EFFECTS OF ANTHROPOGENIC DISTURBANCE ON FOREST BIRDS AND BUTTERFLIES IN THE PHILIPPINES
MARY ROSE CERVANTES POSA (B.SC.)
A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF BIOLOGICAL SCIENCES NATIONAL UNIVERSITY OF SINGAPORE 2007 Acknowledgements
It is my pleasure to be able to acknowledge the people and institutions that have helped me, directly and indirectly, to complete this thesis.
Nothing I did in the field would have been possible without the assistance of the
Subic Bay Metropolitan Authority Ecology Center and the indigenous Aeta
Community of Pastolan Village. I am grateful to Ms. Lilia Alcazar for help with permits and to Chiefs Bonifacio Florentino and Condrado Frenilla for allowing me to conduct research on their land. I am especially indebted to my field assistants Sonny and Samson de la Cruz, Wilges Ignacio, Raul Bautista and Jhoy de la Cruz for their forest know-how, companionship and incredible tree-climbing skills. I also extend my appreciation to Phil Abenoja for various logistical odds-and-ends and coming to hang out in the field; and to Miggy Cervantes for the innovative patch-up job to replace a flimsy camera cable.
Thanks goes to my colleagues in the UP Institute of Biology, especially the junior faculty, for their support; the staff of the Raffles Museum of Biodiversity Research and the Philippine National Museum and Dr. Victor Gapud for help in identifying specimens. Credit goes to Tom Brooks for keeping an optimistic view on the
Philippine biodiversity conservation; and thanks goes to Cagan Sekercioglu for letting me at his great bird database.
Most of my time at NUS has been spent among the denizens of the Conservation
Ecology Laboratory – from the original cohort of graduate students, from whom I
i learned most of what I know about fieldwork and statistics and were there when first manned a mist net and ran a logistic regression, up to the diverse bunch of current inhabitants who make day to day university life brighter – I am grateful to have learned with and learned from them all.
A big thank you goes to all my friends for helping me grow personally, have a life outside of research, and a home away from home: the BioD community of students and staff; the Pinoy mafia — Chico, Arvin and JC — salamat sa pakikisama, mga kuwento, toma, lutong bahay, atbp.; Reuben Clements and Joelle Lai for their personal and professional comradeship; and David Bickford for pushing me to be more active and to “think positive”.
I would like express my deep gratitude to my supervisor, Dr. Navjot Sodhi for letting me join his lab back in 2003. After his initial skepticism of my ability to work alone in the field, he took me on when I kept my poker face. Somehow, I was able to persuade him to take on a research newbie. Without his guidance, critique and grantsmanship, I would not have made it this far.
Lastly, I cannot thank my mom enough for never giving me a curfew, for always coming through for me on field logistics (especially lending me the car!), for her unwavering support, and for trusting me to find my own path in life.
ii Table of Contents
Acknowledgements…………………..…………………..……………………….. i
Summary…………………..…………………..…………………..……………… vi
List of Tables…………………..…………………..……………………………... ix
List of Figures…………………..…………………..…………………..………… x
General Introduction …………………..…………………..……………………... 1
Chapter 1 Overview of biodiversity and conservation in the Philippines
1.1 Biodiversity in the Philippines…………………..……………………………. 3
1.2 Current status and threats…………………..…………………..……………... 4
1.3 Emergence of conservation awareness…………………..……………………. 6
1.4 Effective actions by civil society groups…………………..…………………. 8
1.5 Progress in protected areas and resource management……………………….. 10
1.6 Growth in research and knowledge of species………………………………... 13
1.7 Networking and synthesis…………………..………………………………… 16
1.8 Challenges, priorities and future directions…………………………………... 17
1.9 Conclusions…………………..…………………..…………………..……….. 19
Chapter 2 Effects of land use on forest birds and butterfly communities across a disturbance gradient
2.1 Introduction…………………..…………………..…………………………… 20
2.2 Methodology…………………..…………………..………………………….. 21
2.2.1 Study area…………………..…………………..………………………... 21
2.2.2 Faunal surveys…………………..…………………..…………………… 23
2.2.3. Habitat characterization…………………..…………………………….. 24
iii 2.2.4 Statistical analyses…………………..…………………………………... 25
2.2.5 Analysis of forest bird species response to canopy cover……………….. 27
2.2.6 Analysis of species vulnerability using ecological traits………………... 28
