Note sur le Stenella longirostris (Cetacea, Delphinidae) du golfe d’Aden. Mammalia, 47:237-245. Rudolph, P., Smeenk, C. & Leatherwood, S. 1997. Preliminary checklist of cetacea in the Indonesian Archipelago and adjacent waters. Zoologische Verhandelingen. Leiden, Nationaal naturhistorisch Museum. Rudolph, P. & Smeenk, C. 2002. Indo-West Pacific marine mammals. In: Perrin, W. F., B. Wursig & J. G. M. Thewissen (eds), Encyclopedia of marine mammals. Academic Press, London. Pp. 617-625. Van Waerebeek, K., Gallagher, M., Baldwin, R., Papastavrou, V. and Al-Lawati, S. M. 1999. Morphology and distribution of the spinner dolphin, Stenella longirostris, rough-toothed dolphin, Steno bredanensis and melon-headed whale, Peponocephala electra, from waters off the Sultanate of Oman. J. Cetacean Res. Manage 1: 167-177. Voris, H. K. 2000. Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems and time durations. Journal of Biogeography 27: 1153-1167. Weber, M., 1923. Die cetaceen der Siboga-Expedition. Vorkommen und fang der cetaceen im Indo-Australische Archipel. Siboga-Expeditie 58. E.J. Brill, Leiden. Pp. 1-38, Pls I-III.

33 Chapter 3

34 Abundance estimation of freshwater Irrawaddy dolphins

CHAPTER 4

Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques

The Raffles Bulletin of Zoology, Supplement 10, pp 85-95, 2002

The river was scanned on top of the research vessel at 3,5 m eye-height above the water surface by use of binoculars and naked eye in order to detect dolphins surfacing. One rear observer (not in this picture) checked for dolphins missed by the front observers.

35 Chapter 4

ABSTRACT

On-going monitoring surveys are being conducted on a freshwater Irrawaddy dolphin population, locally referred to as the Pesut, inhabiting the Mahakam River in East Kalimantan, Indonesia. The aim of the study is to provide detailed information on the abundance, distribution, and ecology relevant to conservation of this population. This paper describes results from surveys in February 1999 - July 2000 that relate to population abundance estimates and compares different survey techniques. The primary goal of these investigations is to develop a conservation program for effective management of Indonesia’s only freshwater dolphin population, which is considered to be critically endangered. In this study, both modified strip-transect and direct count survey-methods were employed. Total search effort in the Mahakam River amounted to 4260 km (397 hours). Results of eight sighting surveys indicate that the dolphins in the mainstream Mahakam range from 180 km above the mouth to 480 km upstream, seasonally inclusive of several tributaries and lakes. However, dolphins are reported to sporadically move as far down- and upstream as 80 km and 600 km, respectively. The distribution of the dolphins changes seasonally and is influenced by water levels and variation in prey availability. The middle Mahakam area (MMA) and tributaries between 180 km and 350 km upstream were identified as primary dolphin habitat, based on highest dolphin densities. Sighting rates calculated for medium water levels in the MMA in 1999 and 2000 are nearly similar (c. 0.09 dolphins/ km, CV=25%, 49%). Highest sighting rate for the MMA was recorded at low water levels (0.142 dolphins/ km, CV=51%), indicating that dolphins are congregating in the main river in deeper waters. Lowest sighting rate was recorded at high water levels (0.035 dolphins/ km, CV=33%), suggesting that dolphins have moved upstream into the tributaries. Total mean abundance-estimates, based on density estimates and direct counts, were both 34 dolphins. However, the mean estimate based on density estimates exhibited more variation (CV = 25%), than the mean direct count estimate with associated CV of 5%. Unless a modified density sampling technique has been developed that is appropriate to the river conditions and takes into account dolphins daily migrations between main river and tributaries, direct count studies seem a more useful tool for assessing abundance of this particular freshwater population.

36 Abundance estimation of freshwater Irrawaddy dolphins

RINGKASAN

Pada survei monitoring yang telah dilakukan pada populasi lumba-lumba Irrawaddy air tawar, masyarakat setempat menyebutnya Pesut, mendiami Sungai Mahakam di Kalimantan Timur, Indonesia. Tujuan dari penelitian adalah untuk menghasilkan informasi yang lengkap mengenai jumlah, penyebaran, dan ekologi berkaitan dengan perlindungan populasi pesut. Tulisan ini menjelaskan hasil survei dari Februari 1999 – Juli 2000 yang berhubungan dengan perkiraan jumlah populasi dan membandingkan teknik survei yang berbeda. Tujuan utama dari penelitian ini adalah untuk mengembangkan suatu program konservasi yang efektif demi pengelolaan satu- satunya populasi lumba-lumba air tawar di Indonesia, dimana dapat dianggap terancam kepunahan. Pada penelitian ini, metode yang digunakan adalah strip-transect dan survei penghitungan langsung. Total penelitian di Sungai Mahakam mencapai 4260 km (397 jam). Hasil dari delapan survei pengamatan menunjukkan bahwa pesut berada di jalur utama Mahakam berkisar antara 180 km sampai dengan 480 km ke arah hulu, berdasarkan musim juga termasuk beberapa anak sungai dan danau. Namun, lumba-lumba dilaporkan sesekali bergerak jauh ke hilir dan ke hulu sepanjang 80 km dan 600 km. Penyebaran lumba-lumba berubah sesuai musim dan dipengaruhi oleh ketinggian permukaan air dan ketersediaan makanan yang bervariasi. Di Daerah Tengah Mahakam (DTM) dan anak sungai antara 180 km dan 350 km ke hulu telah diidentifikasikan sebagai habitat utama lumba-lumba, didasarkan pada kerapatan tertinggi lumba-lumba. Penampakan dihitung rata-rata pada permukaan air sedang di DTM pada 1999 dan 2000 hampir sama (c. 0,09 lumba-lumba/km, CV = 25%, 49%). Angka pengamatan tertinggi untuk DTM dicatat pada level air rendah (0,142 pesut/ km, CV = 51%), menunjukkan bahwa lumba-lumba berkumpul pada jalur sungai utama di perairan yang lebih dalam. Penamp = 33%) menunjukkan bahwa lumba- lumba bergerak mudik ke dalam anak sungai. Taksiran tengah total jumlah didasarkan pada perkiraan kerapatan dan perhitungan langsung, keduanya menunjukkan 34 individu. Namun, perkiraan tengah di dasarkan pada taksiran kerapatan menunjukkan lebih banyak perbedaan (CV = 25%), dibandingkan dengan taksiran tengah dari penghitungan langsung dengan hasil CV 5%. Kecuali telah dibuat suatu perubahan teknik pengambilan contoh kerapatan, yang sesuai untuk kondisi sungai dan memasukan perhitungan perpindahan harian lumba-lumba antara sungai utama dan anak sungai, penelitian dengan penghitungan langsung tampaknya menjadi cara yang lebih berguna untuk memperkirakan banyaknya populasi lumba-lumba air tawar yang khusus ini.

INTRODUCTION

River dolphins and porpoises are among the world’s most threatened mammals. The habitats of these animals has been degraded by human activities, in some cases resulting in dramatic declines in their abundance and range (Reeves et al., 2000). In

37 Chapter 4

Indonesia, one facultative freshwater dolphin population of Orcaella brevirostris, or Irrawaddy dolphin (locally referred to as Pesut) inhabits the Mahakam River and associated lakes in East Kalimantan. The species is found in shallow, coastal waters of the tropical and subtropical Indo-Pacific and in the Mahakam, Ayeyarwady and Mekong river systems (Stacey & Arnold, 1999). The status of the Irrawaddy dolphin in the Mahakam River was changed from ‘Data Deficient’ to ‘Critically Endangered’ in the IUCN Red List of Threatened Animals in 2000 (Hilton-Taylor, 2000). The species is protected in Indonesia and has been adopted as a symbol of East Kalimantan. Preliminary investigations on population abundance were made from late February 1997 to early April 1997 (Kreb, 1999). Thereafter, the current research project was undertaken, which began in February 1999 and will continue at least until November 2001. This paper describes research on dolphin abundance and an evaluation of the methods employed during surveys in February 1999 to July 2000. Relatively few published studies exist on the Irawaddy dolphin population in the Mahakam River. Studies so far have focused on their distribution and daily movement patterns in Semayang-Melintang Lakes and, in the region connecting the Pela and Melintang tributaries (Priyono, 1994), and on bio-acoustics (Kamminga et al., 1983). Earlier reports on their abundance were given by the Indonesian Directorate General of Forest Protection and Nature Conservation, which reported the existence of a population of 100-150 individuals for Semayang Lake, Pela River, and adjacent Mahakam River (Hardjasasmita, 1978) and an estimate of 68 individuals by Priyono (1994). However, no methods were presented about how both estimates were derived and these estimates may be merely guesses. A preliminary survey conducted by the author together with the East Kalimantan Nature Conservation Department reported that encounter rates in the middle Mahakam segment were 0.06 dolphins per linear kilometre in 1997 (Kreb, 1999).

METHODS

Study area

The Mahakam River is one of the major river systems of Kalimantan and runs from 118o east to 113o west and between 1o north and 1o south (Figure 1). The climate is characterised by two different seasons, namely dry (from July-October, southeast monsoon) and wet season (November-June, northwest monsoon) (MacKinnon, 1997). However, dry and wet periods, alternate during the wet season as well. The Mahakam River is the main transport system in the central part of East-Kalimantan. The river measures about 800 km from its origin in the Müller Mountains to the river

38 Abundance estimation of freshwater Irrawaddy dolphins

39 Chapter 4

mouth. The Semayang and Melintang Lakes are 10,300 hectares and 8,900 hectares, respectively (MacKinnon et al., 1997). Average widths of the river in the upper segment (from Long Bagun to Muara Benangak), middle segment (Muara Benangak to Muara Kaman) and lower segment (Muara Kaman to Samarinda), are 160 m, 200 m and 390 m, respectively, determined from visual estimates (see Survey methods). Highest mean transparency measured in the main river at low water levels is 24 cm (range 10-35 cm). Mean depths in the upper, middle, and lower segments, and in Semayang and Melintang Lakes were 10 m, 15 m, 12 m, 1.1 m and 1.3 m respectively. Differences in the water levels of the main river between high and low water conditions range as much as 10 m in ‘normal years’ (during extreme drought a maximum difference of 20 m may be recorded), whereas the maximum difference in lakes is c. 5 m. Water levels rise vertically and only slightly horizontally. Large passenger boats are able to navigate up to Long Iram (c. 427 km upstream). These boats of maximum 800 hp are only able to move as far upstream as Long Bagun (c. 560 km upstream) at high water levels. Rapids begin upstream of Long Bagun, which are only navigable by large motorised canoes (minimum 40 hp). These rapids limit dolphins from ranging further upstream. Coal mining, gold digging and logging activities pollute waters throughout the Mahakam. Fisheries in the middle segment of the Mahakam River and Semayang, Melintang, and Jempang Lakes are intensive, with an annual catch of 25,000 to 35,000 metric tons since 1970 (MacKinnon et al., 1997).

Field methods

Survey area Three surveys covering the entire study area were conducted in 1999 at low, medium and high water levels and one survey at medium water levels in 2000. Each one took about 4 weeks. Survey coverage included Ratah, Kedang Pahu, Belayan, Kedang Kepala, Kedang Rantau tributaries, Semayang and Melintang Lakes, as well as connecting tributaries, Pela and Jempang, part of the delta area, and minor tributaries (Figure 1). It was not possible to survey representative transects and extrapolate, because of the unpredictable variation of dolphin densities. Therefore, the entire range of dolphins in the Mahakam was surveyed. Ranges for different seasons were identified during preliminary surveys and from interviewing fishermen about the dolphins’ occurrence and their prey. To study the relation between fish- and dolphin migrations, interviews were held at different locations along the river to identify seasonal fish abundance for 25 species, including those suspected or known to be preyed upon by dolphins. Generally, dolphins did not frequent upstream areas of tributaries, where there was no more connection with freshwater swamp lakes that replenish the river with fish. If during the survey, the water conditions were such that

40 Abundance estimation of freshwater Irrawaddy dolphins

no dolphins were expected to occur in a particular area of a tributary and interviews with fishermen confirmed their absence for that period, the area was not surveyed. Water conditions in upstream areas of some tributaries connected with freshwater swamp-lake habitat, which did not favour dolphin occurrence during particular seasons were flooding, heavy currents in combination with lots of floating tree trunks, aquatic weeds and a high acidity. Also, decreasing water levels caused the dolphins to move downstream in the tributaries together with their prey. During one of the four intensive surveys conducted in May/ June 2001 at medium, decreasing water levels, areas that didn’t seem likely to be visited by the dolphins at that particular water level condition were nevertheless visited to check if this was true. Indeed no dolphins were found in these areas, which represented upstream areas of particular tributaries. Seventeen transect lines were surveyed in different habitats. Table 1 presents only those transects on which at least one sighting was made during one or more seasons. Each transect could be finished within one day. Eight transects were in the main river (c. 66 km), two were in the lakes (c. 48 km), five were in four middle segment tributaries (c. 50 km), and two were in upper segment tributaries (c. 32 km). In addition to transects, narrow tributaries that become accessible during high water levels for boats and potentially for dolphins were also surveyed.

Survey methods Modified strip-transect surveys were conducted, using the width of the river as the strip width for each transect within the identified dolphin distribution area. Modification thus included that strip width was not calculated as a function of perpendicular sighting distance because this distance was not a function of detection probability but of dolphins preferred distribution along the width of the river due to restrictions imposed by river width (see Results, Detection probability). Line-transect surveys were only conducted in Semayang and Melintang Lakes. Parallel line-transects were spaced at 1.5 km apart. Transect lines in the lakes were systematically designed to cover the entire survey area and no prior assumptions were made regarding dolphin distribution. Within the dolphin distribution area, the vessel always travelled in the central part of the river, even in river bends, which was possible because the main river was deep enough to do so. Only in areas where width of river was less than 100m, such as in some tributaries, was the boatsman free to travel near the riverbank. The widest arms of the delta area (width = >400 m) were surveyed following a zigzag pattern. Various environmental random samples, such as depth, clarity, pH and surface flow rate were taken on average five times a day at 3-5 spots along the width of the river, but only depth and clarity samples were analysed at the time of writing and presented in the survey area text. Depth was measured by lowering a rope with attached weight and markings every meter to the bottom of the river. Transparency was measured using a Secchi disk. When taking the depth and clarity measurements, the boat would drift with the flow so that the rope would be hanging in a straight line.

41 Chapter 4

The river was scanned from an elevated platform (eye-height c. 3 m above water level) on top of a motorised boat (12 hp) moving at a speed of c. 10 km/ hr in the central part of the river, covering an average distance of 50 km per day. The observation team consisted of at three observers, who rotated at 30-minute intervals. The first observer scanned the river continuously with 7x50 binoculars. The second observer searched for dolphins with naked eyes and recorded search effort and geographical data every 15 minutes by aid of a GPS. At the same time, environmental data were recorded, such as rain, wind, sun glare, fog conditions, cloud coverage and the extent to which floating tree logs and water weeds impaired sighting ability. Survey effort was suspended when sighting conditions were such that they impaired sighting efficiency, due to heavy rainfall and fog. Sun glare was never so bad that survey effort had to be ended and was anticipated by using a good binocular, head protection and sunglasses. The front observers alternated scanning with binoculars every 15 minutes to keep concentration high. During the first survey, a rear observer was present during the entire survey in primary dolphin habitat. All dolphins that were sighted by this observer involved groups that were located in or just after a river bend. Therefore, during the next surveys a rear observer was only present during and after river bends and confluence areas to allow the third observer at the rear to regain concentration for the next turn at the front observers’ position. Upon sighting dolphins, linear sighting distance and position of the first sighted dolphin along the width of the river was recorded (for calculation of relative perpendicular sighting distances). Dolphin positions were recorded in one of the following three categories. The central part was defined as the nearest area on each side of the transect line that occupies 25% of half the river width. On each side of the transect line, the area in between centre and shore occupied 50% of half the river width. The shore area was defined as an area of 25% on each side of the transect line nearest to the shore. Distance to the dolphins was estimated visually by the observer. A bridge of known distance that crossed the river in Samarinda, was taken as a reference for further distance estimations. During the survey, each fifteen minutes, the river width was estimated and agreed upon by all observers, so that distance estimations became more standardised. In addition, observers now and then referred to floating objects in the river and tried to standardise their estimation. During sightings, for between one half-hour and one hour, dolphins were counted, identified and their group composition was recorded (see Group size and sighting definition). The upper picture in Figure 2 (a) portrays two adults and one calf in the centre. Because of the group’s tight formation calves may easily remain undetected. Therefore observation time was rather long to allow for most accurate group size estimation. By aid of binoculars and naked eye alone observers tried to look for identifiable marks on the dolphin’s body and dorsal fins and drawings were made of the marks. Also, photographs and video footage were taken for identifying individual dolphins, but these analyses are not yet complete. The picture in the centre of Figure 2 (b) shows a typical slow surfacing pattern, which enables observers to notice and

42 Abundance estimation of freshwater Irrawaddy dolphins

Figure 2a. Two adults and one calf in the center swimming in tight formation

Figure 2b. Typical slow surfacing pattern enabling the observers to take notice of natural markings on the dolphin’s body and dorsal fin by aid of the naked eye and binoculars.

Figure 2c. A dolphin surfacing after a deep dive, producing a loud blow. Also, during a ‘normal’ dive- and surfacing pattern, the dolphins regularly produce blows, facilitating detection.

43 Chapter 4

photograph natural markings on the dolphin’s body. General and individual behaviours were recorded in combination with group- dive and surfacing times. Group diving times were collected during 14 sightings and were recorded for c. 30 minutes from the start of a dolphin’s dive and the surfacing of the next dolphin. However, time gaps of less than 3 seconds were ignored to reduce a bias towards short dive time intervals and were included in the duration of group-surfacing, i.e. the time a group is available on the surface for observation. The picture below in Figure 2 (c) shows a dolphin surfacing after a deep dive, producing a loud blow. Also during a ‘normal’ dive and surfacing pattern, the dolphins regularly produce blows, facilitating detection. Finally after all observations were made, the same kinds of samples were taken as those during search effort.

Double counting By the aid of identification of individual dolphins, I attempted to prevent double counting of dolphins on the same transects. Additionally, for the direct count analysis, I tried to reduce double counting of the same group or subgroup (in the case of an aggregation of dolphin groups) encountered on different transects. The following assumptions were made when determining if groups were similar: 1) minimally one individual of the (sub) group was re- identified. 2) similar age-classes. 3) similar (sub) group sizes, i.e., within the range of minimum and maximum group size estimates, as the earlier encountered group. 4) time elapsed between both encounters and distance between both locations should be in accordance with dolphins’ movements (mean speed is < 6 km/ hr). 5) absence or presence of dolphins that are easily recognisable by naked eye in only one of both groups did not favour similarity. 6) in case of any uncertainty, a non-conservative approach was preferred and groups were considered to be different. Preliminary analysis of studies of dolphins that were followed in one confluence area during three periods for on average six consecutive days, revealed that group composition during these days was relatively stable. That is to say, close interactions among different groups never exceeded one hour, which is the time that is spent observing the dolphins during surveys, which aim to identify total abundance of the population. Opposite to the problem of double counting is the problem of dolphins that moved in one direction at night whereas the survey team would move in the other direction and thus miss a sighting. However, replicates of surveys may account for this problem.

Group size and sighting definition For the calculation of sighting rates, mean group sizes are multiplied by number of sightings and divided per linear kilometre of river surveyed. To this aim it is necessary to determine what constitutes a sighting or group. Within this study, dolphins that are leaving the initial observed group of dolphins, i.e. moving outside the visibility of the

44 Abundance estimation of freshwater Irrawaddy dolphins

observers (c. 400 m), which remain close to the initially sighted group, during the observation period (on average one hour), are considered to belong to another group and constitute a new sighting. On the other hand, new dolphins that join the initial group are included in the group size estimate unless they move away from the initially sighted group within the observation period. Although for the new group no sighting distance data are available, the approach of defining group size as described above is preferred for the density and abundance estimates because the chance of a sighting of a group, whose composition remains the same during the observation period, is higher than the chance of encountering an opportunistic aggregation of different dolphin groups. The decision to separate dolphin groups because of their non- or short- duration interaction also makes comparisons of mean group sizes and number of sightings more meaningful among different surveys. Of the 58 sightings and groups of dolphins in total that were used to calculate the abundance and density estimates presented here, three sightings involved dependent sightings of groups that only interacted for a brief time during the observation period. Therefore, they were treated as three different sightings. The following example is given to elucidate what constitutes a dependent sighting: After an initial sighting was made of 3 individuals, which were followed downstream, another group of 9 dolphins was encountered. However, the initial group of 3 dolphins moved downstream away from the new group. While continuing observation on the group of 9 dolphins, another group of 3 dolphins from upstream joined the group for a moment and then moved into a small tributary, whereas the group of 9 dolphins moved upstream. So, instead of considering this as one sighting with 15 dolphins, I consider this as 3 different sightings and groups. The size of the group upon initial sighting includes all dolphins visible to the observers using a best, minimum and maximum estimate. Final decision about the group size estimation was taken by the primary researcher. In most cases at least one- half hour was needed to get a good count (depending on the group size), carefully looking for natural markings to identify individuals and determine if two surfacings were made by the same dolphin.

Availability bias and perception bias To account for undetected dolphins due to the dolphins’ submergence within the observers’ visibility field (availability bias) and reducing observers’ perception bias (those dolphins that surface in the visual range, but are still missed by all observers), a rear observer was present (see Survey methods). An attempt was made to reduce perception bias by suspending survey effort when sighting conditions were such that they impaired sighting efficiency, due to heavy rainfall and fog. Sun glare was anticipated by using a good binocular, head protection and sunglasses. Finally, scanning bouts with binoculars were rather short, i.e. 15 minutes, to keep concentration high. For comparison of increased sighting efficiency, two additional seasonal surveys (besides the four seasonal surveys described in this paper) of higher

45 Chapter 4

observer’ intensity were conducted. Each of these surveys covered the same transects in primary dolphin habitat and one observer was added to the observer team that now consisted of 4 observers (two front observers, one rear observer and one observer stand-by).

Analysis

Mean sighting frequencies were calculated per transect, habitat segment and water level condition. Mean number of sightings and sighting rates were calculated as the mean number of sightings and sighting rates of upstream and downstream surveys per transect and water level condition. Except for one segment representing a line transect in a lake, all transects were replicated per water level condition. For the lake transect that was only surveyed once, the number of sightings recorded were taken as the mean in order to be comparable with the other mean number of sightings, assuming that a replicate survey in the same period conducted in this lake would yield the same results. For the calculation of mean dolphin densities, the mean river width per segment was taken as the mean strip width. Abundance estimates were calculated for each transect as a product of dolphin densities and total transect area completed. Estimates per transect were summed to get total abundance per water level condition. To check for the variation in abundance estimates derived from different surveys, the coefficient of variation was calculated directly from the variance of each seasonal estimate in relation to their mean. Because of the assumption that all groups within the strip width would be detected by either front or rear observers (see Analysis, Detection probability), the fact that there was no group size bias detected, and the entire possible range of dolphins was covered for each survey (except for high water levels), no other components were included in the calculation of CV. Although a considerable variation in group-size was found among different surveys, this is more likely to reflect a biological variation than a size bias related to detection probability. Therefore, instead of calculating the variance of numbers of sightings and group sizes, the CV was directly applied to the abundance estimates. The estimates of the high water level survey were excluded because several transects were not completed. In addition, CVs were calculated per habitat segment, i.e. for the middle-river segment and two tributaries per water level condition to check for the variation of sighting rates among different transects (see formula below). Of the other habitat segments only one transect was completed per water level condition and these segments represented secondary habitat, in which only during specific seasons dolphins were sighted. Therefore, no CVs for seasonal abundance estimate were calculated, but the CV for the middle- river segment may be used as an indication of seasonal variation. Lastly, CVs were calculated for different river segments for the mean abundance estimates of surveys that were both conducted at medium water levels in 1999 and 2000:

46 Abundance estimation of freshwater Irrawaddy dolphins

= g.n R i L

= R i Di Wi

N = ∑ (Di. Ai)

(r − R )2 S (R ) = ∑ j i i − (x j 1)

S.100 CV = R i

Where Ri = mean sighting rate per river segment; r j = mean sighting rate per transect

i = river segment; j = transect

n = number of sightings; L = length of transect completed

D = mean dolphin density; W = mean strip width

N = total abundance within survey area; A = total transect area

S = standard deviation; xj = number of transects completed

CV = coefficient of variation

47 Chapter 4

All sightings are included in the analysis of sighting rates, density- and abundance estimates based on density sampling techniques, except for double counts within one transect and off effort sightings. For direct counts, double sightings on different transects per one-way survey were excluded. In case uncertainty existed about whether two groups consisted of the same dolphins, a non-conservative approach was chosen and these numbers were added in total count. The sightings made by the rear observer are included in total abundance estimate calculations of both survey methods. The percentage of sightings made by front and rear observers are presented in Table 2. Sightings made in one tributary of the upper river segment involved one group of 5 dolphins whose movements were restricted in an area of c. 1 km by two rapids. Sightings made during medium- and high water levels in 1999 are off-effort sightings by other persons than the survey team. The survey team was not able to move upstream of the rapids because of heavy currents due to recent rainfall. However, according to different people in this area, the dolphins have moved upstream of the rapids since October 1998 during a big flood. Because of the overall low sample size these sightings are included in the abundance estimates and because they were confirmed during the next surveys. Correction factors to account for undetected dolphins have been left out because there is a lack of a detailed dive time study. Therefore, it is tentatively assumed that all dolphins will be sighted by front or rear observers within the primary dolphin habitat (middle river-segment, mean width = 200 m, SD = 54), upper river segment (mean width = 161 m, SD = 48), and tributaries (max width of 150 m). Because linear sighting distances only start decreasing after 400 m with 100%, and the survey boat always travelled in the central part of the river, these sighting distances are within the above-mentioned width ranges (see Results). Linear sighting distances of rear sightings and of sightings made in narrow tributaries with many river bends where the average distance between two bends < 400 m were excluded from analysis. Sighting distances of dolphins in river bends are most likely to be restricted as maximal sighting distance is dependent on the distance of two river bends, whereas the sighting distances made by the rear observer may be influenced by the boat’s engine while passing by.

Detection probability Sighting probability was investigated for the following variables: 1) Firstly, linear sighting distances were plotted against the number of sightings made (Figure 3) and tested with chi-squared statistics to investigate if there are significant differences in detection probability of dolphins within the strip width. 2) Relative perpendicular sighting distances were expressed in percentages over three categories in relation to the number of sightings. 3) In addition, the correlation between linear sighting distances and group sizes was investigated and the correlation was measured with the

48 Abundance estimation of freshwater Irrawaddy dolphins

coefficient of determination (r2) (Figure 4). 4) Group dive times were plotted against group size and the Spearman Rank correlation coefficient (rs) was calculated (Figure 5). I preferred to calculate relative perpendicular sighting distances (PSD) because of biases related to the calculation of absolute PSD such as variation in river width between different river segments, and the fact that the vessel cannot maintain a straight course in river bends, leading to biases in calculation of PSDs, whereas many sightings are associated with river bends. In addition, the dolphins restricted and preferred distribution along the width of the river causes both relative and absolute PSD to be of little value to define strip widths as they do not reflect observers’ sighting abilities. Therefore, I did not calculate the probability density function at zero perpendicular distance f(0).

RESULTS Density and abundance estimates

Total search effort in the Mahakam River amounted to 4260 km (397 hours). Actual sightings in the main river segment were confined between Muara Kaman (c.180 km upstream) and Muara Benangak (c.375 km upstream) including tributary Belayan (1 km upstream), tributary Kedang Pahu (max. 80 km upstream), tributary Ratah (480 km upstream main river and 20 km upstream the tributary past a rapid) lake effluent Pela and Lake Semayang (Figure 1). However, depending on water level conditions, the dolphins may move as far downstream in the main river until Loa Kulu (80 km upstream of mouth), whereas their uppermost distribution is limited by the high rapids past Long Bagun (560 km upstream of mouth). Sighting rates for each transect and river segment in which dolphins were sighted are in Table 1. Dolphins were sighted in 6 different habitat segments: middle-river segment (MR, mean width = 200 m, SD = 54); narrow middle-river tributary connected with confluence area with highest dolphin densities (MRT1.1, mean width = 43 m, SD = 13); middle-river tributary in swamp lake area (MRT2.1, mean width = 81m, SD = 13); very narrow upper segment (MRT1.2, mean width = 34 m, SD = 14) of the middle river tributary (MRT1.1), which falls dry in dry season; upper-river tributary with rapids and rock bottom substrate (URT1, mean width = 75 m, SD = 11); Lake Semayang, surrounded by freshwater swamp forest habitat (LS). Mean sighting rates for medium water levels in 1999 and 2000 are nearly similar in the MR segment (0.092 dolphins/ km and 0.096 dolphins/ km with CVs of 25% and 49%). The maximum mean sighting rate for the MR segment was recorded at low water levels (0.142 dolphins/ km, CV = 33%), whereas lowest mean sighting rate in this segment was recorded at high water levels (0.035 dolphins/ km, CV = 51%), indicating that dolphins have moved upstream in the tributaries. Also, the dolphins’ seasonal movements followed changing water levels and seasonal variations in prey availability.

49 Chapter 4 i N

42.7 2000

1.2 i mean D 35 - - 0.48 19.9 0.20 2.9 0.61 8.6 0.61 8.6 5.23 17.1 0 0 0 0 2.29 5.7 49%

i mean R

- 0.096 0.041 0.123 0.123 0.225 0 0 0.172

mean g = 5.7 dolphins MEDIUM WATER LEVELS mean n area; # = no density calculated i N

’99

i mean D 34.7 7.5 34.7 32 - - - 0.71 26.6 3.5 0.71 26.6 0.55 7.6 0.5 0.83 11.4 1.5 0.76 7.6 1.5 0 0 3 - - 0 1.03 3.8 0 1.6 3.8 1 33% 33% = Muara Pahu – Muara Lawa; MRT 1.1

i mean R

- 0.142 0.110 0.165 0.152 0 - 0.084 0.12

MRT

mean g = 3.8 dolphins LOW WATER LEVELS LOW WATER mean n = Ratah tributary = Muara Ratah – rapids; LS = Lake 1 i

N ’99

i mea n D # 2.6 - 14.3 9 14.3 18 0.18 6.5 7 0.19 2.6 2 0.09 1.3 3 0.26 2.6 2 0 0 0 2.5 2.6 - - - 1 1.1 2.6 1 51% 51% where dolphins were sighted. This table only presents those ng was made. Each transect was replicated for each water level = Kedang Pahu tributary;

i 50 mean R 1.1, 1.2 mea n n HIGH WATER LEVELS mean g = 2.6 dolphins 1 0.05 5.5 2.5 0.035 1 0.038 0.5 0.019 1 0.052 0 0 1 0.086 - - 1 0.08

= mean sighting rate; D = mean density; N = total abundance; ; g = average group size

i represent the means of the replicated surveys. N 0 25.5 34 19.2 6.4 4.8 8 3.2 - 0 3.2 25% ’99

i mean D

i mean R 0 0 0.092 0.2 0.092 0.46 0.069 0.34 0.115 0.46 0.042 0.98 - - 0 0 0.1 1.3 mean strip width; L = transect length; - = no data available because of non-surveyed mean g = 3.2 dolphins mean n MEDIUM WATER LEVELS 0 6 2 1.5 2.5 1 - 0 1 = Belayan tributary = Muara Belayan until Tuana Tuha; URT 2.1 ) Km L ( 52 207 69 69 69 76 30 45 33 ) Km 0.200 0.200 0.200 0.200 0.043 0.034 0.081 0.075 mean W (

/ ) ) (MR) 1.1 1.2 2.1 1 1,2,3 1 2 3 condition and number of sightings in this table transects on which during one or more season at least one sighti WATER LEVEL . Sighting rates, density and abundance estimates for each transect TRANSECT N (count) MR MR MR MR MRT MRT MRT 8 (strip) URT (count) LS - N CV(R N = middle river segment = Muara Kaman – Muara Benangak; MRT

1,2,3

River Main Tributary Lake HABITAT because of unknown strip width; CV = coefficient of variation. per water level; i = habitat stratum; W = Semayang; n = mean number of sightings per replicated transect ; R Chapter 4 Table 1 MR =Muara Lawa – Nyawatan; MRT

Abundance estimation of freshwater Irrawaddy dolphins

Mean sighting rate and mean abundance of the combined medium water level surveys is 0.09 dolphins/ km and 19 dolphins (CV = 35%) in the entire MR segment and 0.134 dolphins/ km and 10 dolphins (CV = 97%) in the MRT1.1 segment. No significant differences in mean abundance of dolphins were found between the average abundance of dolphins per transect in the MR segment (mean width = 200 m) and the transect in the MRT1.1 segment (mean width = 43 m) (χ2 = 0.77, d.f. = 1, P > 0.05). Mean abundance in the URT1.1 segment at medium water levels is 4 dolphins (CV = 40%). Total mean abundance estimate of three completed (medium water levels 1999 and 2000 and low water level 1999) and replicated (up-and downstream) surveys based on density estimates (calculated from strip-transects) and direct counts are both 34 dolphins (with respective CVs of 25% and 5%). Mean group sizes of dolphins observed at medium, high and low water levels in 1999 and medium water levels in 2000 are 3.2 dolphins (median = 3; range = 1-7; SD = 2.1), 2.6 dolphins (median = 1; range = 1-6; SD = 2.3), 3.8 dolphins (median = 3; range = 1-8; SD = 2.3) and 5.7 dolphins (median = 5; range = 3-10; SD = 2.4) respectively.

Detection probability

When calculating the percentages of initial sightings in relation to relative perpendicular sighting distances (position along the width of the river), I found that the number of initial sightings peaked near the shore (45% of total n = 49), but not significantly (χ2 = 2.9, df = 2, P>0.05). The remaining sightings were nearly equally spread over the two other segments, i.e. the centre area of the river (29%) and the area in between centre and shore (26%). On the other hand, the number of sightings (total n = 33) were found to decrease sharply with 100% only after 400m linear sighting distance (Figure 3). No significant variation was found among the sighting distances inside of 400 m (χ2 =5.3, df = 5, P > 0.05). Because the maximum mean river width for one of the transects (MR1) within dolphin distribution area is 238 m (range = 120 m – 400 m, SD = 62 m), there is no apparent bias towards undetected dolphins near the shore, because maximum sighting distances are greater than one-half the survey strip. Therefore, I assumed that the probability of sighting dolphins was uniform throughout the survey trip. Because I found no distinct decrease of sightings in relation to perpendicular sighting distances, linear sighting distances (n = 35) were plotted against group size to see if there is any detection bias for any group size (Figure 4). No significant correlation was found between the two variables (r = 0.132, df = 33, P > 0.05) and only 1.7 % of the variation in group-sizes is accounted for by variation in linear sighting distances (r2 = 0.017). Dolphin group dive data were collected only during 14 sightings. However, results presented in Figure 5 seem to indicate that group dive times are negatively related with group size, i.e. small groups have longer mean group dives per sighting

51 Chapter 4

Detection probability 10

2 No. of sightings

0 10 88 166 244 322 More Linear sighting distances (m)

Figure 3. Histogram showing the frequency of sightings per linear sighting distance category.

Group size bias 500 lin.sight.dist 400 Predicted lin.sight.dist 300

distance 200

Linear sighting 100

0 0246810 Group size

Figure 4. Scatter plot of linear sighting distances and group size indicating probability of any detection bias related to group size.

than large groups (rs = 0.665; P < 0.01; n = 14). Mean of all average group dive times per sighting is 72.0 sec (median = 38.3; SD = 69.2; range = 5-240). Mean time that a group of dolphins is visible per surfacing (time between first dolphin’s surfacing and last dolphin’s diving allowing for maximum interval of 3 sec.) is only 2.5 seconds (2-6 sec). Although a lower detection probability is expected for dolphins with a small group size due to longer dive times, no detection bias was found for any given group size in relation to sighting distance as stated earlier (Figure 4). Additionally, single dolphins were frequently observed: 29% of all on effort sightings (n = 49) constitute single dolphins.

52 Abundance estimation of freshwater Irrawaddy dolphins

Group Dive Times and Group Size 300 250 200 150 100 50 Mean dive time (sec.) 0 02468 Group size

Figure 5. Scatter plot showing a negative relation between group size and mean group dive times.

The percentage of sightings during the four surveys covering the entire dolphin distribution range, made by an observer at the front of the boat using binoculars was on average 63 % and that by a front observer without binocular was 31% (total n = 52). On average, during each survey 6 % of all sightings were missed by the front observers, being observed only by the rear observer (Table 2). During two additional one-way surveys at medium to decreasing water levels conducted in the middle-river segment (MR) whereby three transects were completed, observer efficiency was increased from three to four observers (data not presented in table). During each of these surveys, three sightings were made, all by the front observers.

Table 2. Observer perception bias (% sightings made per observer category); n = number of sightings

OBSERVER/ n FRONT FRONT REAR SURVEY PERIOD OBSERVER OBSERVER OBSERVER + BINOCULAR -BINOCULAR Surveys Feb/ March ‘99 14 50 % 36 % 14 % Surveys May ‘99 8 50 % 38 % 12 % Survey Oct ‘99 15 77 % 23 % 0 % Survey May/June 2000 15 75 % 25 % 0 % Total / Average 52 63% 31% 6 %

DISCUSSION

Two different methods, strip-transects and direct counts, were employed to estimate abundance for the Mahakam dolphin population. In this study, a modified form of strip-transect surveys was used. Instead of determining the effective strip width based

53 Chapter 4

on perpendicular sighting distances, the average entire river width was estimated per segment and used as strip width for density calculation. Two things were evident: 1) Dolphin positions along the width of the river at initial sighting peaked near the shore, (although not significantly) and 2) Linear sighting distances start decreasing slightly after 166 m and number of sightings made at 400 m linear distance have not yet dropped to half the number of sightings at 166 m (62%), but dropped to zero beyond 400 m. Because the maximum river width in the dolphin distribution area is 400 m (with strip width as follows 200 m), it seems reasonable to assume that sighting detection probability is not limited by strip width, but is more likely to be influenced by dolphin availability bias and observer perception bias. Also, river width in the Mahakam does not change much throughout seasons and floods almost only vertically instead of horizontally, in contrast to rivers like the Amazon. Sighting distances and river width estimations are visually estimated and are therefore likely to be biased. However, attempts were made to make distance estimations more standardised among the observers of the survey team and among different survey teams (see Survey methods). When comparing total abundance estimates that are calculated on the basis of density estimates calculated from strip-transects and those estimates based on direct counts, I found that the latter analysis method produced more consistent results for the three completed surveys (medium water levels in 1999 and 2000 and low water levels in 1999). Total mean abundance-estimates, based on density estimates and direct counts, were both 34 dolphins. However, the mean estimate based on density estimates exhibited more variation (CV of 25%), than the mean direct count estimate with associated CV of 5%. The higher variation among abundance estimates based on density estimates may arise from the fact that the abundance estimates for different segments, i.e., main river and tributaries, were added together to derive total abundance, whereas dolphins daily migrate between these areas. This problem does not exist for direct count estimates as these daily migrations are taken into account and double counts avoided (see Survey methods). No CVs of total abundance estimates per season were calculated because of the fact that segments other than middle-river segment consisted only of one transect. However, a seasonal CV for abundance was given in the middle-river segment for three completed transects. The highest sighting rate for the middle- river segment (0.142 dolphins/ km) was recorded at low water levels, indicating that dolphins are congregating in deeper waters of the main river. The lowest sighting rate (0.035 dolphins/ km) was recorded at high water levels, indicating that dolphins have moved upstream and into the tributaries. This movement pattern was also confirmed through interviews with local fishermen and coincides with fish-migration at first flooding. At high water levels, only two sightings were recorded in tributaries. However, this is probably not a representative figure, because three other middle-segment tributaries and the narrow upstream areas beyond tributary 1.2 (Kedang Pahu) were not surveyed. During the low water survey no dolphins were found to occur in the upper river segment,

54 Abundance estimation of freshwater Irrawaddy dolphins

although during a prolonged dry season (more than 3 months) dolphins are said to move to the upper river segment as far as Long Bagun (560 km upstream), as currents are less strong than during the other water conditions in this segment. However, the absence of observations in the upper river segment is not representative of the dry season’s low water levels, because of the short duration of the dry season. Also water levels had for a week increased rapidly in the upper segment, due to heavy rainfall. However, data were not included in the high water level category, as this category became of a prolonged period of high water levels and did not extend to the other river segments. The highest sighting rate recorded during low water levels for the middle Mahakam segment (0.14 dolphins/ km), is similar to sighting rates recorded for Irrawaddy dolphins in a segment of the Ayeyarwady River between Bhamo and Mandalay (0.16 dolphins/ km) (Smith & Hobbs, this volume). Average sighting rates during medium water levels in 1999 and 2000 were 0.09 dolphins/ km and similar to encounter rates recorded during a preliminary survey in 1997 in the same river segment and season (0.06 dolphins/ km) (Kreb, 1999). Compared to other freshwater dolphin species, rates are much lower than those recorded for Inia geoffrensis and Sotalia fluviatilis in segments of main channel of Amazon River (0.43 – 0.60 and 0.41 dolphins/ km, respectively) (Vidal et al., 1997; Martin & da Silva, 2000), and those recorded for Platanista gangetica, varying from 0.2 – 1.36 dolphins/ km (Smith, 2000; Smith et al., 2001). Total abundance estimates in this study of 35-42 dolphins are of the same order of magnitude as those for Lipotes vexillifer, of which the ‘best guess’ of current population size is a few tens of animals (Reeves et al., 2000). No significant differences were found between the mean abundance of dolphins at medium water levels (when there are no seasonal dolphin migrations) in two different transects (main river and tributary) within the primary dolphin habitat of different mean width (200 m and 43 m) (χ2 =0.77, d.f. = 1, P > 0.05). However, when comparing densities, a conclusion may be drawn for example that dolphin densities are higher in a narrow river segment than in a wider river segment, whereas sighting rates and abundance are nearly similar in the two segments. For that reason these densities should not be used for comparison between different river segments or with other studies. Instead, sighting rates and direct counts give a much more useful comparison. The following data are in favour of the reliability of the abundance estimates presented here: 1) The dolphin availability bias and observer perception bias seem low, and missed sightings by the front observers are partially anticipated for by using a rear observer. Moreover, in spite of a lower detection probability of dolphins with a small group size due to longer dive times, single dolphins were frequently observed (29% of all on effort sightings (n = 49) constitute single dolphins). In addition, no correlation was found between group size and linear sighting distance and number of sightings only drop sharply beyond 400 m. 2) There seems to be no bias towards undetected dolphins near the shore because most sightings (78%) were made at linear

55 Chapter 4

sighting distances (≥ 166 m) that cover the distance from centre to shore in primary dolphin habitat (mean distance is 200 m). In addition, initial dolphin sightings even peaked near the shore. 3) There is a high similarity of direct count abundance estimates during different surveys (CV =5%). However, with regard to direct counts a potential bias exists with regard to the estimation of best group size estimates. For this reason, absolute counts in the true sense of the word are not possible. The low number of observers may cause an underestimation of numbers and the fact that rear observers were only present in and after river bends and confluence areas, assuming that most dolphins in straight river stretches would be sighted by the front observers. On the other hand, the detection probability analyses plus the two repeated surveys in 2000 and 2001 with increased numbers of observers in the middle river segment suggest that this factor is not likely to influence the estimates significantly. However, the number of sightings was low during these last surveys as only three transects were covered and not the entire river stretch. So, the surveys with an added observer cannot really be compared in terms of the percentage of sightings that are missed by front observers and observed by rear observer due to unequal sample size. Recommendations for future studies are to conduct at least a yearly extensive and intensive monitoring survey during the dry season, covering the entire dolphin distribution range with a standard number of observers, i.e., two front observers, one rear observer and one observer at rest for 30 minutes in between 1,5 hours observing bouts. Photo-identification may also be a valuable tool to determine total abundance. Unfortunately, data collection and analyses are not yet complete at time of writing. Also, a detailed dive time study is required to address the dolphin availability bias more properly and the need to include a correction factor. In conclusion, I would say that for assessing abundance of the dolphin population in the Mahakam, both density- sampling techniques and direct-counts seem appropriate and yield numbers of the same order of magnitude. Nevertheless, the direct counts of different surveys exhibit less variation. A simple direct count also was suggested as the most appropriate method for assessing populations of obligate river dolphins (Smith & Reeves, 2000). However, recommendations for future studies in the Mahakam also include to develop a modified density sampling technique that is appropriate to the river conditions and takes into account the dolphins daily movements between the main river and tributaries.

ACKNOWLEDGEMENTS

I am grateful for the financial support that enabled this long term project to continue and also allowed me to participate in congress meetings to exchange and receive viable information and input. For this support I would first like to thank Ocean Park Conservation Foundation Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Amsterdamse Universiteits Vereniging; Netherlands Program

56 Abundance estimation of freshwater Irrawaddy dolphins

International Nature Management (PIN/ KNIP) of Ministry of Agriculture, Nature Management and Fisheries; Gibbon Foundation; Stichting Fonds Doctor Catharine van Tussenbroek; World Wildlife Fund For Nature (Netherlands, US). I am also grateful to the following non-financial sponsoring institutions: Lembaga Ilmu Pengetahuan Indonesia (Indonesian Institute of Sciences), Balai Konservasi Sumber Daya Alam Kal-Tim (East Kalimantan nature conservation authorities), Directorate Jenderal Perlingdungan Konservasi Alam (General Directorate of Protection and Conservation of Nature), University of Amsterdam, Zoological Museum of Amsterdam, Universitas Mulawarman Samarinda. My special thanks goes to my assistants for their enthusiasm and great efforts during the fieldwork in the Mahakam: Hardy, M. Syafrudin, A. Chaironi, Zainuddin, Arman, M. Syoim, Budiono, Bambang, Sonaji, Syahrani, K.R. Damayanti, Marzuki and Yusri. In addition, I would like to thank the following persons in particular: P. J.H. van Bree, F. R. Schram, A. Arrifien Bratawinata, A. M. Rachmat, Padmo Wiyoso, I. Syarief, S., V. Nijman, A.Ø. Mooers, G. Fredrikson, T. Prins, T. Dunselman, I. Beasley, R.R. Reeves, all boatsmen, fishermen, villagers and colleagues that shared their information, and all those who showed their support in the project. Finally, I would like to thank B. Würsig, P. Rudolph, B.D Smith and T.A. Jefferson for their rigorous and helpful review of earlier versions of this manuscript.

REFERENCES

Hardjasasmita, H.A. 1978. Studi pembinaan habitat dan populasi pesut. Direktorat Perlindungan dan Pengawetan Alam, Bogor. Hilton-Taylor, C. 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, U.K. Kamminga, C., Wiersma, H. & Dudok van Heel, W.H. 1983. Investigations on cetacean sonar VI. Sonar sounds in Orcaella brevirostris of the Mahakam River, East Kalimantan, Indonesia; first descriptions of acoustic behaviour. Aquat. Mamm. 10: 83-94. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Z. Saug. 64: 54- 58. Leatherwood, J.S. 1996. Distributional ecology and conservation status of river dolphins (Inia geoffrensis and Sotalia fluviatilis) in portions of the Peruvian Amazon. Ph.D. Thesis, Texas A & M University, Texas, USA. 232 pp. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press. 152 pp. Martin, A.R. & Da Silva, V.M.F. 2000. Aspects of status of the Boto Inia geoffrensis in the central Brazilian Amazon. Paper, SC/52/SM15, presented at 52nd Annual Meeting of the International Whaling Commission, Small Cetacean Sub-committee.

57 Chapter 4

Priyono, A. 1994. A study on the habitat of Pesut (Orcaella brevirostris Gray, 1866) in Semayang-Melintang Lakes. Media Konservasi 4: 53-60. Reeves, R.R., Smith, B.D. & Kasuya, T. (eds), 2000. Biology and conservation of freshwater cetaceans in Asia. Occasional Paper of the IUCN Species Survival Commission, 23, IUCN, Gland, Switzerland. 152 pp. Smith, B.D., 2000. Review of river dolphins, genus Platanista, in the South Asian subcontinent. Paper, SC/52/SM4, presented at 52nd Annual Meeting of the International Whaling Commission, Small Cetacean Sub-committee. Smith, B.D., Ahmed, B., Ali, M.E. & G. Braulik, 2001. Status of the Ganges River dolphin or shushuk Platanista gangetica in Kaptai Lake and the southern rivers of Bangladesh. Oryx, 35: 61-72. Smith, B.D. & Hobbs, L. 2002. Status of Irrawaddy dolphins Orcaella brevirostris in the upper reaches of the Ayeyarwady River, Myanmar. Raffles Bull. Zool., Suppl. 10: 67-73. Smith, B.D. & Reeves, R.R. 2000. Survey methods for population assessment of Asian river dolphins. In: Biology and conservation of freshwater cetaceans in Asia. Occasional Paper of the IUCN Species Survival Commission 23: 97-115. IUCN, Gland, Switzerland. Stacey, P.J. & Arnold, P.W. 1999. Orcaella brevirostris. Mammal. Spec. 616: 1-8. Vidal, O., Barlow, J., Hurtado, L.A., Torre, J., Cendon, P. & Ojeda, Z. 1997. Distribution and abundance of the Amazon River dolphin (Inia geoffrensis) and the Tucuxi (Sotalia fluviatilis) in the upper Amazon River. Mar. Mamm. Sci. 13: 427- 445.

58 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

CHAPTER 5

Abundance of freshwater Irrawaddy dolphins in the Mahakam River in East Kalimantan, Indonesia, based on mark- recapture analysis of photo-identified individuals

In press: Journal of Cetacean Research and Management, 2004

One photo-identified individual PM 34 with a distinctively shaped dorsal fin. During early 1999 and mid 2002, a total of 59 individuals were identified.

59 Chapter 5

ABSTRACT

From February 1999 until August 2002 c. 9000 km (840 hours) of search effort and 549 hours of observation on Irrawaddy dolphins (Orcaella brevirostris) were conducted by boat in the Mahakam River in East Kalimantan, Indonesia. Intended goal was to generate an estimate of total population size essential for conservation and management of this threatened freshwater dolphin population. An abundance estimate based on mark-recapture analysis of individuals photographed during separate surveys is presented here. Two different analysis methods, i.e. Petersen and Jolly-Seber methods were employed and compared with each other and with earlier estimates derived from strip-transect analysis and direct counts. These comparisons serve to evaluate the biases of each method and assess the reliability of the abundance estimates. The feasibility of video-identification is also assessed. Total population size calculated by Petersen and Jolly-Seber mark-recapture analysis, was estimated to be 55 (95% CL = 44-76; CV=6%) and 48 individuals (95% CL = 33-63; CV=15%). Estimates based on strip-transect and direct count analysis for one sampling period, which was also included in the mark-recapture analysis, were within the confidence limits of the Jolly-Seber estimate (Ncount = 35 and Nstrip = 43). Calculated potential maximum biases appeared to be small, i.e. 2% of N for Petersen and 10% of N for Jolly-Seber method, which is lower than the associated CVs. Also, a high re-sight probability was calculated for both methods varying between 65% and 67%. Video images were considered a valuable, supplementary tool to still photography in the identification of individual dolphins in this study. For future monitoring of trends in abundance using mark/ re-capture analyses, a time interval is recommended between the two sampling periods that is short enough to minimise the introduction of errors due to gains and losses. Also, survey area coverage during photo-identification should be similar to avoid violation of the assumption of equal capture probabilities. The alarmingly low abundance estimates presented here call for immediate and strong action to preserve Indonesia’s only known freshwater dolphin population.