2.3 Results…………………..…………………..………………………………… 29
2.3.1 Community measures for forest species………………………………… 29
2.3.2 Indirect gradient analysis…………………..……………………………. 31
2.3.3 Response of forest birds to canopy cover……………………………….. 32
2.3.4 Ecological traits related to species vulnerability to disturbance………… 32
2.4 Discussion…………………..…………………..…………………………….. 33
2.3.1 Faunal communities in forests…………………..………………………. 33
2.3.2 Faunal communities in modified habitats………………………..……… 34
2.3.4. Ecological traits of vulnerable species………………………………….. 36
2.5 Conclusions…………………..…………………..…………………………… 37
Chapter 3 Effects of land use on predation of nests and caterpillars across a disturbance gradient
3.1 Introduction…………………..…………………..………………………….... 39
3.2 Methodology…………………..…………………..………………………….. 39
3.2.1 Study area…………………..…………………..………………………... 41
3.2.2 Experimental set-ups…………………..…………………..…………….. 41
3.2.3 Predator identification…………………..……………………………….. 42
3.2.4 Statistical analyses…………………..…………………..………………. 43
3.3 Results…………………..…………………..………………………………… 44
3.3.1 Nest predation…………………..…………………..………………….. 44
3.3.2 Caterpillar predation…………………..………………………………… 44
iv 3.3.3 Vegetation variables…………………..…………………………………. 45
3.3.4 Potential predators…………………..…………………………………… 45
3.4 Discussion…………………..…………………..…………………………….. 46
3.4.1 Effects of disturbance on nest predation………………………………… 46
3.4.2 Effects of disturbance on caterpillar predation………………………….. 48
3.5 Conclusions…………………..…………………..…………………………… 49
Chapter 4 Correlates of extinction risk for Philippine avifauna
4.1 Introduction…………………..…………………..…………………………… 50
4.2 Methodology…………………..…………………..………………………….. 51
4.2.1. Response variable…………………..…………………..………………. 52
4.2.2 Clustering variable…………………..…………………………………... 52
4.2.3 Predictors…………………..…………………..………………………... 53
4.2.4 Generalized estimating equations…………………..…………………… 56
4.3 Results…………………..…………………..………………………………… 56
4.3.1 Univariate analyses…………………..…………………..……………… 56
4.3.2 Minimum adequate model…………………..…………………………... 57
4.3.3 Species at risk…………………..……………………………………….. 57
4.4 Discussion…………………..…………………..…………………..………… 57
4.5 Conclusions…………………..…………………..…………………..………. 60
General Conclusions…………………..…………………..……………………… 62
References…………………..…………………..…………………..…………….. 64
Tables……………..…………………..…………………..………………………. 83
Figures……………..…………………..…………………..……………………… 98
v Summary
The Philippines is a top biodiversity “hotspot” owing to its high number of endemics under extreme threat from habitat destruction. A review of biodiversity conservation in the country revealed that progress has been achieved in recent decades towards increasing awareness and developing strategies to sustainably manage resources.
Various sectors of society are taking action to reverse the effects of environmental degradation. However, many challenges remain, one of which is the lack of scientific data. Hence, this study sought to provide empirical information on the effects of forest disturbance on faunal communities and ecological processes and examine if ecological traits confer differential extinction risk.
Field work was conducted in the Subic Bay Watershed Reserve to investigate the effects of different land uses across a disturbance gradient on forest bird and butterfly communities. The two taxa showed dissimilar trends for species richness and population densities across the five habitat types surveyed. The distribution of bird species was related to several habitat characteristics, and over 50% of the forest species observed were significantly affected by canopy cover. Butterfly distribution was not strongly correlated with any of the measured variables. Forest species seemed to be able to tolerate moderate levels of forest disturbance. However, higher levels of disturbance resulted in changes in community composition and decreases in population density, as was most evident in the urban habitat. An analysis of ecological characters indicated that endemicity and traits related to reproduction were important predictors of vulnerability to disturbance for both taxa.
vi The effect of disturbance on reproductive success was assessed by examining patterns of predation on artificial nests and lepidopteran larvae within and among habitats.
Predation levels were significantly higher in rural habitats than in forests for both nests and caterpillars. Nests at 1-1.5 meters were significantly less predated than ground nests. Caterpillar predation did not differ significantly at different heights.
Potential predators were identified through the marks on plasticine models, infrared cameras and live traps. Changes in predator assemblages were observed with disturbance, which may be related to changes in habitat structure affecting visibility and predator diversity.
An analysis to determine possible ecological correlates of extinction risk was made for all resident Philippine avifauna. Single-island endemics, lowland species and habitat and diet specialists were found to be more extinction-prone. This set of traits reflects the impact of habitat destruction on the Philippine fauna as a threat which affects ecologically restricted species that are less able to adapt to rapid and drastic changes.
Increasing levels of disturbance have a negative effect on the Philippine forest fauna, altering community composition, population density and important ecological processes such as predation. Deforestation reduces niche availability, putting habitat specialists and restricted species at greater risk of extinction. More information is needed on the effects of habitat loss and degradation, as results show that taxa have different responses to anthropogenic change. Conservation efforts will benefit from biological knowledge of species and their interactions with their habitats, and
vii knowledge of ecological patterns and processes can form the basis of effective conservation.
viii List of Tables
Table 1. Number of samples, forest species and individuals observed in the five habitat types in the Subic Bay Watershed Reserve...…………………………….84
Table 2. Nonparametric species richness estimators and curve models for asymptotic species richness for the five habitat types in the Subic Bay Watershed Reserve…………………………………………………………………………...85
Table 3. Mean abundance of forest bird and butterfly species detected in the five habitat types in the Subic Bay Watershed Reserve………………………………86
Table 4. Variance in bird and butterfly community composition represented by the final 3 ordination axes in nonmetric multidimensional scaling analysis and Spearman correlation coefficients for the most strongly correlated habitat variables…………………………….……………………………………………88
Table 5. Results of pairwise comparisons between habitats using multi-response permutation procedures………………………………….……………………….89
Table 6. Parameter estimates from univariate general estimating equations on ecological traits used to predict species vulnerability to disturbance……………90
Table 7. QICu values of candidate models for species vulnerability using significant ecological traits as predictors…..……………………………….………………..91
Table 8. Numbers of artificial nests and caterpillar models predated at different habitats and height categories. ……………………………….……………….…92
Table 9. Single-fixed effect models of probability of nest and caterpillar predation with habitat and height as predictors and the inclusion of transect and plot as nested clustering variables to control for spatial autocorrelation………………..93
Table 10. Minimal adequate model of nest predation probability.....………………..94
Table 11. Parameter estimates from univariate generalized estimating equations using traits to predict extinction risk for resident Philippine bird species with family as included as a clustering variable..……………………………….……………….95
Table 12. Minimum adequate model of extinction risk in Philippine birds using significant ecological traits as predictors and family as the clustering variable…………………………………………………………………………..96
Table 13. Resident Philippine birds that possess traits identified as correlates of extinction risk and are not currently listed as threatened……………………...…97
ix List of Figures
Figure 1. Map of the Philippine archipelago showing approximate percentages and distribution of forest cover (including secondary forest and plantations) remaining on the major islands...... ….…………………………………………..99
Figure 2. Numbers and stability of Philippine bird species considered threatened in four global conservation assessments for the IUCN Red List………………….100
Figure 3. Number of publications on Philippine biodiversity and conservation obtained from searching three ISI Web of Knowledge databases……………...101
Figure 4. Growth in attendance at the annual symposium on biodiversity by the Wildlife Conservation Society of the Philippines……………………………....102
Figure 5. Map of the study area showing the five habitat types surveyed in the Subic Bay Watershed Reserve…………………...……………………………………103
Figure 6. Species accumulation curves for forest birds in the five habitat types…..104
Figure 7. Species accumulation curves for forest butterflies in the five habitat types…………………………………………………………………………….105
Figure 8. Rarefaction curves of forest bird species richness in the five habitat types………………………………………………………………………...…..106
Figure 9. Rarefaction curves of forest butterfly species richness in the five habitat types…..………………………………………………………………………...107
Figure 10. Population densities of forest birds in the five habitat types……………108
Figure 11. Population densities of forest butterflies in the five habitat types ……..109
Figure 12. Nonmetric multidimensional scaling ordination joint biplot of sample scores for the entire bird community with an overlay of strongly correlated habitat variables………………………………………………………………………...110
Figure 13. Nonmetric multidimensional scaling ordination joint biplot of species scores for the entire bird community with an overlay of strongly correlated habitat variables…….…………………………………………………………………..111