RINGKASAN

Dari Februari 1999 sampai Agustus 2002 kurang lebih 9000 km (840 jam) penelitian dan 549 jam pengamatan lumba-lumba Irrawaddy (Orcaella brevirostris) dilakukan dengan menggunakan kapal di Sungai Mahakam, Kalimantan Timur, Indonesia. Tujuan utamanya untuk menghasilkan suatu perkiraaan dari jumlah keseluruhan populasi yang digunakan sebagai bahan untuk perlindungan dan pengelolaan lumba- lumba air tawar dari kepunahan. Perkiraan keadaan yang berlebihan didasarkan pada

60 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam analisa penandaan-ulang dari potret individu selama survey terpisah dilaksanakan. Dua metode analisa, yaitu metode Petersen dan Jolly Seber digunakan dan dibandingkan satu dan lainnya dan dengan perkiraan awal yang diperoleh dari analisa strip-transect dan perhitungan langsung. Perbandingan digunakan untuk evaluasi penyimpangan dari masing-masing metode dan mendapatkan taksiran yang dapat dipercaya dari jumlahnya. Kemungkinan identifikasi dengan video juga dipergunakan. Perhitungan ukuran populasi total dengan metode analisa penangkapan kembali Petersen dan Jolly Seber, telah diperkirakan menjadi 55 (95 % CL = 44-76; CV= 6%) dan 48 ekor (95% CL=33-63; CV=15%). Perkiraan didasarkan pada metode strip- transect dan penghitungan langsung untuk periode pengambilan contoh, dimana termasuk juga dalam analisa penandaan dengan penangkapan kembali, berada dalam batas keyakinan dari perkiraan Jolly-Seber (Ncount=35 and Nstrip=43). Perhitungan penyimpangan potensial maksimum kelihatan lebih kecil, yaitu 2% dari N untuk Petersen dan 10% N untuk metode Jolly-Seber yang mana lebih rendah dari CV yang terkait. Juga suatu kemungkinan pengamatan kembali yang diperhitungkan untuk dua metode adalah berbeda antara 65% dan 67%. Gambar-gambar video dianggap sebagai hal berharga, sebagai alat bantu gambar potret dalam mengidentifikasi individu lumba- lumba pada penelitian ini. Untuk pengamatan yang akan datang dari kecenderungan dalam jumlah menggunakan metode penandaan-penangkapan kembali, jarak waktu yang disarankan antara dua periode pengambilan contoh adalah cukup pendek untuk mengurangi kesalahan awal berkaitan dengan pencapaian dan kehilangan. Juga cakupan daerah survei selama identifikasi foto harus sama untuk menghindari kesalahan dari kemungkinan penangkapan yang sama. Rendahnya tingkat perkiraan jumlah yang disajikan di sini adalah untuk dapat dengan segera diambil tindakan yang tegas dalam melestarikan satu-satunya populasi lumba-lumba air tawar di Indonesia yang telah diketahui.

INTRODUCTION

Since 1970, photo-identification studies have proven to be a valuable tool in revealing aspects of population dynamics, social organisation, distribution and movement patterns for many species of cetaceans (Whitehead et al., 2000). The technique involves collecting and cataloguing photographs of dolphins with distinctive marks on the dorsal fins, flukes and bodies that allow for identification of individuals. Photo- identification, when seriously attempted, was found feasible for every cetacean species that is in possession of a distinct dorsal fin (Mann, 2000). But the ease of getting good photo-identification results highly varies among species depending on uniqueness of the marks and behaviour of the species. Easily identifiable cetaceans with nearly complete photo-identification databases for certain populations include killer whales, 2000). For most other species, e.g., hump-backed dolphins, Sousa chinensis (Jefferson

61 Chapter 5

Orcinus orca (Baird, 2000) and humpback whales, Megaptera novaeangliae (Clapham, and Leatherwood, 1997; Jefferson, 2000); Pacific white-sided dolphins, Lagenorhynchus obliquidens (Morton, 2001) and northern bottlenose whales, Hyperoodon ampullatus (Gowans and Whitehead, 2001) only a proportion of the population can be reliably identified because not all fins are characteristically marked. Another factor limiting identification is due to the elusive behaviour of some species of dolphins. Photo- identification of Irrawaddy dolphins, Orcaella brevirostris, commonly described as an elusive species (Lloze, 1973; Dhandapani, 1992; Kreb, 1999) required greater effort, but was shown to be feasible for coastal populations in Australia (Parra and Corkeron, 2001). In addition, freshwater populations of Irrawaddy dolphins that are known to occur in only three major river systems, i.e. the Mahakam River in Kalimantan, the Mekong River in Vietnam, Laos and Cambodia and the Ayeyarwady River in Myanmar (Burma) were reported to be visually identifiable, but photo-identification efforts until now were more or less incidental (Stacey, 1996; Smith, 1997; Kreb, 1999). Freshwater dolphin populations in many cases live in a closed system and have no exchange with coastal populations. Thus, photo-identification and subsequent mark- recapture analysis to determine total population size might be feasible. This study reports on photo-identification of a population of Irrawaddy dolphins in the Mahakam River, Indonesia and is the first attempt to provide a catalogue in which most individuals of an entire freshwater Irrawaddy dolphin population are identified. The Irrawaddy Dolphin Orcaella brevirostris is a facultative freshwater dolphin, occurring both in shallow coastal waters and large river systems in tropical South East Asia and subtropical India (Stacey and Arnold, 1999). Irrawaddy dolphins in Indonesia occur along several coastlines and in one river in East Kalimantan, the Mahakam, where they are referred to as pesut (Kreb, 1999). The species is fully protected by law in Indonesia since 1990 and is adopted as a symbol of East Kalimantan Province. Their IUCN status was raised from ‘Data Deficient’ to ‘Critically Endangered’ based on data related to abundance collected from 1999 until 2000 (Kreb, 2002; Hilton- Taylor, 2000). The objectives are: to present an estimate of total population size based on photo-identification of individual dolphins by using different mark-recapture methods and to compare these with earlier estimates of abundance from strip-transects and direct counts (Kreb, 2002). In addition, the feasibility of digital video recordings as a tool to identify dolphins is evaluated. This photo-identification study is part of a long- term conservation and research project, begun in 1999 to provide a framework to protect the freshwater Irrawaddy dolphin population in the Mahakam River in East Kalimantan, Indonesia.

SURVEY METHODS

During the study period from February 1999 through August 2002, 12 surveys were conducted. Six extensive monitoring surveys (mean duration 20 days; SD= 4 days) covered the entire distribution range and six (mean duration 12 days; SD = 3 days)

62 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

were conducted in areas of high dolphin density (Figure 1). Extensive surveys were conducted with 12-16 m long motorised vessels (between 12 and 21 hp), travelling at an average speed of 10 km/ hr. The average observation time and photographic effort during the extensive monitoring surveys was one hour per sighting. Most intensive monitoring surveys involved attempts to follow one group for an entire day, with daily alternation of groups and using a small, motorised canoe with 5hp outboard engine. Photographic effort was spread out over the observation time (average duration 7 hours; range 1.5 -13 hours). Upon sighting, the group was approached to a minimum distance of 30m in order to take photographs and video images. We always tried to take these photos from similar angles, i.e. perpendicularly to the dolphins’ dorsal fin region. In addition, identification marks were recorded on datasheets. For each sighting, the duration, location, group behaviour, group size, group composition and environmental data were collected. Four age classes were defined: i) “neonates” were individuals of less than 1/2 the average length of an adult, which spent all their time in close proximity to an adult and exhibited an awkward manner of swimming and surfacing; ii) “calves” were animals between 1/2 and 3/4 the average length of an adult and which still spent most of their time in close proximity to an adult; iii) “juveniles” were animals of 3/4 the average length of an adult and which swam independently; iv) “ adults were individuals larger than an estimated 2 meters in length. Photographs were made by the author using a Canon EOS 650 camera body with a Sigma 300mm/ f4.0 lens, occasionally attaching a 1.4 teleconverter, effectively making it a 420mm/ f5.6 lens. Manual focus was always used with shutter speeds of 1/250 to 1/1500 of a second. Some 75% of the photo-id images were made with slide films using Sensia Fujichrome 100 ISO and another 25% were made using Fuji Superia 200 ISO print-films. It was attempted to always photograph every individual within the group irrespective of whether they at first sight appeared to have distinct dorsal fin markings or not. Photographs were generally taken perpendicularly to the dolphins’ dorsal fin region. Additionally, drawings of dorsal fins (made by aid of binoculars) were made by observers who did not take photographs. Dolphin age classes were also noted for each drawing. Direct observations and drawings were matched with a field photo-identification catalogue and assigned an existing or new identification code. One field-assistant was assigned to the task of making simultaneous video footage using a Sony VX 1000 digital camcorder with 10x optical and 20x digital zoom. Nearly always only the10x optical zoom was employed or better image quality. The auto- focus option was usually preferred since manually focusing proved more difficult with the camcorder than with the photo-camera. Information on the number of dead dolphins during the entire study period and in particular between the two sampling periods, was obtained through our own observations and from local, reliable reporters.

63 Chapter 5

Kedang Rantau Muara Kaman Loa Kulu Kota Bangun Kedang Kepala Batuq Semayang empang J Melintang Tepian Ulak Muara Pahu bution area, b) areas of high dolphin density and c) coastal Irrawaddy dolphin area. Muara Jelau Bohoq 64 Rambayan Muyub Ulu Damai Datah Bilang Kedang Pahu Muara Benangak

Long Bagun Ratah Figure 1. Study area with a) total dolphin distri Chapter 5 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

ANALYSIS

Photographs and slides were selected by aid of an 8x loupe for their good image quality, i.e. focus, glare, photographic angle, dorsal fin size coverage in image and catalogued on basis of identifiable features. Distinctive features noted included notches, scars and cuts on the dorsal fin and distinct fin shapes. Pigmentation patterns were only secondarily considered if they could be linked to a distinct fin shape. Pigment spots or areas do not occur symmetrically on both sides of the dorsal fin. In addition, it was found that pigmentation patterns on the bodies of dolphins and likely therefore on dorsal fins were not stable during the study period. Each photograph in the photo-identification catalogue corresponded to an identified individual and held information on the date, time and location at which the picture was taken as well as data on group size and composition. Photographs with distinctive features such as scars, cuts and humps on the dolphin’s bodies were also selected, but catalogued under another identification code. Photographs with distinctive body features alone were only used for mark-recapture analysis if they could be linked to an individual, which was already identified based on its dorsal fin. Identifications that were obtained through direct observation and drawings were kept in a separate database file than the photo-identified dolphins. These identifications were not used for the mark-recapture analysis. For analysis of recorded video-images, each image of a dorsal fin was played in slow-motion and paused. Similar to the photo-identification analysis, only images of good quality were selected. These good images were then compared with individuals from the photo-identification catalogue and these were given an identification code and put into a video-identification catalogue together with related sighting data. Two estimates of total population size (N) were calculated based on two different mark-recapture analysis methods of photo-identification data. Only sampling periods with extensive area coverage were selected, which were suitable for estimating total population size. The first estimate utilized the Petersen method for closed populations, involving one session of catching and marking and one recapture session. Bailey’s modified estimator was applied for sampling with replacement (Equation 1.1) (Hammond, 1986). Sample periods May/ June 2000 and August 2001 were chosen because the photographic effort (i.e. area coverage) was similar in those periods (Table 1). The second method to estimate total abundance was the Jolly-Seber method for open populations, allowing for gains and losses within the sampling periods. Also, capture histories of each identifiable individual were needed since the method requires both knowledge of the number of animals in each sample that were previously marked and information on the most recent previous sample in which each of them was last trapped. The number of marked individuals in four sampling periods, i.e., October 1999, May/ June 2000, January/ February 2001 and August 2001, with extensive area coverage, were higher than the minimum sample size of 10 marked individuals recommended to overcome imprecision of abundance estimates (Table 1) (Sutherland, 1996). Prior to the calculation of an abundance estimate, a goodness-of-fit test was

65 Chapter 5

applied (Sutherland, 1996) to test if animals differed in capture-probabilities, which may cause a serious bias of the estimate. After testing, three sampling periods were chosen to be appropriate for abundance estimation (see results). According to the Jolly-Seber method, no estimates of abundance can be calculated for the first and last sampling period and thus only one estimate is derived from the second sampling period (Equation 2.1). For this last method, it was also possible to calculate the proportion of the population surviving (Φ) from the 1st to the 2nd sampling occasion (Equation 2.3). A correction factor was applied to the population estimates of both methods to correct for the proportion of dolphins that are not identifiable (p) (Jefferson & Leatherwood, 1997). These were neonates and calves, which could not be captured properly on photo because their mothers protect them away from the boat and from a good camera angle and because calves often surface very suddenly (high arch dives). The averages of the proportion of neonates and calves encountered during two (Petersen) and three (Jolly-Seber) sampling periods are 10% and 8% respectively, which represent the proportion unidentifiable dolphins (p). For the Petersen method binomial 95% confidence intervals were calculated for the fraction of marked individuals (m2 + 1)/ (n2 + 1), which were then applied to the formula in Equation 1.1. to obtain the 95% confidence limits for population size (Krebs, 1999). Jolly-Seber confidence limits were calculated using the formula provided by Manly (1984). Coefficients of variation were calculated for both methods according to the formulas in Equation 1.2 and 2.4. Estimated re-sight probabilities for the Petersen estimator are given by m2/ n2 and p2 = m2/ n1 and for Jolly-Seber by ni/ Ni, in which Ni is (only here) the uncorrected abundance estimate for proportion of identifiable dolphins.

( ) Eqn 1.1 n2 + 1 (Petersen method) N = n 1 ( ) ( − ) m 2 + 1 1 p

2 − 1 n ( n + 1 ) ( n − m ) v a r ( 1 − p ) Eqn 1.2 C V ( N ) = N 1 2 2 2 + 2 2 ( ) ( ) ( − ) m 2 + 1 m 2 + 2 1 p

where n1 = number identified on the first occasion n2 = total number identified on the second occasion

m2 = number of identified dolphins found on the second occasion p = proportion of unidentifiable individuals

( ) M i n i + 1 Eqn 2.1 N i = (Jolly-Seber method) ( m i + 1 ) ( 1 − p )

66 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

( ) m i + R i + 1 z i Eqn 2.2 Mi = ( r i + 1 )

Eqn 2.3 Φ M i + 1 i = ( M i − m i + R i )

Eqn 2.4

− ( − ) ⎛ M i m i + R i + 1 ⎞ ⎛ 1 1 ⎞ 1 1 v a r 1 p ⎝ ⎠ ⎝ − ⎠ + − + ( ) ( ) 2 M i + 1 r i + 1 R i + 1 m i + 1 n i + 1 ( 1 − p ) CV ( N i ) = x ⋅ i − ⎛ 0.5 3 n i ⎞ lo g e N i + 0.5 lo g e ⎝ ⎠ 8 N i

Where Ni = population size at the time of the ith sample Mi = number of marked animals in the population when the ith sample is taken (excluding animals newly marked in the ith sample) ni = total number of animals caught in the ith sample Ri = number of animals that are released after the ith sample mi = number of animals in the ith sample that carry marks from previous captures zi = number of animals caught both before and after the ith sample but not in the ith sample itself ri = number of animals that were released from the ith and were subsequently recaptured xi = number of samples

Finally, maximum biases that may affect population size estimates for each method were calculated. A maximum bias using Petersons method, which assumes no losses, was calculated by adding the number of dead dolphins (= 3) in between the two sampling periods, to the number of ‘recaptured’ animals during the second sampling period (m2bias= m2 + 3). This number was also added to the total number caught on the second occasion (n2bias= n2 + 3). When applying this bias one assumes that these dolphins would have been ‘marked’ during the first session and also assumes that they would have been ‘recaptured’ if they hadn’t died. A maximum bias using Jolly-Seber method was related to the fact that one area was not surveyed during the second sampling period of the three sampling periods in total. This area, which is an area in between two rapids and home to a group of six dolphins, was surveyed only during the first and last sampling period. Two and three new individuals were marked during the first and last sampling period, respectively, without any recaptures. The largest deviation from the abundance estimate would apply for a situation in which we assume that this area would have been surveyed during the second sampling period, which four new individuals would be captured and marked and three of which would be recaptured during the third sampling period.

67 Chapter 5

This maximum deviation of the estimate is calculated following equation 2 above by adding three individuals to r2 (number of marked dolphins in the 2nd sample, which were recaptured in the 3rd sample) and four individuals to n2 and R2 (total number caught and released in the 2nd sample). Variable z2 is not affected by the missing survey effort during the second sampling period because the individuals marked in that area were not similar during the first and last sampling period. Conclusively, this maximum bias holds only if the following assumptions are true: None of the two individuals marked during the first sampling period would be recaptured if the ‘missed’ area was surveyed during the second sampling period. Four individuals would be marked during the second sampling period so that r2bias = r2 + 3, n2bias = n2 + 4 and R2bias = R2 + 4.

RESULTS

Estimates of abundance based on photo-identification mark-recapture analysis

During the entire study period from February 1999 until August 2002, a total of 2074 photographs were made during 83 days of which 1499 (partially) portrayed dolphins and 558 (27%) failed, showing merely circles in the water (Table 1). Of the dolphin photographs, 753 photographs (50%) were selected for photo-identification because of good image quality. Some 728 photographs showed identifiable features on dorsal fins, sometimes in combination with other characteristic traits on the dolphins’ bodies, producing an average of almost 9 identifiable dorsal fin photographs per day. An additional number of 25 photographs only showed identifiable features on the dolphins’ bodies. As such, a total of 59 individual dolphins were catalogued based on dorsal fin identification. Four individuals are shown in Plate 1. Within the four initially chosen sampling periods for the Jolly-Seber method, animals appeared to differ significantly in capture-probabilities (G = 10.06; d.f. = 2; P < 0.01), meaning that the underlying assumptions (see discussion) of the method were violated. The bias was consequently rendered insignificant by only using sampling periods, which include a high proportion (i.e. over 50%) of the population. Therefore, sampling period October 1999 was removed from analysis, which included only 31% of the Petersen population estimate. Another G-test for the remaining periods revealed that this time no assumptions were violated (G = 1.8; d.f. = 1; P = 0.17). The number of dolphins identified on photograph for each sampling period (ni) are presented in Table 1. For both Petersen method applies that the number of dolphins that were identified in the first period (May/ June 2000) and recaptured on photograph during the second period (m2) (August 2001) is 22 individuals.

68 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

Table 1. Photo-identification success rate and discovery rate of new individuals.

Year Survey period Survey area No. dolphin No. identified No. different No. of new coverage photographs dorsal fins individuals (ni) individuals 1999 Fe/ Ma E 25 3 2 2 Ap/ May E 25 7 5 5 Oc E 49 28 16 13 2000 May/Jun E 206 90 33 21 Au I 157 83 24 4 Nov I 65 23 16 1 2001 Ja/Fe E 175 82 29 6 Jun/Jul I 267 127 37 1 Au E 178 90 34 3 Oc/No I 89 36 23 1 2002 Ap I 181 102 28 1 Au I 82 54 23 1 Total 12 periods 1499 728 59

E = Extensive monitoring survey in entire dolphin distribution area; I = Intensive monitoring survey in high dolphin density areas.

Plate 1. (left above:) PM 2; (right above:) PM 1; (Left below:) PM 8; (Right below:) PM 3

For Jolly-Seber method applies that m2 is 14 individuals (using periods May/ June 2000 and January/ February 2001). The estimated re-sight probabilities for Petersen method are 65% and 67% and for Jolly-Seber method is 66%. The number of dolphins that were re-captured on photograph in the third sampling occasion (Jolly-

69 Chapter 5

Seber) and identified during earlier occasions (m3) is 28 individuals, illustrating the high re-sight probability over more than two sampling periods. The estimate of total population size using the Petersen two-sample mark-recapture method was 55 individual dolphins (95% CL = 44 – 76; CV = 6%). Calculating a potential maximum bias due to loss of individuals between the sample periods, lowers the estimate to 54 individuals (95% CL = 44 -76; CV = 10%), which is 2% of the estimated population size above. During the 3.5 year study period at least 17 dolphins have died but the specific dolphin identities were not available and thus could not be traced back to the photo-identification catalogue. An estimate of population size using the Jolly-Seber method arrives at 48 individual dolphins (95% CL = 33 - 63; CV = 15%). The proportion of the population surviving from the 1st to the 2nd sampling occasion is 66%. Reported number of dead dolphins between these two sampling periods is 2 individuals (4% of N2). An estimate was also calculated including a maximum bias due to lack of survey effort during one of the sampling periods in one ‘closed’ area that is inhabited by a group of six dolphins. This corrected estimate arrives at 53 individuals (95% CL = 36 – 64; CV = 19%), which is 10% of the unbiased population size.

Selected pictures 70 160 Identified dolphins 60 140 Cumulativ e 120 identified dolphins 50

100 40 80 30 60 20 40 Cumulative numberidentified 20 10 0 0

Oc Au Nov Au Ap Au Ja/Fe Fe/Mar Jun/Jul Oc/ No Ap/May May /Jun 1999- 2000- 2001- 2002- Survey period

Figure 2. Discovery rate of new individuals and number of identified dolphins per survey period in relation to the number of selected pictures

Figure 2 shows the cumulative number of new individuals identified in different survey periods in combination with photographic success in obtaining identifiable pictures of dorsal fins for each sub-period. The cumulative curve begins to level of after the August 2001 survey period and during the next three survey periods only one

70 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

individual was added each time (Table 1). Some 95% of the individuals of the photo- identification catalogue are identified in the period March 1999 until August 2001. After that date a plateau in the number of new identifications is more or less reached, with only a yearly 5% increase of new identifications (three individuals) of the total photo-identification catalogue. With an annual birth rate of 10.5 % of the total population, this yearly 5% increase in new identifications is within this birth rate range and may therefore be attributed to possible neonates. It should be noted though that these neonates can be identified only when they are over one-year of age, since they are otherwise difficult to photograph. So, the new identifications within any one year may include last year’s neonates i.e., one year old calves. The plateau was not a result of low photographic effort, since the number of new individuals added to the catalogue is not correlated with the number of identifiable photographs (r = 0.06; DF = 10). Some 98% of the identified dolphins were recaptured on photograph on at least two different days and 90% were recaptured during at least two different survey periods (Figs 3 and 4). Individual dolphins were recaptured on a mean number of 7.0 different survey days (± SD = 4.7) and 4.5 survey periods (± SD = 2.4). Individual dolphins were recaptured on a maximum of 21 days and 10 survey periods (Plate 2).

Plate 2. Example of a low quality photograph (small dorsal fin image), in which dolphin PM01 can still be identified over larger distances due the distinctiveness of its mark. Dolphin PM01 was photographed during 21 different survey days, on 41 pictures and photographed here on 23/8/00 (upper picture) and 2/7/01 (lower picture).

71 Chapter 5

Number of re-sighted individuals 2

0 12345678910 Number of survey periods

Figure. 3. The number of re-sighted individuals during a number of survey periods, e.g. 14 individuals were re-sighted during four different survey periods.

14 No. of dolphins re-sighted on photograph on x days 12 No. of dolphins re-sighted on video on x days 10

0 1 3 5 7 9 111315171921 Number of survey days

Figure. 4. The number of re-sighted dolphins on photograph and video over a maximum of 21 days, e.g. 14 and 11 dolphins were re-sighted on photograph and video respectively during a period of 2 until 3 days.

72 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

Feasibility of video-identification

Video recordings were made during seven different survey periods and 21 days. Total recording effort to get photo-identification images was 8.8 hours. Identifiable dorsal fins of surfacing dolphins were recorded on 79 video-images, from which 31 different individuals could be identified. On average, 9.0 identification images per hour and 4 images per day recording were produced. Four individuals were identified based on body marks alone. Fifty-two percent of the individuals were encountered on more than one day (mean = 2.1; ± s.d. 1.4; range = 1 – 5) (Figure 4).

DISCUSSION

Estimates of abundance based on photo-identification mark-recapture analysis

Violated assumptions and biases Two methods for analysing mark-recapture results of photo-identified dolphins were used in this study, the Petersen two-sample method and Jolly-Seber method. The first method was found appropriate because during two of the 12 survey-periods the following required condition to obtain an estimate of total population size was met: photographic ‘trapping’ effort was equally spread over the entire dolphin distribution range, so that all animals have the same probability of being identified (assumption 2, see below). Most other survey periods involved intensive monitoring surveys in areas of high dolphin density only. Also, one area in between two rapids was not surveyed during the other extensive monitoring surveys due to bad weather conditions. The second method (Jolly-Seber) was applied because it allowed for gains and losses between the sampling periods. The disadvantages of using these methods are that they rely on underlying assumptions, which, if violated, produce serious biases of the results. For the Petersen method, these assumptions are: 1) the population is closed; 2) all animals have the same probability of being caught; 3) marking does not affect the catchability of an animal; 4) the second sample is a simple random sample; 5) animals do not lose their marks; and 6) all marks are reported on recovery. For the Jolly-Seber method, assumption 2 and 5 from Petersen are also relevant. Additional assumptions for Jolly-Seber are that: 7) every marked animal has the same probability of surviving from the ith to the (i + 1)th sample; 8) every animal caught in the ith sample has the same probability of being returned to the population; 9) all samples are instantaneous (Hammond, 1986). The first and second assumptions are being violated in this study regarding the Petersen method and Jolly-Seber method, respectively, and the effects are discussed below. The first assumption of the Petersen method was violated as three dolphins (identity unknown) had died and four dolphins were born between the sampling periods. However, mortality is not likely to influence n2 (total number caught on the

73 Chapter 5

second occasion), since during each sampling period only 55-57% of total photo- identification catalogue was captured on film. A possible influence of dead dolphins on the m2 (number of ‘marked’ animals recaptured on the second occasion) likely occurred although not unalterably, since the number of ‘recaptured’ animals is not equal (only 64-66%) to the total number of individuals caught on the first and second occasions. So, these dead dolphins of unknown identity could just as well not have been ‘marked’ on the first occasion or, if they were, had not been recaptured. Still, the three dead dolphins may possibly have produced a biased estimate and therefore a correction was calculated for this bias, which decreased the estimate at the most by two individuals. This bias only applies if we assume that these three dolphins were ‘marked’ at the first occasion and presumably would have been caught on the second occasion as well if they hadn’t died. In that case, the abundance estimate would be 54 individuals, within the confidence limits of the abundance estimate of 55 individuals as inferred in the results section. This small difference may be a result of the fact that a high proportion of the estimated population was captured during each sampling period (65-67%), since catching over 50% of the population limits biases that may arise through assumptions being violated (Sutherland, 1996). Similar to mortality, recruitment (dolphins born in the period between the two sampling periods) is not likely to influence the overall number of dolphins caught on the second occasion (n2). Furthermore, neonates will not influence the number of ‘marked’ animals found on the second occasion (m2), since they were born after the first sampling period and were thus not recorded. Neonates and calves have a low chance of being identified at all since they surface very irregularly and briefly during the first few months and are hard to photograph as they swim very close to the mother. Consequently, neonates encountered in the first sampling period will most certainly not have been ‘marked’ and will for that reason also not affect one of the variables of the Petersen formula. Violations of the second assumption due to heterogeneity between dolphins in catchability and trap responses were tested with a goodness-of-fit-test for three sampling periods used within both analysis methods and this revealed that there was no difference in capture probabilities except for the neonates and calves for which a correction factor is applied to calculate abundance estimate (see analysis). This is in contrast to most other cetacean photo-identification studies in which unequal capture probabilities are often the case due to variation in individual behaviour, such as wariness of boats or fluking behaviour, that affect the probability of obtaining good photographs (Whitehead, et. al., 2000). Capture probabilities are more likely to vary for bow-riding dolphins, whereas the dolphins in this study were all photographed from some distance of the boat. Thus, boat-shyness or attraction did not play as much of a role. Since photo-identification is in principal a non-invasive technique, any issues of trap responses are not relevant here. In spite of the fact that dolphins in principal had an equal probability of being photographed, differences in distinctiveness of marks and in survey area coverage may cause capture probabilities (obtaining identifiable

74 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

images) to vary among individuals and cause a bias of the estimate of population size (Gowans & Whitehead, 2001). Although all photographs of good image quality yielded identifiable marks, photographs of less quality (smaller images) were only identifiable for those individuals with very distinct marks (Plate 2). Other markings needed to fill a significant part of the frame for identification and therefore more slides were discarded for use in connection with these features. Another bias in capture probability was related to differences in survey area coverage for each sampling period in the calculation of the Jolly-Seber estimate. However, the G-test result and the high percentage of re-sightings over different survey days and periods (95% and 90% of total identified individuals were re-sighted over two days and periods or more, respectively), indicate that the bias is not large, possibly due to the fact that a large part of the population was caught during both samples, as stated earlier. Nevertheless a maximum bias was calculated that could affect the Jolly-Seber estimate for the difference in area coverage. This bias produced an estimate that only differed with three individuals from the Jolly-Seber estimate. Finally, dolphins in this study were only identified using natural marks, which would be stable over long sampling intervals (such as notches, cuts, scars and fin shapes) to prevent biases when marks are lost (such as pigmentation patterns) as suggested by Gowans & Whitehead (2001). Furthermore, other underlying assumptions of both methods did not seem problematic in this study. The difference between the total number identified dolphins (59) and the estimated total population size (N= 48-55), may be explained by the fact that the first number was derived over a 3.5 year study period, during which 17 dolphins had died. The total number identified dolphins does therefore not represent an abundance estimate. The proportion of the population surviving from the 1st to the 2nd sampling occasion (66%) based on the Jolly-Seber equation whereas the proportion surviving based on the reported number of dead dolphins between these two sampling periods is 96%. The difference may be explained in the fact that the probability of survival within Jolly-Seber is determined by sampling the marked population only and variation in the size of the marked population may occur between two sampling periods for reasons other than mortality and emigration. For example, photographs are not always successful for all sightings within each sampling period due to the dolphins’ group behaviour at that specific moment, which may vary through time for the same group. In this way, some groups may be missed from identification during one period but identified during another period.

Identifiability As stated above, in this study, from all photographs of good image quality of dorsal fins, individual dolphins could be identified. This agrees with a photo-identification study on coastal Irrawaddy dolphins in North Queensland, Australia, although juveniles were reported to lack any distinctive features to allow for identification (Parra and Corkeron, 2001). Of Pacific white-sided dolphins, Lagenorhynchus obliquidens

75 Chapter 5

and Indo-Pacific humpbacked dolphins, Sousa chinensis, only a percentage of dolphin dorsal fins could be identified (Jefferson and Leatherwood, 1997; Morton, 2000). In addition, as in the Australian study, no standardized identification measure could be used such as the Dorsal Fin Ratio (Defran et al., 1990) to identify Irrawaddy dolphins in the Mahakam, since fins lacked clearly distinct top and bottom points. In contrast to the studies mentioned above on other species than Orcaella brevirostris, Irrawaddy dolphins in this study and others could also be identified based on the variation of dorsal fin shapes (Stacey, 1996; Parra and Corkeron, 2001). With regard to possible false matches: I only found three dolphins with more uniform, smooth dorsal fin shapes (although not similar compared to each other). However, each of these dolphins were only re-sighted on 5, 7 and 11 different survey days, which is within the standard deviation of the mean number of days on which all dolphins were re-sighted (mean = 7 days, SD = 4.7). So, the chance seems small that different dolphins were identified as one of these three dolphins. Then I would expect the number of sighting days for these dolphins to be much more numerous. Also, I found these fins still identifiable on basis of overall shape, even though characteristic notches were missing. With regard to identification of calves and juveniles, I found that Irrawaddy dolphins in the Mahakam River did have identifiable features on their dorsal fins. This stands in contrast to Parra and Corkeron (2001), who conducted a photo- identification study of coastal Irrawaddy dolphins in Australia and found that calves and juveniles did not have any distinctive features to allow identification. During each of the extensive sampling periods (covering entire dolphin distribution range), we encountered one group of animals consisting of some six juveniles without adults. Unfortunately, individuals of these groups were never successfully photographed, because of their elusive surfacing-behaviour. Only drawings of dorsal fins, (made by aid of binoculars) and one photograph with distinctive marks on the juvenile’s body were available for these. Juveniles in mixed groups were on the other hand much less shy, in fact they often surfaced near the boat. Since no record was kept in the field of the dolphin age classes of each photograph, it is not possible to trace which identified dolphin is a juvenile and which is an adult on basis of the picture alone. However, occasionally, when drawings were made during the study of several characteristic dorsal fins, age class was also noted and these included both juveniles and calves. The high percentage of individuals that were re-sighted on more than one occasion (98% of 59 identified dolphins) is an indication of the closeness of the Mahakam dolphin population. Percentage of re-sightings were similar (97% and 100%) for resident populations of marine tucuxis, Sotalia fluviatilis in Southern Brasil and of 21 identified bottlenose dolphins, Tursiops truncatus, in the Stono River estuary in South Carolina (Flores, 1999; Zolman, 2002). Resightings of seasonally occurring groups are typically lower; varying percentages of 32%, 50% and 57 % were found of 675 identified individual Pacific white-sided dolphins, Lagenorhynchus obliquidens, in the Broughton Archipelago, Canada, 35 identified Irrawaddy dolphins, Orcaella brevirostris, in Cleveland and Bowling Green Bay in North Queensland, Australia and 213

76 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

identified Indo-Pacific humpbacked dolphins, Sousa chinensis, in Hong Kong waters, respectively (Morton, 2000; Parra, 2001; Jefferson, 2000).

Comparison of different techniques to estimate population abundance The estimates of population size based on two different methods in this mark- recapture study are very much in agreement with each other since both estimates are within the confidence limits of each other (combined between 33 and 76). It may be noted though that the Petersen estimate (N = 55) is somewhat higher than the Jolly- Seber estimate (N = 48), whereas the coefficient of variation is smaller for the first estimate (CV = 6% and 15%). The latter estimate is close to the estimate derived from direct counts and strip-transects in May/ June 2000 (Ncount = 35 and Nstrip = 43) Kreb, 2002) with both estimates within the confidence limits of the Jolly-Seber estimate. Because the low estimates calculated here represent the total population size of dolphins in the Mahakam, immediate conservation measures are required to reduce the high minimum mortality rate of 10.5% dolphins of total population per year. Moreover, intended live-captures of dolphins for display in a local oceanarium to be built in the district’s capital city along the Mahakam should therefore definitely not be allowed for this small population. In order to monitor future trends in abundance, photo-identification may be a valuable tool. However, to increase precision and prevent biases due to gains and losses of individuals I recommend that photographs be taken during two extensive monitoring surveys in sequence covering the entire dolphin distribution range with a minimum time interval. Conclusively, since the results of the mark-recapture studies and direct count and strip-transect studies are very similar, future surveys to monitor trends in abundance of the latter type are feasible, if one needs to be cost efficient. However, surveys in combination with photo-identification are preferable in order to obtain data on long-term social system and migration patterns.

Feasibility of video-identification

The number of identifiable video-images per hour recording in this study (9 images/ hr), was much lower than those recorded in the video-identification study of bottlenose dolphins in South Carolina (Zolman, 2002), which yielded 31 images per hour recording time. This may be a result of the fact that in the latter study only a video was used for identification of dolphins, which may increase the efficacy to make good quality recordings. Another reason is that it may be more difficult to record dorsal fins of Irrawaddy dolphins because of their shy and irregular surfacing pattern (Kreb, 1999). Also, the number of identifiable video images per day were much lower i.e., four identifiable video images per day, in comparison to the still photography in this study, which produced nine identifiable photographs per day. Nevertheless, although the yield of identifiable images may be less than in other studies and in

77 Chapter 5

comparison to still photography, video-identification used as an additional tool has some advantages. First, in most cases the entire movement of the dolphin is visible, during play-back including all the different angles from which a dorsal fin can be seen. This was particularly useful in cases when there were any doubts within the photo- identification catalogue about whether two assumedly different identified dorsal fins belong in fact to one and the same individual. Although dorsal fin pictures were always attempted to be taken perpendicularly to the dolphins body axis close to the dorsal fin region, small deviations from this angle could in some cases cause confusion about the identification. Second, this technique can link body characteristics to individuals, which are initially identified based on dorsal fins alone. Third, for other purposes, such as study of social structure, video-recordings make it possible to record the physical position of individual dolphins with regard to each other. However, disadvantages of the use of a video camera were experienced by author and field- assistants in connection with the slow adjustment between wide-angle and zoom modes. Even though we tried to use a fixed zoom length and estimated where the dolphins would surface, the manoeuvrability of the video camera suffered in comparison with the photo-camera. In addition, the quality of video images for which a digital zoom was used often did not allow for accurate identification. Since the images were analysed by using the slow motion, or pause mode the quality of still video images decreased significantly as a consequence, as did images recorded with the optical zoom. No mark-recapture analyses were performed using video images, since the images were not recorded systematically throughout the study period. The quality of the still video images was found low in comparison to the photographs. Therefore, identifications were not directly based on the video images but were first traced back to the photo-identification catalogue. However, my overall conclusion is that video- identification in combination with photo-identification appeared to be useful for determining identities of individual dolphins.

ACKNOWLEDGEMENTS

I would like to thank Hardy Purnama, Zainuddin (BKSDA), M. Syafrudin, Achmad Chaironi, Ade Rachmad, Arman, M. Syoim, Budiono, Bambang Yanupuspita, Sonaji, Syahrani, Rudiansyah, Ahank, Iwiet, Hendra, Munadianto (Universitas Mulawarman Samarinda), Audrie J. Siahainenia, Ramon (Coastal Resource Management Program/ Proyek Pesisir Kal-Tim), Karen Damayanti Rahadi (Universitas Padjajaran), Pak Sairapi and Pak Muis for their assistance, enthusiasm and hard work. Funding for fieldwork was provided by Ocean Park Conservation Foundation, Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Gibbon Foundation; Netherlands Program International Nature Management (PIN/ KNIP) of Ministry of Agriculture, Nature Management and Fisheries; Van Tienhoven

78 Mark-recapture analysis of photo-identified Irrawaddy dolphins in the Mahakam

Stichting; World Wildlife Fund For Nature (Netherlands); Amsterdamse Universiteits Vereniging; Coastal Resource Management Program/ Proyek Pesisir. I would like to thank the Indonesian Institute of Sciences (LIPI), the East Kalimantan nature conservation authorities (BKSDA), the General Directorate of Protection and Conservation of Nature (PHKA) for allowing me to conduct my research. The University of Mulawarman in Samarinda (UNMUL), Plantage Library, M. Lammertink and Dr. P.J.H. van Bree (University of Amsterdam (UvA), Zoological Museum Amsterdam) are thanked for their support and J. Van Arkel (Institute for Biodiversity and Ecosystem Dynamics (IBED) for producing the map figure. The manuscript was improved thanks to comments from Prof. F.R. Schram, Dr. Vincent Nijman (IBED, UvA), Dr. T.A. Jefferson (Southwest Fisheries Science Center, National Marine Fisheries Service, La Jolla) and one anonymous referee.

REFERENCES

Baird, R.W. 2000. The killer whale: foraging specializations and group hunting. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds.) Cetacean societies. Field studies of dolphins and whales. Pp. 127-153. The University of Chicago Press, Chicago. Clapham, P.J. 2000. The humpback whale: seasonal feeding and breeding in a baleen whale. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds.) Cetacean societies. Field studies of dolphins and whales. Pp. 173-196. The University of Chicago Press, Chicago. Defran, R.H., Schultz, G.M. & Weller, D.W. 1990. A technique for the photographic identification and cataloguing of dorsal fins of the bottlenose dolphin (Tursiops truncatus). Rep. int. Whal.Commn. (special issue) 12: 53-55. Dhandapani, P. 1992. Status of the Irrawaddy River dolphin Orcaella brevirostris in Chilka lake. Journal of the Marine Biology Association of India 34: 90-93. Flores, A.C. 1999. Preliminary results of a photo-identification study of the marine tucuxi, Sotalia fluviatilis, in southern Brazil. Marine Mammal Science 15: 840-847. Gowans, S. & Whitehead, H. 2001. Photographic identification of northern bottlenose whales (Hyperoodon ampullatus): sources of heterogeneity from natural marks. Marine Mammal Science 17: 76-93. Hammond, P.S. 1986. Estimating the size of naturally marked whale populations using capture-recapture techniques. Rep. int. Whal.Commn. (special issue) 8: 253-282. Hilton-Taylor, C. 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, U.K. Jefferson, T.A. and Leatherwood, S. 1997. Distribution and abundance of Indo-Pacific hump-backed dolphins (Sousa chinensis, Osbeck, 1765) in Hong Kong waters. Asian Marine Biology 14: 93-110.

79 Chapter 5

Jefferson, T.A. 2000. Population biology of the Indo-Pacific hump-backed dolphin in Hong Kong waters. Wildlife Monographs 144: 65pp. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Zeitschrift für Säugetierkunde 64: 54-58. Kreb, D. 2002. Density and abundance of the Irrawaddy dolphin, Orcaella brevirostis, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. The Raffles Bulletin of Zoology, Supplement 10: 85-95. Krebs, C.J. 1999. Ecological Methodology. Addison-Welsey Educational Publishers, Inc, US. 620 pp. Lloze, R. 1973. Contributions a l’étude anatomique, histologique et biologique de l’Orcaella brevirostris (Gray -1866) (Cetacea-Delphinidae) du Mekong. Dissertation thesis, Toulouse, France. [In French] Mann, J. 2000. Unraveling the dynamics of social life; long term studies and observational methods. In: J. Mann, R.C. Connor, Tyack, P.L. & Whitehead, H. (eds.) Cetacean societies. Field studies of dolphins and whales. Pp. 45-64. The University of Chicago Press, Chicago. Manly, B.F.J. 1971. A simulation study of Jolly’s method for analysing capture- recapture data. Biometrics 40: 749-758. Morton, A. 2000. Occurrence, photo-identification and prey of Pacific white-sided dolphins (Lagenorhynchus obliquidens) in the Broughton Archipelago, Canada 1984- 1998. Marine Mammal Science 16: 80-93. Parra, G.J. and Corkeron, P.J. 2001. Feasibility of using photo-identification techniques to study the Irrawaddy dolphin, Orcaella brevirostris (Owen in Gray 1866). Aquatic Mammals 27: 45-49. Smith, B.D., Thant, U.H., Lwin, J.M. & Shaw, C.D. 1997. Investigations of cetaceans in the Ayeyarwady River and northern coastal waters of Myanmar. Asian Marine Biology 14: 173-194. Stacey, P.J. 1996. Natural history and conservation of Irrawaddy dolphins, Orcaella brevirostris, with special reference to the Mekong River, Lao P.D.R. Unpublished M.Sc. thesis, University of Victoria, Canada. 123 pp. Stacey, P.J. and Arnold, P.W. 1999. Orcaella brevirostris. Mammalian Species 616:1-8. Sutherland, W.J. (ed) 1996. Ecological Census Techniques. A Handbook. Cambridge University Press, UK. 336pp. Whitehead, H., Christal, J. & Tyack, P.L. 2000. Studying cetacean social structure in space and time: innovative techniques. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds.) Cetacean societies. Field studies of dolphins and whales. The University of Chicago Press, Chicago. Zolman, E.S. 2002. Residence patterns of bottlenose dolphins (Tursiops truncatus) in the Stono River estuary, Charleston County, South Carolina, U.S.A. Marine Mammal Science 18: 879-892.

80 Conservation of riverine Irrawaddy dolphins in Borneo

CHAPTER 6

Conservation management of small core areas: key to survival of a critically endangered population of riverine Irrawaddy dolphins in Borneo

Daniëlle Kreb and Budiono

In press: Oryx, 2004

Dolphins preference for fish-rich but human-crowded areas makes them vulnerable to many human-induced threats. Awareness campaigns therefore form a critical factor in their survival. Photo: Daniëlle Kreb

Chapter 6

ABSTRACT

In order to clarify the previous status of the facultative Irrawaddy River Dolphin, Orcaella brevirostris, in the Mahakam River in East Kalimantan, which was ‘insufficiently known’ following IUCN criteria, we collected data from early 1999 until mid 2002 on abundance, habitat use, population dynamics and threats relevant to the conservation of Indonesia’s only freshwater dolphin population. Our best estimates of total population size varied between 33 and 55 dolphins (95% confidence limits: 31- 76) based on direct counts, strip-transect analysis, and Petersen and Jolly-Seber mark- recapture analyses of photo-identified dolphins. Mean minimum annual birth and mortality rates were nearly similar, i.e. 13.6% and 11.4% and no changes in abundance > 8% were detected over 2.5 years. Dolphins primarily died after gillnet entanglement (73% of deaths). Dolphins’ main habitat includes confluence areas between the main river and tributaries or lakes. Dolphins daily intensively use small areas mostly including confluences, moving up and downstream over an average length of 10 km of river and within a 1.1 km2 - area size. These areas are also primary fishing grounds for fishermen and subject to intensive motorized vessel traffic. Sixty-four percent of deaths (from 1995-2001) with known location (n=36) occurred in these areas. Formal interviews with local residents revealed a generally positive attitude towards the establishment of protected dolphin areas. Because of the dolphins’ dependence on areas that are also used intensively by people, primary conservation strategies should be to increase local awareness and introduce alternative fishing techniques.

RINGKASAN

Dalam usaha memperjelas kondisi lumba-lumba Irrawaddy (Orcaella brevirostris) di Sungai Mahakam Kalimantan Timur, yang mana “belum banyak diketahui” berdasarkan kriteria IUCN, kami mengumpulkan data-data sejak awal tahun 1999 hingga pertengahan 2002 tentang jumlah, penggunaan habitat, perubahan populasi, dan ancaman yang berkaitan dengan upaya konservasi satu-satunya lumba-lumba air tawar di Indonesia. Diperkirakan jumlah populasi Pesut Mahakam berkisar antara 33 hingga 55 ekor (tingkat kepercayaan 95%: 31-76) berdasarkan perhitungan langsung, analisis strip-transek, dan analisis penandaan-ulang Peterson dan Jolly-Seber dari identifikasi foto lumba-lumba. Rata-rata terendah tingkat kelahiran dan kematian pertahun hampir sama yakni 13,6% dan 11,4% dan tidak ada perubahan jumlah lebih dari 8% selama 2,5 tahun. Kematian utama lumba-lumba adalah terperangkap rengge (73%). Habitat utama lumba-lumba termasuk pertemuan antara sungai utama dan anak sungai atau danau. Sehari-hari lumba-lumba secara intensif menggunakan daerah yang kecil kebanyakan merupakan daerah pertemuan sungai, bergerak ke hulu dan ke hilir dengan jarak tempuh rata-rata 10 km dan dalam ukuran areal 1.1 km2. Tempat- tempat ini juga daerah utama penangkapan ikan dan lalu lintas kapal bermotor. Enam puluh empat persen (64%) kematian (dari 1995-2001) dengan lokasi yang diketahui

82 Conservation of riverine Irrawaddy dolphins in Borneo

(n = 36) terjadi di daerah ini. Wawancara formal dengan penduduk lokal umumnya menyatakan sikap positif terhadap pembentukan daerah perlindungan lumba-lumba. Karena ketergantungan lumba-lumba pada tempat yang juga digunakan intensif oleh masyarakat, strategi utama konservasi adalah meningkatkan keperdulian dan memperkenalkan cara alternatif menangkap ikan kepada masyarakat lokal.

INTRODUCTION

River dolphins and porpoises are among the world’s most threatened mammal species. The habitat of these animals has been highly modified and degraded by human activities, often resulting in dramatic declines in their abundance and range (Reeves et al., 2000). Protection of freshwater dolphins and their habitat is a major challenge since river systems are the veins of human activities in terms of transport, fishing, and industrial processes, and are also heavily affected by forest fires, which were more likely to occur near rivers (Fuller & Fulk, 1998) and likely caused a large increase in sedimentation rates together with large-scale illegal logging practices (Anon, 2000) with disrupting consequences for the aquatic ecosystem (e.g. Mackinnon et al., 1997). In Indonesia one representative freshwater dolphin population occurs in the Mahakam River in East Kalimantan, i.e., the facultative river dolphin Orcaella brevirostris, commonly and locally referred to as Irrawaddy dolphin and pesut, respectively. The species is found in shallow, coastal waters of the tropical and subtropical Indo-Pacific but also in three major river systems: the Mahakam in Indonesia, the Ayeyarwady in Myanmar, and the Mekong crossing through Vietnam, Cambodia and Laos (Stacey & Arnold, 1999). These river populations were all identified to consist of less than 100 individuals based on preliminary studies and faced ongoing and pervasive threats to their long-term persistence (Kreb, 2002; Smith et al., 2003). In order to identify and monitor the population status and threats more thoroughly and set a rationale for conservation action, a 3.5-years study from February 1999 until August 2002 was initiated. This article presents the results of this study and an in-depth analysis of habitat preferences, population dynamics, threats and recommendations for future conservation activities of the Irrawaddy dolphin population in the Mahakam River. Since 1990 the species has been fully protected by law in Indonesia and is adopted as a symbol of East Kalimantan Province. Prior to the present study, no systematic data had been collected before on the Mahakam population. A two-month preliminary study in 1997 revealed that sighting rates (0.06 dolphins/ km) in the middle Mahakam river segment (with highest dolphin densities) were very low (Kreb, 1999). Based on data we collected during 1999 and 2000, the IUCN status of this freshwater population was raised from ‘Insufficiently Known’ to ‘Critically Endangered’ (IUCN, 2003)

Chapter 6

Study area

The Mahakam River is one of the major river systems of Borneo and runs from 118º east to 113º west and between 1º north and 1º south (Fig 1). Regional climate is characterised by two seasons, i.e. dry (from July-October, southeast monsoon) and wet (November-June, northwest monsoon) (MacKinnon et al., 1997). The river measures about 800 km from its origin in the Müller Mountains to the river mouth. Rapids start upstream of Long Bagun at c. 600 km from the mouth, which limit the dolphins from ranging further upstream. Three major lakes and nearly all major tributaries and many smaller swamp lakes are connected to the main river system in the Middle Mahakam Area (MMA) between 180 km until 375 km from the mouth. These lakes are very important fish-spawning grounds and replenish the main river seasonally. Therefore, the MMA is an area of intensive fishing activity with an annual catch of 25.000 to 35.000 metric tons since 1970 (MacKinnon et al., 1997). Coal mining and logging companies occur along the entire length of the Mahakam River, especially in the tributaries. A large gold digging company is located in the upper Mahakam River segment together with several small-scale, illegal gold mines. Infrastructure is poorly developed in East Kalimantan and the Mahakam River is the main transport system.

METHODS

Field methods

We searched the Mahakam River from the delta to upper rapid streams (600 km from the mouth) by boat from February 1999 until August 2002 for a total of 8925 km (837 hours), and observed river dolphins for a total of 549 h. We conducted 12 involved extensive monitoring surveys in 6 survey periods, which covered the entire distribution range (average duration 10 days; SD ± 2 days) during all types of water levels (high, low, medium, increasing, decreasing) to invest distribution patterns, annual recruitment and estimate population abundance using strip-transects, direct counts and mark-recapture techniques through photo-identification, which are more described in detail in Kreb (2002 & in press a). The distribution range was divided in 15 strip- transects (main river and tributaries) and 2 line-transects (Melintang and Semayang Lakes). Each transect could be finished in one day. Another six intensive surveys (average duration 12 days; SD ± 3 days) were conducted in areas of high dolphin density to investigate preferred habitat and to locate dolphin groups for further focal group follows (see below) to assess daily home ranges (Figure 1). To monitor abundance and locate groups, surveys were conducted with 12-16 m long motorised vessels (12-21 hp), travelling at an average speed of 10 km/ h.