Figure 14. Nonmetric multidimensional scaling ordination of sample scores for the entire butterfly community.……………………………….…………………….112
Figure 15. Nonmetric multidimensional scaling ordination of species scores for the entire butterfly community.……………………………….…………………….113
Figure 16. Results of simulations showing number of forest bird species present versus amount of canopy cover ……….………………………………………..114
x Figure 17. Proportion of threatened and nonthreatened resident Philippine bird species with ecological traits that were significant correlates of extinction risk………..115
xi General Introduction
The scale of human enterprise now affects the structure and function of all of the earth’s ecosystems (Vitousek et al. 1997). Maintaining the planet’s ecological integrity, which is vital for human well-being, is unquestionably one of the primary challenges that must be met in the coming century. Of paramount concern is the loss and degradation of tropical habitats that threaten numerous species with extinction
(Brook et al. 2003, Sodhi and Brook 2006). In South-East Asia, the impact of these anthropogenic activities on biota is anticipated to be catastrophic (Sodhi et al. 2004a), as the region has a high concentration of endemic species that continue to be under great pressure from high rates of deforestation (Myers et al. 2000, Achard et al.
2002). The Philippine archipelago epitomizes the dire biodiversity situation in the region. The combination of high endemism in many floral and faunal groups coupled with extensive and rapid habitat loss makes the country a particularly critical global priority – a top conservation “hotspot” for both terrestrial and marine ecosystems
(Myers et al. 2000, Roberts et al. 2002). As the impacts of human action on the environment become clearer and more people begin to become aware of the value of biodiversity, conservation efforts have increased in recent decades, but still face immense challenges.
However, effective conservation is hampered by poor understanding of species biology and knowledge of how complex ecological processes are affected by disturbance (Sodhi and Liow 2000). Even for highly visible and charismatic taxa such as birds and butterflies, detailed biological information for many Philippine species is incomplete and ecological studies are scarce (Settele 1993; Kennedy et al. 2000).
1 Given the massive amount of deforestation in the Philippines, research on the effects of habitat loss and degradation on the native fauna is urgently needed.
This study aims to review the state of conservation in the Philippines and contribute to the ecological knowledge on Philippine bird and butterfly fauna to provide information that can form the basis of conservation strategies. Chapter 1 gives an overview of the Philippines as an area of high biological endemism, where human action has caused considerable environmental devastation. It chronicles recent positive progress by various sectors and discusses key priorities and challenges to conservation in the country’s context. One of these obstacles is the lack of ecological studies that can provide scientific data for conservation. Field work was conducted to obtain empirical data on the effects of anthropogenic disturbance on faunal communities and ecological processes. In Chapter 2, the impacts of various levels of disturbance on forest bird and butterfly communities are assessed by comparing measures of species richness and population densities across five habitat types.
Further, it is determined whether habitat variables are related to species distribution and if ecological traits contribute to species vulnerability to forest disturbance.
Delving deeper into how disturbance may affect important ecological processes that maintain biodiversity, Chapter 3 looks into the possible effects on faunal reproductive success. Through the use of artificial nest and caterpillar models. levels of predation and predator assemblages are compared across different habitat and height locations.
Finally, in Chapter 4, data on species biology are utilized to model the extinction proneness of resident Philippine birds. Identifying such ecological correlates of extinction risk can help pinpoint species that may be in critical need of conservation action.
2 Chapter 1 Overview of biodiversity and conservation in the Philippines
1.1 Biodiversity in the Philippines
The Philippines is known as one of the most biologically rich regions in the world, with exceptionally high levels of endemism among its flora and fauna. Situated at the interface of the Indomalayan and Australasian biogeographic regions, the country has a complex geological history that is inextricably linked to its biodiversity. The archipelago, composed of more than 7,100 islands of Sundaland and oceanic origins, is segregated into distinct biogeographic regions concordant with the configuration of the Philippines during the great ice ages of the Pleistocene. Present-day islands were once joined by land bridges that were exposed when seas fell up to 120 m below current levels, only to be isolated again as the ice melted (Heaney 1986). Today, each of the ice-age island amalgamations contains a unique set of biota, and researchers have identified several centers of biodiversity and endemism. Knowledge of these geological processes has become an essential key to understanding the distribution of life in the Philippines (Heaney and Regalado 1998, Brown et al. 2001, Ong et al.