84 Conservation of riverine Irrawaddy dolphins in Borneo

Conservation of riverine Irrawaddy dolphins in Borneo Kedang Rantau Muara Kaman Loa Kulu Kota Bangun Kedang Kepala

Batuq 85

Semayang

Jempang Melintang Tepian Ulak bution area, b) areas of high dolphin density and c) coastal Irrawaddy dolphin area. Muara Pahu Muara Muara Jelau Bohoq Muyub Ulu Rambayan Kedang Pahu Damai

Muara Benangak Datah Bilang Long Bagun Figure 1. Study area with a) total dolphin distri

85 Chapter 6

The photographic effort during the extensive monitoring surveys was one hour per sighting with a total observation effort of 545 h. Durin observation team existed of three active observers: two front observers and one rear observer. The average observation time and g all surveys 2074 photographs were made of dorsal fins. For each sighting, the duration, location, group behaviour, group size, group composition and environmental data, i.e. depth, clarity, surface flow rate, temperature, pH and type of river section (river bend, straight stretch or confluence area) were collected. On average five times a day, similar random samples were collected as those obtained during sightings, whereas type of river section was recorded every fifteen min. In order to assess daily home ranges, 58 groups were followed for 321 h in total and on average 5.5 h daily (range 1.5-13 h) using a motorized canoe of 5 hp outboard engine maintaining an average distance of 50 m. In addition, land-based observations were made in the confluence area of Muara Pahu, c. 300 km upstream, which was frequented daily by different dolphin groups. On average, five sequential days (32 days in total) of land-based observation have been completed by two observers, which overlooked the area some 7 until 10 meter above the water surface (depending on water levels) during six different survey periods for a total of 286 h. When a group of dolphins was sighted, we recorded group size and composition (Kreb, in press a), changes in group-composition, and time spent in the area. Formal interviews were conducted using questionnaires with mainly open questions with residents and fishermen (n = 258) in six important dolphin areas to determine their knowledge and attitude with regards to the dolphins and their conservation. Respondents were questioned separately to ensure independence of data. In order to assess the minimum annual birth rates between November 2000 and November 2001, the total number of different newborns were counted during 5 different surveys (both extensive and intensive surveys), which were more or less equally distributed over the year with an average 2.5-month gap in between the surveys. Newborns encountered during each of these surveys were assumed different than those encountered during an earlier survey. We defined newborns to be of less than one month of age if they complied with all three categories: 1) exhibited an awkward manner of swimming and surfacing, 2) spent all their time in close proximity to an adult and 3) were of less than ½ the average length of an adult, following Bearzi et al. (1997). Mortality was estimated from own observations and semi-structured interviews conducted during a preliminary survey in 1997 and during the surveys from February 1999 until August 2002. Mortality was traced back as far as 1995. Incomplete or untrustworthy accounts with missing locality, date, and traceable eyewitnesses were disregarded (14% of n = 44). Dolphin reactions towards different types of boat traffic were tested by comparing dolphin group surfacing frequencies in presence and absence of different types of boats (see Kreb & Rahadi, in press b, for a more detailed method description)

86 Conservation of riverine Irrawaddy dolphins in Borneo

Analysis

To assess the importance of different river areas in terms of dolphin densities, the river was divided in seven areas where at least one dolphin sighting was made. Total sightings made in each of these different areas during 10 extensive surveys were compared using a chi-square test. Sightings in tributaries within 1 km of the confluence area were considered main river sightings. Sighting rates, densities and abundance estimates based on strip-transects and direct counts were calculated according to the formula described in Kreb (2002). Since no sightings were made in any of the lakes during these extensive surveys, no analysis of the conducted line- transects was required. The mean abundance estimates and coefficient of variation (CVs) of two replicated surveys within each survey period, were added for all survey periods and averaged to obtain the total mean population size (and mean CV). In addition, abundance estimates were calculated per water level condition (combining different years) averaging the estimates of each replicated survey. This was done since there was no trend in abundance (see Results, Trends in abundance), and there was no difference between the variation in abundance estimates per replicated survey within the same time period and the variation in abundance estimates of surveys conducted in different periods but at similar water level conditions. Because all rear sightings (n=9) were associated with the dolphins’ positions in river bends (which is an unpredictable variable), no detection correction factor g(0) was used to calculate abundance and associated CVs. Instead, rear sightings were directly included in the abundance estimates. Also, no seasonal variation was found in sizes of groups (see Results, Population composition) so this component was also not included in the calculation of CVs. Abundance was also estimated using both the Jolly-Seber and Petersen mark- recapture methods based on 728 selected identifiable dorsal fin pictures (Kreb, in press a). Mean population size was calculated as the average of the mean abundance estimates from strip-transects, direct counts and mark-recapture analysis. Mean group size in the Mahakam was based on all on-effort sightings made during nine extensive abundance surveys covering the entire distribution range. Groups were considered different if a group joined after 15 min of observation or groups split during observation time. To detect any trend in abundance, regression analysis was applied to the natural logarithm of 5 mean strip- and direct count abundance estimates over a 2.5 years period (early 1999 until mid 2001). Statistical power was calculated by means of a linear regression program TRENDS (Gerrodette, 1993). The same analysis was performed to detect if there was any trend in mortality using data from 1995 until 2001. Random environmental samples i.e., depth, flow rates, pH, temperature and clarity were compared with samples collected at dolphin locations per water level using a two sample T-test, prior to which a two-tailed F-test was applied to test for similar variances, which were equal for all sample comparisons. Dolphin-preferred areas within the main river were investigated by comparing the percentage of dolphin

Chapter 6

sightings made per water level in straight stretches, river bends and confluences during ten strip-transects surveys that covered the same area using a chi-square test. The relative random availability of straight stretches, river bends and confluences were tested using a chi-square test. When df was 1, Yates correction factor was applied. To identify the year-round significance of a confluence area, for which highest dolphin densities were found, the numbers of identified dolphins per water level were compared using a chi-square test. To prevent biases, the correlation between the number of pictures obtained and number of identified dolphins was tested using the Product Moment Correlation Coefficient. Daily home ranges were estimated by measuring the distance between the two most widely separated sighting locations of the focal group. Minimum annual birth rate was estimated by dividing the total number of newborns encountered in one year (see Methods) with the mean population size. Minimum annual mortality rate was estimated by dividing the number of dead dolphins during the study period (interviews + observations) by years and mean population size.

RESULTS

Abundance and distribution

During ten extensive surveys, we made 76 on effort sightings of Irrawaddy dolphins in the Mahakam (Table 1). The actual dolphin sightings were confined to the area in the main river between Muara Kaman (c. 180 km from the mouth) and Datah Bilang (c. 480 km from the mouth) and the tributaries Belayan, Kedang Rantau, Kedang Kepala, Kedang Pahu, Ratah and Semayang Lake. The cross-shaded area (Figure 1) of 195 km length in the main river from Muara Kaman until Muara Benangak (c. 375 km from the mouth) represents an area of high dolphin densities. The total dolphin distribution area in the Mahakam, based on sightings and interviews with fishermen, starts about 90 km upstream of the mouth at Loa Kulu and ends some 600 km upstream at the rapids past Long Bagun, including several tributaries and two lakes (Figure 1, single- shaded areas). Significant differences were found in sighting density among eight survey areas where we made sightings (X2 = 35.91, df = 7, P < 0.01) (Table 1). The three areas where most sightings were made include several confluence areas with tributaries and lakes. Seasonal variation in distribution pattern is summarized in Table 2 and illustrated in Figure 2. At medium water levels sighting rates in the main river and tributaries are similar. At prolonged high-water levels dolphins were more often found in the main river than in the tributaries, whereas at rising high water levels (data not tabulated since incomplete total area coverage) a lowest mean sighting rate (0.03 dolphins/ km) was recorded in the main river indicating that dolphins had moved upstream into the tributaries. At low water levels no dolphins were sighted in the tributaries.

88 Conservation of riverine Irrawaddy dolphins in Borneo

All but one sighting of Irrawaddy dolphins in and near the Mahakam delta were offshore the delta at low tide (n = 4), whereas one sighting was made 10 km upstream of the delta at high tide. A mean salinity of 21 ppt was measured at dolphin

Table 1 Priority areas for conservation based on the combination of dolphin densities, presence of newborns, observed matings and mortality, with low numbers indicating high priority.

River survey segments of 40 km Priority area Dolphins/ Newborns Mating Deaths length * km < 2 months events ** Muara Kaman – Kota Bangun 2 0.13 12 Kota Bangun – Batuq 3 0.16 1 1 Batuq – Tepian Ulak 5 0.1 1 1 Tepian Ulak – Rambayan – Muara Jelau 1 0.31 8 1 13 Rambayan – Bohoq 6 0.04 2 Bohoq – Muara Muyub Ulu 6 0.04 Ratah 4 0.12 1 Muara Jelau – Damai 6 0.04 1

* Actual proposed conservation areas (1–3) are confined to smaller areas based on frequent sighting locations (see results); ** Dolphins that died between 1995 – 2001 in the survey area; 5 dolphins

died outside the survey areas and 2 dolphins died with unknown location.

Main River Tributaries Lake Rapid area

Number of dolphins 0 Medium Medium-high* High Low Water Level

Figure 2. Seasonal habitat use of dolphins based on strip-transect analysis. The main river includes confluence areas of up to 1 km upstream tributaries. * At medium-high water levels, tributary data is lacking and thus not presented in the graph.

Chapter 6

Coefficient of Variation Abundance (strip – transects) 2 No. of dolphins/ km 0 0 0 0% 0.13 0.67 28 24% 0.15 0.73 30 9% 0.14 3.2 12 62% 3.2 0.14 0.06 1.4 5 1.4 0% 0.06 - - 5 0% - - 4 0% - - 5 0% No. dolphins/ km 90 4.5 0.12 0.58 23 27% Mean group size dance per river section and water level condition No. sightings Mean strip width (m) Total length (km)

based on strip-transects. Number of sightings, dolphin densities and abun

Upper tributary Low-water levels 4 33** 75 4 4.3 75 4 33** levels Low-water Low-water levels 4 304 43 0 4.3 43 4 304 levels Low-water Low-water levels 12 828 200 28 levels 4.3 Low-water High-water levels 2 33** 75 2 4.6 75 2 33** levels High-water High-water levels 2 152 43 2 4.6 43 2 152 levels High-water High-water levels 6 levels 414 200 12 4.6 High-water Middle tributary** Middle main river* Medium-water levels 3 228 43 7 4.5 43 3 228 3 4.5 75 levels 3 33** Medium-water levels Medium-water Medium-water levels 9 levels 621 200 16 Medium-water Survey area Transects Chapter 6 Table 2. ** in cross-shaded river area of Figure 1. ; ** distance from mouth of tributary until rapids, however the dolphins sighted there occupy a 'closed' area in between two rapid streams of only 2 km, so no sighting rates have been calculated.

90 Conservation of riverine Irrawaddy dolphins in Borneo

positions in the delta and is associated with brackish waters. Their most inshore occurrence is about 20km upstream of the mouth at high tide according to interviews with fishermen. Since the coastal dolphins have not been sighted or reported to move further upstream than 20 km from the mouth and only enter the delta at high tide, we consider these to belong to a different, coastal stock than the Mahakam population, which range starts 180 km upstream the mouth according to our sightings. The coastal and freshwater populations thus seem isolated from each other. Total mean abundance estimates for the entire dolphin population in the Mahakam derived from strip-transect analysis (method 1) and direct counts (method 2) made during nine extensive monitoring surveys, arrive at 37 individuals (mean CV = 13%; 95% CL = 33-41) and 33 individuals (CV = 8%; 95% CL = 31-35), respectively (Table 3). Independent abundance analyses based on photo-identification and mark-recapture techniques gave total estimates of 55 dolphins (CV = 6%; 95% CL = 44-76) following Petersen’s method (3) and 48 dolphins (CV = 15%; 95% CL = 35-63) according to the Jolly-Seber method (4). The total number of identified dolphins during the study period is 59 individuals.

Table 3 Total abundance of dolphins in the Mahakam based on strip-transect analysis an direct counts.

Water Total T No. Si/ N levels No. length No. Mean surveys strip- 95% N 95% conditions T (km)* Si G ** transect CV CL count CV CV Medium 15 882 26 4.5 8.7 40 19% 21-58 33 8% 26-40 High 10 599 16 4.6 8 37 17% 1-94 36 12% 1-74 Low 20 1165 32 4.3 8 34 9% 30-39 32 2% 31-33 Combined 45 2646 74 4.4 8.2 37 15% 33-41 33 8% 31-35

* Only including transect length of those sections where dolphins were sighted, excluding search effort in areas where no dolphins were sighted during these extensive monitoring surveys (e.g. lakes and upper and lower river section); ** For medium, high and low water levels 3, 2, and 4 surveys were conducted respectively; T = transects; Si = Sightings; G = group size; N = abundance; CV = Coefficient of Variation; 95% CL = 95% Confidence Limit

Regression analysis of the natural logarithm of five mean strip-transect abundance estimates showed a non-significant 1% increase in abundance (b=0.01, t=0.52, df=3, P>0.05) (Figure 3). Direct count abundance estimates for the same study period revealed no changes in abundance (b=0.001, t=0.18, df=3, P>0.05). Power analysis revealed that in order to detect a 5% change (either positive or negative) with high statistical power (90%) and mean CV of 13% or 6% (mean of the CVs of each replicated strip-transect or direct count estimates per survey period) 28 “strip”-samples or seven “count”-samples are needed. In our study period (with five

Chapter 6 samples) only changes as large as 20% (strip-transect estimates) or 8% (direct counts) would have been detected with 90% power.

Trend in dolphin abundance in the Mahakam River

100 Y = 34.4 + 0.6x 80 r2 = 0.070

Abundance estimates 20

0 Med 99 Low 99 Med 00 High 01 Low Time period

Figue 3. Trend in abundance estimates based on strip-transects. No significant population changes occurred during the study period. The vertical bars represent 95% confidence intervals of the estimates.

Population dynamics

The dolphin population consisted on average of 61% adults, 30% juveniles and 9% calves and neonates (Table 4). The mean group size observed during the extensive monitoring surveys was 4.4 (SD = 2.2; range = 1-10). Population composition and group size did not fluctuate during different water levels (H = 0.17, df = 2). Minimum annual number of newborns during the study period was six dolphins. Newborns (< one month of age) were observed at all water levels and in all months of the year. Birth rates between 11 - 18% may apply (of N = 55 and 33 dolphins, respectively). With six newborns per year, the minimum numbers of breeding females within this population are 12 or 18 individuals if a 2- or 3-year reproduction cycle applies, respectively. During the 3.5 year study period minimally 17 dolphins died an unnatural death (interviews and own observations). Minimum average annual mortality during the study period is five dolphins (SD = 2, range = 3-8), which is between 9 and 15 % of maximum and minimum abundance estimates, respectively. True mortality rates, including deaths from natural causes, are unknown.

92 Conservation of riverine Irrawaddy dolphins in Borneo

Table 4. Group composition during three different surveys with different water level conditions.

Number of dolphins per water level during one survey Group composition Medium 2000 High 2001 Low 2001 Mean numbers Adults 19 58 % 22 61 % 20 63 % 20 61 % Juveniles 10 30 % 11 31 % 9 28 % 10 30 % Calves & neonates 4 12 % 3 8 % 3 9 % 3 9 %

Habitat preferences and home ranges

Environmental characteristics for different river sections at medium water levels are presented in Table 5. All freshwater fish trade production comes from the middle river section (including tributaries and lakes), which has the highest dolphin densities. No significant differences were found between random samples and samples collected at dolphin locations for most parameters and water levels. Only for depth measurements at low water levels in the tributaries of the middle river section did we find a significant difference in mean depth of random samples 7.5 m and of samples at dolphin locations 16.7 m (t = 2.85, df = 16, P < 0.05). This suggests that dolphins prefer to remain in deep water pools, such as confluence areas, during the dry season. Their dependence upon confluence areas in particular during the dry season is indicated in Figure 4. At low water levels, significantly more sightings occurred in these areas than in river bends (X2 = 8.5, df = 1, P < 0.01), in spite of the fact that river bends are significantly more numerous (X2 = 24.3, df = 1, P <0.01). Also, more than half (54%) of the 59 dolphins photo-identified in the Mahakam during the 3.5 years study period occurred in the confluence area of Muara Pahu at low water levels (see Abundance) (Table 6). Although photographic effort was largest at low water, no correlation was demonstrated between the number of pictures obtained and the number of identified dolphins (r = 0.893, df = 2, P > 0.05). Also, during the other water level conditions the number of dolphins identified in this area is still high and no significant differences were found in seasonal presence (X2 = 5.1, df = 3, P >0.05). Dolphins occurred in the confluence area of Muara Pahu on average 42% of daytime during every observation day and at all water levels (Table 7). The highest daily occupancy was found at high water levels (65%). Dolphins still remained nearby (< 10 km) at medium and low water levels, but they spent less time milling in the confluence area itself. On average three different groups (range 2–6 groups) consisting of a combined mean number of 12 individuals (range 5–19 individuals), frequented the confluence area daily. Moreover, 44 (75%) of the 59 photo- identified dolphins were sighted at least once in the confluence area (mean = 6 days, max = 17 days of 49 photo/ observation days in that area).

Chapter 6

Straight stretch River bend Confluence

% of dolphin sightings

0 High Medium Low

Water levels

Figure 4. Dolphins preferred areas within the main river. NB random relative Availability of straight stretches was significantly higher than that of bends and confluences (X2 = 112.3, df =2, P<0.01). Also, bends were significantly more numerous than confluences (X2 = 24.3, df = 1, P<0.01).

Eight individuals were sighted exclusively within a 20 km radius of the confluence area (mean number of sightings per individual = 9; range = 2–13 sightings). The confluence area of Muara Pahu and another confluence area about 10 km upstream of there, in the Kedang Pahu tributary, accounted for 89% of the sightings of newborns observed during boat surveys (n = 9) (Table 1). The majority of deaths (64%) with known location (n = 36) also occurred in confluence areas. Mating was observed within different groups in the confluence of Muara Pahu and at one location between Batuq and Tepian Ulak (Figure 1). The average daily home ranges of 27 groups, which were followed for more than 6 hours, were 10 km long (SD = 8.6 km, range = 1-45 km) and 1.1 km2 in area (SD = 1.8 km, range 0.1–9 km2). One group of six photo-identified dolphins are ‘trapped’ between rapid streams in the Ratah River; they have been living for 3.5 years in a river segment 2 km long and 0.2 km2 in area.

94 Conservation of riverine Irrawaddy dolphins in Borneo

Table 5. Environmental characteristics of the Mahakam River collected during medium water levels Random samples values ± standard deviation Mean Mean Mean clarity Mean width Bottom Total fish depth surface (cm)/ Mean (m)/ Mean substrate production River section (m) flow (m/s) salinity (ppt)* distance off/ (ton) 1999** inshore (km)* Lower River 15 ± 5 0.8 ± 0.4 30 ± 9 370 ± 65 Mud 0 Middle River 17 ± 6 0.8 ± 0.3 26 ± 9 200 ± 54 Mud 23201 Upper River 12 ± 7 1.1 ± 0.3 20 ± 10 161 ± 48 Sand, cobbles 0*** MR tributary 9 ± 4 0.7 ± 0.5 22 ± 15 41 ± 14 Mud - UR tributary 12 ± 8 0.8 18 75 ± 12 Rocky - Lakes 2 ± 0.3 0 56 ± 9 - Mud - Delta 5 ± 4 - 22 ± 9* 3km off- Mud, sand 6931 10km inshore*

* Mean salinity only applies to the delta area; ** Data representing direct catch (excluding aqua-culture) for market sale within dolphin habitat based on data from the Kutai Fisheries Department (Dinas Perikanan Kabupaten Kutai Tenggarong, 2000). N.B. Fish production data strictly applicable to the river sections, tributaries or large lakes alone are not available and are therefore combined within the middle river section to which they are connected; *** Fish production data in the upper river section only available until 425km upstream, whereas actual total dolphin distribution area stretches until 560km upstream.

Table 6. Seasonal occurrence of photo-identified individuals in the confluence area of Muara Pahu during 4 different seasons on 5 days in sequence.

Water levels High Low Medium/ low Medium 2001 2001 2001 2000 Identified dolphins 19 32 25 22 Selected identifiable pictures 28 64 37 46

N.B. Two additional survey periods of 5 continuous days were conducted at medium-high water levels, but these were not included as the number of selected identifiable pictures were too low (≤ 16 pictures) due to camera failure.

Table 7. Daily occupancy of dolphins in the confluence area of Muara Pahu expressed in hours and percentage of daytime.

Water levels Medium- Medium-high High Low Low-low Total Time (hours) low 2000 2000 2001 2001 2001 2001 Search effort (km) 46.2 48.2 35.9 51.2 45.9 58.7 286.1 Dolphins present (hrs) 18.0 23.7 15.4 33.4 17.3 11.8 119.6 % daytime dolphins 39 % 49 % 43 % 65 % 38 % 20 % 42 % present (mean)

Chapter 6

Threats

Between 1995 and 2001, 38 deaths, mostly of adults (86%) were recorded on the basis of interviews and two of our own observations (Figure 5). Most dolphins (74%) died as a result of gillnet entanglement in nets with larger mesh sizes (7.5–17.5 cm). Dolphins are obviously attracted to the gill nets and as we often observed them feeding near these nets. Dolphins are also said to aid fishermen by guiding fish into their nets. Many fishermen use the dolphins’ feeding patterns as indicators of the location and time to set up gillnets and in this way increase the danger of entanglement. However, frequent reports exist also of dolphins that have been successfully released by fishermen from gillnets. Second cause of death involved deliberate kills (10%), which mostly happened in isolated areas where dolphins rarely occur. Five dolphins that incidentally died in gillnets were eaten and the skin of two of them was also used as medicine for skin allergy (allegedly, the patients’ allergy disappeared). Vessel strikes caused 5% of deaths. From 1974 until 1988, 28 dolphins were live-captured and taken to Jaya Ancol oceanarium in Jakarta. Two detailed local accounts of illegal captures in 1997 and 1998

gillnet entanglement killed 8 trapped in shallow water boat collision birth process unknown 6

2 Numbers dead 0 1995 1996 1997 1998 1999 2000 2001 Year

Figure 5. Dolphin mortality during 7 years based on reliable reports and interviews in combination with own observations. Major cause of death was through gillnet entanglement (74% of all deaths n = 38).

of 3 and 4 dolphins respectively, were reported. Their fate and destination remain unknown. In 2002 a request for live captures was submitted at the General Directorate of Protection and Conservation of Nature (PKA) by the Regent of Central Kutai Province, East Kalimantan for a new oceanarium along the Mahakam River (8-12 dolphins) and by Jaya Ancol Oceanarium in Jakarta (4-5 dolphins). Following intensive lobby by local NGOs and the ban on live captures since 1990 by the Ministry of Forestry, the request for the captures has not been granted.

96 Conservation of riverine Irrawaddy dolphins in Borneo

A range decline occurred in 20 years time between the 1980s and 2000 in a river stretch of 120 km, from 60 km until 120 km upstream of the mouth, which is 15% of total dolphin historic range, i.e. 820 km including tributaries based on own observations and semi-structured interviews with local residents. The range decline coincides with increased industrial activities, boat traffic and decreases in fish populations (based on fishing data from 1990-2000 of the Fisheries Department in Tenggarong). A recent habitat decline involves the elimination of the Mahakam lakes as primary areas of occupancy according to Tas'an and Leatherwood (1984). We only found dolphins in the confluence area connecting to one lake (Semayang) and in the southern part of that lake. The disappearance of dolphins from Jempang Lake and dolphins decreased occurrence in the other two lakes (confirmed by residents) is probably due to 1) Reduced depth of the lakes in the middle Mahakam area through sedimentation (data from Environmental Controlling Body, Tenggarong) caused by deforestation of the surrounding shorelines (agriculture, illegal logging and forest fires). 2) High density of gillnets in fish seasons (own observations), which obstruct dolphin movements. Furthermore, other factors, which degrade the main distribution area, include: 1) Noise pollution by frequent passing of high-speed vessels (40-200 hp) (mean = 4.6 boats/ h). These boats cause the dolphins to dive significantly longer within a range of 300 m distance of the dolphins than in their absence (Kreb and Rahadi, in press b). Container barges (in 2001 mean=8.4 boats per day) daily pass through a narrow tributary, Kedang Pahu, which represents primary dolphin habitat and occupy over two-thirds of the width of the river and over half the depth of the tributary during the dry season. Dolphins always changed their direction (if swimming upstream) when they encountered loaded container barges and moved downstream ahead of the boats back to the confluence area. 2) Chemical pollution of mercury and cyanid from leaks in dams that retain chemical wastes from gold mining industries, which occurred in 1997 (pers. comm. A. Faroek) and from many small-scale and illegal operating gold miners (own observations). In addition, the observed amount of coal, which falls into the river during the process of over-loading into containers and while being tugged along a tributary that represents primary dolphin habitat is most likely not negligible. Also, during one survey in 2000 in that area we observed some dolphins to be affected by changes in skin pigment of some parts of their bodies. However, it is not clear whether these represent a skin disease and what caused this. These pigment changes have never been observed on dolphins in other areas. 3) Possible (future) prey depletion due to intensive fishing with gillnets, electricity and poison. Interviews with fishermen (n = 108) revealed that 48% were opposed to using electricity for fishing, but 43% were in favour of using this method. Reasons for favouring electric fishing are that fish can be caught faster and more easily. This group also believed that fish abundance is still high and unlikely to decrease.

Chapter 6

Conservation

Based on interviews (n = 258) we found that the majority of residents along the Mahakam River was positively inclined towards the dolphins, felt they needed to be protected and agreed upon establishment of protected areas (Table 8). The following incident is illustrative of the good will of local residents. In 2002 an entire village in important dolphin habitat helped the provincial wildlife conservation department (BKSDA) with over 50 men to capture and transport a dolphin that was trapped in a shallow lake, which would soon fall dry, back to the main river. Afterwards villagers joined a symbolic meal to mark their commitment to dolphin conservation, whereas the same village helped during earlier live captures for oceanaria. The general positive attitude towards the dolphins may be linked with the local belief that the dolphins have a human origin. An ongoing conservation program, was initiated in November 2000 by a local NGO, Yayasan Konservasi RASI (Conservation Foundation for the protection of Rare Aquatic Species of Indonesia) focuses on the protection of the freshwater Irrawaddy dolphin population in the Mahakam, and its habitat. Main activities that have been conducted since 2000 include a yearly conservation/ awareness campaign aimed at different layers of society; yearly monitoring of the dolphin population; socio-economic survey related to fisheries; attitude assessment surveys with local communities; demarcation of an important dolphin site by placing a large billboard in Muara Pahu (Plate 1); establishment of patrolling teams in several villages consisting of local fishermen who patrol their areas and report illegal fishing activities since 2002.

Table 8. Attitude assessment interview with local residents along the Mahakam River(n=258). Questions Answers of % of Explanation respondents respondents (n = 258) Has the pesut Yes 75% Advantages: indicates good fishing areas (47%); indicates brought any No 4% right time and season for fishing (20%); indicates long term- advantages to you? Don’t know 21% rising and decreasing water levels (9%); is enjoyable to observe (24%) Disadvantages: has no commercial value like fish (100%)

Does the pesut need Yes 99% Reasons: rare mammals species (30%); indicator of good to be protected? No 1% fish seasons (14%); has a tourism value (13%); to prevent them from extinction (12%); regret the rare sightings (6%); symbol of East Kalimantan (2%); preserve for future generations (2%); don’t know (21%)

Would you agree of Yes 74% Agreeing under conditions (27%): no fishing ban (59%); establishing No 4% profitable to residents (4%); positive for development (7%) protected dolphin Don’t know 22% and for tourisme (7%); restricted to tributaries (7%); with areas? approval of fishermen (4%) Disagree: too much disturbance (100%)

Would you regret if Yes 64% Reasons for regret: pride of East Kalimantan (44%); rare pesut became No 30% mammal species (28%); indicator of good fish seasons extinct? Unrealistic 6% (28%) Reasons for no regret: pesut has no value (100%) Unrealistic: still many dolphins, extinction is not possible

98 Conservation of riverine Irrawaddy dolphins in Borneo

Plate 1. Installment of a welcoming billboard to indicate the major dolphin habitat in the confluence area of Muara Pahu creating a local awareness and sense of belonging (photo Hari Mulyono).

DISCUSSION

The low abundance estimates, nearly similar minimum birth and mortality rates, degradation of habitat, depletion of fish resources and dolphins dependency on fish- rich but human-crowded confluence areas, underline the critical situation with which this freshwater dolphin population is faced with. Dolphin’s preference and small daily movement patterns in confluence areas may be explained in terms of their depth, high fish abundance and counter-currents, which causes fish to be momentarily ‘trapped’ in the confluence area. Irrawaddy dolphins in the Mekong River (Stacey, 1996; Baird & Mounsouphom, 1997) and Ayeyarwady River (Smith et al., 1997), and other river dolphin species, i.e. the Amazon dolphin, Inia geoffrensis (McGuire & Winemiller, 1998), Indus and Ganges dolphins, Platanista gangetica gangetica and P. g. minor (Khan & Niazi, 1989; Smith, 1993) and Yangtze dolphin, Lipotes vexilifer (Hua et al., 1989) also preferred confluence or deep areas with counter-current eddies. Based on interviews with fishermen, we found that the dolphins’ seasonal migration pattern coincides with the pattern of fish migration, during which fish migrates upstream tributaries to spawn

Chapter 6

when water levels start rising after the dry season at medium water conditions. At prolonged high water levels fish disperses over a larger water surface area and may enter freshwater swamps, which are inaccessible for dolphins, and they may return to the confluence area, where they spent 65% of day-time (see results). During the medium and low water levels, the dolphins still remain close to the confluence area, but spent less time milling in the confluence area itself. No historic data exist for the Mahakam population on population abundance, densities or on annual birth and mortality rates so we do not know surely whether and how these rates have changed. However, information of local residents indicates a definite decline in most sections, although in the middle Mahakam some residents doubt if numbers decreased or that the dolphins became less visible and more shy, due to the intensive boat traffic. Following the observations of residents, this would mean that the 120 km-range decline of the lower river section did not result in higher densities in other areas. In this way, when multiplying the length of decline with current sightings rates, a population decline of 14 individuals in 20 years may have occurred, i.e. 0.7 dolphins per year. This would mean a 30% and 20% population decline in 20 years compared to the present minimum and maximum population size, respectively. It seems reasonable to assume that the habitat degradation and consequent habitat loss for dolphins in the more downstream areas near Samarinda indeed caused a population decline by causing an increased competition in more upstream areas, which were already occupied. Competition for fish resources may have increased the dolphins’ attraction for gillnets with fatal consequences. Moreover, the live captures between the 70s and 90s may have caused a significant sudden decrease in breeding population and past birth rates may have been higher than current rates. Mortality mostly affects adults and birth rates may decrease due to loss of breeding animals. Present rates are nearly similar to those recorded for the less threatened, but ‘vulnerable’ Amazonian ‘Boto’ river dolphin, Inia geoffrensis, i.e. with annual pregnancy rates of c. 10 – 15% (Martin & Da Silva, 2000). If genuine efforts are made to reduce mortality, stem habitat deterioration and protect the dolphins’ food supply, survival and even recovery of this, Indonesia’s only freshwater population of dolphins, might be feasible. How viable this future population will be in genetic terms remains a question mark since we have no data on the degree of inbreeding. The dolphins’ dependence on small and possibly manageable sites and the generally positive attitude of local residents towards the conservation of the pesut may further enhance prospects for success. Protected deep-water areas that are important for dolphins in the Mekong have also benefited fish populations and fishermen livelihood (Baird, 2001). The pesut fits the definition of a flagship species, which not only internationally, but especially locally, has charisma and thus may effectively facilitate protection of other species and ecosystems with which it is associated (Bowen-Jones & Entwistle, 2002). Recommendations for conservation should benefit the freshwater ecosystem including dolphins and humans (Table 9) (also in Reeves et al., 2003). Without establishment of protected areas, the future of

100 Conservation of riverine Irrawaddy dolphins in Borneo

Indonesia’s only freshwater dolphin and symbol of East Kalimantan Province will become increasingly bleak.

Table 9. Recommendations for freshwater ecosystem protection including dolphins and humans.

No. Major recommendations

1 Establish conservation areas in: 1) the confluence area of Muara Pahu and Kedang Pahu tributary until Bolowan, 2) the confluence area of Muara Kaman and tributary Kedang Rantau, 3) the Pela tributary and southern part of Lake Semayang (Fig 1).

2 In conservation areas: 1) Set a speed limit for boats and 2) exclude large coal-carrying ships, employing smaller barges or transport over land (upgrading an old, existing road)

3 1) Exclude gillnets in these areas, or 2) set regulations on type of gillnets used and on the location, season, and manner of setting, and 3) introduce alternative fishing techniques. Offer alternative employment options for gillnet fishermen.

4 Strict law-enforcement by local government to attain sustainable use of available fish resources and stop illegal fishing, logging, pollution and dolphin-capture.

5 Conduct environmental awareness campaigns to increase concern for the conservation of natural resources at both the political and community level.

6 Continue to monitor the dolphin population and the threats to it.

ACKNOWLEDGEMENTS

We thank the Indonesian Institute for Sciences (LIPI), the provincial wildlife conservation department (BKSDA) and local governments of Central- (KUKER) and West Kutai (KUBAR) for granting permission to conduct field research. All field assistants, particularly Ahang, Arman, Karen Damayanti and Syahrani, boatsmen, and respondents are thanked gratefully. Funding for fieldwork was provided by Ocean Park Conservation Foundation, Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Gibbon Foundation; Netherlands Ministry of Agriculture, Nature Management and Fisheries (PIN/ KNIP); Van Tienhoven Stichting; World Wildlife Fund for Nature (Netherlands); Amsterdamse Universiteits Vereniging. The University of Mulawarman in Samarinda (UNMUL), Achmat A. Bratawinata, Frederick R. Schram, Peter J.H. van Bree, Thomas A. Jefferson and Vincent Nijman are thanked for their support and Martjan Lammertink, Randall R. Reeves, Tony R. Martin, Martin Fisher, Tamara L. Mcguire and Ian G. Baird for their comments on the manuscript.

101

Chapter 6

REFERENCES

Anon. 2000. Illegal logging in Indonesia threatens the future of the country’s forests. Oryx 34: 73-74. Baird, I.G. 2001. Towards sustainable co-management of Mekong River aquatic resources: The experience in Siphandone wetlands. In: G. Daconte (ed.) Siphandone wetlands. Pp 89-111. CES VI, Bergamo, Italy. Baird, I.G. & Mounsouphom, B. 1997. Distribution, mortality, diet and conservation of Irrawaddy dolphins (Orcaella brevirostris Gray) in Lao PDR. Asian Marine Biology, 14: 41-48. Bearzi, G., Notarbartolo-Di-Sciara, G. & Politi, E. 1997 Social ecology of bottlenose dolphins in the Kvarneric (Northern Adriatic Sea). Marine Mammal Science 13: 650- 668. Bowen-Jones, E. & Entwhistle, A. 2002. Identifying appropriate flagship species: the importance of culture and local context. Oryx 36: 189-195. Fuller, D.O. & Fulk, M. 1998. Satelite remote sensing of the 1997-1998 fires in Indonesia: data, methods and future perspectives. WWF-Indonesia, Jakarta. Gerrodette, T. 1993. Trends: software for a power analysis of linear regression. Wildlife Society Bulletin 21: 515-516. Hua, Y., Zhao, Q. & Zhang, G. 1989. The habitat and behaviour of Lipotes vexilifer. In: W.F. Perrin, R.L. Brownell Jr, K. Zhou & J. Liu (eds.) Biology and conservation of the river dolphins. Pp. 92-98. Occasional paper of the IUCN Species Survival Commission, 3, IUCN, Gland, Switzerland. IUCN 2003. 2003 IUCN Red List of Threatened Species. http:/www.redlist.org [accessed 15 July 2004]. Kahn, K.M. & Niazi, M.S. 1989. Distribution and population status of the Indus dolphin, Platanista minor. In: W.F. Perrin, R.L. Brownell Jr, K. Zhou & J. Liu (eds.) Biology and conservation of the river dolphins. Pp. 77-80. Occasional paper of the IUCN Species Survival Commission, 3, IUCN, Gland, Switzerland. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Zeitschrift für Säugetierkunde 64: 54-58. Kreb, D. 2002. Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. The Raffles Bulletin of Zoology, Supplement 10: 85-95. Kreb, D. (in press a) Abundance of freshwater Irrawaddy dolphins in the Mahakam in East Kalimantan, Indonesia, based on mark-recapture analysis of photo-identified individuals. Journal of Cetacean Research and Management. Kreb, D. & Rahadi, K.D. (in press b) Living under an aquatic freeway: effects of boats on Irrawaddy dolphins (Orcaella brevirostris) in a coastal and riverine environment in Indonesia. Aquatic Mammals.

102 Conservation of riverine Irrawaddy dolphins in Borneo

MacGuire, T.L. & Winemiller, K.O. 1998. Occurrence patterns, habitat associations, and potential prey of the river dolphin, Inia geoffrensis, in the Cinaruco River, Venezuela. Biotropica 30: 625-638. Martin, A.R. & Da Silva, V.M.F. 2000. Aspects of status of the Boto Inia geoffrensis in the central Brazilian Amazon. Paper, SC/52/SM15, presented at 52nd Annual Meeting of the International Whaling Commission, Small Cetacean Sub- committee. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press. Priyono, A. 1994. A study on the habitat of Pesut (Orcaella brevirostris Gray, 1866) in Semayang-Melintang Lakes. Media Konservasi, 4: 53-60. Reeves, R.R., Smith, B.D., Crespo, E.A. & Notarbartolo di Sciara, G. 2003. Dolphins, Whales and Porpoises: 2002-2010 Conservation Action Plan for the World’s Cetaceans. IUCN/SSC, Cetacean Specialist Group. IUCN, Gland, Switzerland and Cambridge, UK. Smith, B.D. 1993. 1990 status and conservation of the Ganges River dolphin Platanista gangetica in the Karnali River, Nepal. Biological conservation, 66: 159-169. Smith, B.D., Thant, U.P., Lwin, J.M. & Shaw, C.D. 1997. Investigation of cetaceans in the Ayeyarwady River and northern coastal waters of Myanmar. Asian Marine Biology, 14: 173-194. Smith, B.D., Beasley, I. & Kreb, D. 2003. Marked declines in populations of irrawaddy dolphins. Oryx 37: 401-401. Stacey, P.J. 1996. Natural history and conservation of Irrawaddy dolphins, Orcaella brevirostris, with special reference to the Mekong River, Lao P.D.R. M.Sc. thesis, University of Victoria. Stacey, P.J. & Arnold, P.W. 1999. Orcaella brevirostris. Mammalian Species 616: 1-8. Tas’an & Leatherwood, S. 1984. Cetaceans live-captured for Jaya Ancol Oceanarium, Djakarta, 1974-1982. Reports of the International Whaling Commission 34: 485-489.

103

Chapter 6

104 Effects of boats on Irrawaddy dolphins

CHAPTER 7

Living under an aquatic freeway: Effects of boats on Irrawaddy Dolphins (Orcaella brevirostris) in a coastal and riverine environment in Indonesia

Daniëlle Kreb and Karen D. Rahadi

In press: Aquatic Mammals, 2004

A container barge passing through a narrow and shallow tributary, i.e., Kedang Pahu River, which also represents primary dolphin habitat. During the dry season, these types of boats occupy three-quarters of the total river width causing a great deal of disturbance to the dolphins

105 Chapter 7

ABSTRACT

Interactions between boats, and coastal and freshwater Irrawaddy dolphins (Orcaella brevirostris), were studied in East Kalimantan, Indonesia during 2001. The goal was to determine the conditions under which dolphins reacted to boats and recommend conservation actions. Both coastal and freshwater Irrawaddy dolphins surfaced less in the presence of boats, but the avoidance reaction lasted longer for the river dolphins. River dolphins surfaced significantly less in the presence of motorized canoes (< 40 hp), speedboats (40-200 hp), and container barges (>1000 hp). Coastal dolphins only reacted to speedboats, and only when they approached at a 50 m distance, whereas river dolphins reacted within a maximum distance of 250 m before and 300 m after a speedboat passed. Besides surfacing changes, river dolphins actively avoided container barges. The strength of reactions did not depend on the dolphins’ behavior, group size, or age. Hyper-sensitivity by river dolphins to intensive boat traffic could explain the different responses between coastal and river dolphins. To prevent dolphin displacement from their core areas, an action plan currently is being developed by a non-governmental organization in cooperation with Indonesian governmental institutions and residents. Speedboat owners will be urged to reduce boat speed in areas indicated on sign boards.

RINGKASAN

Hubungan antara kapal, dengan lumba-lumba Irrawaddy laut dan sungai (Orcaella brevirostris) telah dipelajari di Kalimantan Timur, Indonesia selama 2001, Tujuan utamanya adalah untuk mengetahui kondisi-kondisi dimana lumba-lumba memberi reaksi kepada kapal dan menganjurkan kegiatan konservasi. Kedua lumba-lumba laut dan sungai jarang muncul dipermukaan saat kehadiran kapal, tetapi reaksi penghindaran lebih lama pada lumba-lumba sungai. Lumba-lumba sungai menjadi sangat jarang muncul dengan kehadiran kapal motor (< 40 stk), speedboats (40-200 stk) dan kapal penarik ponton (>1000 stk). Lumba-lumba Laut hanya bereaksi kepada speedboat, jika berdekatan pada jarak 50 m, sedangkan lumba-lumba sungai bereaksi dengan dalam jarak maksimum 250 m sebelum dan 300 meter setelah speedboat berlalu. Disamping perubahan kemunculan, lumba-lumba sungai juga biasanya menghindari kapal penarik ponton. Kekuatan reaksi tidak tergantung pada tingkah laku lumba-lumba, ukuran kelompok atau umur. Kepekaan tinggi pada lumba-lumba sungai terhadap lalu lintas kapal menjelaskan perbedaan reaksi antara lumba-lumba laut dan lumba-lumba sungai. Untuk mempertahankan agar lumba-lumba tidak terusir dari tempat asalnya, suatu rencana kegiatan sedang dikerjakan oleh organisasi non pemerintah yang bekerjasama dengan institusi pemerintah Indonesia dan masyarakat. Pemilik speedboat diwajibkan untuk mengurangi kecepatan di daerah yang ditunjukan dengan tanda.

106 Effects of boats on Irrawaddy dolphins

INTRODUCTION

The Irrawaddy Dolphin (Orcaella brevirostris) is a facultative freshwater dolphin, occurring both in shallow coastal waters and large riverine systems in tropical Southeast Asia and subtropical India (Stacey & Arnold, 1999). In Indonesia, Irrawaddy dolphins occur along the coasts and in one river in East Kalimantan, the Mahakam (Kreb, 1999). The International Union for the Conservation of Nature (IUCN) status of this freshwater population recently was clarified and defined as ‘Critically Endangered’ (Hilton-Taylor, 2000). The mean Mahakam population size was estimated at 43 individuals (95% CL = 31 to 76, CV = 8% - 15%) based on direct counts, strip-transect analysis, and Petersen and Jolly-Seber mark-recapture analyses of photo-identified dolphins (Kreb, 2002; in press). The dolphins occur primarily in deep pools located near confluences and meanders, and occasionally in connecting lakes and tributaries. These areas also are primary fishing grounds and subject to intensive motorized vessel traffic (Kreb, D., in press). Between 1995 and 2001 (D. Kreb, in press), at least 37 dolphins were killed by entanglement in gillnets (81%), illegal hunting (8%), and vessel collisions (5%). The impacts of boat traffic on the dolphins were investigated because the dolphins prefer specific confluence areas, where boat traffic is intense. A number of studies have focused on the short-term and long-term reactions of whales and dolphins to boats: Richardson et al.. (1995) and Gordon & Moscrop (1996) reviewed several studies on the behavioral effects of small cetaceans, in particular bottlenose dolphins (Tursiops truncatus) and belugas (Delphinapterus leucas), and their reactions to boat traffic. Short-term behavioral changes involved longer dive times, shorter surface intervals with increased blow rates, changes of direction, ‘freeze’ responses with tight pod formation, increased swimming speed, changes in acoustic behavior (from a reduction in calls to a shift to higher frequency bands). An increase in whistle repetition at the onset of a vessel approach was proposed to be an effective way to reduce signal masking and enhance communication in a noisy environment (Buckstaff, 2003). Additionally long-term changes involved shifting to higher frequencies and greater intensity echolocation signals as an adaptation to a noisier environment, and departure from frequently disturbed areas. Approach (bow-riding dolphins) or neutral (with no apparent change in directional movement) responses to boats were found in two studies on bottlenose dolphins in areas with typical tourist boat densities (Allen & Read, 2000; Gregory & Rowden, 2001). In contrast, Janik & Thompson (1996) found that bottlenose dolphins dived longer and/or moved away when approached by dolphin-watching boats. An initial approach response was found to dolphin-watching boats by Hectors’ dolphins, Cephalorhyncus hectori, which turned into either disinterest or active avoidance after a 70-min encounter (Bejder et al.., 1999). Bottlenose dolphins also were reported to decrease their use of primary foraging habitats during periods of high boat density (Allen & Read, 2000). Subsurface responses of bottlenose dolphins to approaching boats, which were recorded on videotape, involved decreased inter-animal distance, changed heading, and increased

107 Chapter 7

swimming speed (Nowacek et al.., 2001). Observations of river dolphins are sparse and only qualitatively investigated. Smith (1993) reported that Ganges dolphins (Platanista gangetica) avoided motorized boats and Zhou & Li (1989) found that the baiji (Lipotes vexillifer) in the Yangtze River generally dived for longer times and tended not to surface within 50 m of boat traffic. Irrawaddy dolphins in the Mekong River surfaced less often when large motor boats were within 100 m and surfaced closer to slow moving boats (Stacey, 1996). This study was conducted to determine if boat traffic affected the freshwater population of Irrawaddy dolphins in the Mahakam River in East Kalimantan, Indonesia. A coastal Irrawaddy dolphin population in a nearby coastal bay in East Kalimantan, where boat traffic was less intense formed a reference population. Both quantitative and qualitative reactions of dolphins to different types of boats at different distances were measured in both the river and coastal bay habitats. In addition, we examined whether behavioral responses were related to the dolphins’ activities prior to the arrival of boats, group size, or the presence of calves. These comparisons identified conditions that affect dolphin responses. The results facilitated recommendations for boat traffic control in certain areas and for certain boat types; a modest step forward towards the conservation of Irrawaddy dolphins in Indonesia.

METHODS

Study areas

Boat traffic was studied at two sites in East Kalimantan, Indonesia: 1) the middle Mahakam area, 180 km to 375 km from the mouth and 2) Balikpapan Bay (Figure 1). The Mahakam River is one of the major river systems of Kalimantan and runs from 118o to 113o E and between 1o N and 1o S. Regional climate is characterized by two seasons (MacKinnon et al.., 1997), i.e. dry (from July-October, southeast monsoon) and wet (November-June, northwest monsoon), but dry and wet periods alternated during the wet season as well. The river measures approximately 800 km from its origin in the Müller Mountains to the river mouth (MacKinnon et al.,. 1997). The study area was in the middle Mahakam area (from 180 km to 375 km from the mouth) because of the high dolphin densities (Kreb, 2002). Mean river width in this area measured 200 m (SD = 53, range 110-400 m, n = 105), whereas mean water depth at an average water level was 15 m (SD = 6, range 6-37, n = 65). Mean water clarity in the study area (measured with a Secchi disk) at an average water level was 23 cm (SD = 7, n = 27). The middle Mahakam and the connecting lakes systems is an area of intensive fishing activity with an annual catch since 1970 of 25,000 to 35,000 metric tons (MacKinnon et al.., 1997), and the highest dolphin densities are there (Kreb, 2002). Some coal mining and logging activities also occur here, especially in the tributaries. Furthermore, the middle Mahakam area is subject to intense boat traffic with boats passing our stationary observation vessel every 3 min on average, mostly

108 Effects of boats on Irrawaddy dolphins

boats of < 40 hp (Table 1). Infrastructure is poorly developed in Kalimantan and rivers are the main mode of transportation. Hence, in central East Kalimantan, the Mahakam River is the main transport system. Boats most frequently encountered were: small canoes with outboard engines of < 40 hp, boats with inboard engines of < 40 hp and of 40 hp to 800 hp, speedboats with outboard engines ranging from 40 hp to 200 hp, and container-tugboats (>1000 hp). Balikpapan Bay stretches from 116o42 to 116o50 E and 1o to 1o22 S (Figure 1). Water surface area of the bay is approximately 120 km2. Maximum width of the bay is approximately 7 km. During the study period, the Irrawaddy dolphins in the Balikpapan survey area were observed from the observation vessel at locations varying from 2 m to 30 m deep. Average water depth at dolphin sightings within the bay was 14.5 m (SD = 8.0, n = 39) and outside the bay averaged 5.7 m (SD = 3.6, n = 13). Mean water clarity recorded at dolphin sightings in the bay was 170 cm (SD = 57.7, n = 24). Boat traffic was most frequent in the downstream part of the bay, where ferries and speedboats crossed the bay in one lane. Usually, five tankers were present in the bay, but most of the time these were stationary. In one of the tributaries where dolphins occurred daily, speedboats, which frequented an upstream logging company, were encountered. Small fishing boats were in all areas of the bay. Dolphin encounters were more or less equally distributed in the bay.

Table 1. Average number of boats per hours by vessel type in the middle Mahakam area and Balikpapan Bay per daylight hour

Boat types Outboard/ Inboard Speedboat Container All boat types inboard > 40 hp 40-200 hp tugboat Study area < 40 hp > 1000 hp Mahakam 13.0 2.4 4.6 0.7 20.7 Balikpapan Bay 1.7 0.2 1.4 0.1 3.2

Data acquisition

Boat/dolphin interactions were studied in the river during four periods in 2001: 23 June-5 July; 10-24 August; 2-9 September; and 25 October-7 November. Studies on coastal dolphins were conducted between 30 May-10 June and between 2-15 October 2001. These periods were chosen because these included low and medium water levels. The dolphin/boat interactions were hypothesized to be more detrimental in the dry season due to reduced water depth and narrowing of river thereby restricting dolphins’ movements. Therefore, we focused on dolphin/boat interactions during low water levels. The surveys were conducted in high dolphin density areas to maximize the number of group sightings and in three habitat types: main river, tributaries, and confluence areas.