2002).
With a land area of 300,780 km2, the level of diversity in the Philippines is considered to be remarkable, taking into consideration its size (Heaney and Regalado 1998).
Nearly half of the approximately 1,100 terrestrial vertebrates known from the
Philippines are unique to the islands, with endemism in certain groups ranging from
70–90%, while estimates of richness of vascular plants range from 9,000–12,000 species, with 45–60% endemism (WCSP 1997, Ong et al. 2002). More recently, studies have showed that the archipelago is also the epicenter of marine shore fish
3 diversity (Carpenter and Singer 2005) and is one of the richest locations for corals, reef fish, marine snails and lobsters (Roberts et al. 2002). Biological exploration of the archipelago is still incomplete, and surveys continue to discover new species.
Between 1990 and 2005, the new vertebrate taxa documented include ten species of forest rodents (including an arboreal, herbivorous giant cloud rat), seven birds (recent finds are a forest woodcock, a flightless rail, and a single island endemic parrot), 20 forest frogs, 11 snakes, and 11 lizards (including the world’s second known frugivorous monitor lizard). Botanical novelties include the discovery of three new species of Rafflesia, increasing the number of known Philippine species from two to five (Barcelona et al. 2006). Ongoing revisions to taxonomy and systematic reassessments through the use of molecular techniques is expected to further elucidate the relationships of many species groups and uncover more new species. Thus, the true measure of biological wealth in the Philippines is yet to be known.
1.2 Current status and threats
Unfortunately, hand-in-hand with the knowledge that the Philippines is home to a unique and megadiverse biota, it is also apparent that its flora and fauna are among the most threatened in the world. The archipelago, which was historically almost completely covered in forest, has undergone extensive deforestation. From the time of Spanish colonization in the 1500’s when 90% of the land was forested, forest cover was reduced to approximately 70% by the 1900s (Liu et al. 1993). The bulk of deforestation then occurred in the post-World War II era when large-scale logging technologies were utilized, and the national economy was buoyed by the export of logs overseas (Kummer 1992). The rapid deforestation by the timber industry that occurred throughout much of the 20th century has reduced primary forest cover to
4 less than 3% (FAO 2005). Although secondary forest and other wooded land cover around 20% of the land (Figure 1), recent estimates of rates of forest loss continue to be high, reported at 1.4–2.1% annually (FAO 2000, 2005; WRI 2003). Mangroves have declined from half a million hectares to around 12,000 ha from 1918 to 1994 due to exploitation and conversion to fishponds (Primavera 2000). Further, the archipelago’s extensive coral reefs are jeopardized by harmful fishing practices and siltation, with only 5% of reefs considered to be in excellent condition (Gomez et al.
1994).
With the reduction and degradation of available habitats, many of the flora and fauna are now at risk of extinction. Of 1,007 Philippine vertebrates that have been assessed for the IUCN Red List, 20.7% are classified as threatened, as are 215 of the 323 plants that have been evaluated (IUCN 2006). The combination of high endemism in many floral and faunal groups coupled with extensive and rapid habitat loss makes the
Philippines a top “hotspot” for both terrestrial and marine ecosystems (Myers et al.
2000, Roberts et al. 2002). The advanced state of environmental degradation has had serious repercussions for the human population as well. Logging on hillsides has led to flooding and erosion, with landslides claiming many human lives (Vitug 1993).
Loss of soil fertility, pollution from large-scale mining operations, destruction of coral reefs and mangroves, and other such problems now affect the livelihoods of millions of rural people who are dependent on the land (Broad and Cavanagh 1993).
The country also faces many other impediments to conservation, not least of which are the socioeconomic problems prevalent in a developing country. Conservation efforts are hampered by corruption, weak government leadership and opposition by powerful vested interests that make it difficult to change and implement sound
5 environmental policies (Vitug 1993). The remaining natural resources continue to be under pressure from a large, fast-growing and mostly impoverished population (78.6 million in 2002 with a 2.3% growth rate per annum; WRI 2003), and national funds are constrained by external debt servicing and thus rarely diverted into protecting biodiversity (Myers 1988).
In the face of this dire situation, many groups and individuals are working towards striking a balance between human needs and preserving the country’s biological heritage. This chapter chronicles recent positive progress by various sectors of
Philippine society and presents key priorities and challenges to conservation in the country’s context.
1.3 Emergence of conservation awareness
The rise of conservation and environmental activism came at a time of social upheaval in the Philippines. During the dictatorship of Ferdinand Marcos in the
1970’s, deforestation peaked as the government issued cutting licenses to timber companies based on political patronage (Vitug 1993). By then, the unsustainable and inequitable use of natural resources to benefit the elite had severely devastated the landscape, while marginalizing the poor. When the regime was overthrown through the “People Power” revolution in 1986, a revived democracy fostered the emergence of scores of civil society groups, including those with environmental agendas. Non- government organizations (NGOs), grassroots organizations, and indigenous peoples groups thus began to be involved in attempts to reverse decades of environmental plunder (Broad and Cavanagh 1993). Today, the involvement of civil society in the planning, development, and implementation of government policies and programs
6 pertaining to environmental issues has become a salient feature of biodiversity conservation efforts in the Philippines. Environmental groups (numbering over 5,000) now represent a formidable counterforce to the political elite and upper class blocs who wield power to influence political decisions on sensitive issues (Broad and
Cavanagh 1993, Malayang 2000).
The impact of these events on policy-making is evident in the shift from predominantly government-centered to people-oriented policies. Legislations enacted in the 1990s saw an increase in the role of civil society in poverty alleviation, sustainable development, preservation of indigenous peoples rights, and environmental protection. These legislations provide a framework where the participation of local stakeholders in resource management can be best manifested
(Malayang 2000). For instance, both the Local Government Code and the Fisheries
Code provide for the devolution of management and authority of natural resources from the central government to regional, municipal, and community levels.
Government-led efforts to address deforestation have focused on social forestry and alternative livelihood. One major initiative is the community-based forest management program (CBFM) that was developed as a strategy to achieve ecological stability and social equity (Lasco and Pulhin 2006). By awarding tenure and the right to utilize forest resources to organized communities, stakeholders are given security and an incentive to plant trees and defend forestland against illegal logging (Johnson
1999). CBFM activities vary greatly across the country, ranging from protection of biodiversity, regeneration schemes, agroforestry, and plantations for timber. Some of these schemes appear to be approaching sustainability, such as work with
7 communities in the Alcoy reforestation program on Cebu and the Landcare movement on Mindanao (Lasco and Pulhin 2006).
Other laws have been passed to ensure protection of wildlife and areas of biological significance, such as the 2001 Wildlife Conservation and Protection Act. At the international level, the growing valuation of biodiversity in the Philippines is evidenced by the country becoming a signatory to the 1992 Convention on
Biodiversity, and to other agreements such the Convention on International Trade in
Endangerd Species of Wild Flora and Fauna (CITES) and the Ramsar Convention on
Wetlands.
1.4 Effective actions by civil society groups
Aside from raising awareness for conservation and ensuring equitable resource use, efforts by civil society groups also have direct impacts on the conservation of biodiversity. In some communities, a conservation ethic has arisen spontaneously, as seen in Bais Bay, Bohol and Banacon Island, where local residents independently reforested mangroves to stabilize coastal areas and for their own sustained use
(Walters 2003). Grassroots organizations and NGOs are also leading programs that protect threatened species. One such effort that has achieved remarkable success is the
Philippine Cockatoo Conservation Program. Considered as one of the most threatened birds in the world (Collar et al. 1999), the endemic Philippine cockatoo, Cacatua haematuropygia, was historically known from 45 islands but is now extirpated or rare throughout much of its range due to loss of habitat and poaching for the pet trade. An integrated conservation program for the species was initiated in Palawan in the early
1990s, led by government agencies, academic institutions, and a locally-based NGO.
8 Key strategies that were implemented included awareness and education campaigns, nest protection, monitoring, captive breeding, and ecological research. Former poachers were recruited and trained as wardens and the export of birds was restricted, leading to a decline in available birds for the illegal wildlife trade (Boussekey 2000,
Widmann et al. 2006). The local government endorsed the creation of the Rasa Island
Wildlife Sanctuary to protect and manage a resident cockatoo population. Since then, there have been clear signs of recovery and plans are underway to apply these experiences to other areas that still harbor cockatoo populations (Widmann et al.