109 Chapter 7

Figure 1. Map of study area at the Mahakam River and Balikpapan Bay, Indonesia

Dolphin observations were conducted from two types of vessels. The first was a wooden boat with a 26-hp inboard engine and 16 m in length with the observer’s eye- height 3.5 m above the water. The second was a wooden canoe 10 m in length, with a 5- hp outboard motor with observer eye-height 1 m above the water. When using the large vessel, a distance of 100 m away from dolphins was kept when the engine was on. However, for the small vessel a distance of 50 m away was maintained based on preliminary work with shore observations that indicated dolphins did not respond at the closer distances for similar types of boats. The boat driver maintained a constant speed, heading direction, and distances to the dolphins. Observations also were made from an

110110 Effects of boats on Irrawaddy dolphins

elevated land-based platform, 7 to 10 m above the water surface (depending on water levels), which provided an unobstructed view of one important dolphin confluence area and connecting areas; 2 km upstream the main river, 500 m downstream the main river, and upstream in the tributary. Dolphin observations were conducted from two types of vessels. The first was a wooden boat with a 26-hp inboard engine and 16 m in length with the observer’s eye- height 3.5 m above the water. The second was a wooden canoe 10 m in length, with a 5-hp outboard motor with observer eye-height 1 m above the water. When using the large vessel, a distance of 100 m away from dolphins was kept when the engine was on. However, for the small vessel a distance of 50 m away was maintained based on preliminary work with shore observations that indicated dolphins did not respond at the closer distances for similar types of boats. The boat driver maintained a constant speed, heading direction, and distances to the dolphins. Observations also were made from an elevated land-based platform, 7 to 10 m above the water surface (depending on water levels), which provided an unobstructed view of one important dolphin confluence area and connecting areas; 2 km upstream the main river, 500 m downstream the main river, and upstream in the tributary. The observation team consisted of four people; observer 1 recorded dolphin surfacing behavior and boat traffic, and observer 2 drew a spatial distribution of the group and recorded distances among individuals. New drawings were made when the spatial distribution changed and the time at which the change occurred was recorded. Two other observers indicated behavioral displays and when dolphins surfaced and boats arrived. After each observation session, water depth was measured at the main location of the dolphins using a sonar fish-finder (Fish-finder 260, Apelco). Dolphin reactions to boats were indicated by changes in surfacing patterns. Two methods were used; First, we compared the average number of individuals surfacing (based on the number of group surfacings divided by group size) per min during the absence and presence of boats. The following data were recorded continuously for at least 15 min: group size, group composition (i.e., presence of neonates, calves or juveniles), individual- and group- behaviors of dolphins, total number of surfacing per min, type and time of boat entering and leaving an area, and distance to the dolphins. Speedboats, container tugboats, and boats with inboard engines > 40 hp were recorded as “present” if they were < 300 m from the core of the dolphin group. Small boats were recorded as “present” when < 100 m distance to the core of the dolphin group. Dolphins, “far away” from the core group (> 50 m to the nearest individual) were not included in the observation session. Since surfacing reactions were measured per boat type, we assigned multiple boat approaches in one min (27 of 397 encounters) to the boat type category with the largest horsepower. This was done since we found that dolphins reacted stronger, i.e., surfaced less in close presence to boats with larger horse-powered engines. Some observation sessions were extended beyond 15 min when there were many boats passing and continued until 5 boat-absent min were obtained. A new session started when a new group was encountered, if during the 15-min session another

111 Chapter 7

group joined, the group split, group behavior changed or if another daytime period started (three session blocks: from 0800-1100 h; 1101-1500 h and from 1501-1800 h). Data collected for less than 15 min were analyzed if the number of boat-absent min was at least five min (see, analysis). Distance from boats to the dolphins was estimated visually by the observer. Distance estimation training was conducted by the observers by estimating distance from one object along the river bank to another. These estimations were cross- checked by a simple calculation based on the boat speed and time traveled between both objects using a GPS and a stop-watch. In the same way, observers now and then referred to floating objects in the river and estimated river width every 15 min to standardize their estimation and error. The second method compared the number of dolphin surfacings per boat by distance classes. For all boat encounters the distance between each surfacing dolphin and the approaching/leaving boat was recorded. Thereafter, a dolphin’s surfacing or not surfacing was marked for each distance class (see below). Distance between boat and surfacing dolphins were estimated visually. For example, if during a speedboat encounter only two dolphins surfaced and the distance to each individual was measured as 170 m and 190 m, then a response was noted in the 151-250 m distance class. All other distance classes were recorded as no surfacings. Two boat categories were distinguished: 1) small, motorized canoes and low-engine boats < 40 hp (for these boats, dolphin behaviors were recorded of each dolphin at 0-25 m, 26-75 m, and 76-100 m from the boat) and 2) speedboats with outboard engines from 40 hp to 200 hp (hp was on the outboard engine and read using binoculars), boats with inboard engines > 40 hp, and container barges (>1000 hp). Dolphin surfacing was recorded for the following large boat distance ranges: 300-251 m, 250-151 m, 150-51 m, and 50- 0 m. The data set from both methods was based on a group-follow protocol with dolphin groups observed for > 30 min. The first method used a focal-group sampling method for continuous assessment of group behavioral displays, predominant group activity (> 50% of observations), and number of surfacings within 1-min time blocks (Mann, 1999). The second method used a one-zero sampling method to determine whether a dolphin surfaced within a certain boat distance class. For both methods, observations were analyzed if the boat traveled at a steady speed. The second method was applied only for single boat approaches within the distance classes at 300 m for speedboats or 100 m for small canoes < 40hp. Occasional underwater recordings were made when boats approached dolphins to define the maximum distance classes for recording surfacing rates for each type of boat. The distance of the boat to the dolphin at which the noise could be heard clearly by the observers was used as the distance range, although dolphins detect sounds over larger distance ranges. For underwater listening a High Tech Inc. 94-SSQ hydrophone (frequency response: 2Hz - 30 kHz, –168 dB re 1V/µPa) was hung over the observation vessel to a depth of 1.5 m.

112 Effects of boats on Irrawaddy dolphins

Data analysis

Average surfacing rates per individual in the absence/presence of boats were calculated per session. These were calculated as the total number of surfacings per min of the group under observation. The number of surfacings per min were summed for all boat-absent min in the session. Likewise, total surfacings were summed for all boat- present min in the 15-min session. These total surfacings per session for boat-absent vs boat-present categories were divided by the number of boat-absent or boat-present min to obtain the average group surfacings per min. The average group surfacing rate was divided by group size to obtain the average surfacing rate per individual, hereafter abbreviated in the text as ‘surfacing rates’. Surfacing rates in the presence vs the absence of boats were compared within each session in which at least one boat passed, which were 37 and 16 sessions in the river and bay, respectively. Surfacing rates were only calculated for the 15-min sessions during which at least 5 min (not necessarily continuously) were free of boat traffic (on average 3.6 boats in the river and 2.0 boats in the bay were recorded in the 15-min sessions with boats present). These surfacing rates without boats were then compared with surfacing rates of those in which one or more boats passed by during the same session. A minimum of 5 boat-absent min per session was chosen to overcome potential biases associated with fluctuations in surfacing rates per min inherent in the natural pattern of surfacing. This was tested by drawing replicated random selections of different sample sizes from an entirely boat-absent session. All ten replicates of the surfacing rates of this sample size (i.e., 5 boat-absent min) fell within the standard deviation of the average surfacing rate, which was based on 15 min. Smaller sample sizes showed significant deviations from the standard deviation. All tests involved non-parametric statistics (Siegel & Castellan, 1988; Fowler & Cohen 1990). Wilcoxon’s matched pair tests was used to determine the effects of presence and absence of several types of boats within a sampling period (in the results section n is the number of pairs less the number of pairs for which no differences were found, i.e., d = 0). The Mann-Whitney U test was used for differences in surfacing rates among different boat types. Effects of group size on the surfacing rates in absence and presence of boats were tested using Pearson’s product moment correlation test. To test for the correlation between group size and strength of reactions to boats, boat reactions were expressed as negative differences between surfacing rates in the absence and presence of boats and tested with Spearman Rank Correlation test. Because the surfacing rates in the presence of boats already proved significantly lower (see, results), only those data when dolphins showed negative reactions to boats, i.e., had lower surfacing rates in presence than in absence of boats, were used. Then, we tested if group composition was an influencing factor on surfacing rates in the presence and absence of boats. In this case, surfacing rates of groups with calves, which were found only within group sizes of 4-8 individuals, were compared with groups without calves of similar group sizes, using the Mann-Whitney U-test. These comparisons were made in the absence and presence of boats. Reactions

113 Chapter 7

of dolphins as expressed in differences in surfacing rates in the absence and presence of boats were compared with the Kruskal-Wallis and Mann-Whitney U-tests for three habitat types in the river, which were the main river, confluence areas, and tributaries. We used Pearson’s product moment correlation test to test if any correlation existed between surfacing rates and depth both in the absence and in presence of boats. Seasonal differences in surfacing rates were tested for low and medium water levels in order to test if surfacing rates, which were collected during different surveys, could be combined and did not bias the overall results using the Mann-Whitney U-test. It was tested if the predominant (core) group activity (> 50% of time per min) during the sampling session affected surfacing rates, both in the absence and in presence of boats with a Mann-Whitney U test. To test for other reactions to boats than surfacing changes, the number of times general group behavior changed after a boat passed during the next minute of dolphins surfacing were recorded for each boat type. These numbers were then compared with the number of times behaviors did not change using Chi-square tests and applying Yates’ correction. Reactions of dolphins per distance class using one-zero sampling and data set 2 (see, methods) were all tested using Chi-square tests and applying Yates’ correction for one-degree of freedom. The number of encounters during which at least one dolphin of the group surfaced for each distance and for each boat type was compared with the number of times that no dolphin surfaced at all for similar distances. Combined distance class comparisons between surfacing and non-surfacing occasions were also made. Moreover, frequencies of dolphin surfacing among distance classes were compared. All test results were analyzed for two-tailed values and significance was assumed when p ≤ 0.05.

RESULTS

Changes in surfacing rates

Observation effort data are presented in Table 2. Boat traffic in the Mahakam River was greater than recorded for Balikpapan Bay. In the bay, during 75% of the 15-min sessions (total n = 66) no boats entered within the defined dolphin-boat distances for different boat types (i.e., 300 m for speedboats, container-tug boats and boats with inboard engines > 40 hp, and 100 m for small vessels of < 40 hp), whereas in the river only 15% of the sessions (total n = 58) were completely boat-absent. Boat traffic in the Mahakam River (mean = 20.7 boats/h) is 6.5 times more intensive than in the Balikpapan Bay (mean = 3.2 boats/h). In the river, canoes with outboard engines of < 40 hp were most frequent and secondly speedboats. In the bay, vessels with inboard engines of < 40 hp and speedboats were equally common or rare (Table 2). Most encounters between dolphins and boats involved boats with inboard engines < 40 hp or speedboats of 40 hp (Table 2).

114 Effects of boats on Irrawaddy dolphins

Table 2. Observation effort and number of Irrawaddy dolphin/ boat encounters in the Mahakam area and Balikpapan Bay

Study area Session n sessions n sessions n sessions n boat/ dolphin time (h) with boats without boats encounters River 14 58 49 9 343 Bay 13 66 16 50 54 Total 27 124 65 59 397

In the river, the mean surfacing rate per individual was significantly greater in the absence of boats (0.97 surfacing/min) than in their presence (0.74 surfacing/min) (Wilcoxon matched pairs, t = 98, n = 33, p < 0.002). In contrast, in the coastal bay, there were no significant differences in surfacing rate with 0.89 surfacing/min for both conditions (Wilcoxon matched pairs, t = 62, n = 16, p > 0.1). Dolphins in the river significantly surfaced less in presence of boats of < 40 hp (0.67 surfacing/min, t = 61, n = 25, p = 0.02), speedboats of 40-200 hp (0.55 surfacing/ min, t = 9, n = 13, p < 0.02), and boats tugging large containers of > 1000 hp (0.47 surfacing/min, t = 1, n = 7, p = 0.05), when testing per boat type (all Wilcoxon matched pairs). In decreasing order, river dolphins surfaced least often in the presence of container tugboats, speedboats, and motorized canoes, but the group differences were not significant (Kruskall-Wallis, H = 4.75, df = 2, p > 0.05). On the other hand, differences in the medians (all sessions) of surfacing rates (per session) between the first and last boat types were found to be significant (Mann-Whitney U-test, U = 30, z = 2.15, n1 = 6, n2 = 24, p = 0.015). However, river dolphins did not surface less in the presence of vessels traveling at medium speed with inboard engines of > 40 hp (1.16 surfacing/min, t = 26, n = 10, p > 0.05). Our observation vessel (with outboard motor of 5 hp) approached within the defined distances on one session and dolphins were found to surface at the same rate as in the absence of boats (0.93 surfacing/ min). Dolphins within the bay did not surface more or less frequently in the presence of speed boats of 40-200 hp (0.92 surfacing/min, t = 22, n = 11, p > 0.1) or boats with inboard engines of < 40 hp (0.74 surfacing/min, t = 16.5, n = 9, p > 0.1). These were the only boats that were commonly encountered besides ferries, of which no encounter with dolphins was observed. Surfacing rates in the presence of boats were compared among three boat types; there was a significant difference in presence and absence of boats in the river environment. Dolphins surfaced least often in the presence of container tugboats, then speedboats and motorized canoes of < 40 hp. Group differences were not significant (Kruskall-Wallis, H = 4.75, df = 2, p > 0.05), but differences in the medians (all sessions) of mean individual surfacing rates (per session) between the first and last boat types were significant (Mann -Whitney U-test, U = 30, z = 2.15, n1 = 6, n2 = 24, p = 0.015). Surfacing rates per min in the absence of boats for river and coast were similar, i.e., 0.97 and 0.98 times (CV = 18%, range 0.32-

115 Chapter 7

2.0, n = 36; CV = 21%, range 0.2-2.5, n = 70) during milling and slowly swimming behaviour.

Surfacing reactions in relation to distance between dolphins and boats

Table 3 and 4 present the number of dolphin/ boat encounters by boat type and dolphin reactions by distance class (also visualized in Figure 2 & 3). Reactions for speedboats with different horsepower engine are combined as dolphins showed significant reactions to all types of speedboats: Dolphins significantly did not surface at all when approached at < 300 m by speedboats of 40 hp (X2 = 31.2, df = 1, p < 0.0001); 85 hp (X2 = 26.4, df = 1, p < 0.0001), 115 hp (X2 = 53.2, df = 1, p < 0.0001) and 200 hp (X2 = 25.1, df = 1, p < 0.0001). Speedboats of 40 hp were most frequently encountered (34 times), then 115 hp (20 times), 200 hp (11 times) and 85 hp (10 times). In the coastal bay, only speedboats of 40 hp were encountered (26 times). Coastal dolphins significantly did not surface when speedboats passed within a ≤ 300 m radius to the group (in relation to the group’s last observed position) (X2 = 22.3, df = 1, p < 0.0001). River dolphins did not surface at all in the presence of small motorized canoes, speed boats, and other boats (> 40 hp) (X2 = 32.3, df = 1, p < 0.01; X2 = 136.4, df = 1, p < 0.01; X2 = 1.8, df = 1, p < 0.01). Exceptions were dolphin reactions to tugboats and the observation vessel for which no significant differences were found between the number of times that even one dolphin surfaced in the presence of these boats.

Table 3. Irrawaddy dolphin reactions to different types of boats of > 40 hp by distance class.

Boat Boat Speedboat 40-200 hp Inboard > 40 hp >1000 hp Position Distance (m) River; n = 75 Coast; n = 26 River; n = 38 River; n = 10 A 300-250 n.s. n.s. n.s. n.s. A 250-150 s n.s. n.s. n.s. A 150-50 s n.s. n.s. n.s. A & L 50-0; 0-50 s s s n.s. L 50-150 s s s n.s. L 150-250 s s n.s. n.s. L 250-300 s s n.s. n.s. Total all distances s s s n.s. A = approaching boat; L = leaving boat; n.s. = non-significant reaction, i.e., for most encounters at that distance class at least one dolphin of the group surfaced during the encounter; s = significant reaction to boats (P < 0.05), i.e., for most encounters at that distance class no dolphin surfaced during the encounter

116 Effects of boats on Irrawaddy dolphins

Table 4. Irrawaddy dolphin reactions to boats of < 40 hp by distance class for river vs coastal habitats

<40hp Distance (m) River; n = 216 Coast; n = 28 A 100-75 s* n.s. A 75-25 n.s. n.s. A/L 25-0; 0-25 s s L 25-75 s s L 75-100 n.s. n.s. Total all distances s n.s.

Surfacing reactions for river dolphins are presented in Figure 2 for each distance class for small canoes of < 40 hp. The impact from boats of < 40 hp in the river and bay for both situations increased with shorter distances and was largest within 25-0 m distance of the dolphins when a boat was near and within 0-75 m after a boat passed. River dolphins surfaced more often before than after boats passed (X2 = 12.86, df = 1, p < 0.01).

Figure 2. Irrawaddy river dolphins surfacing reactions to boats of < 40 hp. Distance classes from left to right correspond to the sequence of the approach of a boat.

117 Chapter 7

Figure 3. Irrawaddy river dolphin surfacing reactions to speedboats. Distance classes from left to right correspond to the sequence of the approach of a boat

Figure 3 shows surfacing reactions of river dolphins to speedboats by distance class. The number of occasions when at least one dolphin surfaced increased stepwise for each increasing distance class. No significant differences were found in dolphin surfacings prior to or after boats passed. River dolphins react significantly when speedboats enter an area within 250 m until they leave the area within 300 m distance of the dolphins, whereas coastal dolphins only react significantly when the boat approached with a 50-m distance and until it has left the area at 300 m distance. For other boats of > 40 hp no significant differences were found in dolphin surfacing among the distance classes. These boats caused most strong reactions in dolphin surfacing behavior by not surfacing at all when approaching at a 50-m distance until leaving an area of 150 m distance to the dolphins.

Other factors than boat presence influencing surfacing rates

Surfacing rates when boats were absent in the river were significantly higher with greater group size (Product Moment Correlation, r = 0.36, df = 39, p = 0.05) (Figure 4). Surfacing rates in the presence of boats on the other hand, did not correlate with group size, and these were similar for small or large groups (Product Moment Correlation, r = 0.13, df = 35, p > 0.05). No significant relation was found between

118 Effects of boats on Irrawaddy dolphins

group size and strength of reactions to boats (expressed as differences between individual surfacing rates in the absence and presence of boats) (Spearman Rank Correlation, rs =0.149, n = 28, p > 0.05). Mean dolphin group size over all sessions was 4.6 (n = 48, SD = 2.0, range 1 - 10).

Surfacing rates in absence of boats

al al 2,5 Predicted surfacing rates in absence of 2 boats

1,5 /min) n ( 1

0,5 ean surfacing rates per individu per rates surfacing ean

M 0 024681012

Group size

Figure 4. Irrawaddy river dolphin group size and mean surfacing rates per min per individual.

We also checked if group composition was an influencing factor, by comparing surfacing rates of groups with calves (0.95 surfacing/ min), which were found only within group sizes of 4-8 individuals, to groups without calves of similar group sizes (1.22 surfacing/ min). In the absence of boats, no significant differences were found between the means of the two samples (Mann-Whitney U, U = 88, z = -0.93, n1 = 13, n2 = 11, p > 0.05). Also, in the presence of boats, groups with vs without calves did not have significantly different average individual surfacing patterns, i.e., with 0.70 surfacing/ min and 0.79 surfacing/ min, respectively (Mann-Whitney U, U = 79, z = - 0.41, n1 = 13, n2 = 11, p > 0.05). Dolphin reactions (differences between surfacing rates in presence and absence of boats) also were compared between three types of habitat within the river, i.e., main river, confluence area, and tributary, of which the mean width for the first and last were 200 m (SD = 54 m) and 43 m (SD = 13 m), respectively. No significant differences in reactions were found between each habitat type (Kruskal-Wallis, H = 0.67, df = 2, p > 0.05). No correlation was found between depth and the surfacing rates of dolphins in the river and in the bay, both in absence (Product Moment Correlation, r = 0.54, df = 3, p > 0.05 & r = 0.155, df = 38, p > 0.05) or presence of

119 Chapter 7

boats (Product Moment Correlation, r = 0.121, df = 47, p > 0.05 & r = 0.466, df = 8, p > 0.05). Median surfacing rates were similar during low and medium water levels both in absence and presence of boats (Mann-Whitney U, U = 128, z = -5.68, n1 = 10, n2 = 10, p > 0.05 & U = 118, z = -5.1, n1 = 10, n2 = 10, p > 0.05). When testing the influence of behavior on surfacing rates, dolphins engaged in slowly swimming or milling activities appeared to surface similarly in the absence of boats (Mann-Whitney U, U = 117, n1 = 20, n2 = 16, p > 0.05). Likewise, behavior did not influence surfacing rates when boats passed by (Mann-Whitney U-test, U = 13, n1 = 20, n2 = 16, p > 0.05). Finally, we tested if dolphins showed other reactions in response to boat traffic than changes in surfacing rates. For all boat types and cases when the predominant group behavior changed after a boat passed within the next minute were compared with the number of times behaviors did not change. Apparently, dolphins did not change their predominant group behaviors when boat approached within our pre- defined distances (see, methods). Only in 4 out of 130 boat encounters recorded, did group behavior change after a boat passed.

Potential impacts of the observation vessels

The presence of the large observation vessel (at > 100 m distance) and the small observation vessel (at > 50 m distance to the dolphins) did not influence dolphin surfacing rates; i.e., there were no significant differences in median surfacing rates when recorded from the observation vessel or when recorded from the shore (Mann- Whitney U, U = 143, n1 = 25, n2 = 11, p > 0.05).

Qualitative responses to boat traffic

Unfortunately, data on dolphin surfacing in the presence of large container tugboats were only collected during seven sessions. This had to do with the fact that the boats were mostly encountered in one narrow tributary (and adjacent confluence area) where sessions at most times had to stop for safety reasons. Therefore, most encounters between dolphins and container ships were documented according to protocol from an observation bridge in the confluence area of Muara Pahu in primary dolphin habitat. Nevertheless, a number of observations were made other than recording the number of surfacings, which are worthwhile mentioning. During medium water levels, on average four empty container boats and 4 heavily-loaded large, container boats each day moved up and upstream the narrow tributary, which represents primary dolphin habitat. During low water levels, about ten smaller, container tugboats moved up and downstream each day. When a container boat passed the confluence area to move up or downstream of the tributary, dolphins usually anticipated by swimming a short distance away from the boat’s heading

120 Effects of boats on Irrawaddy dolphins

direction. Because the distance was too short to define this behavior as slowly swimming, the general behavior on the data sheet was still regarded as milling. Reactions to container boats in the narrow tributary of Kedang Pahu (mean width of river = 45 m at medium water levels; mean depth = 5 m at the locations where, and in the period when the boat encounters were recorded), are even more conspicuous than in the deeper confluence area (average depth of 15 m). Dolphins changed their swimming direction or waited to allow a container boat, which was moving in the same direction as the dolphins to pass by and then resumed their swimming direction. However, there also were cases when dolphins increased their swimming speed if they were moving downstream with a container-tugboat following from behind, in this way arriving ahead of the boat in the confluence area of main river and tributary. When the dolphins moved upstream and encountered more than one container-tugboat in sequence that moved downstream, they also moved downstream ahead of the first container boat. During one occasion, one group of five dolphins moving downstream encountered a container-tugboat moving upstream and they reacted by swimming fast, surfacing almost continuously, producing loud blows, displacing much water, and dove until the boat approached at a 10-m distance away. In the narrow tributary with many river bends, the noise of an empty container- tugboat was heard underwater using a hydrophone at 150 m distance and dolphins were still heard vocalizing. However, at a 100-m distance no dolphins were heard, and at a 80-m distance the sound was becoming uncomfortably loud for humans. According to Gordon & Moscrop (1996) belugas are supposed to suffer discomfort at received levels of about 140-160 dB. When the container boat passed, the noise level immediately dropped and dolphins were heard vocalizing again at 70-m distance from the boat. Increasingly uncomfortable noise levels for the human ear were caused by speedboats in the same tributary at 300 m distance. At the same time no dolphin was heard vocalizing until the boat passed at a 150-m distance when noise levels also dropped. Whether they could still echolocate was undetermined. On a few occasions speedboats appeared after a river bend and this caused startled responses and immediate dives by the dolphins.

DISCUSSION

Methodological constraints

One of the shortcomings of the analyses occurred in cases when there were two boats passing by during the same minute (7% of all recorded encounters) and the average individual surfacing rates per min were entered in only one boat type category. The largest horsepower boat type category was chosen because dolphins surfaced least often in the presence of container tugboats, then speedboats of > 40 hp and motorized canoes of < 40 hp (see, results, changes in surfacing rates). Another shortcoming was that surfacing rates during the 15-min sessions were counted for

121 Chapter 7

each minute using real time. So, if a boat was present within the defined distance in the first or last thirty seconds of a minute, this was recorded as a boat-minute, which may cause differences between surfacing rates in absence and presence of boats to be diluted.

Interpreting results

Surfacing rates decreased most significantly during encounters with container boats, secondly with speedboats and then small, motorized canoes of < 40 hp. Although the impacts of the first two boat types are more intense, boats of the last category also caused significant reactions in surfacing rates. Motorized canoes are by far the most frequent and are in this respect an important factor of disturbance. A reason why these small boats of < 40 hp evoked significant reactions, while larger boats > 40 hp, other than speedboats and container boats, did not is probably that the small boats use outboard engines that produce high-frequency sounds, e.g., 5 kHz, and hearing sensitivity of small cetaceans improves with increasing frequency (Richardson et al.., 1995). Coastal Irrawaddy dolphins in Australia produced whistles between 1 and 8 kHz and were mostly heard during both foraging and socializing behaviors (Van Parijs et al.., 2000). The frequency of the noise produced by outboard vessels is within this same range (see, above) and therefore most likely is a disturbing factor in the dolphins foraging and socializing activities. Another likely factor for outboard vessels of < 40 hp is that these boats often move fast and make sudden changes in speed and direction. Differences in dolphin reactions to boats (expressed in differences in average individual surfacing rates in the absence vs presence of boats) between coastal and river habitats could be a result of habituation to noise for the last group since boat traffic was almost 7 times more intensive in the river. Gordon & Moscrop (1996) suggested that dolphins either become habituated to the sound and show less response, or show an increasing level of disturbance with exposure.

Implications for conservation

This is the first detailed quantitative study on boat disturbances of freshwater dolphins. Boat traffic in the Mahakam River was intense with 20.7 boats per hour passing on average, and 6.5 times more frequent than in the Balikpapan Bay, with only 3.2 boats passing on average per hour. The greatest disturbers were speedboats and container tugboats. These boats were particularly dangerous as they moved in a narrow tributary representing major dolphin habitat, in which dolphins may experience great difficulty in evading both their physical presence and noise produced by these vessels. When speedboats pass in confluence areas and river bends dolphin/ vessel collisions can occur, boats and their sounds may appear too suddenly and at

122 Effects of boats on Irrawaddy dolphins

short distance after these bends for dolphins to avoid a collision. During this study one juvenile dolphin was found dead with wounds thought to be inflicted by a boat’s propeller. Other disturbers are the canoes with outboard engines of < 40 hp because of their high rate of trespassing and sudden changes in courses and speed. In the most important core area of the river dolphins, i.e., the confluence area of Muara Pahu (Kreb, unpublished data), boat traffic of all types discussed here, was particularly intense. Besides the time, which was spent to record dolphin/boat encounters (Table 1), an additional 30 observation days (from 0800 to 1800 h) was spent in this confluence area to study habitat use (Kreb, in press), where the highest density of dolphins was recorded, during six periods from January 2000 until November 2001 and at different water levels. On average, three different groups (range between two and six groups) of dolphins frequented the confluence area daily for an average of 42 % of observation daytime (Kreb, in press). Dolphins’ continuous presence in this intensive boat traffic area does not mean that they are not disturbed, but it rather underlines the importance of that area to the dolphins (Brodie, 1989). However, frequent interruption of dolphin feeding, resting, or socializing through active boat avoidance, disrupted echolocation signals for safe navigation and active hunting, and masking of acoustic cues to hunt passively and to locate group members for maintaining social cohesion and coordination, may induce stress of which the long-term physiological effects are still unknown (Richardson et al.., 1995). Prolonged exposure to sound-induced stress for terrestrial mammals has reportedly led to harmful effects in digestive and reproductive organs and similar effects are suggested to be likely for cetaceans (Gordon & Moscrop, 1996). Another threatened area is a narrow and shallow tributary, i.e., Kedang Pahu River that also represents primary dolphin habitat (Kreb, in press). During the dry season, when container barges pass through this tributary, they occupy three-quarters of the total river width. When dolphins encounter these boats in this limited area they are possibly exposed to noise levels that may cause a temporary hearing threshold shift, perhaps even permanent hearing loss. Gordon & Moscrop (1996) report that for belugas, noise levels of 140 dB re 1µ Pa at their most sensitive frequencies are assumed sufficient to cause threshold shifts. Most dominant echolocation clicks for Irrawaddy dolphins are between 50-75 kHz (Kamminga et al.., 1983), but these boats are not likely to produce noise in that frequency range. Although small cetaceans are more likely to be affected by high-speed vessels that produce high-frequency noise, the noise produced by tugboats is clearly annoying for the human ear and at least interferes with vocalizations (medium frequencies) for social communication. Similar interference likely applies to speedboats. Noise levels of ca. 142 dB re 1µ Pa by an outboard engine of 70 hp at 50 m distance were estimated in the frequency range of 400 Hz-4 KHz (Gordon & Moscrop, 1996). From 1998 until 2001, at least two juvenile dolphins died in the Mahakam as a result of injuries due to a vessel collision, most likely with speedboats (Kreb, in press). Also in other species, vessel collisions were described: Damage from ship propellers reportedly accounted for 6.5 % of baji deaths (Chen Peixun, 1989). Wells & Scott

123 Chapter 7

(1997) documented vessel collisions between small boats and bottlenose dolphins. Young belugas were found to be less likely than adults to react to boats approaching at high speed (Richardson et al.., 1995). Although our results show that dolphins generally are aware and anticipate approaching high-speed vessels by long dives, very occasionally (1 out of 75 cases) one dolphin surfaced within 0-25 m distance of the speedboat. Startle responses also were observed if a speedboat appeared around a river bend at close range of the dolphins; they made immediate dives. Another dangerous situation occurs when dolphins were engaged in activities with many surface behaviors such as mating; on such occasions the dolphins did not attempt to avoid fast moving vessels. Shane (1990) also stated that dolphins least commonly altered their behavior in response to boats when they were actively socializing.

CONCLUSIONS

In conclusion, Irrawaddy dolphins in the Mahakam showed significant changes in their surfacing pattern in reaction to various approaching boat types. Some likely ultimate consequences such as stress, temporary or permanent hearing loss, occasional injuries or death by vessel collisions, are serious threats to this freshwater dolphin population. A possible positive development is that there are plans to improve a previously existing, but hardly used road along the river, which leads to most coal- and gold-mining companies upstream. Thus, road traffic may become an alternative to river traffic. An ongoing conservation program to protect important dolphin habitat was initiated by a local non-governmental organization in cooperation with local government by the end of 2000. The conservation program also plans to place board signs for boat speed reduction in important dolphin areas in close cooperation with local residents and speedboat owners. Hopefully, in the near future, reduction of noise, and physical harassment by boats will prevent displacement of dolphins from these biologically important areas.

ACKNOWLEDGMENTS

We are grateful for the cooperation and permission to conduct these surveys from the Indonesian Institute of Sciences (LIPI, Jakarta), the Directorate General for Nature Conservation and Forest Protection (PKA), the Nature Conservation Department (BKSDA) in Samarinda and the University of Mulawarman (UNMUL). We thank Prof. A.A. Bratawinata (UNMUL) as a counterpart of the project. We are grateful for the enthusiasm and hard work of the field assistants, Ir. Budiono, Ir. Arman, Ir. Syahrani, Ir. Ahang, Marzuki, Deni and Yusri. A special thanks for our boatsmen Pak Muis. In particular we thank Prof F.R. Schram (Institute for Biodiversity and Ecosystem Dynamics, Amsterdam), Dr. P.J.H. van Bree (Zoological Museum, Amsterdam), Dr. E.N. Megantara, Ir. B. Suhendar, M.Sc. (Padjadjaran University) and

124 Effects of boats on Irrawaddy dolphins

Dr. T.A. Jefferson (NMFS/ SFSC, La Jolla, USA) for providing scientific guidance. T. Prins and T. Dunselman (Zoological Museum) provided administrative support. The Plantage Library in Amsterdam is thanked for their help in accessing all literature. Earlier versions of this manuscript greatly improved by the comments of Prof. F.R. Schram, Dr. M. Genner, Dr. V. Nijman (all from the Institute for Biodiversity and Ecosystem Dynamics, Amsterdam), Dr. R.S. Wells, Dr. J.A. Thomas and an anonymous reviewer. Financial support during this period was provided by Ocean Park Conservation Foundation in Hong Kong, Martina de Beukelaar Stichting, Stichting J.C. van der Hucht Fonds, Gibbon Foundation, and the Netherlands Program International Nature Management (PIN/KNIP) from the Ministry of Agriculture, Nature Management and Fisheries.

REFERENCES

Allen, M. C. & Read, A. J. 2000. Habitat selection of foraging bottlenose dolphins in relation to boat density near Clearwater, Florida. Marine Mammal Science 16: 815- 824. Bejder, L., Dawson, S. M. & Harraway, J. A. 1999. Responses by Hector’s dolphins to boats and swimmers in Porpoise bay, New Zealand. Marine Mammal Science 15: 738-750. Brodie, P. F. 1989. The white whale, Delphinapterus leucas (Pallas, 1776). In: S. H. Ridgeway & Sir R. Harrison (eds.), Handbook of marine mammals: River dolphins and the larger toothed whale (Volume 4). Pp. 119-144. Academic Press, London. Buckstaff, K. C. 2003. Effects of watercraft noise on the acoustic behavior of bottlenose dolphins, Tursiops truncatus, in Sarasota Bay, Florida. M.Sc. thesis, University of California, Santa Cruz. 41 pp. Chen Peixun 1989. Baiji, Lipotes vexilifer Miller, 1918. In: S. H. Ridgeway & Sir R. Harrison (eds.), Handbook of marine mammals: River dolphins and the larger toothed whale (Volume 4) Pp. 25-43. Academic Press, London. Fowler, J. & Cohen, L. 1990. Practical statistics for field biology. Open University Press, Philadelphia. Gordon, J. & Moscrop, A. 1996. Underwater Noise Pollution and its significance for whales and dolphins. In: M. P. Simmonds, J. D. Hutchinson (eds.), The conservation of whales and dolphins. Pp. 281-19. John Wiley & Sons Ltd., University of Greenwich, UK. Gregory, P. R. & Rowden, A. A. 2001. Behaviour patterns of bottlenose dolphins (Tursiops truncatus) relative to tidal state, time-of-day, and boat traffic in Cardigan Bay, West Wales. Aquatic Mammals 27: 105-113. Hilton-Taylor, C. 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, U.K. Janik, V.M. & Thompson, P. M. 1996. Changes in surfacing patterns of bottlenose dolphins in response to boat traffic. Marine Mammal Science 12: 597-602.

125 Chapter 7

Kamminga, C., Wiersma, H. & Dudok van Heel, W. H. 1983. Investigations on cetacean sonar VI. Sonar sounds in Orcaella brevirostris of the Mahakam River, East Kalimantan, Indonesia; first descriptions of acoustic behaviour. Aquatic Mammals 10: 83-94. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Zeitschrift für Säugetierkunde 64: 54-58. Kreb, D. 2002. Density and abundance of the Irrawaddy Dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. The Raffles Bulletin of Zoology, Supplement 10: 85-95. Kreb, D. (in press) Conservation management of small core areas: key to survival of a critically endangered population of riverine Irrawaddy dolphins in Borneo. Oryx. Lesage, V., Barette, C., Kingsley, M. C. S. & Sjare, B. 1999. The effect of noise on the vocal behavior of belugas in the St. Lawrence River Estuary, Canada. Marine Mammal Science 15: 65-84. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press. Mann, J. 1999. Behavioral sampling methods for cetaceans: a review and critique. Marine Mammal Science 15: 102-122. Nowacek, S. M., Wells, R. S. & Solow, A. R. 2001. Short-term effects of boat-traffic on Bottlenose dolphins, Tursiops truncatus, in Sarasota Bay, Florida. Marine Mammal Science 17: 673-688. Richardson, W. J., Greene, Jr., C. R., Malme, C. I. & Thomson, D.H. 1995. Marine mammals and noise. London: Academic Press. Shane, S.H. 1990. Behaviour and ecology of the bottlenose dolphin at Sanibel Island, Florida. In: S. Leatherwood & R. R. Reeves (eds.), The bottlenose dolphin. Pp. 245- 265. London: Academic Press. Siegel, S., & Castellan, Jr., N. J. 1988. Non-parametric statistics for the behavioral sciences (Second edition). McGraw-Hill Book Company. Smith, B.D. 1993. 1990 status and conservation of the Ganges river dolphin Platanista gangetica, in the Karnali River, Nepal. Biological Conservation 66: 159-169. Stacey, P.J. 1996. Natural history and conservation of Irrawaddy dolphins, Orcaella brevirostris, with special reference to the Mekong River, Lao P.D.R.. M.Sc. thesis, University of Victoria. Stacey, P.J., & Arnold, P. W. 1999. Orcaella brevirostris. Mammalian Species 616: 1-8. Van Parijs, S. M., Parra, G. J. & Corkeron, P. J. 2000. Sounds produced by Australian Irrawaddy dolphins, Orcaella brevirostris. Journal Acoustical Society of America 108 (4): 1938-1940. Wells, R.S. & Scott, M. D. 1997. Seasonal incidence of boat strikes on bottlenose dolphins near Sarasota, Florida. Marine Mammal Science 13: 475-480. Zhou, K., & Li, Y. 1989. Status and aspects of the ecology and behaviour of the baiji, Lipotes vexilifer, in the Lower Yangtze River. In: W .F. Perrin, R. L. Jr., Brownell, K. Zhou, J. Liu (eds.), Biology and conservation of the river dolphins. Occasional Papers of the IUCN Species Survival Commission, 3. Pp. 86-91. IUCN, Switzerland.

126 Marked declines in populations of Irrawaddy dolphins

CHAPTER 8

Marked declines in populations of Irrawaddy dolphins

Brian D. Smith, Isabel Beasley, Daniëlle Kreb

Oryx 37 (4), pp. 401, 2003 (with additions)

Coastal Irrawaddy dolphins held in captivity in Oasis Seaworld, Laem sing, Thailand. Live captures pose a serious threat to the fragmented populations in South East Asia. In Indonesia, Irrawaddy dolphins from the Mahakam River were held for live-display in Jaya Ancol, Jakarta between 1974 and 2000. At present no dolphins have remained alive. Photo: Daniëlle Kreb

127 Chapter 8

ABSTRACT

Recent population assessments of Irrawaddy dolphins Orcaella brevirostris in South East Asia were conducted in three major rivers, i.e., the Mahakam (Indonesia), Mekong (Vietnam, Cambodia, Laos) and Ayeyarwady (Myanmar) and Songkhla Lake (Thailand). All populations are faced with drastic declines in numbers and ranges and face several threats but death through gillnet entanglement is the main cause for these declines. In all areas, site-based research and conservation projects have been initiated by local and international NGOs in cooperation with government agencies. Generally, throughout their range, humans are positively inclined towards the dolphins, which may aid in their conservation.

RINGKASAN

Pengamatan terkini lumba-lumba Irrawaddy (Orcaella brevirostris) di Asia Tenggara dilakukan pada tiga sungai besar, yaitu Mahakam (Indonesia), Mekong (Vietnam, Kamboja, Laos), Ayeyarwady (Myanmar) dan Danau Songkhla (Thailand). Seluruh populasi dihadapkan pada penurunan yang drastis dalam jumlah dan daerah hidup dan menghadapi sejumlah ancaman. Namun kematian yang disebabkan oleh terperangkap jala adalah penyebab utama dari penurunan jumlah tersebut. Di seluruh daerah, penelitian didasarkan pada tempat dan proyek perlindungan telah diprakarsai oleh LSM lokal dan internasional bekerjasama dengan perwakilan pemerintah. Pada umumnya, untuk seluruh ruang lingkupnya, manusia setuju untuk melindungi lumba- lumba, yang mana hal ini dapat membantu usaha konservasi.

Conservation of Irrawaddy dolphins in South East Asia

Irrawaddy dolphins Orcaella brevirostris occur in some of the larger rivers and marine appended lakes in South-east Asia, as well as in coastal waters of the Indo-Pacific (Figure 1). Although the species is categorized as Data Deficient on the IUCN Red List, recent population assessments conducted in the Mahakam, Mekong and Ayeyarwady rivers and Songkhla Lake indicate alarming declines in their numbers and ranges and ongoing and pervasive threats to their long-term persistence. Based on an intensive set of surveys conducted over 1999–2002, the population of Irrawaddy dolphins in the Mahakam River, Indonesia, was estimated to be < 50, generally confined to a 200 km segment of the mainstream plus connecting tributaries. Meanwhile, the population has been subject to a mean annual mortality rate of > 10%, with 80% of deaths attributed to gillnet entanglement. Prey depletion by non-selective fishing practices such as poisoning and electrocution may increase the temptation for dolphins to prey on fish caught in gillnets. Large coal barges towed by tugboats also

128

Marked declines in populations of Irrawaddy dolphins

physically block dolphin movements in tributaries, where many of the remaining animals occur. On the basis of similarly intensive surveys in the Mekong River from February 2001 to April 2003, the population was estimated to be < 100, confined during the dry season to a 190 km segment between Kratie, Cambodia, and Khone Falls just upstream of the Laos-Cambodian border. Eleven dolphin carcasses were recovered in the past 10 months, at least six of which were attributed to entanglement in gillnets. A survey of the entire length of the Ayeyarwady River, Myanmar, in December 2002 recorded only eight dolphin groups and 37 individuals, in a 373 km long segment. This represents > 50% decline from their reported historical range. Meanwhile, > 1,200 gillnets were documented in the river with a significantly higher net density recorded in areas where dolphins have apparently been extirpated. Almost 900 gold mining operations were also recorded within the area of current dolphin distribution. These operations use mercury to amalgamate the gold. Accidental introduction of this element into the river could have profound toxic effects on the animals, especially given the bio-accumulative properties of this trace metal and the dolphin’s position at the top of the aquatic food chain. Excessive noise from these operations also interferes with the ability of dolphins to navigate and detect and catch their prey. In addition, the operations introduce, break-up, and redistribute large quantities of sediment, causing major changes in the geomorphologic and hydraulic attributes of the river. In May 2000 and February 2001 extensive surveys of Songkhla Lake recorded only four dolphin groups, with the largest composed of eight individuals. Between 1990 and 2001, 28 dolphin carcasses were recovered. At least 13 of these died from entanglement in gillnets. Six dolphins were also found dead in the first 6 months of 2003. One of these was pregnant and had its flukes cut off, probably to extract it from a gill net. These findings indicate that freshwater populations of Irrawaddy dolphins are at a critical juncture and that immediate conservation action is required to prevent their extirpation. The greatest threat to these animals, and probably all small cetaceans, is entanglement in gillnets. However, despite the grim situation facing these populations there are reasons for hope. In collaboration with local NGOs and government agencies the Whale and Dolphin Conservation Society and Wildlife Conservation Society, among others, are establishing site-based research and conservation projects in all three rivers where the dolphins occur and in Songkhla Lake. There has also been a great increase in local awareness of the conservation importance of these animals. The Irrawaddy dolphin has been adopted as the official mascot of the Phattaloug Province, which borders Songkhla Lake, and the Queen of Thailand has declared the animals a Royal Protected Species. People living in the Mahakam Basin attribute the animals to a human origin and the species participates in a cooperative fishery with throw-net fishermen in the Ayeyarwady River. Throughout their freshwater range in South-east Asia local people generally revere Irrawaddy dolphins. The challenge is to channel these positive sentiments into effective conservation action.

129 Chapter 8

130 and own observations. Map with Irrawaddy dolphin distribution in South East Asia. Black dots representing actual records from literature

Figure 1.

Chapter 8

130 Marked declines in populations of Irrawaddy dolphins

REFERENCES USED

Beasley, I., S. Chooruk, N. Piwpong, N. 2002. The status of the Irrawaddy dolphin, Orcaella brevirostris, in Songkhla Lake, Southern Thailand. The Raffles Bulletin of Zoology, Supplement 10: 75-83. Kreb, D. 2002. Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. The Raffles Bulletin of Zoology, Supplement 10: 85-95. Kreb, D. (in press) Conservation management of small core areas: key to survival of a critically endangered population of riverine Irrawaddy dolphins in Borneo. Oryx. Smith, B.D. and Hobbs, L. 2002. Status of Irrawaddy dolphins Orcaella brevirostris in the upper reaches of the Ayeyarwady River, Myanmar. Raffles Bulletin of Zoology, Supplement 10: 67-74. Smith, B.D. and Jefferson, T.A. 2002. Status and conservation of facultative freshwater cetaceans in Asia. The Raffles Bulletin of Zoology, Supplement 10: 173-187. Smith, B.D., Beasley, I., Buccat, M., Calderon, V., Evena, R., Lemmuel de Valle, J., Cadigal, A., Tura, E., Visitacion, Z. 2004. Status, ecology and conservation of Irrawaddy dolphins Orcaella brevirostris in Malampaya Sound, Palawan, Philippines. Journal of Cetacean Research and Management 6(1): 41-52.

131 Chapter 8

132

Social dynamics of facultative river dolphins

CHAPTER 9

Social dynamics of facultative Irrawaddy river dolphins (Orcaella brevirostris) in Borneo: Impacts of habitat

Submitted manuscript

Coastal Irrawaddy dolphins mating in Balikpapan Bay. This particular mating event lasted about 1,5 h and involved several groups totaling 13 individuals. The individual in the center has its belly turned up.

133 Chapter 9

ABSTRACT

Information on social structures and breeding strategies within cetaceans and especially (facultative) river dolphin species is sparse. The impacts of habitat on social structures of coastal and freshwater populations of Irrawaddy dolphins, Orcaella brevirostris, in East Kalimantan were studied between 1999 and 2002. River and bay habitats differed in three aspects, respectively; 1) constrained vs open geographical shape; 2) year-round vs seasonal high food abundance and 3) clumped vs scattered food resources. These conditions favored a lesser degree of sociality of “bay” dolphins indicated by smaller groups and less frequent inter-group interactions; seasonal breeding and a lesser degree of sexual dimorphism, which may result in less competition among males and probably in one, polyandrous mating system. Inter- group interactions of “bay” dolphins were mostly functional, i.e., feeding, whereas river dolphin inter-group interactions were of varying nature, i.e., feeding, traveling, socializing, agonistic interactions. River dolphins display year-round breeding; congregation in a few preferred feeding sites; lower site fidelity of males; overall high association values among individuals but fluid association patterns amongst sexes; existence of preferred male companionships; and a higher degree of sexual dimorphism. These result in high competition among males and two (polyandrous) mating strategies seem to apply: 1) roving-male strategy (single and in alliances) for early detection of females in estrus and to abduct them from other males, and 2) residing-male strategy in feeding sites of females and joining the open, direct competition if a female is in estrus whereas the female ultimately chooses (a) mating partner(s).

RINGKASAN

Informasi pada struktur sosial dan strategi perkembangbiakan dalam cetacean dan khususnya lumba-lumba sungai masih kurang. Pengaruh habitat pada struktur sosial dari populasi lumba-lumba Irrawaddy (Orcaella brevirostris) di laut dan air tawar, telah dipelajari sejak tahun 1999 dan 2002 di Kalimantan Timur. Habitat sungai dan teluk dibedakan menjadi 3 bagian, yaitu 1) bentuk geografis yang tertutup dan terbuka 2) jumlah makanan yang banyak sepanjang tahun dan musiman dan 3) sumber makanan yang terkumpul dan tersebar. Berbagai keadaan memperjelas semakin sedikitnya tingkat sosial dari lumba-lumba “Teluk” yang ditunjukkan dengan adanya kelompok- kelompok yang lebih kecil dan kurangnya interaksi antar kelompok, perkembangbiakan sesuai musim dan sedikitnya tingkat perbedaan seksual, yang bisa berakibat pada rendahnya kompetisi diantara jantan dan mungkin, dalam sistem perkawinan dengan banyak pasangan. Interaksi antar kelompok pada lumba-lumba “teluk” lebih banyak karena keperluan seperti mencari makan, sedangkan lumba- lumba sungai berinteraksi antar kelompok berubah-ubah secara alami, mencari makan, perjalanan, sosialisasi, interaksi yang berat. Lumba-lumba sungai

134 Social dynamics of facultative river dolphins

berkembangbiak sepanjang tahun, berkumpul di beberapa tempat yang mereka sukai untuk makan, bagi jantan tingkat kesetiaan pada suatu tempat rendah, secara keseluruhan nilai hubungan antar individu adalah tinggi tetapi nilai hubungan antara jenis kelamin lebih rendah. Adanya hubungan dengan jantan yang disukai, dan tingginya tingkat perbedaan jenis kelamin. Akibatnya adalah kompetisi yang tinggi antara para jantan dan dua strategi perkawinan dengan banyak pasangan terjadi yaitu: 1) seekor atau lebih jantan mengambil betina dari kelompok pada awal masa reproduksi untuk dijauhkan dari jantan lain, dan 2) jantan yang ikut dalam kompetisi langsung dan terbuka dalam mencari pasangan dimana pada akhirnya betina akan memilih satu atau lebih jantan sebagai pasangannya.

INTRODUCTION

Social structures of odontocetes are usually characterized by long-lasting associations among individuals (mostly mother/ calf, females, and (sub)adult male bands) and mating systems are typically described to be promiscuous (in this study further, more properly referred to as polyandrous) (Berta & Sumich, 1999). Best studied groups of oceanic dolphins (Delphinidae) in terms of social ecology are bottlenose dolphins, Tursiops truncatus, spinner dolphins, Stenella longirostris, Hectors’ dolphins, Cephalorhynchus hectori, Indo-Pacific Humpback dolphins, Sousa chinensis and killer whales, Orcinus orca. Least studied of species in terms of social organization include all river dolphin species. Long-term studies on the social organization of bottlenose dolphins were described for Florida (Wells, 1991), Western Australia (Smolker et al., 1992), Scotland (Wilson, 1995) and the Northern Adriatic Sea, near Croatia (Bearzi et al., 1997). School composition in bottlenose dolphin societies in a short-term is predominantly fluid (fission-fusion), but many associations are relatively long-term (Connor et al., 2000a). Common mating strategies involve (super) male alliances attending or even sometimes abducting adult females until estrus occurs (Connor et al., 1999; 2001). Mating within spinner dolphins is seasonal and polyandrous and also male coalitions do occur (Johnson and Norris 1994). Hectors’ dolphins and Indo-Pacific hump-backed dolphins also appeared to have a typical fission-fusion society and they are hypothesized to have either a polyandrous (Bräger, 1999; Slooten et al., 1993), or mate searching mating system (Karczmarski, 1999; Jefferson, 2000), respectively. Resident killer whales live in small stable social pods of two to nine individuals. These pods are matrilineal consisting of an older mature female and her first and second generation’s offspring (Olesiuk et al., 1990). Matrilineal groups may travel together in larger pods of three to 59 individuals. Inter-pod matings occur during pod encounters in shared or resting areas and seem to be seasonal (Ford et al., 1994). Transient killer whales also associate in fairly stable pods, equivalent to a single matrilineal group with one to two generations present existing of one to four individuals, and individual associations

135 Chapter 9

over fifteen years have been documented (Baird, 1994; Baird & Whitehead, 2000; Baird & Dill, 1996). Dispersal by members of the native group has been recorded of both males before they reach, and females once they reach sexual maturity (Baird, 1994). One-third of groups encountered consisted of single, male individuals (Baird & Dill, 1996). River dolphins are least documented as to their individual associations and mating strategies. The mating system of the boto (Inia geoffrensis) was hypothesized to be monogamous based on the relatively small testis size and lack of sexual dimorphism (Best & Da Silva, 1984), although Connor et al. (2000) found a moderate sexual dimorphism (1: 1.11) for the same species. Male-biased sexual dimorphism (length, weight) was found in hump-backed dolphins in South Africa in contrast to hump- backed dolphins in Hong Kong for which no evidence was found for sexual dimorphism (Jefferson, 2000; Cockcroft, 1989). In Hectors’ dolphins, females are larger than males (Slooten et al. 1993). Sexual dimorphism in both bottlenose and spinner dolphins in length is minimal, although males may be heavier up to 39% in large bottlenose dolphins (Read et al., 1993; Tolley et al., 1995). In contrast, a high degree of sexual dimorphism occurs in killer whales. Social structures (group formation) in general are affected by prey availability and predator and parasites avoidance (Connor, 2000). The most distinct intra-species differences of social structures due to prey resources are demonstrated for killer whales. Fish-eating “resident” killer whales do not disperse from their maternal pod, whereas some female and male mammal-eating “transient” killer whales do disperse from their natal pod of which the males associate with other individuals to a lesser degree than females do (Baird & Whitehead, 2000). In this study, social structures are compared of two populations of the same delphinid species, Orcaella brevirostris, commonly described as an elusive species (Lloze, 1973; Dhandapani, 1992; Kreb, 1999), i.e. a coastal population in Balikpapan Bay and a freshwater population in the Mahakam River, both occurring in East Kalimantan, Indonesia. These habitats not only differ in prey availability and presence of predators, but also in geographical shape, which affect dolphin movement patterns and bioacoustics, and which in their turn may affect social structures. For example, the chance that different dolphin groups meet unintended is much higher in a river habitat than in an open bay habitat due to constrained area size in rivers and clumped food resources in confluence areas (Kreb, 2002; Smith, 1993). Also, the confined river shape and many river bends limits the free passage of dolphin sounds over large distances and thus favors physical nearness of individuals to maintain social relationships. In contrast, clicks of sperm whales, Physeter macrocephalus may travel over several kilometers and individuals may keep track of each other over larger distances (Berta & Sumich, 1999). A study on bottlenose dolphins living at the southern extreme of the species’ range showed that ecological constraints were important factors in shaping social interactions within cetacean societies as they showed in contrast to bottlenose dolphins in other areas, a temporally stable community structure (Lusseau et al., 2003).

136 Social dynamics of facultative river dolphins

The objectives of the present study are to asses the impacts of ecological differences on shaping social structures within coastal and river populations of the same dolphin species in terms of group size and composition, and social interactions among groups. In addition, I wanted to determine which breeding strategies might apply for both populations based on the degree of sexual dimorphism, extent of home ranges, dolphin distribution, inter-group interactions, breeding seasonality. Finally, individual association patterns and site fidelity are being analyzed for the river dolphin population. Individual associations of the Irrawady river dolphin population will be analyzed using the simple ratio-index as recommended by Ginsberg and Young (1992), when association is defined by presence in the same group. This population is listed as Critically Endangered following IUCN criteria (Reeves et al., 2003) and the population is estimated to exist of less than 75 individuals (Kreb, in press). A better understanding of their social ecology may aid in their conservation. Other studies on association patterns, which used association indices, only involved coastal dolphins. Most studies were on Hectors dolphins in New Zealand (Slooten et al., 1993; Bejder et al., 1998; Bräger, 1999). Association patterns of bottlenose dolphins were studied in Texas and New Zealand (Bräger et al., 1994; Lusseau et al., 2003). Transient and resident killer whales were studied in Canada and Alaska (Baird & Whitehead, 2000; Matkin, 1999). Indo-Pacific Humpback dolphins were studied in Hong Kong and South Africa (Jefferson, 2000; Karczmarski, 1999). Half-weight indices were most frequently used, than the simple-ratio association index, whereas Cole’s index and point correlation coefficient index were only used in one study.

METHODS

Study area

Balikpapan Bay stretches from 116o42’ to 116o50’ E and 1o to 1o22’ S (Figure 1). Water surface area of the bay is approximately 120 km2. Maximum width of the bay is approximately 7 km, and the surrounding shorelines within the bay consisted mainly of mangrove vegetation. Dolphin densities were more or less equally distributed over all the stratums in the bay (Kreb, unpublished data). Food ecology is regulated seasonally and four different seasons can be distinguished based on wind prevalence. December until February is governed by northern wind with varying but mainly low fish availability, and many waves. From March until May an eastern wind prevails with high fish abundance. From June until August southern wind is dominant, with many waves at sea, but high abundance of fish in the bay. Finally, September until November is characterized by the highest waves, with low fish abundance in the bay due to fish spawning in mangrove vegetation. The Mahakam River is one of the major river systems of Borneo and runs from 118º east to 113º west and between 1º north and 1º south (Figure 1). Regional climate is characterised by two seasons, i.e., dry (from July-October, southeast monsoon) and

137 Chapter 9

Area 4

Area3

Core area 1 Core area 2

Figure 1. Map of both study areas, the Mahakam River and Balikpapan Bay. Four dolphin distribution areas in the river are encircled and high dolphin density areas are indicated by square boxes named “core’’.

138138 Social dynamics of facultative river dolphins

wet (November-June, northwest monsoon) (MacKinnon et al., 1997). However, dry and wet periods alternate during the wet season as well. The river measures about 800 km from its origin in the Müller Mountains to the river mouth and crosses two districts, West and Central Kutai. Average river width between Samarinda (80 km from the mouth) and Long Bagun (c. 560 km from the mouth) at medium water levels is 200 m. Three major lakes, Semayang (10,300 ha), Melintang (8,900 ha), and Jempang (14,600 ha) are connected to the main river system in the Middle Mahakam Area (MMA) between 180 km and 375 km from the mouth. In addition, nearly all the major tributaries connect with the main river in the MMA, together with many smaller swamp lakes that are connected to these (some only seasonally through flooding). These lakes are very important fish-spawning grounds and replenish the main river seasonally. Rapids start upstream at c. 600 from the mouth, which limit the dolphins from moving further upstream. In the dry season, dolphins congregate in deep, confluence areas. At the onset of the rainy season, fish spawn upstream tributaries and dolphins migrate accordingly (Kreb, 2002).

Field techniques

From February 1999 until August 2002, we surveyed the Mahakam River for a total of 750 hr with transects totaling 7933 km. In total, 12 extensive surveys (= six replicated up-and downstream surveys) were conducted that covered the entire distribution range (mean duration = 10 days; SD ± 2 days) during all types of water levels (high, low, medium, increasing, decreasing). Another six intensive surveys were conducted in areas of high dolphin density (average duration 12 days; SD ± 3 days). River dolphins were observed for a total of 545 h. Four coastal surveys were conducted in the Balikpapan Bay survey area from May 2000 until October 2001, and a total distance of 1360 km during 127 h was covered in 36 days. Coastal dolphins were observed during 60 h. Of the river dolphin population, 775 photographs were made representing identifiable dorsal fins and a total number of 66 dolphins identified. Observation- and survey procedures, and photo-identification techniques are described in Kreb (2002; in press). Along the river, 58 focal groups distributed over the entire dolphin range were followed for 321 h in total (hereafter referred to as “focal follows”) and each group on average 5.5 h (range = 1.5 – 13 h). In the bay, 21 coastal focal groups were followed each for on average 3 h (range = 1.5 – 8.5 h) and for a total of 64 h. General group behaviour was recorded from the start of observation and ended when changes occurred in group behaviour or when dolphins left or joined the group. Then, a new recording session would start. Group behaviour involved general behaviour in which > 50% of dolphins were engaged. When new individuals joined the focal group after the first 15 minutes of observation time, these were recorded as inter-group interactions and lasted until these individuals left the focal group or until the end of

139 Chapter 9

observation time. If the focal group was observed to split during observation time without any new individuals joining the group at first then the largest group was still followed and regarded as focal group. The observation time until the groups split was then recorded as inter-group interaction time. Sex was determined visually from the presence or absence of a neck crest. Males have prominent neck crests and females have smooth necks with no crest. This distinction was based on a preliminary study in the Mahakam (Kreb, 1999) during which individual differences in sizes of neck crests irrespective of body length were noticed. Also, individuals with calves always had a smooth, thin neck shape as had one dead female dolphin. To double check if differences in neck sizes were related to sex, a control group of 12 captive Irrawaddy dolphins of known sex in Oasis Seaworld, Thailand were examined. All these dolphins were six years or older. Four dolphins with no crest at all were females. Males always had a neck crest, varying from slight to conspicuous. However, crests of these Irrawaddy dolphins, which were caught from the Gulf of Thailand, never attained the same size as those observed for dolphins in the Mahakam River.

Data analysis

All tests involved non-parametric statistics since sample populations were not always normally distributed nor had similar variances. Only once, a parametric Z-test was used to compare group sizes between both populations, because a large data set was available that complied with the set restrictions for parametric tests (Fowler & Cohen, 1990).

Group size and composition Mean group size in the Mahakam was based on all on-effort sightings made during 9 extensive abundance surveys covering the entire distribution range. Likewise, mean group size in Balikpapan Bay employed on-effort sightings made during 4 surveys with equal survey effort in all segments of the bay. Groups were considered different if a group joined after 15 min of observation or groups split during observation time. The proportion of adult males in groups were determined only for those groups for which the number of individuals and males could be determined with a high certainty, i.e., was agreed upon by all observers who had been assigned the special task to determine group composition (2-3 observers). In order to find the maximum, optimum group sizes for various general behaviors only interacting groups were used.

Movement patterns and site fidelity Overall home ranges of photo-identified dolphins were only determined for river dolphins, since we had no photo-identification data for the coastal population. Home ranges were estimated by measuring the distance between the two most widely

140 Social dynamics of facultative river dolphins

separated sighting locations for each individual photographed during surveys that covered the entire dolphin distribution range. This linear distance was then multiplied by the average width for the different river sections. Since we did not always get a positive photo-identification during each survey period for each of the photo- identified dolphins and the photographic effort was not equally spread over the different years, we calculated overall home ranges based on sightings made during the entire study period (3.5 years) and not per year. To assess site fidelity, the locations of photo-identified river dolphins were mapped according to different zones in the river (Figure 1). The core areas 1 & 2 represented high dolphin density regions, which included confluence areas of tributaries, lakes and main river (Kreb, 2002). In addition, residence indices (RI) were calculated for 13 females and 9 males, excluding known juveniles and sightings made < 3 days, following a formula, which was thought to reflect best the extent of site fidelity for the river dolphin situation (Eqn. 1). Residence indices were calculated for each individual by distracting the number of periodical sightings in the area where most sightings were made (xi), with the combined number of periodical sightings in other areas (xΣi). The distribution areas of dolphins to this end were divided in 40 km river strips and only sighting was calculated per area per survey period (= periodical sighting). This value then was divided by the total number of survey periods, in which each individual was sighted (s). Index values could range between -1 (low site fidelity) and 1 (high site fidelity).

xi - xΣi Eqn 1. RI = s

Social interactions and behaviours Interactions between groups were described either as non-interactions when dolphins moved in different directions (zero interaction time), or low level-interactions if groups were traveling in the same direction but kept a distance of > 50 < 500 m. In a pure sense dolphins of the latter category probably interacted to some extent with each other through their acoustics, since click trains of dolphins in the river could be detected at least over 500 m at straight stretches using a High Tech Inc.- 94-SSQ hydrophone (frequency range: 2Hz - 30 KHz at –168 dB re 1V/µPa), which hung 1.5 m deep into the water). When assessing the mean interaction time between the focal and newly encountered groups, only those interactions were included in which the new groups were encountered during the focal follow time and also separated from the focal group before the end of the follow time. However, these interactions were included when assessing how many group interactions lasted > 1 hour.

141 Chapter 9

Individual associations In this study, associations were analyzed using SOCPROG 1.3, a program developed in MATLAB (The Mathworks, Inc., Natick, Mass., U.S.A.) by Hal Whitehead for analyzing social structure. In this study, the MATLAB 6.5.1 version was used.

Individuals were considered associated if they were seen together in the same group (total no. groups = 95). Only photo-identified individuals were used (n = 50 of total identified n = 66), which were sighted in 3 or more groups and 2 or more sampling periods (years). Groups were newly defined when a new sighting started and every 3 h when groups were followed > 3 h. Minimally 2 individuals were photo-identified within each group. Simple-ratio association indices (Equation 1) were used to measure associations, which are thought most appropriate when association is defined by presence in the same group, here referred to simply as association index (AI). Also, since sampling biases vary among pairs of individuals, the association indices which use arbitrary weightings only reflect the direction, but not the extent of sampling biases (Ginsberg & Young, 1992). The simple ration indices provided values between 0 (if individuals were never seen together) and 1 (if two individuals were always seen together).

x Eqn 2 AI = x + yA + yB

Where x = number of sightings in which both dolphin A and B were seen in the same group, yA = total number of sightings of individual A excluding individual B, and yB = number of sightings in which only individual B was sighted and not A. Association patterns were displayed in two ways: 1) by an average-linkage cluster analysis showing the average level of association between hierarchically formed clusters and 2) by a principal coordinates analysis, which makes a metric scaling and produces an arrangement of points, each representing an individual, so that the distance between them is inversely proportional to the square-root of their association. Real association values were also randomly permuted 50,000 times following Bejder et al. (1998), Manly (1995) and a modification of Whitehead (1999) in order to test if the mean, SD, and proportion of non-zero association indices were different than would be expected from random associations. For detection of diads (pairs) that have significantly large or small associations, a 2-sided significance level of 0.05 was selected. Two types of random permutations were performed: 1) permuting groups within samples, which test accounts for situations in which not all individuals are present in each sampling interval (because of birth, death, migration etc.) and tests both long-term (between sampling periods) as well as short-term (within sampling

142 Social dynamics of facultative river dolphins

periods) preferred companions, and 2) permuting associations within samples, testing only for preferred companionships between sampling periods.

RESULTS

Group sizes

Mean group size of dolphins in the Mahakam (covering the entire dolphin distribution range) is 4.4 and most dolphins (29% of total n) occurred in group sizes of 5 individuals (n = 75; SD = 2.2; range = 1-10). Mean group size in Balikpapan Bay was significantly smaller, i.e., 3.2 (n = 79; SD = 2.1, range = 1-9; Test equal variances, F = 1.13, P > 0.05; Z-test, z = 3.43, P < 0.01). Frequencies of group size occurrences are in Figure 2. Most coastal dolphins moved about solitarily (27% of total n). Within the river dolphin population, lowest mean group sizes of interacting groups were found when groups were feeding (mean = 6.4, SD = 2.4) or traveling (mean = 6.6, SD = 2.6) together (Table 1). Interacting groups, which were feeding or

25 Coastal dolphins River dolphins

10 Number of sightings 5

0 12345678910 Group sizes

Figure 2. Frequency of group size occurrence of coastal (n = 75 groups) and freshwater (n = 79 groups) Irrawaddy dolphins

143 Chapter 9

Table 1. Predominant inter-group activities, interaction time and group sizes for coastal and river Irrawaddy dolphin populations

*Mean interaction % of total interactions time (min) Mean group size Predominant inter-group Bay River Bay River Bay River activities (n = 20) (n = 76) (n = 8) ( n = 32) (n = 20) (n = 76) Heading in one direction 10% 18% - 167 9 6.6 Feeding 75% 34% 30 58 6.1 6.4 Intensive socializing 10% 25% 181 168 9.5 8.5 Agonistic displays 0% 8% - 56 - 8.5 Low level interaction 5% 8% 19 - - 8.2 No interaction 0% 7% - 0 - 10.2

- = only one or no interactions available; * only interaction used, which formed and split during observation time traveling, were significantly smaller than groups engaged in any of the other behaviors (median group sizes of each behavioral category were compared with one another using Mann-Whitney U test; all p < 0.05 except for one pair, i.e., travel and low level interaction p > 0.05). The mean group sizes associated with feeding or traveling are considered the optimal group sizes for these types of behaviors. If interacting groups are larger in size than the standard deviation, the interaction is characterized by intensive socializing or gets a less desirable character. Groups either avoid each other (low level interaction and no interaction), or become agonistic if groups just happened to encounter each other unintended, whereas they actually intended to feed. For the bay population we found that the median optimum group size for interacting feeding groups (mean = 6.1, SD = 2.1) was not significantly different from the median feeding group size in the river population (Mann-Whitney U-test, U= 196.5, P > 0.05).

Group composition

A positive correlation was found between group size and the relative percentage of calves within the river dolphin population (r = 0.81, df = 8, P < 0.01), but not within the coastal population (r = 0.11, df = 7, P > 0.05). The largest relative proportion of river calves in relation to juveniles and adults occurred within group sizes of 8-10 individuals. The relative proportions of coastal calves were normally distributed over different group size classes with the largest proportion of calves found in groups of 4- 5 individuals. Significant relatively larger proportions of calves for the coastal population in comparison to the river population were found in small groups of two to three individuals (G = 28.4, df = 2, P < 0.01). Most river dolphin groups sighted (54 % of n = 75) involved adults with their offspring; 29% were adult groups; 4% involved groups consisting of juveniles exclusively (6-7 individuals); 13% were single

144 Social dynamics of facultative river dolphins

individuals. Proportions differed significantly from those in the coastal population with percentages in the same sequence of 34%, 30%, 0% and 21% (G = 7.906, df = 3, P < 0.05). Especially more single individuals were recorded in the coastal area and there were no juvenile groups.

Sexual dimorphism

In both the Mahakam River and Balikpapan Bay, Irrawaddy dolphins that lacked neck crests and dolphins with slightly to more prominent neck crests were observed, irrespective of body length. Individuals with calves had smooth necks with no crest. However, neck crests of coastal Irrawaddy dolphins never attained the same size as those observed for dolphins in the Mahakam River. The proportion of adult males and females within groups for coastal and river populations are presented in table 2, single groups excluded. The proportion of males in both coastal and river dolphin groups significantly less often exceeded the proportion of females in groups (X2 = 7.922 & X2 = 18.54, df = 2, P < 0.05 & P < 0.01). No differences in the proportion of adult males in groups was found between both populations (G = 3.59, df = 2, P > 0.05). Two cases in the coastal population where the proportion of males exceeded females in groups involved groups that exclusively consisted of males (n = 3 & 5). Two single coastal and river ‘groups’ of which the sex could be determined involved males. Groups of river and coastal dolphins consisting of minimally one adult female and maximally 7 females were mostly escorted by single adult males (56 - 60% of groups n = 36 & 25) and up to a maximum of 4 adult males. Two river dolphin groups (6%) consisted each exclusively of two adult females. Five coastal groups (20%) involved single-parent groups of one adult female and her offspring, whereas no single-parent groups were found in groups of river dolphins. Coastal adult female groups were significantly more often unescorted by males in comparison to river dolphins (Gadj = 11.75, df = 1, P < 0.01). A total of 66 river dolphins were photo-identified during the study period for which the sex could be identified for 27 individuals (41%); 15 were females and 12 were males, of which 1 female juvenile and 2 male juveniles.

Spatial distribution

Home ranges In order to assess daily home ranges, 58 river and 21 coastal focal groups were followed for a total of 321 and 64 h and on average 5.5 and 3.0 h daily (range 1.5 - 13 h; 1.5 - 8.5 h), respectively. Daily home ranges of 27 river focal groups, which were

145 Chapter 9

Table 2. Proportion of adult males in groups

Coastal bay River Proportion of adult No. of No. of males per group groups % groups % Males > females 2 8% 3 8% Males = females 13 50% 10 27% Males < females 11 42% 24 65% Total groups 26 37

followed > 6 h were small, i.e., 10 km river strip (SD = 8.6 km, range 1- 45 km) and 1.1 km2 in area (SD = 1.8 km, range 0.1 – 9 km2). Daily ranges of 3 coastal focal groups, followed > 6 h were all in a mangrove river strip and were very small, i.e., 2 km length and 0.4 km2 (n = 3), respectively. The average daily home ranges of 27 groups, followed > 6 h were 10 km in length (SD = 8.6 km, range 1- 45 km) and 1.1 km2 in area (SD = 1.8 km, range 0.1 – 9 km2). Home ranges throughout the year were only calculated for the 53 river dolphins that were photo-identified during the 3.5 years study period during a mean number of 12.5 sightings (SD = 9.5, range = 2 – 39), 6.2 different survey days (SD = 3.7, range = 2 – 20) and 4.4 survey periods (SD = 2.0, range = 1 - 9). No correlation was found between the number of survey days or periods and home range length (r = 0.071, df = 51, P > 0.05 & r = 0.085, df = 51, P > 0.05). These dolphins moved freely in a river strip of on average 61 km in length (SD = 44 km, range = 4 – 181 km) and of 10 km2 of river area (SD = 9.1 km2, range = 0.3 – 35.5 km2). Female overall home ranges (n = 15) were significantly smaller (mean = 44 km; range = 4-103 km) than male overall home ranges (n = 12; mean = 103 km; range = 15-163 km) (U = 142, P < 0.05) (Figure 3).

Site fidelity of river dolphins Two center areas of high dolphin density could be distinguished in the river with their peripheral zones (Figure 1). Sighting locations were mapped for all photo-identified individuals sighted on more than one day (n = 53 individuals) and on average 7 days (range = 2 – 22 days) and 4 survey periods (range = 2-10 periods). Eight individuals were exclusively sighted in area 1, and 30 individuals in area 2. Individuals exclusively occurring in either area 1 or 2 with known sex (n = 16), were mostly females (81%). The remaining 12 individuals had an overlap of sightings with peripheral zones of other areas, of which the sex could be identified for 10 individuals. Ten percent of individuals, which showed overlap with peripheral zones of other areas, were females and 90% were males. Three individuals were sighted in both center areas of areas 1 and 2 and were identified to be males (Table 3).

146 Social dynamics of facultative river dolphins

6 males females

1 Number of males and fermales

0 4316497130163 Overall home range sizes (km river strip)

Figure 3. Male and female home ranges sizes (km river strip)

One group of 6 individuals occurred in an isolated area, being ‘trapped’ in between two rapids since a big flood in 1998 until the end of the study period (August 2002). Residence indices for females (median = 0.88) were significantly higher than for males (median = 0.37) (U = 16.5, n1= 13 & n2 = 10, P < 0.01), indicating a higher degree of site fidelity by females.

Table 3. Distribution of photo-identified individuals and adult sexes in different river areas

Number of Sex of > 75% River areas1 individuals of N (N) Area 1 (core + periphery) 8 F Area 2 ( core + periphery) 30 F Core 1- periphery 22 3 M Core 2- periphery 1 5 M Cores 1 & 2 3 M Peripheries 1 & 2 1 M Core 2- periphery 3 2 M Area 1,2,3 1 M Total photo-id sightings3 53

1 Corresponding to areas indicated in Figure 1; 2 Example; most sightings made in core area of area 1, but also sightings made in periphery of area 2; 3 Only individuals used that were sighted > 1 day F = females; M = males

147 Chapter 9

Inter-group interactions

Number and duration of interactions In total 58 river and 21 coastal focal groups were followed each for > 90 min (mean = 5.5 h, range = 1.5 – 13 h & mean = 3 h, range = 1.5 – 8.5 h) and for a total of 321 h and 64 h, respectively. A high percentage (71%) of 58 river focal groups had interactions with other groups on 74 occasions during the follow time. Each group encountered other groups for on average 1.8 times (SD = 0.95; range = 1 - 4). Coastal dolphin focal groups interacted less frequently than river dolphins, i.e., only 48% of 21 coastal focal groups had inter-group interactions (mean number of interactions per group = 2, SD = 0.94; range-1 - 4), but not significantly (Gadj = 3.45, critical value = 3.84). Number of interactions was positively correlated with the focal group follow time in the river population (r = 0.34, df = 40, P = 0.03), but no correlation existed in the coastal population (r = 0.13, df = 8, P > 0.05) (Figure 4). River dolphins followed for an average 8 h (range = 6 - 10 h) encountered a mean number of two other groups (range = 1 – 4 h). Inter-group interactions of river dolphins lasted on average 62 min (SD = 67 min; range = 10 - 325 min), and more than half the number of total inter-group interactions (53%) lasted > 1 hour. Mean duration time of interactions of coastal dolphins is 47 min (SD = 56 min; range = 10 – 181 min). Significantly more interactions lasted longer than 1 hour in the Mahakam river as compared to Balikpapan Bay where 6 interactions (30%) of 20 interactions lasted > 1 hour (Gadj = 16.02, df = 1, P < 0.01). Interactions that lasted longest (nearly 3 h) were socializing in both populations and traveling (Table 1). Feeding lasted shortest (< 1 h) and significantly shorter in the bay than in the river (X2 = 8.3, df = 1, P < 0.01). Agonistic encounters lasted quite long, nearly 1 h.

Nature of interactions Six main behavioral categories were distinguished of interacting groups: 1) traveling as primary activity, 2) (combination of) feeding (and traveling), 3) intensive socializing (in combination with traveling), 4) agonistic behavioral displays, 5) low level interaction (mean distance between two groups > 50 m < 500 m), 6) no interaction (groups heading in different directions). Predominant inter-group activity (> 50% of interaction time) in both the Mahakam River and Balikpapan Bay was feeding together (34% & 75% of encounters, respectively), which was observed significantly more often than the other behaviors (X2 = 28.99, df = 5, P < 0.01) (Table 1). The proportions of different inter-group activities varied significantly between coastal and river populations (Gadj = 33.26, df = 5, P < 0.01). Although feeding was the dominant activity, in the river populations other behaviors, such as intensive socializing and traveling were also often displayed, whereas in the coastal population these activities were less common. In Balikpapan Bay, dolphins only socialized during 10% of encounters among groups. Also, nearly one-fourth (23%) of river group interactions

148 Social dynamics of facultative river dolphins

involved agonistic-, low level- and non-interactions, whereas these behaviors were only displayed in 5% of total number of interactions in coastal groups.

Breeding

Mating events (n = 3) in the Mahakam were observed during the months July and August at low water levels. However, behaviors associated with mating were observed throughout the year and at all water levels. Mating events took place between 2 to 3 subgroups with total group sizes ranging between 5 until 12 adult individuals per mating event. Mating most probably only occurs among different groups, since each time after mating interactions had finished (duration time = 9-80 min) the groups were observed to split into subgroups. Matings were characterized by vocal and behavioral dominance displays: loud blows, fast swimming, rolling along the axis of the dolphin’s body, swimming side-wards and with belly up, group swimming in small circles and speeding up (chases), jumps, many fin and fluke waves and slaps, and intensive body contact. Behaviors associated with mating were not as intensive as mating events, less speeding, rolling activity and body contact compared to the mating events. After a mating event groups were observed to split and mating behavior continued in one group. Newborns (< 1 month of age) were observed in all different months and water levels, so it is likely that matings also take place throughout the year. During all mating events, several males were identified, which during the study period occurred in both core areas of area 1 and 2. In Balikpapan Bay, one mating event and behaviors associated with mating were only observed during the month October. The mating event lasted 1.5 h and involved 13 adult individuals (and 1 juvenile). After 79 min, the group split in two subgroups of 5 and 9 individuals. The latter group continued mating until they started moving fast and disappeared. Newborns were only observed during the months June and July in 2001 and 2002, respectively. The weather conditions in these months are governed by a southern wind, which causes a high fish availability in the bay due to high waves in the open sea, whereas waves in the bay are moderate.

Individual associations of the river dolphin population

Random permutations of real association values of groups and associations within samples (testing both short- and long-term companionships) were performed for all identified individual dolphins, that were sighted in 2 or more sampling periods (n = 50 = 79% of N identified) in a total of 95 groups. Individual dolphins showed clear preferences for association with certain individuals and had long-term preferred companionships both indicated by a significantly higher SD of the real association indices than the random value (P = 0.99). The mean real simple-ratio association index (AI) value of non-zero elements was 0.42 (SD = 0.18), which was also

149 Chapter 9

Individuals

Association index

Figure 4. Average linkage dendogram of the Irrawaddy dolphin population in the Mahakam River showing associations among individuals in between the period of February 1999 until August 2002. On the y-axis are all individuals, which are identified in more than 2 groups and during 2 or more sampling periods (years). The letters M and F in the ID names correspond with male and female respectively, whereas the numbers with which the ID names begin correspond to the areas in which the individuals have been sighted. The three clusters described in the text are indicated.

150 Social dynamics of facultative river dolphins

significantly higher than the random mean value (P = 0.99). The significantly lower proportion of non-zero associations (p = 0.37) than randomly excepted indicates that some individuals avoided others (P < 0.01). This can be seen from figure 4, where 3 clusters can be distinguished, the first and second existing of individuals with a high site fidelity in areas 1 and 2 (mostly females), respectively mixed with males with overlapping ranges in both areas, and a third cluster representing individuals, which are ‘trapped’ in between two rapids in area 4 (Figure 1). Within sampling periods 20 significant diads were found, whereas between sampling periods 30 significant long- term diads were detected. Mean individual association values (including zero-elements) were found significantly higher for individuals exclusively sighted in one area (AI = 0.18) than for those individuals, which moved in between 2 or more areas (AI = 0.13) (Mann-whitney U-test, U = 142; n2 = 35 & n2 = 13; P < 0.05). Also, the mean, maximum association values were significantly higher for individuals, which occurred in one area (mean AImax = 0.75) than for individuals occurring in both areas (mean AImax = 0.65) (U = 80; P < 0.05). All females occurring in area 2 (n = 11) associated with each other relatively intensively (mean AI = 0.41; SD = 0.16), but no preferred companionships were detected as indicated by a significantly lower mean and SD of the real association indices compared to the random values (P < 0.05). The proportion of non-zero associations (p = 0.85) was significantly higher than that of the random values (P = 0.95) indicating that most females associated with other females within sampling periods and also maintained these associations between sampling periods, although not significantly (P = 0.92). The proportion of 12 individual males that associated with other males (p = 0.55) was much lower compared with the proportion of associations among females (p = 0.85). Mean association values (including zero-elements) were also significantly smaller for males (AI = 0. 20) than for females (AI = 0.36) (U = 120; n1 = 11 & n2 = 12; P < 0.01), whereas maximum association values (mean AImax = 0.55) were also significantly smaller than those of females (mean AImax = 0.68) (U = 103; n1 = 11 & n2 = 12; P = 0.01) (Figure 5). No significant short- or long-term preferred companionships were found, except for one diad within one sampling period (Figure 6). Finally, most associations among sexes were fluid both within and between sampling periods as no significant deviations of real association indices from random values were found, which should indicate short- or long-term preferred companionships. However, the mean association indices of non-zero elements were just as high as found within intra-female associations (mean = 0.41; SD = 0.17), indicating that most males and females associated with each other. One significant male-male diad was found within one sampling period and 3 female-male diads, 2 male-male diads and one female-female diad were found in between sampling periods. The proportion of individuals, which associated with others (p = 0.59) was significantly lower than randomly expected (P < 0.01), indicating that some individuals avoided others.

151

Chapter 9 12 Males mean A.I. Females mean A.I. Males max A.I. Females max A.I.

a 10

2 Number of males/ fem

0 0-0.14 0.15- 0.30- 0.45- 0.60- 0.74- 0.90-1 0.29 0.44 0.59 0.74 0.89 Intra-sex association indices

Figure 5. Mean and maximum simple-ratio indices of female-female (n = 11) and male-male (n = 12) associations (A.I.)

Figure 6. Principal coordinates analysis plot showing strong (small circles close together, such as trio 2-3MP27, 2MP10 & 1-2MP42) and one looser male consort-ships through time (between sampling periods). Males 2MP10 and 1-2MP42 formed a significant alliance within one sampling period (year).

152 Social dynamics of facultative river dolphins

DISCUSSION

Methodological constraints

Regarding the calculation of association indices, the estimates of the total number of sightings of individual A excluding individual B (and vice versa) may not be accurate, since a successful photo-id picture is not always obtained from each individual in the group. So, dolphin pairs may actually occur together during a sighting, but only one individual may be photo-identified. No photo-identification data for the coastal dolphin population was available, since only a low proportion was found to have characteristic dorsal fins during a preliminary survey.

Social organization

Group sizes are described to be mainly affected by prey availability and predator avoidance (Connor, 2000). In this study, both populations did not seem to be affected by predators. In the sheltered bay area, there are no sharks, only marsh crocodiles, Crocodylus porosus, which do not seem to pose a real threat to the dolphins as inferred from the absence of predator defense mechanisms (formation of large groups) as a high proportion of ‘groups’ exists of single individuals. Similarly, in the river two species of crocodiles, the false gavial, Tomistoma schlegeli, and the Siamese crocodile, Crocodylus siamensis, do not seem to pose a threat to the dolphins because of their rounded body shape, which would be difficult to handle for these relatively small crocodiles. Differences found in coastal and river group sizes are most likely to be related to the distribution of food resources, which are more clumped in the river and more equally distributed in the bay. The high percentage of more or less unwanted interactions, i.e., agonistic, low- level or non-interactions (23% of total interactions) are typical for the Mahakam situation in which groups compete for confluence areas and many unintended encounters are more a result of the same preferences for specific feeding areas, mostly confluence areas with tributaries and lakes (Kreb, 2002). These types of interactions are unknown for botos, which also occupy a limited range for part of the year but for which no avoidance or aggressive behaviors were found (Schnapp & Howroyd, 1992). A similar large proportion of social interactions (25% of total interactions) as the “negative” interactions take place among different river dolphin groups. Since social interactions among groups may last for 2.8 h; this indicates that these interacting groups are not an aggregation of loose groups but represent a true social unit. In this aspect the Mahakam dolphins, although not as comprehensively studied, resemble the pods of their closest relatives (resident) killer whales (LeDuc et al., 1999), which are composed of several intra-pod groups and form stable relationships based on high

153 Chapter 9

association indices within intra-pods in resident killer whales (Matkin et al., 1999). They also have similar group sizes consisting of 4 individuals on average (Bigg et al., 1990). In the Mahakam, three distinct pods could be distinguished based on individual site fidelity corresponding with downstream core area 1 and upstream core area 2, and the trapped group occurring in Ratah tributary. The Mahakam population also resembles a population of bottlenose dolphins living in a fjord in New Zealand in their geographical isolation, their small population size of both less than 75 individuals, and similar temporally stable community structure with high association values among all members in the population. However, they were dissimilar in group sizes amongst others, which were much larger for these bottlenose dolphins with 17 individuals on average. Association values among individuals of studies on humpback dolphins in Hong Kong and South Africa (Jefferson, 2000; Karczmarski, 1999), bottlenose dolphins in Galveston Bay, Texas (Bräger et al., 1994). Hectors’ dolphins in New Zealand were all low and dolphins appeared to have relatively fluid associations, associating with many individuals rather than with very close associates (Brager, 1999; Slooten et al., 1993). Transient female killer whales had a high average but low maximum association rates indicating that they are not only gregarious, but also socially mobile. Male transient killer whales have no strong or long-term relationships with nay individuals (Baird & Whitehead, 2000). Coastal Irrawaddy dolphins were very similar to transient killer whales with respective small group sizes of 3 and 2 individuals on average. Also, single-individual encounters occurred just as frequent in 27% and 31% of total number of encounters with coastal Irrawaddy dolphins and transient killer whales, respectively (Baird and Dill 1996). A possible explanation for the apparent smaller social units of coastal Irrawaddy dolphins may be the fact that social contact among pods may be maintained over longer distances by means of acoustics, instead of physical contact by the less restrained geographical environment.

Breeding strategies

Matings of dolphins in the Mahakam were also similar to the resident killer whales because in both species matings occurred outside the intra-pod groups and possibly among pods when pods encounter each other in shared foraging or resting areas (Berta & Sumich, 1999; Bigg et al., 1990; Ford et al., 1994). Inter-pod matings in Balikpapan Bay occur seasonally and estrus periods of females may be more or less synchronized, which could explain the less marked sexual dimorphism in coastal Irrawaddy dolphins. Since distribution of females in the bay was found to be equally spread over different segments in the bay and estrus of females is seasonal, males may also be more distributed over the bay (daily home ranges were found to be rather small) and competition for males may not be too fierce, favoring a polyandrous or “multi-mate” mating system like those found for other species, e.g. spinner dolphins in Hawaian waters, which mate seasonal (Johnson & Norris, 1994). However, also in

154 Social dynamics of facultative river dolphins

bottlenose dolphins, which are poly-estrous, matings are suggested to be polyandrous (Berta & Schumich, 1999). River dolphins in the Mahakam on the other hand bred throughout the year and estrus females are easily detected by males since most groups co-occur in favorite feeding areas and encounter other groups twice on average during day time. Female distribution is most likely defined by resource availability, and male distribution is most likely affected by female availability (Davies, 1991). Since the groups are clumped in one area and frequently interact, it seems unlikely that males monopolize females. Rather, a roving male or resident male strategy may be more prevalent. In the roving strategy, males wander from group to group to find receptive females such as was found in Hector’s dolphins and bottlenose dolphins for example (Slooten et al., 1993; Connor et al., 1992). Strongly bonded male groups were also found in spinner dolphins (Őstman, 1994), spotted dolphins (Pryor & Schallenberger, 1991), resident killer whales (Baird, 2000), and northern bottlenose whales (Gowans, 1999). The following data confirms the existence of roving males. 1) Ten out of 75 sightings in the Mahakam during the study period occurred of single individuals, which in two cases could be identified and appeared to be males (one juvenile and one adult). 2) Three male diads were identified, of which two had long-term preferred companionships and one pair associated within one sampling period. 3) Overall, males maintained frequent but fluid associations with females. In addition, of the 15 identified males, 9 males had overlap with peripheral zones of other areas, and 3 males were sighted in both core areas of dolphin distribution, whereas females always remained in their core areas. The goal of roving males, whether single or in alliances, may be an early detection of estrus females by males and to herd or guard them, such as observed in bottlenose dolphins (Connor et al., 2000b). The roving technique may apply to juveniles and less dominant males, whereas resident males may be dominant adults, with a high success rate in female mating access. Three identified males were exclusively sighted in one core area during the entire study period and maybe resident males. A receptive female is likely to be soon detected by other males (alliances) since groups frequently meet and a direct competition may occur, in addition to which resident males participate, and in which the female may ultimately choose her mate. Female choice may be based on the multi-variety in behavioral displays observed during the matings in this study (see, results) similarly to many bird species (e.g. Foster, 1981). Also, active mate avoidance behavior of females was observed in this study by speeding up away from other individuals and rolling around the axis of their bodies. It has been suggested that female cetaceans, in contrast to terrestrial female mammals, may more easily avoid unwanted matings due to the three-dimensional fluid structure of the marine or river environment (Whitehead & Mann, 2000). Because three distinct pods were identified in the Mahakam, which display an apparent, site fidelity for preferred dolphin sites, such as confluence areas where they have frequent social interactions, a site-protection-based conservation strategy is recommended in three core areas. This study attempted to fill in the gap in our knowledge of social ecology of river dolphins in general and more specifically of

155 Chapter 9

facultative river dolphin species that consist both of freshwater as well as coastal dolphin populations.

ACKNOWLEDGEMENTS

I would like to thank the Indonesian Institute for Sciences (LIPI), the provincial wildlife conservation department (BKSDA) and local governments of Central- (KUKER) and West Kutai (KUBAR) for granting permission to conduct field research. All field assistants, particularly Ahang, Arman, Budiono, Karen Damayanti and Syachrani, and boatsmen are thanked gratefully as well as local fishermen that participated in the interviews. Funding for fieldwork was provided by Ocean Park Conservation Foundation, Hong Kong; Martina de Beukelaar Stichting; Stichting J.C. van der Hucht Fonds; Gibbon Foundation; Netherlands Program International Nature Management (PIN/ KNIP) of Ministry of Agriculture, Nature Management and Fisheries; Van Tienhoven Stichting; World Wildlife Fund for Nature (Netherlands); Amsterdamse Universiteits Vereniging; Coastal Resource Management Program/ Proyek Pesisir. The University of Mulawarman in Samarinda (UNMUL), Plantage Library, Achmat Ariffien Bratawinata, Frederick R. Schram, Peter J.H. van Bree, Thomas A. Jefferson, Martjan Lammertink and Vincent Nijman are thanked for their support throughout the study. Hal Whitehead is thanked for allowing the use of the SOCPROG software and The Mathworks for providing an evaluation copy of MATLAB 6.5.1. Bernd Würsig and two anonymous reviewers are thanked for their comments on the manuscript. Fieldwork complied with the current laws in Indonesia, where the study was conducted.

REFERENCES

Baird, R.W. 1994. Foraging behaviour and ecology of transient killer whales. PhD thesis, Simon Fraser University, Burnaby, B.C. Baird, R.W. 2000. The killer whale: Foraging specializations and group hunting. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead H (eds), Cetacean societies. Field studies of dolphins and whales. Pp 127-153. The University of Chicago Press, Chicago. Baird, R.W. & Whitehead, H. 2000. Social organization of mammal-eating killer whales: group stability and dispersal patterns. Can J Zool 78: 2096-2105. Baird, R.W. & Dill, L.M. 1996. Ecological and social determinants of group size in transient killer whales. Behav Ecol 7: 408-416. Bearzi, G., Notarbartolo-di-Sciara, G. & Politi, E. 1997. Social ecology of bottlenose dolphins in the Kvarneric (Northern Adriatic Sea). Mar Mamm Sci 13: 650-668. Bejder, L., Fletcher, D. & Bräger, S. 1998. A method for testing association patterns of social animals. Anim Behav 56: 719-725.

156 Social dynamics of facultative river dolphins

Berta, A. & Sumich, J.L. 1999. Marine mammals: evolutionary biology. Academic Press, London. Best, R.C. & Da Silva, V.M.F. 1984. Preliminary analysis of reproductive parameters of the boutu, Inia geoffrensis, and the tucuxi, Sotalia fluviatilis, in the Amazon River system. Rep Int Whal Comm Spec Issue 12: 361-69. Big, M.A., Ellis, G.M., Ford, J.K.B. & Balcomb, K.C. 1987. Killer whales: A study of their identification, genealogy, and natural history in British Columbia and Washington State. Phantom Press, Nanaimo, B.C. Big, M.A., Olesiuk, P.F., Ellis, G.M., Ford, J.K.B. & Balcomb, K.C. (1990) Social organization and genealogy of resident killer whales (Orcinus orca) in the coastal waters of British Columbia and Washington State. Rep Int Whal Comm Spec Issue 12: 383-405. Bräger, S., Würsig, B., Acevedo, A. & Henningsen, T. 1994. Association patterns of bottlenose dolphins (Tursiops truncatus) in Galveston Bay, Texas. J Mamm 75: 431- 437. Bräger, S. 1999. Association patterns in three populations of Hector’s dolphin, Cephalorhynchus hectori. Can J Zool 77: 13-18. Cockcroft, V.G. 1989. Biology of Indopacific humpback dolphins (Sousa plumbea) off Natal, South Africa. Abstracts of the Eight Biennial Conference on the biology of Marine Mammals. Society of Marine Mammalogy, p 13. Connor, R.C., Smolker, R.A. & Richards, A.F. 1992. Dolphin alliances and coalitions. In: Harcourt, A.H. & De Waal, F.B.M. (eds) Coalitions and alliances in humans and other animals. Oxford University Press, Oxford. Pp 415-443. Connor, R.C., Heithaus, M.R. & Barre, L.M. 1999. Super alliance of bottlenose dolphins. Nature 397: 571-572. Connor, R.C. 2000. Group living in whales and dolphins. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds), Cetacean societies. Field studies of dolphins and whales. Pp 199-218. The University of Chicago Press, Chicago. Connor, R.C., Wells, R.S., Mann, J. & Read, A.J. 2000a. The Bottlenose dolphin: Social relationships in a fission-fusion society. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds), Cetacean societies. Field studies of dolphins and whales. Pp 91-126. The University of Chicago Press, Chicago. Connor, R.C., Read, A.J. & Wrangham, R. 2000b. Male reproductive strategies and social bonds. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds), Cetacean societies. Field studies of dolphins and whales. Pp 247—269. The University of Chicago Press, Chicago. Connor, R.C., Heithaus, M.R. & Barre, L.M. 2001. Complex social structure, alliance stability and mating access in a bottlenose dolphin 'super-alliance'. Proceedings of the Royal Society of London series B – Biological Sciences 268 (1464): 263-267. Davies, N.B. 1991. Mating systems. In: Krebs J.R., Davies, N.B. (eds), Behavioral ecology- and evolutionary approach. Pp 263-294. Blackwell, Oxford. Dhandapani, P. 1992. Status of the Irrawaddy River dolphin Orcaella brevirostris in Chilka lake. J Mar Biol Assoc India 34: 90-93.

157 Chapter 9

Ford, J.K.B., Ellis, G.M. & Balcomb, K.C. 1994. Killer whales: The natural history and genealogy of Orcinus orca in British Columbia and Washington State. University of Washington Press, Seatle. Fowler, J. & Cohen, L. 1990. Practical statistics for field biology. Open University Press, Philadelphia. Gowans, S. 1999. Social organization and population structure of northern bottlenose whales in the Gully. PhD thesis, Dalhousie University, Halifax, Nova Scotia. Ginsberg, J.R. & Young, T.P. 1992. Measuring association between individuals or groups in behavioural studies. Anim Behav 44: 377-379. Jefferson, T.A. 2000. Population biology of the Indo-Pacific hump-backed dolphin in Hong Kong waters. Wildlife Monographs 144. Johnson, J.H. & Norris, K.S. 1994. Social behavior. In: Norris KS, Würsig B, Wells RS & Würsig M (eds), The Hawaian spinner dolphin. Pp 243-286. University of California Press, Berkeley. Karczmarski, L. 1999. Group dynamics of humpback dolphins (Sousa chinensis) in the Algoa Bay region, South Africa. J Zool Lond 249: 283-293. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Z Säugetierk 64: 54- 58. Kreb, D. 2002. Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. Raffles Bull Zool, Supplement 10: 85-95. Kreb, D. (in press) Abundance of freshwater Irrawaddy dolphins in the Mahakam in East Kalimantan, Indonesia, based on mark-recapture analysis of photo-identified individuals. J Cetacean Res Manage. LeDuc, R.G., Perrin, W.F. & Dizon, A.E. 1999. Phylogenetic relationships among the delphinid cetaceans based on full cytochrome b sequences. Marine Mammal Science 15: 619-648. Lloze, R. 1973. Contributions a l’étude anatomique, histologique et biologique de l’Orcaella brevirostris (Gray -1866) (Cetacea-Delphinidae) du Mekong. PhD thesis, Toulouse France [In French]. Lusseau, D., Schneider, K., Boisseau, O.J., Haase, P., Slooten, E. & Dawson S.M. 2003. The bottlenose dolphin community of Doubtful Sound features a large proportion of long-lasting associations. Can Geographic isolation explain this unique trait? Behav Ecol Sociobiol 54: 396-405. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. (eds). 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press, UK. Manly, B.F.J. 1995. A note on the analysis of species co-occurrences. Ecology 76: 1109- 1115. Matkin, C.O., Ellis, G., Olesiuk, P. & Saulitis, E. 1999. Association patterns and inferred genealogies of resident killer whales, Orcinus orca, in Prince William Sound, Alaska. Fish Bull 97: 900-919.

158 Social dynamics of facultative river dolphins

Őstman, J. 1991. Changes in aggressive and sexual behavior between two male bottlenose dolphins (Tursiops truncatus) in a captive colony. In: Pryor K, Norris KS (eds), Dolphin societies: Discoveries and puzzles. Pp 305-317. University of California Press, Berkeley. Pryor, K. & Shallenberger, I.K. 1991. Social structure in spotted dolphins (Stenella attenuata) in the tuna purse seine fishery in the eastern tropical Pacific. In: Pryor, K., Norris, K.S. (eds), Dolphin societies: Discoveries and puzzles. Pp 161-196. University of California Press, Berkeley. Read, A.J., Wells, R.S., Hohn, A.A. & Scott, M.D. 1993. Patterns of growth in wild bottlenose dolphins, Tursiops truncatus. J Zool 231: 107-123. Reeves, R.R., Smith, B.D. & Kasuya, T. 2000. Biology and conservation of freshwater cetaceans in Asia. Occasional Paper of the IUCN Species Survival Commission 23. IUCN, Gland, Switzerland. Schnapp, D. & Howroyd, J. 1992. Distribution and local range of the Orinoco dolphin (Inia geoffrensis) in the Rio Apure, Venezuela. Z Saugetierk 57: 313-315. Slooten, E., Dawson, S.M. & Whitehead, H. 1993. Associations among photographically identified Hectors dolphins. Can J Zool 71: 2311-2318. Smith, D. 1993. Status and conservation of the Ganges river dolphin Platanista gangetica in the Karnali River, Nepal. Biol Conserv 66: 159-169. Smolker, R.A., Richards, A.F., Connor, R.C. & Pepper, J.W. 1992. Sex differences in patterns of associations among Indian Ocean bottlenose dolphins. Behaviour 123: 38-69. Tolley, K.A., Read, A.J., Wells, R.S., Urian, K.W., Scott, M.D., Irvine, A.B. & Hohn, A.A. 1995. Sexual dimorphism in wild bottlenose dolphins (Tursiops truncatus) from Sarasota, Florida. J Mammal 76: 1190-1198. Wells, R.S. 1991. The role of long-term study in understanding the social structure of a bottlenose dolphin community. In: Pryor K & Norris KS (eds), Dolphin Societies. Discoveries and puzzles. Pp 199-225. The University of California Press, London. Whitehead, H. 1999. Testing association patterns of social animals. Anim Behav 57: 26-29. Whitehead, H. & Mann, J. 2000. Female reproductive strategies of cetaceans: Life histories and calf care. In: Mann, J., Connor, R.C., Tyack, P.L. & Whitehead, H. (eds), Cetacean societies. Field studies of dolphins and whales. Pp 219-246. The University of Chicago Press, Chicago. Wilson, D.R.B. 1995. The ecology of bottlenose dolphins in the Moray Firth, Scotland: a population at the northern extreme of the species’ range. PhD thesis, Aberdeen University, Aberdeen Scotland.

159 Chapter 9

160 Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

CHAPTER 10

Impacts of habitat on the acoustic behaviour of coastal and freshwater Irrawaddy dolphins, Orcaella brevirostris in East Kalimantan, Indonesia

Daniëlle Kreb and Junio Fabrizio Borsani

One group of six dolphins was trapped since 1999 until present in between two rapids in a habitat of only 2 km of length. The dolphins acoustic behaviour showed a low diversification and no whistles were heard.

161 Chapter 10

ABSTRACT

Vocal repertoires of one coastal and one river population of Irrawaddy dolphins, Orcaella brevirostris, in Balikpapan Bay and Mahakam River, respectively, were studied in 2002 in order to reveal the impact of habitat on acoustic behaviour and in particular of social communication. Vocalizations were most varied and frequent in one core area of dolphins in the river, in which a well-identified sub-population with a high site- fidelity occurred and with the highest degree of social exchange among groups in comparison to two other areas in the river and the coastal bay population. These vocalizations included single- (“jaw-claps”) and multiple- broadband (“click trains”) clicks; broadband- (“squeaks” and “creaks”) and narrow-band- tonal pulsed sounds (“grunts”, “moans” and “quacks”); narrow-band frequency-modulated sounds, i.e., 2 types of calls and whistles of up to 5 frequency modulations. Pod-specific whistle- dialects exist among coastal and riverine populations, but also within sub-pods within the river, which differ in the number of modulations, duration, minimum and maximum frequencies. Call sharing occurred among neighbouring areas but not among remote areas. There is also evidence for individual “signature” whistles and “contact” whistles. Vocal repertoire (sound types) was more similar between the likely more genetically related, coastal and freshwater populations in East Kalimantan than between coastal populations of Australia and East Kalimantan. Vocal repertoire was less varied for coastal Irrawaddy dolphin populations in East Kalimantan and Australia compared with the Mahakam River and may be determined by ecological conditions. The whistles and vocalizations rates (numbers per time unit) seem to be determined by social structures. Larger groups with (more) calves whistled less often than smaller groups, which may be caused by the fact that there is less need for contact whistles. Whistle frequencies were significant higher upon approach of (speed) boats of > 40 hp and lasted longer than in their absence.

RINGKASAN

Tipe suara dari satu populasi lumba-lumba Irrawaddy (Orcaella brevirostris) laut dan sungai, telah dipelajari pada 2002 untuk mengungkapkan pengaruh habitat pada tipe suara dan khususnya komunikasi sosial. Suara lebih banyak berubah-ubah dan lebih sering dalam satu tempat utama lumba-lumba di sungai, dalam suatu sub populasi yang telah diketahui dengan tingkat kesetiaan pada tempat yang tinggi dan dengan tingkat pertukaran antar kelompok tertinggi dalam perbandingan dengan populasi di dua daerah lain di sungai dan laut dalam teluk. Suara ini termasuk “jaw-clap” tunggal dan banyak “click” dengan frekuensi yang luas dan dalam waktu lama; frekuensi “squeak dan creak” luas dan suara dengan frekuensi rendah (menggumam, melenguh dan berkotek), dua (2) tipe dari panggilan dan bunyi siulan dari 5 atau lebih modulasi frekuensi. Juga terdapat siulan dengan dialek khusus untuk kelompok di antara populasi laut dan suangai, namun juga antara sub-kelompok di sungai, dimana berbeda

162

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

dalam jumlah modulasi, durasi, frekuensi minimum dan maksimum. Saling memanggil terjadi diantara daerah-daerah yang bertetangga tetapi tidak dengan daerah yang jauh. Didapatkan juga bukti “siulan khas” dari setiap individu dan “siulan kontak”. Tipe suara hampir sama antara individu yang berhubungan secara genetis, populasi laut dan air tawar di Kalimantan Timur lebih banyak memiliki kesamaan dibandingkan dengan populasi laut di Kalimantan Timur dan laut di Australia. Tipe suara lebih sedikit macamnya untuk populasi lumba-lumba laut di Kalimantan Timur dan Australia dibandingkan populasi di Sungai Mahakam dan mungkin ditentukan oleh kondisi ekologi. Rata-rata siulan dan vokalisasi (jumlah per unit waktu) nampaknya ditentukan oleh struktur sosial. Kelompok yang lebih besar dengan (lebih banyak) anak lebih jarang bersiul dibanding kelompok yang lebih kecil, yang mungkin disebabkan karena tidak dibutuhkan banyak kontak siulan. Frekuensi siulan lebih tinggi saat ada dengan tenaga > 40 stk dan berakhir lebih lama dibandingkan pada saat tidak ada kapal.

INTRODUCTION

General background

River dolphins and porpoises are among the world’s most threatened mammal species. The habitat of these animals has been highly modified and degraded by human activities, often resulting in dramatic declines in their abundance and range (Reeves et al., 2000). In Indonesia, one representative freshwater dolphin population is known to inhabit the Mahakam River and associated lakes system in East Kalimantan, i.e., the facultative river dolphin species Orcaella brevirostris, commonly and locally referred to as the Irrawaddy Dolphin or pesut, respectively. The species is found in shallow, coastal waters of the tropical and subtropical Indo-Pacific and in the following major river systems: Mahakam, Ayeyarwady, Mekong (Stacey & Arnold, 1999). The species is protected in Indonesia and adopted as symbol of East Kalimantan. Based on monitoring surveys conducted from 1999 until 2002 in the Mahakam, which indicated a population abundance of less than 50 individuals (Kreb, 2002), the IUCN has raised the status of the Mahakam Irrawaddy dolphin population from ‘Data Deficient’ to ‘Critically Endangered’ in 2000 (Hilton-Taylor, 2000). The only information on the acoustic behaviour of the Irrawaddy dolphins in the Mahakam pertains to a study of these dolphins in a captive environment (Kamminga et al., 1983). The dolphin was described to be a lively, fervently vocalizing animal similar to the Amazonian freshwater Boto, Inia geoffrensis, but no audible whistles or pure tones were observed. More recent studies include an unpublished study on freshwater Irrawaddy dolphins in the Mekong River (Borsani, 1999), during which whistles, jaw-claps, pulse-trains, and single clicks could be heard. Another study was conducted on coastal Irrawaddy dolphins in Australia, which revealed that the dolphins exhibited a varied repertoire consisting of broadband clicks, pulsed sounds

163 Chapter 10

that resemble the freshwater tucuxi, Sotalia fluviatilis and nonwhistling delphinids (e.g., Cephalorhynchus spp.). Whistles were found simple in form and low in frequency (1 to 8 kHz) (Van Parijs et al., 2000). In the present study, we have observed the acoustic behaviour of two populations of free-ranging freshwater and coastal Irrawaddy dolphins in the Mahakam River and Balikpapan Bay in East Kalimantan, respectively. Our objectives are 1) to identify and compare the vocal repertoire and amount of vocalizing of both populations but also within different identified sub-pods in the Mahakam River, 2) to relate vocalizations to behaviours displayed, 3) compare the vocalizations with those from Irrawaddy dolphins in Australian coastal waters and in the Mekong River to investigate whether the acoustic behaviour of the Irrawaddy dolphin follow an ecological (freshwater/ coastal) and/or geographical separation (Asia/ Indonesia/ Australia), 4) to investigate whether whistle shapes and frequencies are more determined by ecological, genetic or social factors by identifying whether whistles within different species of river dolphins and within sub-pods of one population of Irrawaddy dolphins in the Mahakam are more or less similar to each other, than those whistles of ecologically different, but more geographically, nearby living populations of the same species that are most likely more genetically related. The comparison of the acoustic behaviour of the freshwater and coastal Irrawaddy dolphin populations may be valuable in terms of determining whether the Mahakam population is an isolated, single breeding population that needs careful management to maintain a viable population.

METHODS

Study areas

Acoustic recordings of freshwater Irrawaddy dolphins in the Mahakam River were made at three different study sites (Figure 1). Core areas 1 and 2 were areas of high dolphin densities in the Middle Mahakam River, from 180 km to 375 km from the mouth, which included confluence areas of the main river with tributaries and lakes. These areas were chosen because dolphins in each area had a high site fidelity based on a photo-identification study (Kreb, in press a). Mean river width in this area measured 200 m (SD = 53 m, n = 105), whereas mean water depth at an average water level was 15 m (SD = 6 m, n = 65). Mean water clarity in the study area (measured with a Secchi disk) at an average water level was 23 cm (SD = 7 cm, n = 27). Bottom substrate is muddy. The middle Mahakam is an area of intensive fishing activity (MacKinnon et al., 1997). Some coal mining and logging activities occur here, especially in the tributaries. Furthermore, this area is subject to intensive boat traffic with boats passing every 3 minutes on average, mostly constituting boats of less than 40 hp (Kreb, in press b). The transport infrastructure is poorly developed in East Kalimantan and the Mahakam River is the main transport artery. The Ratah tributary

164

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

Area 3

Core area 1 Core area 2

Figure 1. Map of both study areas, the Mahakam River and Balikpapan Bay. Three study areas in the river are indicated by square boxes.

(core area 3) joins the Mahakam at 500 km from the mouth. The dolphin habitat is an area of only 2 km length with a rocky bottom and shore substrate in between two rapid streams. Only long, motorized canoes (40 hp and higher) pass the area irregularly. Acoustic recordings of coastal Irrawaddy dolphins were made in Balikpapan Bay, which stretches from 116o42’ to 116o50’ E and 1o to 1o22’ S (Figure 1). Water surface area of the bay is ca. 120 km2. Maximum width of the bay is ca. 7 km. Shorelines within he bay consisted mainly of mangrove vegetation. Average water depth at dolphin sightings within the bay was 14.5 m (SD = 8.0 m, n = 39).

165 Chapter 10

Mean clarity recorded at sighting locations in the bay is 170 cm (SD = 58 m, n = 24). Boat traffic was most frequent in the downstream part of the bay, where mostly ferries and speedboats crossed the bay in one lane. In one of the mangrove tributaries where most recordings were made, we encountered mainly speedboats that frequented a logging company upstream. Small fishing boats could be found in all areas of the bay and shrimp fishing was the most common activity. Bottom substrate was mainly muddy/ sandy. Dolphin densities were more or less equally distributed throughout the bay.

Data acquisition

The acoustic study was conducted during the months April, May 2002 at medium water levels and repeated in August 2002 during low water levels. The high water level season was excluded because it was assumed that the higher flow rates would cause more background noise and would cause difficulties in making recordings from the small research canoe. Total effective recording effort consisted of 16 h during 21 days, excluding searching and travel days. In the Mahakam, 12 h recordings were made over 14 days and in Balikpapan Bay 4 h over 7 days. Recordings were made of group in between 2 and 8 individuals in the Mahakam and between 2 and 10 individuals in Balikpapan Bay. Recordings were made from different groups of dolphins (different group sizes and/ or composition) in several different areas of the river and bay. Recording time per group was more or less equally spread over the entire length of the day by strategically visiting those areas where a high chance existed to encounter a group of dolphins soon after searching commenced. These areas were already identified during several abundance monitoring surveys in 1997 and from 1999 until 2002 (Kreb, 1999; Kreb, 2002). When a group of dolphins was encountered we attempted to follow it for the entire day until 1800 h at maximum and continually made recordings (average group follow time = 7 h; SD = 3 h; range = 1.5 -13 h). For each recording session, starting and ending time of the session (in h/min/s), group size and composition (presence of neonates, calves or juveniles), general group behaviour (feeding, milling, socializing, travelling or a combination between these categories), and spatial group distribution were recorded. A recording session lasted as long as good recordings could be made and no change in group size/ composition or in general group behaviour occurred. The survey team existed of four persons: one sound recorder; one data recorder, who wrote down individual behaviours displayed for each minute; a second data recorder and observer, who drew the spatial distribution of the group and recorded distances between individuals and towards the hydrophone each minute; one observer, who would inform the “behaviour” data recorder about the individual behaviours displayed by the dolphins.

166

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

A High Tech Inc.- 94-SSQ hydrophone was used for recording (frequency range: 2Hz - 30 kHz at –168 dB re 1V/µPa) that hung 1.5 m deep into the water. The hydrophone was connected to a Sony TCD-100 DAT walkman-recorder set at a sampling rate of 48 kHz (providing an effective frequency range of 20 Hz to 24 kHz). SONY DAT tapes (60-90 min) were used. Dolphin observations were conducted from two types of vessels. The first was a wooden boat with inboard engine of 26 hp and 16 m length with observer eye-height 3.5 m above the water. This boat was used if the dolphins were milling in one area so that the boat could stand-by with engine off, which provided better observation positions than the second boat we employed that was a wooden canoe of 10 m length with an outboard motor of 5 hp and with observer eye-height 1 m above the water. This small boat was used if the dolphins spent a great deal of the time traveling and did not remain for a long time in one area.

Analysis

Sound recordings were down-loaded at a 48 kHz sampling rate in a PC using a Windows spectral analysis program Syrinx 2.2 l (Burt, 2003). The sounds were stored to wave sound files that could be linked to a database with time and date of the recording, location, group size and composition, and spatial distribution. All files were grouped per general behavior category. Recording sessions (with continuous sound recording) that were longer than 4 minutes (~ 22 MB) were split into two or three files for better handling in analysis as otherwise a lot of memory was needed to open and display the files. This resulted in a total of 149 sound files with a total duration of 4.5 hours recording (mean duration = 1.8 min; SD = 1.4 min) for the Mahakam and 26 files with a total duration of 0.7 hours (mean duration = 1.7 min; SD = 1.6 min) for the Balikpapan Bay study areas. Only those sections from sessions were down-loaded that were analyzable, leaving bad recording sections out with too much background noise. For the Balikpapan Bay area silent passages, where dolphins did not vocalize in spite of their close presence, were included in analyses to quantify number of vocalizations per behavior and duration category. The numbers of sounds within a sound type were counted for all files combined per behavior category, except for broad-band clicks (pulse trains). These were so numerous that pulse trains in the river were counted until a number of 304 samples were obtained and 26 samples in Balikpapan Bay

167 Chapter 10 es, 157-12313 404-2740 0-1434 0-557 0-9932 574-4822 range D freq 2810 1994 mean D freq = continuous frequency & modulated 7

532-6318 1243 91-5582 257-13569

459-8544 1538-4320 388 0-1847 388 1538-4320 22000 22000 = rep rate >500< 1500; mean max freq < 5000 Hz; 3 4453 3819 mean max freq freq range . > 6856 10482 3469-17704 3874 579-11708 1643 > > 22000 22000 > 5998 6866 5998-13902 1097 398-2341 1825 5863 5714 1907-9252 1585 231-5204 6193 8234 6193-15848 2317 332-6836 6132 8001 6132-18442 1868 0-9845 mean min freq 3472 3884 3472-9320 415 3472-9320 3884 3472 5068 5187 5068-11783 127 5068-11783 5187 5068 226-1244 3101 4035 3101-6484 1332 349-1511 132-1199 6075 8793 6075-16704 2638 range rep. rate 447-1783 m River and described according to their duration, repetition rat 168 463 899 mean rep. rate = continuous frequency & modulated calls.; 6 0.912 0.047 mean duration (sec) 2-97 2.472 35 >22000 0.75 404 0.75 0.69 905 0.69 e differences between these frequencies = rep rate < 500Hz; mean max freq >22000 Hz; 2 n 16 0.174 16 0.174 1859 1538 65 0.05 532 1775 396 213-736 = wavering calls; 5

jaw-clap 0.044 134 whistle 0 mod down 16 0.264 whistle 1 mod down 16 0.262 whistle 2 mod down 5 0.174 quack whistle 0 mod up 45 0.235 "common name" sound type whistle 1 mod up 86 0.262 whistle 2 mod up 25 0.308 whistle 3, 4, 5 mod 25 0.487 call 1 2 0.371 19 call 0.295 (CF) 64 call whistle 5 6 7 squeak 50 squeak 118 creak 18 grunt 97 moan 1 2 3 4 minimum and maximum frequencies and th Vocalizations identified for Irrawaddy dolphins in the Mahaka

narrowband frequency sounds modulated narrowband tonal pulsed sounds broadband tonal pulsed sounds narrowband tonal pulsed sounds narrowband frequency sounds modulated narrowband frequency sounds modulated loud single broadband click broadband tonal pulsed sounds = rep rate > 500< 1500; mean max freq >22000 Hz; = rep rate < 500Hz; mean max freq < 5000 Hz; 5 4 1 2 multiple broadband clicks pulse click trains >304 3

whistles

Chapter 10 Table 1. Categories sound types 1 4

168 Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

Quack Mew Moan Frequency (kHz) 22

2 0.000 kHz S0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Duration (seconds)

Creak Broad-band clicks Squeaks Frequency (kHz) 22

2 0.000 kHz S0 0.5 1 1.5 2 Duration (seconds)

Figure 2. Spectrographic representation (fast Fourier transforms, sample rate 48 kHz, FFT size = 512) of three types of broadband tonal, pulsed sounds (top figure) and two types of narrowband tonal, pulsed sounds, i.e., Creak and Squeaks (figure below). In the center of the lower figure is a pulse train of broad-band clicks with increasing, decreasing and increasing click rates (65, 11 & 58 clicks per sec). The spectrogram is limited by the recording equipment, which was only able to record up to 24 kHz.

169 Chapter 10

Balikpapan Bay. These were considered enough to calculate average duration and click rates. In the Mahakam, good quality recording time was distributed over different behaviour categories as follows: feeding (43% of total recording time), slow swimming (34%), socializing (14%), milling activity (8%), and fast swimming (1%). In Balikpapan Bay, activities during which most recording sessions were made included feeding (57%), slowly swimming (39%), milling (3%), and fast swimming (1%). The term social communication is explained in this article in terms of vocalizations, which are directed to other individuals to transfer different kinds of information (individual identification, i.e., so called “signature” calls first described by Caldwell & Caldwell (1965), or to establish or maintain contact (Tyack, 1987; Sayigh et al., 1990; Smolker et al., 1993). To compare social communication for different habitats we looked at whistles, which have been described to fit the mentioned earlier meaning. We used similar group sizes and behaviours, and excluded whistles made during approaching boats to compare 1) the numbers of whistles in coastal and freshwater habitat and within different core areas in the river, 2) minimum and maximum frequencies, and 3) whistle duration.. The measurements of whistles were compared using the non-parametric Mann-Whitney U-test (Fowler & Cohen, 1990). To compare the numbers of whistles in different areas, the numbers of whistles per behaviour in each area were divided by the recording time for that behaviour, and were then added and divided by the number of behaviour categories. The resulting average number of whistles per time unit was multiplied with the mean recording effort of all areas to obtain the total number of whistles per area with equal recording effort. The numbers of whistles were compared between the areas using a Chi-square test and applying Yates’ correction when there was only 1 degree of freedom. We also tested whether there was a correlation between the number of whistles per min recording effort and group size, as well as the number of calves in a group, using the Product Moment Correlation Coefficient (r). To test the impact of approaching (speed)boats of > 40hp on whistle duration and frequencies, we compared whistles with the same modulations, same behaviour categories, same group sizes and in one area, in the absence and presence of these boats using the Mann-Whitney U-test.

RESULTS

Repertoire

Sounds produced by dolphins in the Mahakam River could be distinguished broadly in five categories: 1) loud single broadband clicks, 2) multiple broadband clicks, 3) broadband tonal pulsed sounds, 4) narrow-band tonal pulsed sounds, 5) narrow-band frequency modulated sounds (Table 1). Each different type of vocalization is graphically represented in a spectrogram (Figures 2 & 3). After pulse trains (> 304), whistles are the most common sounds produced by the dolphins (n = 282), then jaw claps (n = 134), creaks (n = 118), moans (n = 97), “metallic” quacks (n = 65), what we

170

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

Whistles (no. of modulations)

Call 1 Call 2 1 mod. Jaw clap 0 mod. 5 mod. Frequency (kHz) 22

2 0.000 kHz S0 0.5 1 1.5 2 2.5 3 Duration (seconds)

Figure 3. Spectrographic representation of narrowband frequency-modulated sounds. Calls of type 1 had a wavering contour, whereas calls of type 2 were of continuous frequencies or modulated as shown here. Whistles with up to 5 modulations were produced in the Mahakam River. In this graph, the whistle of 1 modulation was followed by a loud, single broadband click (“jaw-clap”). The whistle with 0 modulation and large frequency increase (from 892 Hz to 6243 Hz) in this graph preceded a boat approach.

termed call 1 (long, weep-like sound; n = 19), grunt (cow-like sound; n = 18), and call 2 (short, bird-like cries; n = 16). Whistles with up to five frequency modulations were found, but the most common whistle had one modulation (Table 2). Most whistles (64%) commenced with an initial raising frequency, whereas 23% maintained a continuous frequency and 13% commenced with a decreasing frequency. Distinctive whistles were identified based on their spectral contours. On several occasions identically shaped whistles were found during one sighting, but also during sightings on other days in the same area. The number of distinctive whistles during some sessions exceeded the group size by twice the number. Most calls of type 2 had zero- modulations (63%), then one modulation (25%), and 3 modulations (12%). In total, 6 different calls of type 2 were distinguished based on their spectral contours. Vocalizations produced by coastal Irrawaddy dolphins in Balikpapan Bay could be distinguished in the same five categories as those in the Mahakam River (Table 3).

171 Chapter 10 , range D freq 513 358 117-673 368 234-594 1466 0-3906 3085 904-5266

mean D freq

continuous frequency & modulated 4 7065 2438-11692 92 0-185 >22000 >22000 >22000 mean max freq range freq

1589 2102 9884 10804 2699-16081 919 280-2083 6983 7805 1589-10880 822 513-1317 9297 10407 8612-10695 1110 411-2083 >22000 mean min freq wavering calls; 3 249-359 range rep. rate 309 n Bay and described according to their duration, repetition rates 172 mean rep. rate 0.101 mean duration (sec) e differences between these frequencies. 2 0.302 503 426-580 1914 4999 903-8192 4 0.144 6916 7128 1639-11616 4 0.342 1890 2258 1043-3862 25 0.059 445 148-2019 1431 2898 484-5344 rep rate < 500Hz; mean max freq < 5000 Hz; 2 pulse click trains >26 5.92 53 20-90 Whistles 2 down 1 0.054 Whistle 1 mod up 9 0.183 Whistle 2 mod up 3 0.179 Whistle 0 mod up 4 0.147 quack creak 9 moan 1 call (CF) whistle 2 0.243 6972 Whistle 0 mod down "common name" sound type n jaw-clap 8 0.039 jaw-clap 0.039 8 3 4 2 minimum and maximum frequencies and th 1 Vocalizations identified for Irrawaddy dolphins in Balikpapa

narrowband frequency modulated sounds narrowband frequency modulated sounds narrowband tonal pulsed sounds loud single broadband click multiple broadband clicks broadband tonal pulsed sounds rep rate < 500Hz; mean max freq >22000 Hz; 5 4 1 2 3 1 Table 2.

Chapter 10 Categories sound type whistles

172 Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

Table 3. Whistle types and their abundance per study area

AREAS Recording Total Whistles effort number (min) whistles CF Mod = 0 Mod = 1 Mod = 2 Mod = 3,4,5 Mahakam 271 281 23 % 22 % 36 % 10 % 9 % Balikpapan 166 24 8 % 33 % 38 % 17 % 4 % Area 11 76 21 24 % 14 % 62 % - - Area 21 183 260 23 % 22 % 34 % 11 % 10 %

1 = Both areas are in the Mahakam River

The repertoire of individual sound types was less varied though, and the most common sounds produced in decreasing order are pulse trains (n > 26), quacks (n = 25), whistles (n = 24), creaks (n = 9), jaw-claps (n = 8), type 1 calls (wavering calls; n = 4), moans (n = 2). Whistles were produced with up to four frequency modulations, and the most common whistle had one modulation (Table 3). Most whistles (71%) commenced with an initial raising frequency, whereas 21% commenced with a decreasing frequency and 8% of whistles had a continuous frequency. Eighteen whistles were distinguished based on their spectral contours, of which some were repeated during the same or different sightings. Differences between repertoires were found among three core areas in the Mahakam River. “Upstream” area 2 had the most varied repertoire with all sounds produced as in table 1and whistles of up to 5 modulations. “Downstream” area 1 lacked the calls of type 2 and whistles consisted only of zero to one modulation. In the rapid stream area 3, no moans, calls of type 1 and whistles were heard at all. Also, calls of type 2 in area 3 were of continuous frequency, whereas in area 2 these calls had up to 3 modulations. In both areas 1 and 2, whistles with one modulation of frequency were most common (Table 3). In area 1, we also found “whistle trains”, individual whistles produced in fast sequence consisting mostly of 4 whistles with mean interval of 0.15 sec (SD = 0.036; n = 8).

Vocalizing time per (sub)population

173 Chapter 10

2 2 Sounds per min n sounds Total 118 799 4.1 799 118 Creak % Whistle % Squeak % = average number of sounds per minute of different behaviour 2 2 24 2.5 9 72 Moan % 174 - - - 100% - 16 8 - 16 - - - 100% 5 - 10 - - - 100% 1% 12% 13% 23% 20% 277 2.4 23% 277 13% 20% 1% 12% - - - 50% - 10 0.15 - 10 - - - 50% 4% 9% 2% 48% 3.1 8% 286 - 3% - 24% - 29 0.31 - 29 - 24% - 3% 0.4% 21% 4% 26% 20% 174 4.7 4% 26% 174 21% 20% 0.4% - 4% - 9% 39% 23 4.6 23 - 9% 39% - 4% 4% 7% 4% 46% 9% 46 2.2 Grunt %

Jaw clap % Call2 % Call1 % 1 Quack % effort (min) 269 min 65 19 16 134 18 97 50 282 65 19 16 134 min 269 167 min 25 4 - 8 - 4 - 8 25 min 167

Sound types per behavioral category in the Mahakam River and Balikpapan Bay. n n = percentage of occurrence of each sound type within one behavior category.; categories Total 2 - - - - fast Swim Swim fast 2 - - - - fast Swim Play 37 3% 4% 1% 21% 5 - - 4% Milling 44% Total Balikpapan Chapter 10 Table 4. Mahakam Recording 92 slow 2% 1% 15% 11% Swim 117 Feeding - - 30% 65 9% 3% 3% 16% 20% slow - 14% Balikpapan Swim 14% 95 Feeding 45% Milling 21 7% - 4% 19% 1

174 Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

all whistles. Next, most frequent sounds were heard during play and included in decreasing order mostly whistles, moans, and jaw-claps. During slow swimming, dolphins produced different sound types more or less equally often. During feeding and milling, whistles and jaw-claps were most frequent. In Balikpapan, sounds were just as frequently heard as in the river during the behaviour activities, fast swimming, and milling. During fast swimming, dolphins exclusively produced whistles, whereas during milling quacks and creaks were most common. A low number of vocalizations per minute were heard during slow swimming and feeding. During feeding, quacks, whistles, jaw-claps, and calls were emitted most frequent in decreasing order, whereas during slow swimming, first whistles, then jaw-claps, and finally barks prevailed.

Vocalizing time per (sub)population

All sound types Dolphins in the Mahakam River were significantly more vocal than their coastal relatives, where 4 sounds other than pulse trains (and including whistles) per minute recording were produced on average in contrast to 2.5 sounds per minute in Balikpapan Bay (X2 = 78; df = 1; P < 0.01) (Table 4). Most sounds per minute recording effort were produced by river dolphins during fast swimming, which were all whistles. Next, most frequent sounds were heard during play and included in decreasing order mostly whistles, moans, and jaw-claps. During slow swimming, dolphins produced different sound types more or less equally often. During feeding and milling, whistles and jaw-claps were most frequent. In Balikpapan, sounds were just as frequently heard as in the river during the behaviour activities, fast swimming, and milling. During fast swimming, dolphins exclusively produced whistles, whereas during milling quacks and creaks were most common. A low number of vocalizations per minute were heard during slow swimming and feeding. During feeding, quacks, whistles, jaw-claps, and calls were emitted most frequent in decreasing order, whereas during slow swimming, first whistles, then jaw-claps, and finally barks prevailed. Within the different core areas in the Mahakam River, differences in the amount of vocalizing were also found. Most sounds per minute were produced by dolphins in area 2, namely 4.4 sounds per minute, which was significantly different and higher than in the “rapid stream” area 3 and “downstream” area 1, where on average 1.7 sound per min and 1.1 sounds per min were heard (X2 = 232; df = 2; P < 0.01).

Whistles Most whistles in the Mahakam were produced during fast swimming (65%), then slow swimming (12%), playing (10%), milling (8%) and feeding (5%). Likewise, in Balikpapan Bay, most whistles were heard during fast swimming (90%). In comparison to Balikpapan Bay, significantly more whistles than expected were produced in the Mahakam River (X2 = 62; df = 1; P < 0.01). In the Mahakam River

175 Chapter 10

and Balikpapan bay, 2.4 whistles per minute and 1.4 whistles per minute were produced. When comparing the two core areas within the Mahakam River, we found that in areas 1 and 3 significantly less whistles were produced within the same recording time than in area 2 (X2 = 392; df = 2; P < 0.01). In areas 1 and 2, 0.2 whistles and 2.5 whistles are produced per minute, respectively. In area 3, no whistles were heard at all. There was a negative correlation (although not significant) between the number of whistles per minute recording effort and group size and number of calves per group (r = 0.48; df = 14; P = 0.06 & r = 0.46; df = 14; P = 0.07). Larger groups with (more) calves whistled less often than smaller groups.

Whistle and call characteristics per (sub) population

After combining all types of whistles irrespective of number of modulations, we found that whistles in the Mahakam had a mean duration of 0.23 sec (SD = 0.27; range = 0.02 – 3.09 sec), mean minimum frequencies of 5.8 kHz (SD = 2.9 kHz; range = 0.8 – 14.5 kHz) and mean maximum frequencies of 7.4 kHz (SD = 3.8 kHz; range = 1.4 – 18.4 kHz). All whistles combined in Balikpapan Bay had a mean duration of 0.18 sec (SD = 0.08 sec; range = 0.05 – 0.35 sec), mean minimum frequency of 8.7 kHz (SD = 4.1 kHz; range = 1.6 – 15.2 Hz) and mean maximum frequency of 9.4 kHz (SD = 4.3 kHz; range = 1.6 – 16.1 kHz). We found that whistles of one modulation commencing with raising frequencies in coastal and freshwater habitat were significantly different in minimum and maximum frequency, as well as in the differences between minimum and maximum frequency (U = 209, 280 & 54; n1 = 24 & n2 = 9; P < 0.01). Namely, whistles of this type in the bay had higher minimum and maximum frequencies (mean = 9.9 kHz & 10.8 kHz) but lower delta frequencies between minimum and maximum frequencies (0.9 kHz) than those produced in the river (mean minimum = 6.1 kHz; maximum = 8.0 kHz, Delta Frequency = 1.9 kHz). No significant differences in the duration of this type of whistle were found. Whistles types with zero modulations and initial raising or decreasing frequencies were not different between both habitats. Continuous frequency whistles of “downstream” area 1 in the Mahakam River were significantly lower in minimum and maximum frequencies (2.1 kHz & 2.3 kHz) than those in “upstream area” 2 (5.9 kHz & 6.1 kHz) (U = 50 for both min and max frequencies; n1 = 10 & n2 = 5; P < 0.01). Whistles of one modulation with initial raising frequencies were significantly different between both areas. Whistle duration in area 2 was significantly longer than in area 1 (mean = 0.15 sec & 0.26 sec, respectively) (U = 27.5; n1 = 16 & n2 = 10; P < 0.01). Minimum and maximum frequencies were significantly higher in area 2 (8.2 kHz & 10.3 kHz) than in area 1 (2.4 kHz & 3.1 kHz) (U = 45.5 & U = 136.5; n1 = 16 & n2 = 10; P < 0.05 & P < 0.01). Also, the differences

176

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

between the minimum and maximum frequencies of whistles were significantly different (Delta frequency = 0.7 kHz in area 1 & 2.1 kHz in area 2) (U = 18; n1 = 16 & n2 = 10; P < 0.01). Duration, minimum and maximum frequencies, and number of modulations of calls of type 1 were compared for core areas 1 and 2, and river and bay, whereas calls of type 2 were compared for areas 2 and 3. We found no significant differences for calls of type 1 between areas 1 and 2, and between the river and bay, although the mean minimum and maximum frequencies were much lower in the bay (mean minimum = 1.9 kHz; SD = 1.0 kHz; mean maximum = 2.2 kHz; SD = 1.2 kHz) than in the Mahakam (mean minimum = 3.5 kHz; SD = 2.4; mean maximum = 3.9 kHz; SD = 2.5 kHz). Since N for area 3 was only 3 calls no test could be applied, but the mean minimum and maximum frequencies were distinctively lower in area 3 (mean = 0.5 kHz: SD = 0.56 kHz) than those in area 2 (means = 1.7 – 2.1 kHz; SD = 0.9 & 1.2 kHz).

Impact of speed boats

Frequencies of continuous-frequency whistles were significantly higher in the presence of (speed) boats of > 40hp (mean = 5568 Hz) than in their absence (mean = 5247 Hz) (U = 72; n1 = 10 & n2 = 10; P = 0.05). Whistles with zero and one modulations with both initial raising frequencies all had higher minimum (means = 7508 Hz & 6355 Hz) and maximum frequencies (means = 10016 Hz & 83221 Hz) in presence of (speed)boats > 40hp than in their absence (means minimum frequencies = 6351 Hz & 6000 Hz; means maximum frequencies = 8123 Hz & 8326 Hz) (zero-modulation whistles: U = 66 & 109; n1 = 6 & n2 = 15; P = 0.05 & P < 0.01; one-modulation whistles: U = 205 & 296; n1 = 13 & n2 = 20; both P < 0.01). Duration of whistles of one modulation was also significantly longer in presence of these boats (mean = 0.365 sec) than in their absence (mean = 0.285 sec) (U = 174; n1 = 13 & n2 = 20, P = 0.05).

DISCUSSION

The acoustic behaviour of Irrawaddy dolphins in the Mahakam River and Balikpapan Bay differed in a number of aspects: vocal repertoire, amount of time spent vocalizing (all sound types and whistles) and shape, duration and frequencies of whistles. The freshwater dolphins had a more varied repertoire, spent more time vocalizing, and produced more complex whistles, which seems to fit with their turbid environment in which they have to rely on acoustics instead of vision and also reflects a dynamic, social system. Their whistle frequencies were lower, but they had higher delta frequencies of minimum and maximum frequencies than their coastal relatives.

177 Chapter 10

The riverine Irrawaddy dolphins’ whistle frequency ranges (0.8 -18.4 kHz) resemble most those whistles of the riverine Indus dolphin, Platanista g. gangetica and Yangtze dolphin or baji, Lipotes vexilifer, which had frequency ranges of 0.8 to 16 kHz (Andersen & Pilleri, 1970; Pilleri et al., 1971) and 3 to 18.4 kHz (Jing Xianying et al., 1981; Xiao Youfu & Jing Rongcai, 1989), respectively. Also, the dominant frequencies of the baji of 6 kHz corresponded with those of the pesut, which were between 5.8 and 7.3 kHz. In the Mekong River in Laos, whistles of one pod of Irrawaddy dolphins were described (Borsani, 1999), which all corresponded to one type of whistles of two modulations with initial raising frequency and with mean minimum and maximum frequencies of 3 kHz and 8 kHz. On the other hand, in the Mahakam whistles of this type had higher mean, minimum frequencies, i.e., 6.0 kHz, and similar mean, maximum frequencies, i.e. 8.8 kHz. Dominant whistle frequencies were highest for tucuxi, Sotalia fluviatilis (10-15 kHz) of all other (facultative) river dolphin species and whistles also had a higher, maximum frequency range, i.e. 23.9 kHz (Da Silva & Best, 1994; Wang Ding et al., 1995). However, their whistles were simpler as they only produced simple, zero-modulated, rising whistles and whistles of one modulation that lasted shorter (less than 0.5 sec) in contrast to the whistles of pesut that lasted up to 3.1 sec and had up to 5 modulations. Frequency ranges (0.2 – 5.2 kHz) and dominant frequencies of the bouto, Inia geoffrensis (1.8 – 3.8 kHz) had much lower, maximum frequencies than those of the pesut (Wang Ding et al,. 1995). It has been suggested that the low frequency whistles of the bouto have better refractive capabilities, important to species whose habitats are rivers, which have higher noise levels than pelagic environment and carry more suspended material (Evans & Awbrey, 1988). This may explain the lower frequencies of Irrawaddy dolphins in the Mahakam in comparison to Balikpapan Bay, but still does not explain why frequencies of Irrawaddy dolphins in the Mahakam and Mekong, the tucuxi, Indus dolphin and baji are still much higher than those of Inia since they share similar environments. Dolphins in downstream area 1 vocalized less often, produced less and simpler whistles with significant different measurements than in upstream area 2 and had only one type of call. This may suggest that these sub-populations have their own group dialect reflecting their own social structure. In core area 2 daily 2 to 6 different groups were encountered in one and the same confluence area and interactions of mean duration of one hour may occur with two other groups during day time (Kreb, 2004). In area 1, each group only encountered one other group on average during day time and has a smaller sub-population. Dolphins of area 1 in that respect more resemble those killer whales, which live in stable, social groups and mostly produce simple, group-specific whistles (Ford, 1991; Strager, 1995). The dynamic social situation in area 2 then more reflects a “fission-fusion” system although to a lesser extent than those of coastal bottlenose dolphins (e.g. Wells, 1991). In contrast, the group of 6 dolphins trapped in the rapid stream area of 2 km in length, where they have been since 3 years at the time of study have a less varied vocal repertoire, vocalize less frequently, and did not whistle, which may imply that they do not need to establish contact and identify themselves. Similar low vocalization rates and low variety of

178

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

repertoire were found for a group of freshwater Irrawaddy dolphins in an isolated area in the Mekong River that during the 8 days of study had no exchange with other groups and which also spent a lot of their time listening. However, they still produced whistles that resembled one another, which might appear to indicate relatedness (Borsani, 1999). So, a low number of whistles produced per unit time seem to suggest a low, social diversification, which is also found in studies of other cetacean species (Tyack, 1991). This also seems to correspond with the lower mean number of whistles per unit time in Balikpapan Bay, where interactions with other groups occurred less frequently and were mostly functional in terms of feeding together and less frequently socializing. However, in the Mahakam, interactions were of different kinds of nature, i.e., intensive socializing, agonistic or avoidance interaction, low level interactions, feeding and travelling together (Kreb, 2004). However, dolphins in Balikpapan Bay may also spent less time vocalizing because they spent most of their more time listening, presumably to the sounds of grouper fishes and shrimps, which could be heard for a great deal of the time. This is similar to a pod of Irrawaddy dolphins in the Mekong River, who spent a great deal of the time passively listening to prey fishes as suggested by Borsani (1999). Also, the fact that clarity was much higher in the bay (mean = 170 cm; SD = 58 cm; n = 24 cm) than in the river (mean = 23 cm; SD = 7 cm; n = 27 cm), may explain the differences, if coastal dolphins make more use of their sighting- than vocal capabilities. Calls produced in the Mahakam were different for each of the 3 areas: Core areas 1 and 2 share calls of type 1, whereas areas 2 and 3 share calls of type 2. Calls of each type in different areas have their own characteristics in terms of number of modulations. This seems to correspond with the dialects of repetitious calls, which have been demonstrated to be pod-specific for killer whales (Ford, 1991). When comparing whistle types of coastal Irrawaddy dolphins in Balikpapan Bay with whistle types described for Irrawaddy dolphins in coastal waters of Queensland, Australia, the first appeared to have more whistle types, i.e., continuous frequency whistles and whistles with 1, 2 or 4 modulations with raising or decreasing initial frequencies. Only two whistle types were described of the Australian Irrawaddy dolphins, of which whistle type 1 had one modulation and initial raising frequency and whistle type 2 had zero modulations and initial decreasing frequency (Van Parijs et al., 2000). Also, the minimum and maximum frequencies of Australian whistles of types 1 and 2 were all lower than similar types of whistles of Balikpapan and from the Mahakam (same modulations and initial raising or falling pattern). Mean minimum and maximum frequencies in Australia of whistle type 1 are 3.2 kHz and 4.2 kHz, whereas in Balikpapan these are 9.9 kHz and 10.8 kHz and in the Mahakam 6.1 kHz and 8.0 kHz. The duration of whistles, however, is more or less similar in all areas. Whistles of type 2 in Australia had mean minimum and maximum frequencies of 3.1 kHz and 4.2 kHz, whereas in Balikpapan these were 6.9 kHz and 7.1 kHz. Nevertheless, whistles of this type in the Mahakam had similar mean frequencies, i.e., 3.1 kHz and 4.0 kHz. Also, whistle duration was nearly similar, i.e., 0.26 sec and 0.3

179 Chapter 10

sec in Balikpapan and Australia, respectively and was shorter on average in Balikpapan Bay, i.e., 0.14 sec. In Balikpapan Bay, the mean, minimum frequencies of 7.0 kHz resembled more those in the Mahakam and the mean, maximum frequencies of 7.8 kHz were similar to both Balikpapan and Mekong River. Unfortunately, there are no whistles of this type available from Australia for comparison. Type 2 whistles in the Mahakam have closer resembling frequencies with those in Australia than with those from Balikpapan Bay and in the Mahakam River also significant differences exists in whistle frequencies within different sub-populations. However, the whistle types in terms of numbers of modulations and raising and falling patterns of the Mahakam population are much more varied in contrast to those in Australia and in this aspect more similar to the Balikpapan population. When comparing the frequency ranges and dominant frequencies of coastal whistles in Balikapan Bay (1.6 – 16.0 kHz & 8.7 – 9.4 kHz) with those of their most closely related relative the killer whale, Orcinus orca, as suggested by Le Duc et al. (1999), we find that the latter are very similar (1.5 -18 kHz & 6 – 12 kHz) (Steiner et al. 1979; Ford & Fisher, 1983; Morton et al., 1986). Whistles in this study may be categorized in several whistle “types” according their modulations, and their initial rising or decreasing frequencies, such as described for bottlenose dolphins in McCowan & Reiss (1995). However, a greater number of distinctively shaped whistles than number of whistle types occur in the Mahakam and Balikpapan Bay, which may be repeated during similar or different sighting occasions and may correspond with the individual signature whistles as first reported by Caldwell & Caldwell (1965; Caldwell et al., 1990; Sayigh et al., 1999) for captive bottlenose dolphins, Tursiops truncatus. However, since the number of different shaped whistles during some recording sessions exceeded the group size with more than twice the number, whistles are not only used as signature whistles, but likely also to establish or maintain contact (Tyack, 1987; Sayigh et al., 1990; Smolker et al., 1993; Janik & Slater, 1998). A cautious, premature conclusion of all comparisons between and within genera may be that whistles types (in terms of their spectral shapes) and whistle frequencies may have a high plasticity. Although whistles are to some extent determined by ecological factors (similar environments), they may even vary within one population (group dialects). Possibly, social structures have a more determining impact on whistle types and frequencies. This coincides with the findings of Ding et al. (1995), who compared whistles of several bottlenose populations and found that although there may be differences between whistles from different or the same individuals within the same population, there are still some characteristics that are unique for each population. Sounds other than whistles and broadband clicks were more similar between Balikpapan Bay and in the Mahakam in comparison with those recorded in Australia, of which only the squeak corresponded with similar sounds of the Mahakam. This may indicate that the specific vocal repertoire (specific sounds types) is determined more by genetic relatedness than habitat factors. The vocal repertoires in terms of

180

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

numbers of different sound types in Balikpapan Bay and in Australia are both less varied in comparison to the freshwater dolphins in the Mahakam and may indicate that vocal variety may be shaped by ecological conditions. Finally, we may conclude that although acoustic behaviour does not answer the question of whether or not the freshwater Irrawaddy dolphin populations represent different (sub)species or merely geographical forms, acoustic behaviour may help to define stocks as separate management units. An explanation for the fact that whistles were significantly higher and lasted longer in approach of boats may be that dolphins try to overcome the vessel noise to give a clear, sign to other dolphins of each other’s position. Pacific humpback dolphins, Sousa chinensis, in Australia significantly increased their whistle rate in response to passage of a boats suggesting that the noise from transiting vessels affects dolphins’ group cohesion (Van Parijs & Corkeron, 2001). These humpback dolphins also whistled more when there were more calves present, and these whistles were suggested to function as contact calls in contrast to the Mahakam population where no significant positive relation existed. Moreover, a tendency existed that dolphins whistled less frequently when calves were present in the group. A most likely explanation is that mother and calves already maintain close contact as they were always found to swim close together and do not need to whistle often.

ACKNOWLEDGEMENTS

We would like to thank the Indonesian Institute of Sciences (LIPI), Mulawarman University of Samarinda (UNMUL), East Kalimantan nature conservation authorities (BKSDA), Zoological Museum Amsterdam (ZMA), Plantage Library, for providing permits, for their cooperation and support. We would like to thank field assistants Arman, Budiono, Syahrani, Ahank, and Munadianto (UNMUL) and our boatsmen. We would like to thank Van Tienhoven Foundation for their financial support for this acoustic behaviour study. In addition, we would like to thank the following persons in particular: A. Ariffien Bratawinata, F. R. Schram, P. J. H. van Bree, H. P. Nooteboom, V. Nijman, M. Lammertink, T. A. Jefferson, G. Parra.

REFERENCES

Andersen, S. & Pilleri, G. 1970. Audible sound production in captive Platanista gangetica. Invest. Cetacea 2: 260-263. Borsani, J.F. 1999. The Irrawaddy dolphins (Orcaella brevirostris, Gray 1866) of Lao P.D.R.: A visual-acoustic survey. Technical report. 22 pp.

181 Chapter 10

Burt, J. 2003. Syrinx-PC version 2.2 n. A windows program for spectral analysis, editing and plackback of acoustic signals. http://www.syrinxpc.com/jbhome.htm. Caldwell, M. C. & Caldwell, D. K. 1965. Individual whistle contours in bottlenose dolphins (Tursiops truncatus). Nature 207: 434-435. Caldwell, M. C., Caldwell, D. K. & Tyack, P. L. 1990. Review of the signature whistle hypothesis for the Atlantic bottlenose dolphin. In: Leatherwood, S. & Reeves, R.R. (eds.), The Bottlenose dolphin. Academic Press, San Diego. pp. 653-660. Da Silva, V.M. & Best, R.C. 1994. Tucuxi- Sotalia fluviatilis (Gervais, 1853). In: Ridgeway, S.H. & Harrisson, R. (eds.), Handbook of Marine Mammals. Volume 5: The First Book of Dolphins. Academic, New York. Ding, W., Würsig, B & Evans, W., 1995. Whistles of bottlenose dolphins: Comparisons among populations. Aquatic Mammals 21: 65-77. Evans, W. E. & Awbrey, F., 1988. High frequency pulses produced by free-ranging Commerson’s dolphin (Cephalorhynchus commersonii) compared to those of phocoenids. IWC Special Issue 9: 173-181. Ford, J.K.B. 1991. Vocal traditions among killer whales (Orcinus orca) in coastal waters of British Colombia. Can. J Zool. 69: 1454-1483. Ford, J.K.B. & Fisher, H.D. 1983. Group specific dialects of killer whales (Orcinus orca) in Britosh Colombia. In: Payne, R. (ed.) Communication and Behaviour of whales. Pp. 129-161. Westview Press, Boulder, Colorado. Fowler, J. & Cohen, L. 1990. Practical statistics for field biology. Open University Press, Philadelphia. Hilton-Taylor, C., 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, U.K. Janik, V.M. & Slater, P.J.B. 1998. Context-specific use suggest that bottlenose dolphin calls are cohesion calls. Animal Behaviour 56: 829-838. Jing, X., Xiao, Y. & Jing, R., 1981. Acoustic signals and acoustic behaviour of Chinese river dolphin (Lipotes vexilifer). Sci. Sin. 2: 233-239. Kamminga, C., Wiersma, H. & Dudok van Heel, W.H., 1983. Investigations on cetacean sonar VI. Sonar sounds in Orcaella brevirostris of the Mahakam river, East Kalimantan, Indonesia; first descriptions of acoustic behaviour. Aquatic Mammals 10 (3): 83-94. Kreb, D. 1999. Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Z. Saugetierkunde 64: 54-58. Kreb, D. 2002. Density and abundance of the Irrawaddy Dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. Raffles Bull. of Zool., Suppl. 10: 85-95. Kreb, D. (in press a) Abundance of freshwater Irrawaddy dolphins in the Mahakam in East Kalimantan, Indonesia, based on mark-recapture analysis of photo- identified individuals. Journal of Cetacean Research and Management.

182

Acoustic behaviour of coastal and freshwater Irrawaddy dolphins

Kreb, D. & Rahadi, K.D. (in press b) Living under an aquatic freeway: effects of boats on Irrawaddy dolphins (Orcaella brevirostris) in a coastal and riverine environment in Indonesia. Aquatic Mammals. Kreb, D. 2004. Conservation and social ecology of freshwater and coastal Irrawaddy dolphins, Orcaella brevirostris in Indonesia. Ph.D. dissertation, University of Amsterdam, Amsterdam. LeDuc, R. G., Perrin, W. F., & Dizon, A. E. 1999. Phylogenetic relationships among the delphinid cetaceans based on full cytochrome b sequences. Marine Mammal Science 15: 619-648. MacCowen, B. & Reiss, D. 1995. Quantitative comparison of whistle repertoires from captive adult bottlenose dolphins (Delphinidae, Tursiops truncatus): A re-evaluation of the signature whistle hypothesis. Ethology 100: 194-209. MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. 1997. The ecology of Kalimantan. The ecology of Indonesia series 3. Oxford University Press. 152 pp. Morton, A.B., Gale, J.C. & Prince, R.C. 1986. Sound and behavioral correlations in captive Orcinus orca. In: Kirkevold, B.C. & Lockard (eds), Behavioral biology of killer whales. Pp. 303-333. Alan R. Liss, New York. Pilleri, G., Kraus, C. & Gihr, M. 1971. Physical analysis of the sounds emitted by Platanista indi. Invest. Cetacea 3: 22-30 + plates. Reeves, R.R., Smith, B.D. & T. Kasuya, (eds). 2000. Biology and conservation of freshwater cetaceans in Asia. Occasional Paper of the IUCN Species Survival Commission 23. IUCN, Gland, Switzerland. Sayigh, L. S., Tyack, P. L., Wells, R. S. & Scott, M. D. 1990. Signature whistles of free- ranging bottlenose dolphins, Tursiops truncatus: mother-offspring comparison. Behav. Ecol. Sociobiol. 26: 247-260. Sayigh, L. S., Tyack, P. L., Wells, R. S., Solow, A. R., Scott, M. D. & Irivine, A. B. 1999. Individual recognition in wild bottlenose dolphins: a field test using playback experiments. Animal Behaviour 57: 41-50. Smolker, R., Mann, J. & Smuts, B. 1993. Use of signature whistles during separations and reunions between bottlenose dolphin mothers and infants. Behav. Ecol. Sociobiol. 33: 393-402. Stacey, P.J. & Arnold, P.W. 1999. Orcaella brevirostris. Mammalian Species 616: 1-8. Steiner, W.W. 1981. Species-specific differences in pure tonal whistle vocalizations of five western North Atlantic dolphin species. Behav. Ecol. Sociobiol. 9: 241-246. Strager, H. 1995. Pod-specific call repertoires and compound calls of killer whales, Orcinus orca, in the waters of northern Norway. Can. J. Zool. 73: 1037-1047. Tyack, P.L. 1987. Do untrained dolphins imitate signature whistles to call each other? In: Abstr. 7th Bienn. Conf. Biol. Mar. Mamm., Miami, FL, Dec. 1987. P. 71. Society for Marine Mammology. Tyack, P., 1981. Interactions between singing Hawaiian humpback whales and conspecifics nearby. Behav. Ecol. Sociobiol. 8: 105-116. Van Parijs, S.M., Parra, G.J., & Corkeron, P.J. 2000. Sounds produced by Australian Irrawaddy dolphins, Orcaella brevirostris. J. Acoust. Soc. Am. 108 (4): 1938-1940.

183 Chapter 10

Van Parijs, S. M. & Corkeron, P. J. 2001. Vocalizations and behaviour of Pacific Humpback dolphins Sousa chinensis. Ethology 107: 701-716. Wells, R., 1991. The role of long-term study in understanding the social structure of a bottlenose dolphin community. In: Pryor, K. & Norris, K. S. (eds.), Dolphin Societies. Discoveries and Puzzles. Pp. 199-225. Univ. of California Press, Berkeley, Los Angeles and Oxford. Xiao Youfu & Jing Rongcai. 1989. Underwater acoustic signals of the baiji, Lipotes vexilifer. In: W.F. Perrin, R.L. Brownell, Jr., Zhou Kaiya & Liu Jiankang (eds), Biology and conservation of the river dolphins. Occas. Pap. IUCN Species Surv. Comm. 3. Pp. 129-136. Int. Union. Conserv. Nat., Gland, Switzerland. 173 p.

184

River vagrancy and allopatric speciation

CHAPTER 11

Freshwater distribution of Irrawaddy dolphins based on river “vagrancy” or allopatric “speciation”?

The pesut has adapted to the riverine habitat and has become skilful in catching various prey items from large catfishes to smaller cyprinid fishes and bottom-dwelling river shrimp.

185 Chapter 11

ABSTRACT

Freshwater and coastal Irrawaddy dolphins in East Kalimantan are currently separated from each other, which is obvious from a hiatus in their distribution. This raises the issue of whether separation is a recent process due to degradation of intermediate habitat, or results from a historic process due to allopatric speciation during the last glacial maximum. If we hypothesize that separation is a recent process, it is assumed that prior to the habitat degradation gene flow still existed between coastal and riverine populations and that the riverine population is built up of groups or individual coastal “river vagrants”. To determine the likelihood of each hypothesis the most relevant variables were examined in terms of finding differences between both populations, i.e., dolphin distribution, morphology, social ecology, vision, physiology, acoustics and comparisons with other “facultative” river dolphins. Most evidence supports a separation of coastal and riverine populations over a larger time scale, although one cannot be definite about the time of separation unless there is genetic data. Collection of tissue samples of coastal Irrawaddy dolphins therefore has a high priority. Nevertheless, whatever hypothesis reflects the actual process of separation, it is clear that this riverine population is critically endangered and in need of conservation. If the river vagrant theory prevails, conservation effort should focus on restoring degraded intermediate habitat; if the population appears to be obligate riverine the entire riverine habitat is a priority. It is recommended to continue collecting genetic tissue sample to assess the genetic variation of the riverine population and to determine whether it is necessary to translocate isolated groups in the river to join the main breeding population.

RINGKASAN

Lumba-lumba air tawar dan laut di Kalimantan Timur terpisahkan satu dan lainnya, dimana ternyata terdapat kekosongan dalam penyebarannya. Hal ini menjadi pertanyaan apakah proses pemisahan ini terjadi sebagai proses baru karena adanya penurunan kualitas habitat antara, atau sebagai hasil proses sejarah karena untuk menyelidiki proses perpisahan seiring penurunan habitat lanjutan atau penyebab dari proses sejarah pemisahan selama jaman es terakhir. Jika kita membuat hipotesa bahwa pemisahan adalah proses yang baru terjadi, ini diasumsikan karena penurunan kualitas habitat, kesamaan gen tetap ada antara populasi lumba-lumba laut dan sungai dan bahwa populasi sungai terbentuk dari individu lumba-lumba laut yang masuk ke areal sungai (penjelajah sungai). Untuk menentukan kesamaan dari tiap hipotesis, variabel yang dapat dipercaya diuji dengan tujuan untuk menemukan perbedaan antara kedua populasi, antara lain penyebaran lumba-lumba, morfologi, ekologi sosial, daya pandang, fisiologi, akustik dan perbandingan dengan lumba-lumba sungai fakultatif lainnya. Banyak bukti lainnya mendukung pemisahan populasi laut dan sungai terjadi dalam jangka waktu lama, meski tidak ada satupun yang memastikan tentang waktu

186 River vagrancy and allopatric speciation pemisahan kecuali ada bukti genetik. Oleh karena itu pengumpulan data contoh jaringan lumba-lumba lumba-lumba Irrawaddy laut memiliki prioritas tinggi. Meskipun begitu, apapun hipotesa menggambarkan proses pemisahan yang sesungguhnya, hal ini jelas bahwa populasi sungai ini dalam situasi kritis yang membahayakan dan membutuhkan perlindungan. Jika teori “penjelajah sungai” adalah yang benar, upaya konservasi harus fokus pada tingkat perbaikan habitat antara, dan bila populasi ternyata adalah memang berasal dari sungai, maka seluruh habitat sungai adalah prioritas. Hal ini menganjurkan untuk terus mengumpulkan data genetis untuk memperkirakan variasi genetik dari populasi sungai dan untuk menentukan apakah perlu untuk memindahkan kelompok terasing di sungai untuk bergabung dalam perkembangbiakan populasi utama.

DISCUSSION

The taxonomic separation of a freshwater and coastal form within the Irrawaddy Dolphin has been debated since Gray first described the species Orcaella brevirostris in 1866 based on a specimen described by Owen in Gray, 1866, as Phocaena brevirostris. This description was based on a specimen from the east coast of India in the harbour of Vishakhapatnam. Anderson described a freshwater form of the species in 1871 based on a specimen collected in 1868 at 1500 km from the mouth in the fomerly named Irrawaddy River in Burma (presently named Ayeyarwady River, Myanmar) and assigned it a different species name Orcaella fluminalis. Later, in 1878 he reduced the status of the river form to a subspecies, O. b. fluminalis, but Loze (1973), and Pilleri and Gihr (1974) found no consistent differences between both freshwater and marine specimens stating that both forms belong to the same species Orcaella brevirostris. Until now, this is the general concensus (Rice, 1998). Analysis of the skull morphology of Irrawaddy dolphins throughout much of their range indicated specific or sub-specific differences in the height of the temporal fossa, number and width of nasal bones/ ossicles, development of mesethmoid plate, and pterygoid hamuli between animals from Australia and South Asia (Beasley et al., 2002). However, data was insufficient to investigate the question freshwater and coastal separation. Nevertheless, in this study we found that at least at present the freshwater Irrawaddy dolphin population in the Mahakam River is seperated from the coastal population based on a hiatus of 160 km in their distribution between the mouth and further upstream the river due to degradation of habitat (see chapter 6, results, abundance & distribution), and based on ecological differences related to salinity. Coastal Irrawaddy dolphins were always associated with brackish water and they only entered the delta at high tide. The most in-shore observation was made at 10 km upstream the delta but they may move until 20 km upstream the mouth according to interviews with fishermen. The freshwater population, on the other hand was never sighted lower than 180 km upstream from the mouth, although some local residents claimed that they sighted dolphins very occasionally 90 km from the mouth. In the early 1980s the dolphins were still

187 Chapter 11

commonly sighted by residents in Samarinda at 60 km from the mouth, after which they became increasingly rare, and they vanished completely from the area in the mid 1990s. In this study, I investigate whether the separation of coastal and freshwater Irrawaddy dolphins in East Kalimantan has occurred recently due to degradation of intermediate habitat, or has an evolutionary origin related to the changing sea levels during the Pleistocene resulting in an allopatric speciation process. The possible short- term and long-term separation processes and the origin of the freshwater population may be explained by the following two hypotheses: Hypothesis 1- “river vagrancy”: The freshwater population has its origin in the accumulation of individuals or small groups of occasional, coastal “river vagrants” (see chapter 1), which have adapted to the frehswater ecosytem. The process of river vagrancy may have a historic origin but is an ongoing process and would still be continued if the intermediate habitat between the coastal and freshwater populations had not been degraded, because the behaviour is supposed to be inherent to the dolphins exploring and adaptive biology. Conservation and upgrading intermediate habitat may then be very essential in terms of maintaining genetic exchange. Hypothesis 2- “allopatric speciation”: The freshwater population has its origin in the last glacial maximum during the Pleistocene when sea levels were low and Sundaland, the landmass of South East Asia, was above sea level delineated by the 200- m isobath (Fig. 1) (Tomascik et al., 1997). During this period, land was continuous from peninsular Southeast Asia across to Sumatra, Java and Borneo, and the western shallow part of the South China Sea as far east as the Natuna Islands (Hutchison, 1989). Shelf seas, e.g., the Java Sea, had disappeared and freshwater aquatic species may have dispersed throughout the major ancient rivers that connected different islands (Fig. 1) (Haile, 1975; Verstappen, 1975). Coastal Irrawaddy dolphins of South and West Borneo, North Java, East Sumatra, and South East Asian continental mainland may also have dispersed in this aquatic habitat intermediate of present islands, which became sea again c. 10,000 years ago when water levels started to rise (MacKinnon et al., 1996) and coastal dolphins re-colonized the shallow coastal areas. The three major river systems, Ayeyarwady, Mekong, and Mahakam form an exception. These rivers claimed the shallow waters off their mouths during the last glacial period and directly opened into the deep waters (> 200 m). Coastal Irrawaddy dolphin populations, which were adapted to the brackish, shallow delta areas are more likely to have adapted to the frehswater habitat than they would to the deep, offshore waters, which would have required a greater adaptation of their biology. Therefore, I hypothesize that the current riverine populations in the three major rivers in South East Asia have been seperated from their coastal conspecifics at least since the Late Pleistocene during Sunda Shelf formation. Current coastal populations along the coast of East Kalimantan most likely represent migrants after the glacial period from South Borneo. Conservation of freshwater populations have a high priority because they represent unique, isolated populations.

188 River vagrancy and allopatric speciation

Figure 1. The Sunda Shelf showing present coastlines (unshaded), the area of Sundaland exposed at times of lowest sea level (dark shade) delineated by the 200-m isobath during the last glacial about 12,000 years ago, and past and present river systems (After Tjia, 1980 and MacKinnon, 1996).

Step by step we will look for support for each of both hypotheses based on Irrawaddy dolphin distribution, morphology, social ecology, vision, physiology, acoustics and comparisons with other “facultative” river dolphins. Although genetic material is available of the Mahakam population, unfortunately no genetic material of coastal Irrawaddy dolphins in East Kalimantan is available. Looking at the Irrawaddy dolphin distribution, we find that they are distributed along shallow coasts, bays, brackish water lakes (lagoons), and in river mouths throughout much of the tropical to sub-tropical Indo-Pacific (Stacey and Leatherwood, 1997). In addition, they occur in three major river systems, i.e., the Ayeyarwady (Myanmar), the Mekong (Vietnam,

189 Chapter 11

Cambodia and Laos) and the Mahakam (Kalimantan, Indonesia). If river vagrancy would be inherent to the dolphins biology, we would expect to find Irrawaddy dolphins in more river systems. And indeed, the dolphins were also observed to move further upstream the mouth of some rivers, i.e., the Rajang River in Sarawak, where they were regularly seen in Tg Mani, 33 km from sea (Mörzer Bruyns, 1966). Several sightings were made by crew of a coal barge and one carcass was retrieved between 72 km and 85 km the mouth in the Brisbane River, and a bit upstream of the junction with the Bremen River at 77 km from the mouth (Paterson et al., 1998). However, all these sightings were regarded as unusual by the crew because the dolphins were rarely observed further upstream than 10 km from the mouth of the Brisbane River. The dolphins observed in the Rajang, Brisbane, and Bremer Rivers are nevertheless not typical freshwater vagrants since their occurrence was still within the limit of tidal influence. Similarly, if coastal Irrawaddy dolphins were to explore the riverine environment of the Mahakam, one would expect that they would do so at high tide, but the dolphins presence in the lower areas is associated with high or medium water levels according to semi-structured interviews with local residents, which makes it more likely that these belonged to the freshwater stock. Lloze (1973) suggested that the Irrawaddy population in the Mekong river is not an isolated population. He mentioned that the dolphins left the lakes (e.g. the Tonle Sap) at low water levels together with some large fishes (ranging from 1 to 3 metres) and joined the Mekong River. Further he suggested that after the dolphins arrived there, some individuals would swim upstream and others move downstream and progressively gain the delta regions of south Vietnam and the South China Sea. The problem is that the dolphins have only been followed downstream until the Vienamese border. Therefore it is not known whether they move further downstream and whether any interbreeding exists between these dolphins and the Irrawaddy dolphins in the delta region, whose presence have been recorded by Gruvel (1925). Another sighting, which may suggest the existence of “true” river vagrants (exceeding the tidal limit) was reported by Mörzer Bruyns (1971) and was made 110 km up the Pussur River in Bangladesh. Since there was no date specified or any further information, nor are there any other records in the river, the record is identified as tentative. In South Kalimantan Delsman (1922) reported the occurrence of Irrawaddy dolphins at c. 380 km upstream the Barito River below Puruk Cahu. However, no further details were given about his sighting. According to Westerman (1939) no positive records through observations or interviews exist, which indicates the presence of dolphins in the larger rivers of South Kalimantan including the Barito. Therefore, the sighting below Puruk Cahu most likely represented an occasional river wandering. In this study we found that the river dolphins had a high site fidelity during the 3.5 years study period and that their overal home ranges were limited to a maximum of 180 km river strip (see chapter 9, results, spatial distribution). Since clear evidence for regular river vagrancy is lacking, we find the allopatric speciation hypothesis more plausible to explain the origin of freshwater Irrawaddy dolphin populations. Although occasional river wanderings may occur, this does not seem to be regularly enough to explain the

190 River vagrancy and allopatric speciation existence of relatively large river populations. With regards to their morphology, we observed in the field that the river dolphins seemed to have larger body sizes, displayed a greater sexual dimorphism as expressed in a large neck crest for males, which are absent in females (see chapter 9, results, sexual dimorphism). Lastly, body skin of coastal dolphins had a much rougher texture with many scratches. The first two differences between coastal and river populations favour hypothesis 2, since these variables may have evolved over a longer time-scale. The last characteristic more reflects differences in habitat ecology and at least implies that dolphins remain in their own habitat during their life-time. The social ecology differs in a number of aspects, i.e., groupsize, nature and frequency of inter-group interactions, spatial distribution, breeding period, mating strategies (see chapter 9). As the term social ecology implies, the differences in social structures reflect differences in their ecology and are adaptations to their environments, i.e., the geographical shapes of habitat, the temporal abundance and spatial distribution of food resources. Although ecological adaptations do not necessarily require a historic time-scale, adaptations to fit in another social system probably exceed a life-time and makes river vagrancy therefore a less suitable hypothesis because of the presumed lack of successful adaptations. However, river vagrancy of Irrawaddy dolphins may still occur in other rivers, which are not occupied by a well-adapted riverine population. Other differences, which may have arisen as a result of the separation may be reflected in their visual capacities such as is the case with obligate river dolphin species such as the Indus susu, Platanista gangetica gangetica, Ganges susu, Platanista gangetia minor, the Amazon River dolphin (boto), Inia geoffrensis and the Yangtze dolphin (baiji), Lipotes vexilifer for which an increasing regression of the eye has been demonstrated as an adaptation to a turbid and fluviatile environment (Herald et al., 1969; Purves & Pilleri, 1973; Best & Da Silva, 1989; Zhou et al., 1980). The first two (sub)species are blind, whereas the eyes of the boto and the baiji, although reduced, still appear functional (Leatherwood & Reeves, 1983). Irrawaddy dolphins do not have reduced eyes and like other delphinids appear to have good eyesight, which was also observed in the field during this study when river dolphins were observed to spy-hop for a great deal of the time. However, some reduced use of eyesight has been described for captive freshwater Irrawaddy dolphins from the Mahakam, which would not take a fish in plain view that was thrown in the basin less than 100-150 cm away as if it was unnoticed. Instead they would first scan the near environment by sonar and only then capture their prey (Kamminga et al., 1983). This habit would indicate that they have probably spent their entire life in a turbid environment (mean clarity in the Mahakam is 23 cm) and have developed feeding strategies that rely exclusively on the use of their sonar. However, in order to use this argument to prove that coastal and freshwater Irrawaddy dolphin populations are seperate, similar tests are required with coastal Irrawaddy dolphins. Since clarity in the bay is much higher, i.e., 170 cm, one would expect differences in feeding strategies. In this study, we found that coastal Irrawaddy dolphins did not continuously emit click trains while fishing, which may indicate that they would partially

191 Chapter 11

prey on sight or by listening (see chapter 10). Nevertheless, differences in use of eyesight have not been satisfactorily determined yet to use as an argument for a short-term or long-term coastal/ freshwater separation. The physiology of the freshwater form of the Irrawaddy dolphin does not seem to support a long evolutionary separation event from coastal populations. For four specimens of Orcaella brevirostris in the Mekong River in Cambodia it has been found that each kidney contains about the same number of adrenal bodies as Delphinus delphis, i.e., 380 and 400 respectively. This large number is in contrast with the low number of 80 adrenal bodies found in Platanista gangetica (Lloze, 1973; Harrison & King, 1968). Lloze (1973) therefore suggests that the small ‘lobulation’ as is apparent from the low number of adrenal bodies in Platanista is characteristic for odontocetes that are living in freshwater environments. He also explains the small lobulation in terms of the absence of any necessity for dolphins in freshwater to separate the salt and freshwater from seawater, which marine dolphins gulp down together with fishes. According to Slijper (1962), there is a direct relation between the degree of lobulation and the degree of marine life. The large lobulation for the freshwater Irrawaddy dolphin thus seems to indicate that this should not be an obstacle for changing its river habitat for a marine habitat. For further research it may be useful to compare the number of adrenal bodies, representing the degree of lobulation, of the freshwater forms with that of the marine forms of the Irrawaddy dolphin. If a lower number is found for the freshwater populations, then this may indicate that there is no gene-flow between both populations as a result of a long historical separation process. Based on the current data both hypotheses may hold since lobulation in freshwater populations may be a remnent trait of their marine descent. With regards to the acoustic behaviour of the coastal and freshwater Irrawaddy dolphin populations in this study we found many differences (see chapter 10): pod- specific whistle dialects; vocalizing intensity; size of vocal repertoire. Vocal repertoire was less varied for coastal Irrawaddy dolphin populations in East Kalimantan and Australia compared with the Mahakam River and may be determined by ecological conditions. The vocalizations rates (numbers per time unit) seem to be determined by social structures. Since these two “plastic” variables may involve ecological and social adaptations to each environment or stock, this does not favour hypothesis 1 but instead favours hypothesis 2 since vagrants may not be able to succesfully adapt to other stocks because its dialect may not be recognized. On the other hand, sound types including whistle structures were more similar between the likely more genetically related, coastal and freshwater populations in East Kalimantan than between coastal populations of Australia and East Kalimantan. This does not necessarily imply that there is still geneflow (although it is possible) if sound types are formed during a slow evolutionary process (and is a non-plastic variable) and may only indicate a genetic relatedness although at distance. In addition, I would like to investigate whether the basic sonar system, i.e., echolocation clicks of freshwater dolphins exhibits some degree of adaptation to the riverine habitat and whether this is an irreversible adaptation of having lost their ability

192 River vagrancy and allopatric speciation of using low-frequency-signals. This information would be very important for conservation purposes, as the lost of this capability would restrict these dolphins to their riverine habitat. When looking at dominant frequencies of echolocation signals, however, this present study cannot make any comparisons due to limitation of the recording equipment, which only made effective recordings of sounds up to 24 kHz. It has been generally argued that both pelagic and river odontocetes have basic or one- component sonar signals while the littoral and estuarine species use two-component sonar signals. Fast swimming pelagic species require low frequency sonar signals for long range navigation in the open water, whereas river species require high frequency sonar signals in order to detect close objects. The littoral and estuarine species require the use of both signals as they are faced with both kinds of situations (Dudok van Heel, 1981). Estuarine harbour porpoises, Phocaena phocaena, the beluga, Delphinapterus leucas, and estuarine form of Sotalia fluviatilis guianensis, which were studied in captivity, all displayed the two component signals (Dudok Van Heel, 1981; Kamminga & Wiersma, 1981; Wiersma, 1982). However, in captivity no need rises to use this low frequency signal for long-range navigation. A study on free-ranging cetaceans showed that botos produced echolocation clicks with dominant frequencies around 100 kHz, whereas the sympatric riverine tucuxi produced clicks of 80 to 95 kHz. However, captive studies on the riverine tucuxi revealed that they produced clicks with dominant frequencies of 8- 15, 30 and 95 kHz. Also clicks of high-frequency (95 kHz) and low-frequency (30 kHz) were recorded simultaneously (Norris et al., 1972). These lower frequencies between 8 and 15 kHz were also recorded in an earlier captive study, in which a hydrophone was used, which could only detect frequencies up to 20 kHz (Caldwell & Caldwell, 1970). The results would indicate that lower frequencies can still be used by riverine Sotalia, although it is not clear whether these same dolphins, when released in the wild, would emit a sonar signal of similar dominant frequencies. As the use of the lower frequencies is generally associated with long range-navigation, the recordings of lower frequencies as well is rather surprising for these riverine dolphins which were held in a pool in which the emittance of low frequencies would seem of no use just as in the studies with belugas and harbour porpoises mentioned earlier. Possibly, the emittance of echolocation clicks of low frequency might also serve another function, e.g. communication. Conclusively, coastal and riverine Sotalia can use both high and low frequencies for echolocation, although in the wild only dominant high frequency sounds are used. The high-frequency component of riverine Orcaella brevirostris is similar to that of Sotalia fluviatilis and its vocalising intensity has similarities with Inia geoffrensis (Kamminga et al., 1983). The main sonar signal of Orcaella held in captivity after being caught in the Mahakam River was elementary, consisting of only a few cycles of a dominant frequency of around 60 kHz with small deviations. No low-frequency components could be distinguished (Kamminga et al., 1983). No recordings of echolocation clicks, which used equipment covering the entire bandwidth are available on estuarine or coastal Orcaella. It could be argued that the dolphins from the Mahakam have either lost the ability of using low frequencies for echolocation or that there was no need for them

193 Chapter 11

emitting low frequencies in the small holding pool (in which they contrast Sotalia, if this is the case). An alternative hypothesis might be that all Irrawaddy dolphins, coastal as well as riverine, are only able to use high frequency sonar signals, which would seem unlikely though when comparing with other estuarine species. Generally, it can be argued that the river habitat does not call for a complicated sonar but rather for basic one-component high frequency signals suitable for short-range navigation and functioning in a turbid environment (Kamminga et al. 1983). Although acoustics do not solve the question if the freshwater Irrawaddy dolphin populations represent different (sub) species or merely geographical forms, acoustics may help to define stock s as separate management units. We have evidence that the riverine population seems very well adapted to its environment as it is able to use high frequency signals. Also, the distinction in vocalization intensity, vocal repertoire, and whistle dialects makes a strong case for a historic separation of coastal and freshwater dolphin populations (hypothesis 2) as it would seem likely that river vagrancy would diminish these differences in vocalizations. When examining the origin of two other “facultative riverine” species, i.e., the tucuxi, Sotalia fluviatilis, and finless porpoise, Neophocaena phocaenoides, it appears that coastal and river populations are well distinguished based on body size, number of teeth in the upper row, and skull characters in the tucuxi (Borobia, 1989; Hendriks, 1984; Monteiro-Filho et al., 2002), and skull morphometric, meristic data, and mitochondrial DNA polymorphism in the finless porpoise (Pilleri & Gihr, 1972; Gao, 1991; Gao & Zhou, 1995a, b, c; Jefferson, 2002). Altough their genera are considered monospefic, subspecies or seperate freshwater populations are recognized (Rice, 1998; Borobia & Sergeant, 1989). Another evidence of support for a separate freshwater population of the facultative riverine tucuxi may be the fact that the tucuxi (regarded as relatively recent immigrant of the freshwater system) and obligate riverine boto share rather similar clicks with regard to dominant frequencies and time duration, which are attributed to the adaptation of their riverine environment (Kamminga et al., 1993). Since there appears to be no gene flow between coastal and river populations, which would have reduced the distinct differences, river populations of both species more seemed to have evolved because of allopatric speciation rather than through river vagrancy or sympatric speciation, which increases the likelihood that a similar case prevails (hypothesis 2) within the facultative Irrawaddy dolphin. After examining all variables, only one variable related to visual capacities possibly may be in favour of hypothesis 1. However, since eyesight of the boto and the baiji are still functional, as well as that of two other facultative river dolphins, the tucuxi and finless porpoise, which are considered to have distinct river populations (see below), perhaps this variable is not such a good character upon which to make any distinctions between (sub)species or geographically variable populations. Also, the Mahakam is likely to have greatly increased in turbidity in only a few decades due to increased erosion and sedimentation. Most evidence is in support of a separation of coastal and freshwater populations over a longer time scale. We recommend further sample collections for DNA analysis to reconstruct the time of separation. Conclusively, the

194 River vagrancy and allopatric speciation term facultative as proposed in (Leatherwood & Reeves, 1994) may be a bit misleading as it implies that these dolphins may have a choice to move between coastal and freshwater habitat, which they more than likely do not have. Although genetic analysis may be able to shed more light on which historic or actual processes are involved in the existence of riverine populations of Irrawaddy dolphins or species with a similar ecology, it is clear that many species or populations depend on the river habitat and are in need of protection whether it is because they are restricted to this habitat for their entire or part of their life, or because of their occurrence in estuaries. Each hypothesis to explain the coastal and freshwater dolphin separation in occurrence, implies a need for conservation of riverine habitat in order to maintain viable populations or subspecies. Recommendations for conservation are in chapter 6. In addition, we recommend further sample collections for DNA analysis to determine the genetic variation within the freshwater population. If the genetic variation is very low, a translocation of an isolated group, “trapped” between two rapid streams may be considered an option to increase the genetic variation and breeding population.

REFERENCES

Anderson, J. 1871. Description of a new cetacean from the Irrawaddy River, Burma Orcaella fluminalis Anderson. Proceedings of the Zoological Society of London 39: 142-144. Beasley, I., Arnold, P. & Heinsohn, G. 2002. Geographical variation in skull morphology of the Irrawaddy dolphin, Orcaella brevirostris (Owen in Gray, 1866). The Raffles Bulletin of Zoology, Supplement 10: 15-34. Best, R.C. & da Silva, V.M.F. 1989. Amazon River Dolphin, Boto, Inia geoffrensis (de Blainville 1817). In: Handbook of Marine Mammals; Volume 4, River Dolphins and the larger toothed whales. Edited by S.M. Ridgeway and R. Harrison, Academic Press, London, New York: 1-23. Borobia, M. & Sergeant, D. 1989. Variation in skull morphology of South American dolphins of the genus Sotalia. Paper presented at the Fifth Int. Ther. Cong., Rome 2: 4 Caldwell, D.K. & Caldwell, M.C. 1970. Echolocation-type signals by two dolphins, genus Sotalia. Quarterly Journal of the Florida Academy of Sciences 33 (2): 124-131. Dudok van Heel, W.H. 1981. Investigations on cetacean sonar III. A proposal for an ecological classification of odontocetes in relation with sonar. Aquatic Mammals 8 (2): 65-68. Gao, A. 1991. Morphological differences and genetic variations among the populations of Neophocaena phocaenoides. Ph.D. thesis, Nanjing Normal University, PRC (in Chinese with English summary). 116 pp. Gao, A. & Zhou, K. 1995a. Geographical variations of external measurements and three subspecies of Neophocaena phocaenoides in Chinese waters. Acta Theriol. Sinica 15: 81-92 (in Chinese with English summary). Gao, A. & Zhou, K. 1995b. Geographical variations of skull among the populations of Neophocaena in Chinese waters. Acta Theriol. Sinica 15: 161-169 (in Chinese with English summary).

195 Chapter 11

Gao, A. & Zhou, K. 1995b. Geographical variations of postcranial skeleton among the populations of Neophocaena in Chinese waters. Acta Theriol. Sinica 15: 246-253 (in Chinese with English summary). Gray, J. E. 1866. Catalogue of the seals and whales in the British Museum, 2nd edition. British Museum, London. 402 pp. Gray, J.E. 1871. Supplement to the catalogue of seals and whales in the British Museum. British Museum, London. 103 pp. Gruvel, A. 1925. L’Indochine– ses richesses marines et fluviales. Soc. Ed. Geo. Marit. Col. Paris. Haile, N.S. 1975. Postulated late Cainozoic high sea level in the Malay peninsula. J. Malay. Brit. Roy. Asiatic. Soc. 48: 78-88. Harrison, R.J. & King, J. E. (1968) Marine mammals Hutchinson Librairie. London: 192 p. Hendriks, B.R. 1984. Clustering en classificatie van incomplete data. (Onderzoek aan schedelgegevens van Sotalia fluviatilis). Technische Hogeschool Delft. Unpublished report. Herald, E.S., Brownell, Jr. R.L., Frye, F.L., Morris, E.J., Evans, W.E. & Scott, A.B. 1969. Blind river dolphins: First side-swimming cetaceans. Science, N. Y. 166: 1408- 1410. Hutchison, C.S. 1989. Geological evolution of South-east Asia. Oxford Monographs on Geology and Geophysics 13. Clarendon Press, Oxford, 368 pp. Jefferson, T.A. 2002. Preliminary analysis of geographic variation in cranial morphometrics of the finless porpoise (Neophocaena phocaenoides). The Raffles Bulletin of Zoology, Supplement 10: 3-14. Kamminga, C. & Wiersma, H. 1981. Investigations on cetacean sonar II. Acoustical similarities and differences in odontocete sonar signals. Aquatic Mammals 8: 41-62. Kamminga, C., Wiersma, H. and Dudok van Heel, W.H., 1983. Investigations on cetacean sonar VI. Sonar sounds in Orcaella brevirostris of the Mahakam river, East Kalimantan, Indonesia; first descriptions of acoustic behaviour. Aquatic Mammals 10 (3): 83-94. Kamminga, C., Hove van, M.T., Engelsma, F.J. and Terry, R.P. (1993) Investigations on cetacean sonar X: A comparative analysis of underwater echolocation clicks of Inia spp. and Sotalia spp. Aquatic Mammals 19 (1): 31-43. Leatherwood, S. & Reeves, R.R. (eds.). 1983. The Sierra Club handbook of whales and dolphins. 302 pp. Leatherwood, S. & Reeves, R.R. 1994. River dolphins: a review of activities and plans of the Cetacean Specialist Group. Aquatic Mammals 20: 137-154. Lloze R. 1973. Contributions a l’étude anatomique, histologique et biologique de l’Orcaella brevirostris (Gray -1866) (Cetacea-Delphinidae) du Mekong. Dissertation thesis, Toulouse France [In French]. MacKinnon K, Hatta G, Halim H & Mangalik A (eds.). 1996. The ecology of Kalimantan. The ecology of Indonesia series 3. Periplus Editions (HK) Ltd., Singapore. 802 pp.

196 River vagrancy and allopatric speciation

Monteiro-Filho, E.L.A., Monteiro, L.R. & dos Reis, S.F. 2002. Skull shape and size divergence in dolphins of the genus Sotalia: a tridimensional morphometric analysis. Journal of Mammology 83: 125-134. Mörzer Bruyns, W.F.J. 1966. Some notes on the Irrawaddy Dolphin, Orcaella brevirostris (Owen, 1866). Zeitschrift für Säugetierkunde 31: 367-370. Westerman, J.H. 1938. Natuur in Zuid en Oost Borneo. In: 3 jaren Indisch Natuurleven. Opstellen over landschappen, dieren en planten, 11de jaarverslag (1936-1938). Pp. 334-397. Batavia. Nederlandsch-Indische Vereniging tot natuurbescherming, Batavia. Norris, K.S., Harvey, G.W., Burzell, L.A. & Krishna Kartha, T.D. 1972. Sound production in the freshwater porpoise Sotalia fluviatilis Gervais and Deville and Inia geoffrensis Blainville, in the Rio Negro, Brazil. Invest. Cetacea 4: 251-260. Paterson, R.A., Van Dyck, S.M. & Gynther, I.C. 1998. Irrawaddy dolphins Orcaella brevirostris (Owen in Gray) from Southern Queensland. Memoirs of the Queensland Museum 42 (2): 554. Pilleri, G. & Gihr, M. 1972. Contributions to the knowledge of the cetaceans of Pakistan with particular reference to the genera Neomeris, Sousa, Delphinus and Tursiops and a description of a new Chinese porpoise (Neomeris asiaeorientalis). Investigations on Cetacea 4: 107-162. Pilleri, G. & Gihr, M. 1974. Contributions to the knowledge of the cetaceans of southwest and monsoon Asia (Persian Gulf, Indus Delta, Malabar, Andaman Sea and Gulf of Siam). Investigations on Cetacea 5: 95-153. Purves, P.E. & Pilleri, G. 1973. Observations on the ear, nose, throat, and eye of Platanista indi. Investigations on Cetacea 5: 13-59. Rice, D.W. 1998. Marine mammals of the world: systematics and distribution. Special publication No. 4 of the Society for Marine Mammology. Allen Press, Lawrence, Kansas. 231 pp. Slijper, E.J. 1962. Whales (A general review). London. Stacey, P.J. & Leatherwood, S. 1997. The Irrawaddy dolphin, Orcaella brevirostris : A summary of current knowledge and recommendations for conservation action. Asian Marine Biology 14 : 195-214. Tomascik, T., Mah, A. J., Nontji, A. & Moosa, M.K. (eds.) 1997. The ecology of the Indonesian Seas, Part 1. Periplus Editions (HK) Ltd., Singapore. 642 pp. Tjia, H.D. 1980. The Sunda Shelf, Southeast Asia. Z. Geomorph. 24: 405-427. Verstappen, H.T. 1975. On palaeo-climates an landform development in Malasia. Mordern Quarternary in Southeast Asia (Barstra, G.J. & Caspari, W.A., eds.). Pp. 3-35. Balkema. Wiersma, H. 1982. Investigations on cetacean sonar, IV. A comparison of wave shapes of odontocete sonar signals. Aquatic Mammals 9: 57-67. Zhou, K., Pilleri, G. & Li, Y. 1980. Observations on Baiji (Lipotes vexilifer) and Finless Porpoise (Neophocoena asiaorientalis) in the lower reaches of the Chang Jiang- with remarks on physiological adaptations of Baiji to the environments. Scientia Sinica 23 (6).

197 Chapter 11

198 Predicting long-term survival of riverine Irrawaddy dolphins

CHAPTER 12

Predicting long-term survival of riverine Irrawaddy dolphins (Orcaella brevirostris) in East Kalimantan using population viability analysis

Saving two to three individuals of a mean number of 5 dead dolphins per year may prevent the extinction of the pesut in the Mahakam River with a maximum probability of 76% and 97%, respectively.

199 Chapter 12

ABSTRACT

Population viability analysis was applied to a small, isolated population of Irrawaddy dolphins, Orcaella brevirostris, in the Mahakam River in East Kalimantan, Indonesia. Simulations of population survival were performed using the individual-based VORTEX program for initial population sizes of N = 55 and N = 76. Besides, the impact of inbreeding, harvesting, catastrophes, varying carrying capacities, reductions in mortality and supplementation involving translocation of an isolated sub- population to the main population were modelled. Without any conservation action, the population only has a 1% to 4% chance of survival to the next century. The key to survival lies in mortality reductions: preventing the deaths of 2 individuals of a total of 5 dead dolphins per year may help to set this population back on the road to recovery and prevent the dolphins from extinction with a 50% to 75% probability, whereas saving 3 individuals yearly causes a survival probability of near 100%. Since 80% of deaths occur through gillnet-entanglement, conservation efforts should primarily focus on finding ways to prevent death through entanglement. Gillnet restrictions in certain confluence areas, increased frequency of checking nets, compensating fishermen for damaged nets in the process of releasing entangled dolphins may all make a significant contribution to the survival of Indonesia’s only freshwater dolphin population.

RINGKASAN

Analisa kelangsungan hidup populasi lumba-lumba Irrawaddy dilakukan dalam suatu kelompok kecil, terasing di Sungai Mahakam, Kalimantan Timur, Indonesia. Simulasi tentang kelangsungan hidup populasi telah dilakukan dengan menggunakan program VORTEX didasarkan pada individu dengan mengetahui ukuran populasi dari N=55 sampai N=76. Disamping itu, akibat dari perkawinan antar individu sedarah (inces), penangkapan, kerusakan parah pada lingkungan, kapasitas habitat yang berbeda-beda, penurunan tingkat kematian dan pembuatan model termasuk penambahan individu dengan pemindahan sub-populasi yang terasing ke dalam populasi utama. Tanpa adanya usaha pelestarian, populasi hanya mempunyai 1 % ke 4 % kemungkinan untuk dapat bertahan sampai abad mendatang. Kunci untuk mempertahankan hidup terletak pada pengurangan jumlah kematian : mencegah kematian dari 2 atau 3 individu setiap tahunnya akan membantu mengatur kembali populasi ke arah pemulihan dan mencegah lumba-lumba dari kepunahan dengan tingkat kemungkinan 50%-70 %, dimana penyelamatan 3 individu per tahun dapat membuat kemungkinan bertahan hidup mendekati 100% Dikarenakan 80% kematian terjadi disebabkan oleh rengge atau jala, usaha pelestarian sebaiknya dititikberatkan pada pencarian cara agar mencegah kematian akibat terperangkap. Larangan pemakaian jala dibeberapa daerah tertentu, peningkatan frekuensi pemeriksaan jala, memberikan ganti rugi kepada nelayan yang mana mengalami kerugian akibat proses pembebasan lumba-lumba yang

200 Predicting long-term survival of riverine Irrawaddy dolphins

terperangkap kemungkinan dapat memberikan kontribusi yang nyata terhadap kelangsungan hidup satu-satunya populasi lumba-lumba air tawar Indonesia.

INTRODUCTION

Population viability analysis (Gilpin & Soulé, 1986; Schaffer, 1981, 1987) is widely used to provide estimates of the likelihood of extinction usually by either estimating the time to extinction, or the probability of extinction within a given period, typically 100 years. Although extinction is commonly believed to be inevitable for small, isolated populations due to demographic stochasticity, inbreeding depression, and catastrophic environmental and epizootic events (Soulé & Wilcox, 1980; Gilpin & Soulé, 1986; Lynch, 1996), this is not the expectation either from either the theory (Mills & Smouse, 1994; Frankham, 1995), or the empirical evidence. Many species have persisted for long periods at low population sizes (Haig et al., 1993) or have flourished from small founder populations (Van Aarde, 1979). However, greater problems may arise when a large population is suddenly reduced because most likely genetic and demographic processes interact so that as populations decline it becomes increasingly harder to bring about recovery (Gilpin & Soulé, 1986; Lynch et al., 1995). The Irrawaddy dolphin is a “facultative” river dolphin species of which separate coastal and freshwater populations exist. The freshwater dolphin population in the Mahakam is one of three riverine populations of Irrawaddy dolphins, which also occur in the Mekong River in Vietnam, Laos, Cambodia, and the Ayeyarwady River in Myanmar (Stacey & Arnold, 1999). The Irrawaddy dolphin population in the Mahakam is the only one to be listed as ”critically endangered” on the IUCN Red List based on preliminary results of this study (Kreb, 2002), although preliminary research in the two other rivers indicate a similar critical situation as in the Mahakam (Smith et al., 2003). Irrawaddy dolphins are also patchily distributed in small populations in shallow, primarily estuarine, tropical and subtropical marine waters of the Indo-Pacific from northeastern India east to Malampaya Sound, Philippines (Dolar, et al., 2002) and south to northeastern Australia (Stacey & Arnold, 1999). In order to take the right conservation measures for this population, which upper limit of size estimation only reached 76 individuals based on mark-recapture analysis of photo-identified individuals (Kreb, in press a), population viability analysis seems to be the way to proceed. Through this type of analysis one is forced to reflect on demographic parameters, identify gaps in the knowledge, consider how further data can be collected and understand the consequences of changes in parameter values, e.g. harvesting or changes in mortality rates (Akcakaya & Burgman, 1995). So, in order to study the effects of deterministic forces as well as demographic, environmental, and genetic stochastic (or random) events on the dynamics of a small, isolated population of Irrawaddy dolphins, Orcaella brevirostris in the Mahakam River in East Kalimantan, Indonesia, a population viability analysis was conducted (PVA) using the individual- based simulation program VORTEX (Lacey et al., 2003)

201 Chapter 12

METHOD & ANALYSIS

Basics of the VORTEX model

Data on population dynamics of the freshwater Irrawaddy dolphin population were collected during 718 hours of observation in 13 different survey periods of dolphins in the Mahakam River between early 1999 and mid 2002. A detailed description of survey procedures related to abundance and population dynamics, are described in Kreb (2002; in press a; in press b). For the population viability analysis (PVA) version 9.21 of the VORTEX program was used (Lacey et al., 2003). In VORTEX several scenarios can be set with their own population characteristics (see life table settings and scenarios), which can be simulated (I simulated each scenario 500 times) over a certain time period (I choose 100 years) to obtain information on the probability of extinction, mean deterministic and stochastic growth rates across all years of the simulation, mean population size (of the 500 simulations or ”populations”) of the surviving population at the end of the simulated time period, mean genetic diversity remaining in the extant populations, mean number of founder alleles remaining within extant populations, and the mean time of extinction. Vortex models demographic stochasticity by determining the occurrence of probabilistic events such as reproduction, sex determination, and death with a pseudo-random number generator.

Life table settings of the Mahakam population

Prior to running the Vortex simulations, I set the life table variables and definitions to comply with my population data set (see below) and for two different initial population sizes N = 76 (scenario 1) and N = 55 (scenario 2). The same settings of these scenarios are used for all other scenarios in which some changes were made in one of the variable values or definitions to look for the effects of these specific parameters on the probability of population survival while keeping the other variables constant (see scenarios). The abundance estimates above were chosen as the best and maximum estimated population sizes, respectively, based on the results of a Petersen’s mark-recapture analysis of photo-identified dorsal fins (Kreb, in press a). Estimates based on this analysis had the highest precision. These estimates excluded an isolated group of six dolphins, which is trapped between two rapid streams. The effects of translocation of this group of six individual dolphins on the population’s survival probability are also investigated in the VORTEX simulation (see below: Scenarios: Supplementation) I entered the following values for each life-table variable and set the following definitions for basic scenarios 1 & 2: Scenario settings: 1) number of iterations = 500 times, 2) simulation time = 100 years, 3) extinction definition = only one sex remains.

202 Predicting long-term survival of riverine Irrawaddy dolphins

Species description and catastrophes: 1) no inbreeding depression, 2) set environmental concordance of reproduction and survival, representing the annual variation in the probabilities of reproduction and survival that arise from random variation in environmental conditions. The model is constrained such that that good years for reproduction are also good years for survival, which makes sense in the Mahakam River where fish abundance not only fluctuates seasonally but also annually, 3) no catastrophes; Reproductive system: 1) polygamous reproduction system, which is based on this study (see chapter 9), 2) reproductive age for females and males is 9 years and 12 years, respectively. This is based on the fact that near adult size was reached in 6 years (Marsh et al., 1989), so that adult size and adult reproductive capabilities are hypothesized to occur at the age of 8 and giving birth at age nine at the earliest. Males are assumed to be physically mature at the same age (capable of producing sperm) but are only socially mature (capable of effectively competing for females) some years later (Connor et al., 2000), 3) maximum age of reproduction (defined in Vortex as reproductive until death) = 30 years of age based on Irrawaddy dolphins from northeastern Australia (Marsh et al., 1989), 4) maximum number of progeny in one “litter” is one, 5) sex ratio at birth = 50 % for cetaceans (Berta & Sumich, 1999), 6) no density dependent reproduction. Since it was found that males rove between different female core areas, it does not seem likely that difficulties in finding mates at low densities would apply. Reproductive rates: 1) % adult females breeding in any year = 50 %, since Irrawaddy dolphins are fully weaned at two years of age (Marsh et al., 1989) and pregnancy may occur during the second year of lactation, resulting in births by one female with 2 years of interval minimally, 2) environmental variation in breeding (SD in yearly breeding) = 0 because of lack of data due to short-term character of study in which only information of newborns for two years is available, during which in both years a similar number of newborns was observed, 3) females breeding have a 100% chance to produce a maximum of 1 offspring after each pregnancy. Mortality rates: 1) For N = 76: age 0-1 = 16% age1-2 = 0%; age 2-8 = 3% per yearly age class; age 9-30 = 12%. 2) For N = 55: age 0-1 = 16% age1-2 = 0%; age 2-8 = 5% per yearly age class; age 9-30 = 17%. For each class the standard deviation of rates were defined as 37%. Males and females were assumed to have equal rates since data on sex-specific deaths were incomplete. To calculate mortality rates, the average number of dead individuals per age class were divided by the number of individuals alive in each age class. Mortality was estimated from own observations and semi- structured interviews conducted during a preliminary survey in 1997 and during the surveys from February 1999 until August 2002. Mortality was traced back as far as 1995 and covered 7 years. Incomplete or untrustworthy accounts with missing locality, date, and traceable eyewitnesses were disregarded (14% of n = 44). Only for neonates and calves (0-1 years of age), mortality rates were obtained from a life-table of an in-depth study of bottlenose dolphins, Tursiops truncatus (Stolen & Barlow, 2003) because only once in 7 years a still-birth of a neonate was reported, which is very likely

203 Chapter 12

an under-estimate. Death of neonates may be more likely to remain undetected as their cause of death may be related to disease or birth complications, whereas death of calves (1-2 years of age) have a higher chance to be detected as their main cause of death may also just like juveniles and adults involve gillnet entanglement. Therefore, since no calves (1-2 years) have been reported, the number of deaths in this age class is assumed to be zero within this model. The number of individuals alive in each age class was based on the average yearly composition of 15% newborns and calves (0-1 years), 10% calves (1-2 years), 25% juveniles (2-8 years) and 50% adults (9-30 years) found during three different extensive monitoring surveys, which covered the entire dolphin distribution area (see also population dynamics in Kreb, in press b). The average number of newborns during each of these surveys was only one individual, but since the total number of newborns (0-2 months of age) per year was recorded to be 6, this last total number was used for age class 0 to 1 years of age, when calculating the percentage of population composition. This population composition in percentage was maintained and multiplied with different initial population sizes. The standard deviation was calculated based on the mean of the total number of individuals, which minimally died yearly. Because the standard deviation was unknown for the mortality rates obtained from literature, a mean standard deviation of yearly rates was calculated combined for different age classes. Mate monopolization: 50% of males are assumed in breeding pole based on polygamous breeding system (see also chapter 9).The actual % of breeding males are unknown. Initial population size: for N = 76: specified age distribution in each yearly age class: age 0-1 = 6; age 1-2 = 4; age 2-6 = 2; age 6-26 = 1; age 26-30 = 0. For N = 55: age 0-1 = 4; age 1-2 = 3; age 2-3 = 2; age 3-22 = 1; age 22-30 = 0. The age distribution was based on the percentage population composition for different age classes (described earlier for the mortality rates) multiplied with initial population size and divided with the number of years in each age class. Since only complete values were used in the analysis, in some of the last years of the adult age class a zero value was entered so that the initial population size was not exceeded. Carrying capacity: K = 200; SD in K due to environmental variation = 20. In the VORTEX model if N exceeds K, additional mortality is imposed across all age and sex classes in order to reduce the population back to its upper limit. A carrying capacity of 200 was chosen as it this was assumed to be applicable due to the limited remaining suitable habitat with sufficiently abundant fish resources. If dolphins are forced to move to secondary habitat (which they probably have to when N = 200), this may result in decreased fecundity or survival resulting in a reduction in population numbers.

204 Predicting long-term survival of riverine Irrawaddy dolphins

Scenarios

Scenarios 1 and 2 represent future population survival of populations with initial population sizes of N = 55 and N = 76 with mean mortality rates of 5 individuals per year (mean number derived from Kreb, in press b). I investigated the impacts on population survival if mortality rates can be reduced yearly with one, two, or three individuals per year. Scenarios 3, 5, and 7 had initial population sizes of N = 76 but annual mortality was reduced with one, two and three individuals, respectively. Likewise, scenarios 4, 6, and 8 had initial population sizes of N = 55 and reduced annual mortality with one, two and three individuals, respectively. To test these scenarios with mortality reductions the values in the parameter mortality rates were changed as follows: mortality reductions of one, two or three individuals expressed in percentages of mean mortality of five individuals per year are 20%, 40% and 60%, respectively. These percentages were inversely multiplied with the mortality rates and standard deviation per age class of scenarios 1 and 2 in order to obtain the reduced mortality rates. Default scenarios 1 and 2 also formed the basis for other alternative scenarios (which were not numbered but used descriptive terms mentioned further in the text) in which some changes were made in parameter values or definitions to look for the effects of these specific parameters on the probability of population survival while keeping the other variables constant. In this way, the effects of changes in inbreeding depression, catastrophic events, harvesting and supplementation were investigated. To investigate the impacts of inbreeding depression, the option inbreeding depression within the Species description and catastrophes was ticked and lethal equivalent were set to be 3.14 (the median of 40 mammalian populations surveyed by Ralls et al. 1988) with 50% of that due to lethal alleles. To investigate the impacts of catastrophes, two types of catastrophes that may realistically occur within the river dolphin population are simulated: The first is a catastrophe caused by pollution that may cause a reduction of 50% in breeding during catastrophic years and have no direct impact on survival. The second catastrophe is caused by sudden and extreme drought, which may cause entrapment in shallow areas and death of one group of mean group size (n = 4) and also causes a decrease in direct population survival with 5 % and 7% of N = 76 and N = 55, respectively. To investigate the impacts of habitat improvements, i.e. increasing the carrying capacity K on population survival for N = 76 and N = 55, different values for K were simulated between 250 and 500 individuals with increments of 50 individuals and with SDs = 10% of K. To investigate the impacts of harvest or live-captures on the population survival, section Harvest within VORTEX was ticked and the following harvest values were entered for N = 76 & 55, and for 1, 2, and 3 individuals reductions in mortality rates: harvest = 7 adult females and 2 adult males totalling 9 individuals as requested for live captures by local authorities. I also investigated the effects of a possible supplementation involving translocation of an isolated group of 6 dolphins (4 females and 2 males), which are trapped between two rapids since 1998 and which cannot

205 Chapter 12

exchange with the main population. The effects of supplementation were analysed for N = 76 & 55, and for 1 and 2 individuals reductions in mortality rates. The harvest values entered were: supplementation = 4 females and 2 males. In other scenarios I also simulated supplementation when taking place at later time events, after 10, 20, 30, 40 or 50 years from now.

RESULTS

The outcomes produced by population viability analysis of scenarios with initial population sizes of N = 55 or N = 76 and with reductions of mortality of 0 to 3 individuals per year for 100 years are presented in Table 1. Scenarios 1 and 2 with stable mean mortality rates of 5 individuals per year have a very high probability of extinction (PE), i.e. 94% and 99% respectively, and the mean time to extinction is within two or three decades. Reductions in mortality of one individual per year are not sufficient as the probabilities of extinction are still 83% and 97% for initial population sizes N = 76 and N = 55, respectively (scenarios 3 and 4). The probability of extinction decreases drastically with 70% and 51% for initial population sizes N = 76 and N = 55, respectively, and arrives at 24% and 48% after yearly reduction of mortality with 2 individuals (scenarios 5 and 6). Mortality reductions of 3 individuals per year result in a near 100% probability (96% and 97%) of survival for the next 100 years (scenarios 7 and 8). Additionally, stochastic growth rates become positive only when mortality has been reduced with 2 individuals. Only for scenario 2 with initial population size N = 55 and with no mortality reductions was the population in deterministic decline. Likewise, mean population sizes after 100 years only exceed their initial sizes when mortality has been reduced with 2 or 3 individuals yearly, i.e., scenarios 5 to 8 (Figure 1). Final genetic diversity and number of founder alleles remaining in the population increased steadily overall with increased reductions of mortality rates. The simulations of scenarios 1 to 8 show a high, final degree of genetic variation and indicates that the extent to which the population is inbred according to these models is low. Simulations of alternative scenarios based on scenarios 1 and 2 but with assumed inbreeding depressions (which do not have scenario names) showed nearly similar probabilities of extinction as those without assumed inbreeding depressions with only 2% or no differences in probabilities for N = 76 and N = 55, respectively. The impacts of habitat improvements as expressed in an increase in carrying capacity K with sustained population sizes between 200 and 500 individuals, did not have any impact on population survival probability for initial N = 76 and N = 55, respectively. Catastrophes in the form of one-time events have a far greater impact on the probability of extinction when they affect direct survival instead of rate of reproduction. To study the impact of catastrophes simulations were run of the scenarios in which mortality was reduced with 2 individuals and with probabilities of extinction of 0.24 and 0.48 of N = 55 and 76 respectively, since probability of

206 Predicting long-term survival of riverine Irrawaddy dolphins

Table 1. Population viability analysis outcomes for different initial population sizes and reductions in mortality as expressed in numbers of individuals.

INPUT OUTPUT Prob. Mean Reduce Determ Stoch. of Genetic time of Initial mortal. growth growth SD Extinct. N SD Diversity SD N SD extinct. Scenario N (n) (det-r) (stoc-r) (stoc-r) (PE) extant (N ext) (GD) (GD) Allel (Allel) (years) 1 76 0 0,013 -0,022 0,352 0,94 64 61 0,858 0,096 16 11 35 2 55 0 -0,022 -0,049 0,402 0,99 124 0 0,850 0 10 0 26 3 76 1 0,03 -0,003 0,317 0,83 86 68 0,864 0,102 18 11 41 4 55 1 0,001 -0,03 0,376 0,97 103 92 0,807 0,158 15 11 30 5 76 2 0,043 0,017 0,256 0,24 112 74 0,860 0,135 20 11 57 6 55 2 0,028 0,001 0,282 0,48 99 68 0,842 0,145 17 9 50 7 76 3 0,064 0,051 0,18 0,04 161 53 0,922 0,069 29 11 61 8 55 3 0,05 0,034 0,193 0,03 133 58 0,900 0,081 22 9 23

extinction of the default scenarios 1 and 2, but also 3 and 4 were already too high to analyse the impacts of catastrophes. When the severity of the catastrophe was set at 5% reduction of population survival during one year (of 100 years in total), the probability of extinction increased with 50% and 36% for N = 76 and N = 55, respectively compared to a situation with no catastrophes. On the other hand, when severity of a catastrophe was set to reduce breeding by 50% in any given year, the probability of extinction was increased only with 6% and 5% for N = 76 and N = 55 compared to a situation without catastrophes. The impacts on harvesting the population with the minimum number of animals (n = 9) requested by local authorities for live-display in an oceanarium (see chapter 6, threats) were applied to default scenarios 1 to 8. Three out of 8 default scenarios (scenarios 5, 7 and 8) had > 50% probability of survival when harvesting occurred. If no harvesting takes place, 4 out of 8 scenarios (scenarios 5 to 8) have >50% probability of survival. Also, survival probability of all scenarios was reduced with 5% on average (range = 1% – 15%). Moreover, in the small initial population size of N = 55, genetic diversity decreased from 0.850 prior to recruitment to 0.0 after harvesting. A 4% reduction of gene diversity occurred at initial population size of N = 76. When investigating the impacts of supplementing the population (using default scenarios 1 to 6, because 7 and 8 already cause a 100% survival) with an isolated group of 6 individual dolphins, which is trapped between two rapid streams, it appears that translocation in this case does not make a significant contribution to the population’s survival: only 1% positive difference in survival probability was achieved (range = 1% - 3%). Small, positive increases in survival probabilities were found when this one- time supplementation event occurred at a later time period; a 13% increase of survival probability occurred when supplementation took place after 50 years.

207 Chapter 12

Scenario 7

Scenario 8

Scenario 5

Scenario 6

Scenario 1 Scenario 3

Scenario 2 Scenario 4

Figure 1. Graph presenting mean population sizes through time for several scenarios corresponding with table 1. Mean population sizes are based on 500 simulations per scenario. Scenarios 1 to 4 have between 83% and 99% probability of extinction. In scenarios 1 and 2 no reductions in mortality occur, whereas in scenarios 3 and 4 mortality is reduced with one individual per year. In scenarios 5 and 6, and 7 and 8 mortality is reduced with 2 and 3 individuals yearly and with 3% and 48% probabilities of extinction within 100 years.

DISCUSSION

Population viability analysis has proven very useful in this case-study in determining conservation actions just as it has in other case-studies (Lindenmayer et al., 1993; Green et al. 1996). For example, by studying the impacts of harvesting and supplementation of the population, it is now clear that no live-captures should be allowed. In this study, results of a one-time harvesting event was already drastic in terms of decreased genetic variation and marked decline in survival probability, whereas it is very likely that captures will be repeated in the future to substitute

208 Predicting long-term survival of riverine Irrawaddy dolphins

individuals that have died in captivity. The impacts of repeated captures will undoubtedly cause higher extinction probabilities. Also, the main key to survival is not relocation of an isolated sub-population to the main river dolphin population, but rather involves reducing direct mortality to a minimum of 2 to 3 individuals per year. The huge impact of these seemingly insignificant numbers (3% to 5% of total N = 55 and 76) were also found by Fujiwara and Casswell (2001) for North Atlantic right whales, Eubalaena glacialis, in that prevention of the deaths of 2 to 3 adult females yearly could save the population from extinction. Since 80% of deaths in this study were found to occur through gillnet entanglement (Kreb, in press b), it is essential to focus conservation efforts on ways to prevent these entanglements. In the years after the study had finished in 2002, meetings of fishermen were organized in several villages in primary dolphin habitat by a local NGO, RASI Conservation Foundation in cooperation with the Wildlife Conservation Department of East Kalimantan Province and the Fisheries Department, to provide information on more sustainable fishing techniques, to demonstrate theoretically how to release entangled dolphins, to ask for fishermen’s support to release entangled dolphins, and to refund their nets if damaged while in the process of releasing a dolphin. Also, in one major dolphin area where 60% of the population is thought to occur (see chapter 6), daily patrol is conducted by a locally employed fisherman to check for dangerously placed nets and to relocate these. Through frequent patrols, entangled dolphins may also be sooner detected and prevented from drowning. However, the best option would be to restrict use of gillnets inside important dolphin areas, such as some confluence areas, which are daily frequented by dolphins (Kreb, in press b). Finally, some caution ought to be taken when drawing conclusions based on the simulations conducted because short-term or small data sets may produce very different results, i.e., extinction probabilities compared with a long-term data set (Brook & Kikkawa, 1998). Also, we found that inbreeding depression did not have a great impact on the population. However, in VORTEX inbreeding depression is defined as a reduction in the survival of offspring during the first year of life and likely underestimates its effect since it may also depress other components of fitness such as adult survival, fecundity, and/ or success in competition for mates (Miller & Lacy, 2003). Nevertheless, although parameter values may not reflect actual values, the results of this simulation may help since they hint at the direction where an effort at conservation is most required and indicate which events determine the viability of the population.

REFERENCES

Akcakaya, H.R.& Burgman, M. 1995. PVA in theory and practice. Conservation Biology 9: 705-707. Berta, A. & Sumich, J.L. 1999. Marine mammals: evolutionary biology. Academic Press London.

209 Chapter 12

Brook, B.W. & Kikkawa, J. 1998. Examining threats faced by island birds: a population viability analysis on the Capricorn silvereye using long-term data. Journal of Applied Ecology 33: 491-503. Connor R.C., Read A.J., Wrangham R. 2000. Male reproductive strategies and social bonds. In: Mann J, Connor RC, Tyack PL, Whitehead H (eds) Cetacean societies. Field studies of dolphins and whales. Pp 247-269. The University of Chicago Press, Chicago. Dolar, M.L.L., W.F. Perrin, J.P. Gaudiano, A.A.S.P. Yaptinchay & Tan, J.M.L. 2002. Preliminary report on a small estuarine population of Irrawaddy dolphins Orcaella brevirostris in the Philippines. Raffles Bulletin of Zoology, Supplement 10: 155-60. Frankham, R. 1995. Conservation genetics. Annual Review of Genetics 29: 305-327. Fujiwara, M. & Casswell, H. 2001. Demography of the endangered North Atlantic right whale. Nature 414: 537-541. Gilpin, M. & Soulé, M. 1986. Minimum viable population processes of species extinction. In: Soulé, M. (ed.) Conservation Biology. Pp. 19-34. Sunderland, Sinauer. Green, R.E. Pienkowski, M.W. & Love, J.A. 1996. Long-term viability of the reintroduced population of the white-tailed eagle Haliaeetus albicilla in Scotland. Journal of Applied Ecology 33: 357-368. Haig, S.M., Belthoff, J.R. & Allen, D.H. 1993. Population viability analysis for a small population of red-cockaded woodpeckers and an evaluation of enhancement strategies. Conservation Biology 7: 289-301. Kreb, D. 2002. Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. The Raffles Bulletin of Zoology, Supplement 10: 85-95. Kreb, D. (in press a) Abundance of freshwater Irrawaddy dolphins in the Mahakam in East Kalimantan, Indonesia, based on mark-recapture analysis of photo-identified individuals. Journal of Cetacean Research and Management. Kreb, D. (in press b) Conservation management of small core areas: key to survival of a critically endangered population of riverine Irrawaddy dolphins in Borneo. Oryx. Lacey, R.C., Borbat, M. & Pollak, P.J. 2003. VORTEX: A Stochastic Simulation of the Extinction Process. Version 9. Brookfield, IL, Chicago Zoological Society. Lindenmayer, D.B., Clark, T.W. Lacey, R.C. & Thomas, V.C. 1993. Ecotourism: A Guide for Planners and Managers. The Ecotourism society. North Bennington, VT. Lynch, M. 1996. A quantitative-genetic perspective in conservation issues. In: Avise, J.C. & J.L. Hamrick (eds.), Conservation Genetics. Pp. 471-75. Chapman & Hall, New York. i-xvii + 512 pp. Lynch, M., Conery, J. & Burger, R.C. 1995. Mutation accumulation and the extinction of small populations. American Naturalist 146: 489-518. Marsh, H., Lloze, R., Heihnsohn, G.E. & Kasuya, T. 1989. Irrawaddy dolphin Orcaella brevirostris (Gray, 1866). Pp. 101-118, in Handbook of Marine Mammals (R.J. Harrison & S. Ridgeway, eds.). Academic Press, New York, 4: 1-442.

210 Predicting long-term survival of riverine Irrawaddy dolphins

Miller, P.S. & Lacy, R.C. 2003. VORTEX: A Stochastic Simulation of the Extinction Process. Version 9.21 User’s Manual. Apple Valley, MN, Conservation Breeding Specialist Group (SSC/IUCN). Mills, L.S. & Smouse, P.E.1994. Demographic consequences of inbreeding in remnant populations. American Naturalist 144: 412-431. Ralls, K., Ballou, J.D. & Templeton, A.R. 1988. Estimates of lethal equivalents and the cost of inbreeding in mammals. Conservation Biology 2: 185-193. Schaffer, M.L. 1981. Minimum population size for species conservation. Bioscience 31: 131-134. Schaffer, M.L.1987. Minimum viable populations: coping with uncertainty. In: Soulé, M. (ed.) Viable Populations for Conservation. Pp. 69-86. Cambridge University Press, Cambridge. Smith, B.D., Beasley, I. & Kreb, D. 2003. Marked declines in populations of Irrawaddy dolphins. Oryx 37: 401. Soulé, M.E. & Wilcox, M. (eds.). 1980. Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer Associates, Sunderland, Massachusetts. 395 pp. Stacey, P.J. & Arnold, P.W. 1999. Orcaella brevirostris. Mammalian Species 616: 1-8. Stolen, M.K.& Barlow, J. 2003. A model life table for bottlenose dolphins (Tursiops truncates) from the Indian River lagoon system, Florida, U.S.A. Marine Mammal Science 19: 630-649. Aarde, van R.J. 1979. Distribution and density of feral cat Felis catus on Marion Island. South African Journal of Antartic Research 9: 14-19.

211 Chapter 12

212 Appendix 1

APPENDIX 1

Chapter 6.2. Geographical populations. In: R.R. Reeves, B.D. Smith, E.A. Crespo and G.N. di Sciara (ed.), 2002-2010 Conservation Action Plan for the World’s Cetaecans. Dolphins, whales and porpoises. Pp. 88- 89. IUCN, Gland, Switzerland and Cambridge, UK, 2003.

Irrawaddy dolphins in the Mahakam River, Indonesia

History: The population of Irrawaddy dolphins in the Mahakam River was recently listed as Critically Endangered, based on surveys in 1999 and 2000 that estimated the population of mature individuals to be fewer than 50 (Kreb 2002). Between 1995 and 2001, at least 37 dolphins died, primarily from entanglement in gillnets, but also from vessel collisions and illegal hunting (D. Kreb, pers. comm.; Kreb 2000). A proposal is being promoted to build an aquarium in the provincial capital, Tenggarong, and to stock it with dolphins from the Mahakam. This dolphin population is further threatened by a recent increase of large coal-carrying ships transiting through the core area of their distribution (D. Kreb, pers. comm.). Such boats occupy over three- quarters of the river’s width and therefore force prey fish into shallow areas that are seasonally inaccessible to the dolphins. The vessels also affect dolphin movements and often collide with the tree-lined banks, causing extensive damage to root systems where fish lay eggs.

Management issues and conservation progress: Irrawaddy dolphins are protected from killing and live capture according to Indonesian law, but monitoring and enforcement are minimal. There is also little enforcement of laws against destructive fishing methods (e.g., the use of electricity and poisons) and the logging of riparian forests, which causes sedimentation and destroys fish spawning sites. In early 2001, a local NGO, the Conservation Foundation for the protection of Rare Aquatic Species of Indonesia (Yayasan Konservasi RASI), together with conservation authorities of the East Kalimantan government, initiated the Pesut Mahakam Conservation Program. The primary aims of this community-based

213 2002-2010 IUCN Conservation Action Plan

program are to establish effective protection for dolphins in their core habitat and to create an informed and concerned constituency to support dolphin conservation.

Conservation recommendations: While stressing the importance of population monitoring and further evaluation of threats (Project 1, Chapter 5), the CSG recommends that : • Immediate action be taken to eliminate or drastically reduce human-caused mortality. At a minimum, alternative employment options for gillnet fishermen should be promoted so that accidental killing is reduced (IWC 2001a). Regulations that prohibit the intentional killing of dolphins, destructive fishing methods, and the logging of riparian forests should also be enforced. This will require the development of a reporting network among local villagers so that authorities become aware of infractions in a timely manner and can take appropriate action. • Permanent removals for captive display facilities have the same effect as hunting or by catch on the dolphin population. Therefore, no exceptions should be allowed to the national law that prohibits dolphin captures (IWC 2001a). • Because of concern about the habitat degradation and physical displacement of dolphins and their prey caused by large coal-carrying ships, alternative means of coal transport, such as smaller, less destructive barges, should be employed.

214 Appendix 2

APPENDIX 2

Chapter 5.1. Project 1. In: R.R. Reeves, B.D. Smith, E.A. Crespo and G.N. di Sciara (ed.), 2002-2010 Conservation Action Plan for the World’s Cetaecans. Dolphins, whales and porpoises. P. 56. IUCN, Gland, Switzerland and Cambridge, UK, 2003.

Monitor and evaluate ongoing threats to the Irrawaddy dolphins in the Mahakam River of Indonesia

The Critically Endangered Irrawaddy dolphins in the Mahakam River of East Kalimantan range in the mainstem from about 180 km to 600 km upstream of the mouth, seasonally entering several tributary rivers and lakes (Kreb 2002). The total population was estimated to number fewer than 50 individuals based upon eight surveys of their entire range conducted in 1999 and 2000. The dolphins were found primarily in deep pools located near confluences and meanders, which also primary fishing grounds and subject to intensive motorized vessel traffic. Between 1997 and 1999, 16 deaths were recorded (ten from gillnet entanglement, three probably from vessel strikes, and three deliberate) (D. Kreb, pers. comm.). From 1997 to 1998, at least seven dolphins were also illegally live-captured from the river and taken to oceanaria, and plans exist to capture more animals for a new oceanarium to be built in Tenggarong (D. Kreb, pers. comm.). Intensive fishing with gillnets, electricity, and poison, and the accidental introduction of an exotic piscivorous fish, locally known as ikan toman, may have depleted the dolphins’ prey (D. Kreb, pers. comm.). The high density of gillnets used in Semayang and Melintang lakes causes physical obstruction to dolphin movements, thereby reducing available habitat This problem, together with sedimentation caused by devegetation of the surrounding shorelines, has probably resulted in the elimination of these lakes as primary habitat as reported by Tas’an and Leatherwood (1984). Leaks from dams in the upper reaches that retain mining wastes, including mercury and cyanide, occurred in 1997 and resulted in a massive fish kill (D. Kreb, pers. comm.). An ongoing program, started in 1997 and conducted jointly by the University of Amsterdam and the East Kalimantan Nature Conservation Authority

215 2002-2010 IUCN Conservation Action Plan

(Balai Konservasi Sumber Daya Alam Kaltim) has involved extensive monitoring of the Mahakam dolphins. This program should be continued and expended to include toxicological and genetic analyses of tissues obtained from stranded or incidentally killed dolphins, investigations of factors that continue to degrade dolphin habitat, and further efforts to monitor abundance. The involvement of local scientists is vital. Because of this population’s Critically Endangered status, every effort should be made to prevent any further catches (including live capture) and improve the quality of the riverine environment (Chapter 6).

216 Facultative River Dolphins

SUMMARY

Facultative river dolphins: Conservation and social ecology of freshwater and coastal Irrawaddy dolphins in Indonesia

Irrawaddy dolphins Orcaella brevirostris are facultative river dolphins, of which both coastal as well as freshwater populations exist. These separate populations are most likely to have evolved during a historical, evolutionary process and possibly through allopatric speciation during the last glacial maximum. The species is found in shallow, coastal waters of the tropical and subtropical Indo-Pacific from eastern India to southern Philipines and northern Australia, including most of the Indonesian Archipel. The dolphins also occur in three major river systems: the Mahakam in Indonesia, the Ayeyarwady in Myanmar (former Burma), and the Mekong crossing through Vietnam, Cambodia and Laos. Since 1990 the species has been fully protected by law in Indonesia and is adopted as a symbol of East Kalimantan Province. Prior to the present study, no systematic data had been collected before on the Mahakam population or “pesut” as they are referred to locally. In order to fill in the gap of knowledge on the population’s status, dynamics and threats, as well as on the species’ biology, data were collected during a two-month preliminary study in 1997 and during 3.5 years intensive research from early 1999 until mid 2002. The research basically exists of a fundamental and applied part. The more applied, conservation part of the study attempted to identify and monitor the population status, dynamics, and threats thoroughly and set a rationale for conservation action of the riverine population. The more fundamental part of the study involves the study of the impacts of habitat on the social structures and acoustic behaviour of coastal and freshwater dolphin populations. Survey techniques, which were used to obtain data to fulfil these objectives involved: direct counting, density sampling techniques (strip-transects and line- transects), photo- and video-identification study, focal group-follows, boat-based and land-based observations, collection of skin (cells) samples for genetic analysis, semi- structured and formal interviews, and collection of random, environmental samples. During the fieldwork we always opted for a minimum invasive approach. Results of the applied part of the study of direct importance to conservation are the following: In the Mahakam River, best estimates of total population size varied between 33 and 55 dolphins (95% confidence limits: 31-76 individuals) based on direct counts, strip-transect analysis, and Petersen and Jolly-Seber mark-recapture analyses of photo-identified dolphins. These abundance estimates have a narrow range considering the wide-range of methods that were used. Precision of estimates of population size of direct counts and Petersen mark-recapture analysis was highest and nearly similar but the latter method is preferred since it is less biased and potential

217 Summary

biases may be calculated. Mean, minimum annual birth- and mortality rates were nearly similar, i.e., 13.6% and 11.4% and no changes in abundance > 8% were detected over 2.5 years. Smaller changes could not be detected due to the limitation of the study period. Dolphins primarily died after gillnet entanglement (73% of deaths). Dolphins’ main habitat includes confluence areas between the main river and tributaries or lakes. Dolphins daily intensively use small areas mostly including confluences, moving up and downstream over an average length of 10 km of river strip and within a 1.1 km2 - area size, and exhibit a high degree of individual site- fidelity. These areas are also primary fishing grounds for fishermen and subject to intensive motorized vessel traffic. Sixty-four percent of deaths (from 1995-2001) with known location (n = 36) occurred in these areas. River dolphins surfaced significantly less in the presence of motorized canoes (< 40 hp), speedboats (40-200 hp), and container barges (>1000 hp) and they actively avoided container barges. Formal interviews with local residents revealed a generally positive attitude towards the establishment of protected dolphin areas in small, manageable sites. Because of the dolphins’ dependence on areas that are also used intensively by people, primary conservation strategies in these sites should be to increase local awareness, to introduce alternative fishing techniques, set gillnet restrictions, promote increased frequency of checking nets, compensating fishermen for damaged nets in the process of releasing entangled dolphins. Without any conservation action, the population has a 1% to 4% chance of survival to the next century based on population viability analysis. Key to survival lies in mortality reductions: Preventing the deaths of 2 individuals yearly may help to set this population back on the road to recovery and prevent the dolphins from extinction with a 50% to 75% probability, whereas saving 3 individuals yearly causes a survival probability of near 100%. Since 73% of deaths occur through gillnet-entanglement, conservation efforts should primarily focus on finding ways to prevent death through entanglement following measures stated above. Based on data collected during 1999 and 2000, the IUCN (International Union for Conservation of Nature and Natural Resources) Red List of Threatened Animals status of this freshwater population was raised from ‘Insufficiently Known’ to ‘Critically Endangered’. Irrawaddy dolphin populations in the Mekong River and Ayeyarwady River fare no better as the populations consisted of less than 100 individuals based on preliminary studies, and they faced ongoing and pervasive threats to their long-term persistence. Their status may follow that of the Mahakam population after enough data is collected on their abundance and threats. The following results were obtained during the more fundamental part of the research related to the differences in social structures and social communication of freshwater and coastal dolphin populations: Based on individual recognition through photo-identification, it appeared that different groups of individual dolphins congregated in two major core areas in the river, one more downstream and one more upstream with a high site fidelity, particularly of females. Based on differences in sizes of neck crests, males and females could be identified in the field (although this sexual dimorphism was less pronounced for coastal dolphins) and it appeared that only

218 Facultative River Dolphins

(some) males were found to move in between these two areas. Breeding occurred through the year, this in contrast with the Irrawaddy dolphins in Balikapan Bay, which bred seasonally. Both coastal and freshwater populations are hypothesized to have polyandrous mating systems, although mating strategies of riverine dolphins may exist of both open competitors with a high site fidelity and roving males, which attempt to mate before estrous females are detected by other males. Coastal dolphins were less social in that they interacted less often with other groups than river dolphins and their interactions were mostly functional in term of feeding together, whereas river dolphin inter-group interactions were of varying nature and feeding, travelling and socializing were the most prevailing. Agonistic and avoidance encounters were also relatively common. River dolphins associated with each other relatively intensively and had long-term preferred companionships. Environmental factors thought to impact social structures and explain differences between coastal and riverine dolphins are the differences in degree of restrictedness of habitat shape, seasonality of food abundance, and clumping or scattering of food resources. Habitat differences, but also social structures and genetic relatedness also impacted on acoustic behaviour of coastal and freshwater populations. Vocalizations were most varied and frequent in one core area of dolphins in the river in which a well-identified sub-population with a high site-fidelity occurred and with the highest degree of social exchange among groups in comparison to two other areas in the river and the coastal bay population. Pod-specific whistle-dialects exist not only among coastal and riverine populations, but also within sub-pods within the river, which differ in the number of modulations, duration, and minimum and maximum frequencies. There is also evidence for individual “signature” whistles and “contact” whistles. Vocal repertoire (sound types) was more similar between the likely more genetically related, coastal and freshwater populations in East Kalimantan than between coastal populations of East Kalimantan and Australia (see chapter 10). Vocal repertoire was less varied for coastal Irrawaddy dolphin populations in East Kalimantan and Australia compared with the Mahakam River, which may be determined by ecological conditions. Food resources and dolphins are more scattered in the coastal habitat causing a lesser degree of sociality, which influences their vocal repertoire. The whistles and vocalizations rates (numbers per time unit) seem to be determined by social structures. Larger groups with (more) calves whistled less often than smaller groups, which may be caused by the fact that there is less need for contact whistles. Whistle frequencies were significant higher upon approach of (speed) boats of > 40 hp and they lasted longer than in their absence. Conclusively, acoustic behaviour may help to define stocks as separate management units. Research recommendations involve continuing to collect genetic sample tissues to assess the genetic variation of the riverine population and the time of separation between coastal and riverine populations. Additionally, monitoring should be continued to detect any trends in abundance, to update the photo-identification catalogue, to assess the stability of preferences of core areas, and obtain the latest information on threats and on mortality rates.

219 Summary

220 Facultatieve Rivier Dolfijnen

SAMENVATTING

Facultatieve rivierdolfijnen: Bescherming en sociale ecologie van zoet- en zoutwaterpopulaties van Irrawaddy dolfijnen in Indonesië

Irrawaddy dolfijnen Orcaella brevirostris zijn facultatieve rivierdolfijnen, waarvan zowel kust- als zoetwaterpopulaties voorkomen. Deze gescheiden populaties hebben zich waarschijnlijk ontwikkeld tijdens het laatste glaciale maximum. De soort komt voor in ondiepe wateren van het tropische en subtropische Indo-Pacifische gebied, vanaf de oostkust van India tot aan de zuidelijk eilanden van de Filippijnen en noord Australië, en omvat het grootste deel van de Indonesische archipel. De dolfijnen komen ook voor in drie grote riviersystemen: de Mahakam in Indonesië, de Ayeyarwady in Myanmar (voormalig Birma) en de Mekong die Vietnam, Cambodja en Laos doorkruist. Sinds 1990 is de soort volledig beschermd in Indonesië en is zij geadopteerd als symbool van de provincie Oost Kalimantan. Voorafgaand aan deze studie waren geen systematische gegevens verzameld over de populatie in de Mahakam, die lokaal “pesut” genoemd wordt. Om het gebrek aan kennis omtrent de populatie-status en dynamiek uit te breiden, alsook de bedreigingen en de biologie van de soort in kaart te brengen, zijn gegevens verzameld tijdens een twee maanden durend vooronderzoek in 1997 en tijdens intensief onderzoek vanaf begin 1999 tot medio 2002. Het onderzoek bestond uit een fundamenteel en een toegepast gedeelte. Het toegepaste, op natuurbescherming gerichte deel van de studie, probeerde de status, dynamiek en bedreigingen van de rivierpopulatie zo volledig mogelijk te identificeren om zo de basis te scheppen voor beschermingsactiviteiten. Het fundamentele gedeelte van het onderzoek bestond uit het bestuderen van de effecten van habitatverschillen op sociale structuren en de akoestiek van rivieren kustgebonden populaties. De volgende onderzoeksmethoden werden gebruikt: rechtstreekse tellingen, dichtheden-bemonstering (strook- en lijntransecten), foto- en video-identificatie, het volgen van individuele dolfijnengroepen, boot- en landgebonden observaties, verzamelen van huidmonsters voor genetische analyse, semi-gestructureerde- en formele interviews, en het verzamelen van op willekeurige tijden en plaatsen genomen milieumonsters. Tijdens het veldwerk werd altijd getracht zo min mogelijk verstorend te werk te gaan. De volgende resultaten waren gevonden van het toegepaste gedeelte van de studie, die van direct belang zijn voor bescherming. In de Mahakam rivier variëren schattingen van totale populatiegrootte tussen de 33 en 55 dolfijnen (95% betrouwbaarheidsgrenzen: 31-76 individuen) gebaseerd op rechtstreekse tellingen, strook-transectanalyses, en merk- en terugvanganalyses volgens de Petersen en Jolly-

221 Samenvatting

Seber methodes gebaseerd op foto-identificaties van individuele dolfijnen. Deze aantalschattingen liggen dicht bijeen ondanks het feit dat de methoden zeer gevarieerd zijn. De precisie van schattingen van populatiegroottes was het hoogst voor de rechtstreekse tellingmethode en de Petersen methode, maar de laatste methode is te prefereren omdat er minder vooronderstellingen worden gemaakt. De gemiddelde minimale jaarlijkse geboorte- en sterftecijfers lagen dicht bijeen, namelijk 13.6% en 11.4%, en er was geen stijgende of dalende lijn door 8% of meer veranderingen in aantallen geconstateerd gedurende 2,5 jaar. Minder duidelijke schommelingen in populatieaantal konden door de korte studieduur niet worden waargenomen. Dolfijnen stierven voornamelijk na verstrikking in drijfnetten (73%). Het hoofdhabitat omvat samenstromingsgebieden tussen de hoofdrivier en zijrivieren of meren. Dolfijnen maakten voornamelijk gebruik van kleine gebieden en grotendeels in de samenstromingsgebieden, terwijl ze gemiddeld stroom open afwaarts zwemmen in een strook van 10 km rivier en in een gebied van 1,1 km2. Individuele dolfijnen vertonen een hoge mate van plekgebondenheid. Deze gebieden zijn ook primaire visgebieden voor vissers en zijn onderhevig aan intensief, gemotoriseerd bootverkeer. Van alle dolfijnen die stierven tussen 1995 en 2001 en waarvan de locatie bekend is (n = 36), was 64% gestorven in de samenstromingsgebieden. Rivierdolfijnen kwamen significant minder boven om te ademen in nabijheid van kleine motorbootjes (< 40 pk), speedboten (40-200 pk), en geladen containerboten (> 1000 pk) en dolfijnen ontweken de laatsten ook actief. Formele interviews met lokale inwoners gaf een overwegend positieve houding aan ten opzichte van het instellen van beschermde dolfijngebieden in kleine, controleerbare gebieden. Door de afhankelijkheid van dolfijnen van gebieden die ook intensief door mensen gebruikt worden, kunnen primaire beschermingsstrategieën het beste gericht zijn op het verhogen van het algemene bewustzijn, het introduceren van alternatieve vistechnieken, het stellen van regels voor het plaatsen van drijfnetten, het aanmoedigen van reguliere controles van netten en vissers te vergoeden voor netten die beschadigd zijn geraakt door het bevrijden van verstrikte dolfijnen. Zonder enige natuurbeschermingsactiviteiten, heeft de populatie slechts 1% tot 4% kans op overleven tot de volgende eeuwwisseling gebaseerd op Populatie Levensvatbaarheid Analyse (PVA). De sleutel voor het overleven is het reduceren van de mortaliteit: het voorkomen van de dood van twee individuen jaarlijks kan de populatie weer op de weg naar herstel zetten en behoedt de populatie voor uitsterven met een 50% tot 75% waarschijnlijkheid terwijl het voorkomen van de dood van drie individuen jaarlijks bijna een 100% kans op overleving oplevert. Omdat 73% van het aantal dode dolfijnen wordt veroorzaakt door verstrikking in drijfnetten, moeten beschermingsmaatregelen zich voornamelijk richten op het vinden van manieren om deze doodsoorzaak door verdrinking te voorkomen. Gebaseerd op de gegevens die in 1999 en 2000 werden verzameld heeft de Internationale Unie voor Bescherming van Natuur en Natuurlijke Bronnen (IUCN) de status van de Mahakam populatie gewijzigd van “Onvoldoende Bekend” naar “ Zeer Ernstig Bedreigd”. De populaties van de Irrawaddy dolfijn in de Mekong rivier en

222 Facultatieve Rivier Dolfijnen

Ayeyarwady rivier gaat het even slecht daar, gebaseerd op voorstudies, de aantallen niet boven de 100 individuen uitkomen. Ook zij staan continu bloot aan serieuze bedreigingen in hun voortbestaan op lange termijn. Hun status zal waarschijnlijk die van de Mahakam populatie volgen wanneer voldoende gegevens zijn verzameld omtrent hun precieze aantallen en bedreigingen. De volgende resultaten zijn gevonden gedurende het fundamentele gedeelte van het onderzoek aangaande de verschillen in sociale structuur en sociale communicatie tussen rivieren kustdolfijnenpopulaties. Op grond van individuele herkenning door middel van foto-identificaties bleek dat verschillende groepen van individuele dolfijnen voornamelijk samenkwamen in twee hoofdgebieden in de rivier. Een bevond zich tussen 180-200 km afstand van de riviermonding en een ander c. 350 km stroomopwaarts. In beide gevallen vertoonden de dolfijnen een hoge mate van plekgebondenheid, in het bijzonder de vrouwtjes. Gebaseerd op verschillen in een duidelijke aan- of afwezige verdikking van de nek konden mannetjes en vrouwtjes worden geïdentificeerd in het veld (al was deze seksuele dimorphie minder duidelijk in de kustpopulatie) en het bleek dat alleen (enkele) mannetjes in beide gebieden voorkwamen. Voortplanting gebeurde gedurende het hele jaar, in tegenstelling tot de kustdolfijnen, waar de voortplanting seizoensgebonden plaatsvindt. Zowel de kust- als de rivierpopulatie hebben vermoedelijk een paarsysteem dat zich kenmerkt door polyandrie, maar binnen de rivierpopulatie lijken mannetjes twee strategieën te hebben ontwikkeld: sterke mannetjes met een hoge mate van plaatsgebondenheid die in een open competitie vrouwtjes beconcurreren en jongere, minder competitieve mannetjes die proberen te paren met vruchtbare vrouwtjes voor ze ontdekt worden door andere mannetjes en minder plaatsgebonden zijn. Kustdolfijnen waren minder sociaal dan rivierdolfijnen en hadden minder vaak interacties met andere groepen. Ook waren deze groepsinteracties voornamelijk functioneel en bestonden hoofdzakelijk uit gemeenschappelijk jagen, terwijl groepsinteracties bij rivierdolfijnen van gevarieerdere aard waren en naast het jagen, samen een bepaalde afstand zwemmen en sociale interacties veelvuldig voorkwamen. Ook waren agressieve of ontwijkende interacties algemeen. Individuele rivierdolfijnen associeerden in hoge mate met elkaar en tussen individuen bestonden langdurige relaties. Omgevingsfactoren die mogelijk van invloed zijn op sociale structuren en de verschillen tussen kust- en rivierdolfijnen verklaren, zijn de mate van open- of beslotenheid van de habitat, de mate van seizoensgebondenheid van voedselaanbod en de geconcentreerdheid of wijdverspreidheid van voedselbronnen. Verschillen in habitat, maar ook sociale structuren en genetische verwantschap hadden invloed op de akoestische gedragingen van kust- en zoetwater dolfijnpopulaties. De vocalisaties waren het meest veelvuldig en gevarieerd in een hoofdgebied van dolfijnen in de rivier waar zich een goed geïdentificeerde sub- populatie bevond met een hoge mate van plaatsgebondenheid. Hier kwamen ook meer sociale groepsinteracties voor vergeleken met twee andere gebieden in de rivier en met de kustpopulatie. Kust- en rivierpopulaties, maar ook subpopulaties in de rivier, kennen “fluit” dialecten, die verschillen in het aantal modulaties, tijdsduur,

223 Samenvatting

minimum en maximum frequenties. Ook zijn er aanwijzingen voor individuele “gesigneerde fluitjes” en “contact fluitjes”. Het vocale repertoire (geluid types) was meer gelijk tussen de kust- en rivierpopulaties in Oost Kalimantan die vermoedelijk genetisch meer verwant zijn aan elkaar, dan tussen kustpopulaties van Irrawaddy dolfijnen in Oost Kalimantan en Australië (zie hoofdstuk 10). Het vocale repertoire van kustdolfijnen in Oost Kalimantan en Australië was minder gevarieerd vergeleken met de populatie in de Mahakam rivier , hetgeen veroorzaakt kan worden door ecologische omstandigheden. Mogelijk bestaat in de kusthabitat een meer wijdverspreid voedselaanbod waardoor dolfijnen niet in zulke hoge dichtheden voorkomen op bepaalde plekken als in de rivier en waardoor ze dus minder sociaal zijn hetgeen weer van invloed is op het vocale repertoire. De frequentie (aantal per tijdseenheid) van fluiten en het produceren van andere vocalisaties wordt meer bepaald door sociale structuren. Grotere groepen met (meer) kalfjes floten minder vaak dan kleinere groepen, hetgeen veroorzaakt kan worden door het feit dat er minder “contact fluitjes” noodzakelijk zijn. De frequentie van fluiten was hoger bij nadering van (speed)boten > 40pk en duurde ook langer dan in afwezigheid van deze boten. Akoestische gedragingen dragen bij om (sub)populaties als verschillende management eenheden te definiëren. Aanbevelingen voor verder onderzoek bestaat uit het vervolgen van het verzamelen van huidmateriaal om genetische variatie binnen de rivierpopulatie vast te stellen, alsook de tijd waarin kust- en rivierpopulaties van elkaar zijn gescheiden. Tevens dient het onderzoek voortgezet te worden om de tendens in populatieaantallen in kaart te brengen, om de foto-identificatie catalogus bij te stellen, om de stabiliteit van voorkeur voor bepaalde gebieden te bepalen en om de meest recente gegevens te verwerven van sterftecijfers en bedreigingen.

224 Lumba-Lumba Air Tawar

RINGKASAN

Lumba-lumba air tawar: Perlindungan dan ekologi sosial dari populasi lumba-lumba Irrawaddy pada sungai dan pesisir laut di Indonesia

Lumba-lumba Irrawaddy Orcaella brevirostris adalah lumba-lumba air tawar yang fakultatif, dimana populasinya dapat ditemukan pada laut dan sungai. Populasi yang terpisah ini kemungkinan besar terjadi dikarenakan proses evolusi selama jaman es terakhir. Jenis ini ditemui pada perairan dangkal, pesisir pantai tropis dan sub tropis Indo-Pasifik dari daerah timur India sampai selatan Philipina dan utara Australia, termasuk kepulauan Indonesia. Lumba-lumba ini juga terdapat pada 3 jalur utama sungai : Mahakam di Indonesia, Ayeyarwady di Myanmar (eks. Burma), dan Mekong yang membelah Vietnam, Kamboja dan Laos. Sejak 1990 jenis ini telah dilindungi sepenuhnya oleh Undang-undang di Indonesia dan diambil sebagai lambang bagi Propinsi Kalimantan Timur. Didasarkan pada studi terakhir, tidak ada data yang lengkap dikumpulkan sebelumnya untuk populasi di Mahakam, atau yang lebih dikenal sebagai “pesut”. Bertujuan untuk mengisi kekosongan pengetahuan dalam status populasi, kehidupan dan ancaman, demikian juga dengan biologi jenis, data dikumpulkan selama 2 bulan penelitian awal pada 1997 dan 3,5 tahun penelitian intensif mulai awal 1999 sampai dengan 2002. Penelitian pada dasarnya terdiri dari penelitian mendasar dan terapan. Bagian penelitian yang lebih dapat diterapkan adalah pada bagian konservasi untuk mengetahui dan mengawasi status dari populasi, pertambahan dan pengurangan, dan ancaman yang nyata dan menyusun kegiatan yang rasional untuk perlindungan populasi sungai. Bagian yang lebih mendasar adalah meneliti dampak habitat terhadap struktur sosial dan sifat suara dari populasi lumba-lumba laut dan sungai. Teknik survei yang digunakan untuk melengkapi data agar tujuan tercapai adalah penghitungan langsung, teknik pengambilan contoh kerapatan (strip transect and line transect), penelitian dengan foto dan video, mengikuti satu kelompok secara intensif, penelitian di atas kapal atau dari darat, pengambilan contoh kulit (sel) untuk analisa genetik, wawancara informal dan formal, dan koleksi secara acak, contoh yang berhubungan dengan lingkungan. Selama dilapangan kami selalu melakukan pendekatan yang tidak terkesan mengganggu. Hasil dari bagian penelitian yang bersifat terapan antara lain : di Sungai Mahakam, perkiraan terbaik untuk ukuran total populasi adalah berubah-ubah antara 33 dan 55 lumba-lumba (95% tingkat kepercayaan: 31-76 individu) didasarkan pada perhitungan langsung, analisa strip-transect, dan analisa penandaan dan penangkapan kembali Petersen dan Jolly-Seber dari lumba-lumba yang telah difoto-identifikasi. Jumlah perkiraan ini memiliki ruang terbatas mempertimbangkan metode yang memiliki jangkauan luas yang telah digunakan. Ketepatan perkiraan ukuran populasi dengan

225 Ringkasan

penghitungan langsung dan metode Petersen adalah yang tertinggi dan hampir serupa, namun metode yang terakhir adalah yang lebih sering digunakan karena kurang terjadi penyimpangan dan penyimpangan yang terjadi dapat diperhitungkan. Rata-rata tengah, kelahiran pertahun terendah, dan rata-rata kematian hampir seimbang yaitu 13,6% dan 11,4% dan jumlahnya tidak berubah > 8% yang diketahui selama lebih dari 2,5 tahun. Perubahan kecil tidak dapat diketahui karena waktu penelitian yang terbatas. Lumba-lumba umumnya mati setelah terperangkap rengge (73% kematian). Habitat utama lumba-lumba termasuk pertemuan sungai utama dan anak sungai atau danau. Sehari-harinya lumba-lumba sering menggunakan daerah yang kecil termasuk daerah muara anak sungai, bergerak ke hulu dan ke hilir sepanjang rata-rata 10 km dan luas daerah 1.1 km2, dan menunjukkan tingkat kesetiaan individu yang tinggi terhadap suatu tempat. Daerah-daerah ini juga merupakan tempat utama bagi nelayan menangkap ikan dan lalu lintas kapal bermotor. Sebesar 64% kematian (dari 1995- 2001) dengan lokasi yang diketahui (n = 36) terjadi di daerah ini. Lumba-lumba sungai tidak banyak muncul dipermukaan dengan kedatangan kapal bermotor (<40 stk), speedboat(40-200 stk), dan kapal penarik ponton (>1000 stk) dan mereka biasanya menghindari ponton. Wawancara formal dengan masyarakat lokal menyatakan sikap positif terhadap pembuatan daerah perlindungan lumba-lumba di tempat-tempat yang kecil dan dapat dikelola. Karena lumba-lumba bergantung pada suatu daerah yang juga biasa digunakan oleh masyarakat, strategi utama perlindungan di daerah-daerah ini sebaiknya adalah meningkatkan kepedulian, memperkenalkan teknik lain penangkapan ikan, pembatasan penggunaan rengge, meningkatkan frekuensi pengawasan jala, memberikan kompensasi kepada nelayan atas kerusakan jala pada proses pembebasan lumba-lumba yang terperangkap. Tanpa adanya usaha perlindungan, populasi hanya memiliki 1% sampai 4% kemungkinan untuk dapat bertahan sampai abad yang akan datang didasarkan pada analis kelangsungan hidup populasi. Kunci keberlangsungan hidup terletak pada pengurangan tingkat kematian ; mencegah kematian dua individu tiap tahun dapat membantu populasi ini untuk memperbaiki populasi dan mencegah kepunahan dengan tingkat kemungkinan 50% sampai 75%, dimana menyelamatkan tiga individu per tahun dapat menyebabkan kemungkinan kelangsungan hidup mendekati 100%. Sejak 73 % kematian dikarenakan oleh terperangkap rengge, usaha perlindungan seharusnya dititik beratkan pada menemukan cara untuk menghindari kematian karena terperangkap dengan mengikuti perhitungan yang disebutkan di atas. Didasarkan pada data yang dikumpulkan selama 1999 dan 2000, menurut Persatuan Perlindungan bagi Alam dan Sumberdaya Alam Internasional (IUCN) dalam Daftar Merah bagi Satwa Terancam, status dari lumba-lumba air tawar ini ditingkatkan dari “Tidak cukup diketahui” menjadi “Terancam Kepunahan. Populasi lumba-lumba Irrawaddy di sungai Mekong dan Ayeyarwadi tidak lebih baik dimana populasinya terdiri kurang dari 100 individu didasarkan pada penelitian awal, dan mereka menghadapi ancaman secara terus menerus dan mendalam bagi kelangsungan hidup mereka. Status mereka bisa saja mengikuti status dari populasi di Mahakam bila data yang diperlukan mengenai jumlah dan ancamannya telah dikumpulkan.

226 Lumba-Lumba Air Tawar

Hasil berikut didapatkan selama bagian penelitian yang lebih mendalam mengenai perbedaan dalam struktur sosial dan komunikasi sosial dari populasi lumba-lumba air tawar dan laut. Didasarkan pada pengenalan individu melalui identifikasi foto, nampaknya kelompok berbeda dari lumba-lumba berkumpul pada dua daerah inti yang besar di sungai, satu lebih ke hilir dan satu lebih ke hulu dengan tingkat kesetiaan terhadap satu daerah yang tinggi, khususnya betina. Didasarkan pada beda ukuran tengkuk, jantan dan betina dapat di identifikasi di lapangan (walaupun perbedaan jenis kelamin ini kurang terlihat pada populasi lumba-lumba laut) dan terlihat juga bahwa hanya (beberapa) jantan ditemukan mengarungi kedua daerah ini. Kelahiran terjadi sepanjang tahun, hal ini bertolakbelakang dengan lumba-lumba Irrawaddy di Teluk Balikpapan, yang berkembang biak secara musiman. Kedua populasi laut dan air tawar diprakirakan memiliki sistem perkawinan dengan banyak pasangan, walaupun strategi kawin dari lumba-lumba air tawar mungkin terjadi baik dengan kompetisi terbuka maupun pengambilan betina oleh jantan keluar dari kelompoknya, dimana mencoba untuk mengawini sebelum betina yang siap untuk kawin ditemukan oleh jantan lain. Lumba-lumba laut lebih sedikit bersosialisasi dan mereka lebih sedikit berinteraksi dengan kelompok lain dibandingkan dengan lumba-lumba sungai dan interaksi mereka lebih banyak berdasarkan fungsinya dalam mencari makan bersama. Sedangkan interaksi antara kelompok pada lumba-lumba air tawar memiliki beberapa tujuan selain mencari makan, berkeliling dan bersosialisasi adalah yang paling umum. Perkelahian dan menghindari kelompok lain juga sering terjadi. Lumba-lumba sungai berhubungan satu dengan yang lain relatif secara intensif dan mempunyai hubungan dengan jangka waktu lama. Faktor-faktor lingkungan diperkirakan mempengaruhi struktur sosial dan menjelaskan perbedaan antara lumba-lumba laut dan sungai adalah pada derajat keterbatasan bentuk habitat, jumlah makanan musiman, dan berkumpul atau berpencarnya sumber makanan. Perbedaan habitat, tapi juga struktur sosial dan hubungan genetis turut mempengaruhi suara antara populasi lumba-lumba laut dan sungai. Vokalisasi lebih banyak berbeda dan sering, dalam satu daerah utama di sungai, dimana satu sub populasi yang telah diketahui dengan tingkat kekerabatan yang tinggi, dan dengan derajat pertukaran tertinggi antara kelompok dalam perbandingan dengan dua daerah populasi lain di sungai dan laut. Siulan khusus kelompok tidak hanya ada diantara populasi laut dan sungai, tapi juga antara sub kelompok di sungai, dimana berbeda pada jumlah modulasi, panjangnya waktu, dan frekuensi maksimum dan minimum. Juga didapatkan bukti adanya siulan khusus sebagai tanda individu dan siulan untuk berhubungan. Tipe suara lebih mirip antara yang berhubungan genetis, seperti populasi di sungai dan laut Kalimantan Timur dibandingkan dengan populasi laut Kalimantan Timur dan laut Australia. Tipe suara kurang bervariasi dalam populasi lumba-lumba Irrawaddy laut di Kalimantan Timur dan Australia (lihat bab 10) dibandingkan dengan yang ada di Sungai Mahakam, yang mungkin disebabkan adanya perbedaan kondisi ekologi. Sumber-sumber makanan dan lumba-lumba lebih tersebar di habitat pesisir, yang menyebabkan tingkat sosial berkurang dan yang mempengaruhi tipe suara. Rata-rata siulan dan vokalisasi (jumlah per unit waktu) tampaknya

227 Ringkasan

ditentukan oleh struktur sosial. Kelompok yang lebih besar dengan (banyak) anakan bersiul lebih sedikit dibandingkan dengan kelompok yang lebih kecil, hal ini mungkin dikarenakan kurang perlunya siulan kontak. Frekuensi siulan secara nyata lebih banyak terjadi sewaktu mendekatnya speedboat dengan mesin >40 stk dan lebih panjang dari pada biasanya. Akhirnya, kita dapat menyimpulkan bahwa walaupun sifat akustik tidak menjawab pertanyaan apakah populasi lumba-lumba Irrawaddy memiliki perbedaan (sub) jenis atau hanya karena perbedaan bentuk geografi, tapi sifat akustik dapat membantu membedakan mereka sebagai suatu unit manajemen yang berbeda. Rekomendasi dari penelitian adalah meneruskan pengumpulan contoh genetis untuk memperjelas variasi genetis pada populasi di sungai dan waktu terjadinya pemisahan populasi laut dan sungai. Sebagai tambahan, pengawasan seharusnya diteruskan untuk mengetahui jumlah, untuk memperbaharui daftar identifikasi foto, untuk memperoleh kestabilan daerah utama, dan mendapatkan informasi terbaru tentang ancaman dan pada rata-rata tingkat kematian.

228 Curriculum Vitae

CURRICULUM VITAE

Daniëlle Kreb was born on 5 September 1971 in Emmeloord in the Noordoostpolder, The Netherlands. She attended secondary high school, Gymnaseum at the Prof. Ter Veen Lyceum in Emmeloord and started with her study in Biology in 1992 at the University of Amsterdam. During the differentiation phase she choose subjects, which were amongst others related to Tropical Ecology, Nature Conservation, Animal Behaviour, Systematics and Biogeography, as well as Marine Biology, a subject which she attended at the University of Groningen. During her specialization phase in 1996 she conducted fieldwork for eight months as an assistant in a radio-tracking field study of the Wildlife Conservation Research Unit at the University of Oxford, on the ecology of the Scottish wildcat, Felis silvestris grampia in the Angus Glens in Scotland. Thereafter in 1997, she conducted a two-months’ preliminary survey on the freshwater Irrawady dolphin in the Mahakam River in East Kalimantan, Indonesia. In 1997 she obtained her Master of Science degree at the Faculty of Biology of the University of Amsterdam. In 1998 she assisted in surveys conducted by the Ocean Park Conservation Foundation on the distribution of the Indo-Pacific hump-backed dolphin, Sousa chinensis in the South China Sea near Hong Kong, as a volunteer to train in basic techniques for monitoring of cetaceans. Fieldwork in East Kalimantan was continued between February 1999 until September 2002 and involved surveys upstream the Mahakam River in order to study the freshwater Irrawaddy dolphins, as well as surveys for cetaceans along the coast of Kalimantan. Concerned with the “pesut Mahakam” population she co-founded a local NGO, Yayasan Konservasi RASI (Conservation Foundation for Rare Aquatic Species of Indonesia) in September 2000 of which she became scientific advisor and developed a conservation program to protect the pesut population and its habitat. Conservation activities focusing on increasing awareness at different layers of society are being implemented currently. She also designed and supervised a one-year pilot research project in 2002 and 2003 on a dugong population, which she discovered during her surveys in Balikpapan Bay. From December 2001 until December 2003, she fulfilled a Ph.D. candidacy position at the Research Group Systematics and Zoogeography of the Institute for Biodiversity and Ecosystem Dynamics/ Zoological Museum, Science Faculty, University of Amsterdam. Currently she is developing a research program directed at marine cetaceans near the Berau Islands in Northeast Kalimantan while continuing monitoring and conservation activities on the freshwater Irrawaddy dolphin population in the Mahakam River. Daniëlle is married to Budiono and has one daughter: Rhaudatul Jannah.

229 Curriculum Vitae

(INTER)NATIONAL CONGRESSES/ WORKSHOPS

December 1999 - Participating in 13th Biennial Conference on the Biology of Marine Mammals, Maui, Hawai. June 2000 - Invited participant in the two-week 52nd meeting of the International Whaling Commission within the Standing sub-Committee on Small Cetaceans, Adelaide, Australia. Presenting working paper. June 2001 - Official presentation at the governor’s office in Samarinda during a two days’ workshop in June 2001 on management of the Mahakam lakes together with local governmental organizations related to forestry and fisheries, NGO’s and the Universities of Samarinda and Yogyakarta. December 2003 - Ph.D. days 2003 in Amsterdam, The Netherlands. Oral (powerpoint) presentation – “Conservation and population biology of the freshwater Irrawaddy dolphin in Borneo”. Also, poster presentation, which received the prize for the best Ph.D poster “Getting the numbers right: evaluation of different census methods in abundance estimation of Irrawaddy river dolphins in Borneo”.

PUBLICATIONS

Kreb, D. (1999) Observations on the occurrence of Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River, East Kalimantan, Indonesia. Z Säugetierk 64:54- 58 Kreb, D. (2002) Density and abundance of the Irrawaddy dolphin, Orcaella brevirostris, in the Mahakam River of East Kalimantan, Indonesia: A comparison of survey techniques. Raffles Bull Zool Supplement 10:85-95 Kreb, D. (in press) Abundance of freshwater Irrawaddy dolphins in the Mahakam in East Kalimantan, Indonesia, based on mark-recapture analysis of photo-identified individuals. J Cetacean Res Manage Kreb, D. & Rahadi, K.D. (in press) Living under an aquatic freeway: effects of boats on Irrawaddy dolphins (Orcaella brevirostris) in a coastal and riverine environment in Indonesia. Aquatic Mammals. Kreb, D. (in press) Conservation management of small core areas: key to survival of a critically endangered population of riverine Irrawaddy dolphins in Borneo. Oryx Smith, B.D., Beasley, I. & Kreb, D. (2003) Alarming news on the status of freshwater populations of Irrawaddy dolphins. Oryx 37(4): 401.

230