2006).
Another success story is the case of the Critically Endangered Philippine crocodile,
Crocodylus mindorensis, which is regarded as the most threatened crocodilian in the world. The discovery of a small breeding population in Luzon’s Sierra Madre mountain range in 1999 led to a conservation program involving local communities, government agencies and academic institutions. The local government banned the killing of crocodiles and established a crocodile sanctuary. To date, it is the only in situ program for conserving the species, as all past efforts have focused on captive breeding (van der Ploeg and van Weerd 2004). Similar successful initiatives are unfolding in many other regions with highly threatened unique biodiversity; examples include the Polillo Ecology Stewardship Project (http://mampam.50megs.com/ polillo/), the Cebu Biodiversity Conservation Foundation (http://www.fauna- flora.org/asia_pacific/cebu.html), Negros Rainforest Conservation Project
(http://www.coralcay.org/expeditions/forest/ph2/ overview.php), Philippine Endemic
Species Conservation Project (http://pescp.org/index. html), and the Silliman
9 University-Angelo King Center for Research and Environmental Management
(http://su.edu.ph/suakcrem/main.htm).
1.5 Progress in protected areas and resource management
Protected areas that are dedicated to protecting and maintaining biological diversity are vital to conservation, particularly where pristine habitat is becoming increasingly scarce. The growing awareness of the need to conserve Philippine biodiversity precipitated a restructuring of the country’s existing protected areas through the enactment of the National Integrated Protected Areas System (NIPAS) Act in 1992.
The NIPAS Act replaced the antiquated national parks law of 1932, which generally ignored the protection of species and habitats. The NIPAS incorporates scientific, cultural, and socio-economic dimensions in its framework to assess the biodiversity value of existing national parks and establish new protected areas, both marine and terrestrial. It also includes a mechanism in which stakeholder participation is guaranteed through representation in site-specific Protected Area Management Boards
(PAMBs). Protected areas declared under NIPAS are guaranteed by the Constitution to remain as such and there are currently over 300 parks in various categories that are under evaluation for inclusion in system (DENR 2003). Of these, 160 (roughly 8% of
Philippine land area) fall under the IUCN categories I–V for terrestrial protected areas
(WRI 2003). The realization of an integrated parks system is crucial to conservation in the Philippines, and because of the dependence of many endemic fauna on forest habitats, the establishment of strict forest reserves remains imperative.
However, while the NIPAS Law and its policy framework are progressive, their actual implementation is convoluted and problematic. The implementing government
10 agencies are often strapped for funding, resources, and technical capability. Political maneuverings by interest groups and bureaucratic red-tape create conflicts in the management of areas. Complications also inevitably arise because protected areas are rarely free of human inhabitants, and with numerous stakeholders dependent on limited natural resources, effective management is more than a problem of simple environmental education or law enforcement (Custodio and Molinyawe 2001, White et al. 2002). Consequently, these factors, plus the archipelagic nature of the country, have engendered strategies that favor the decentralization of protected area management to local authorities and communities.
In many cases, the active involvement of local communities in conservation programs holds the key to their success (Malayang 2000, Sodhi et al. 2006). In the Philippines, this has been especially evident in coastal management programs that have achieved the combined goals of the protection of near-shore habitats for biodiversity, as well as increasing fishery yields and enabling locals to utilize resources in a sustainable manner (Russ and Alcala 1999). Celebrated examples include marine protected areas
(MPAs) on Apo Island, Balicasag, Pamiclan, San Salvador, and Mabini (White and
Vogt 2000). The strong involvement of stakeholder communities in the enforcement of protected area regulations, which builds the confidence of people to manage their own resources, was crucial to the success of these MPAs (Russ and Alcala 1999,
White et al. 2002).
Ideally, programs should be balanced multi-sectoral collaborations, combining community participation with environmental education, economic incentives, and legal mandates in a manner suited for a particular area, together with long-term
11 institutional support from the government, NGOs, academe, and other institutions
(White et al. 2002). One success story that was a result of fruitful networking is the
Tubbataha Reef National Marine Park, a 332 km2 reef complex in the Sulu Sea and a
UNESCO World Heritage Site. The unique characteristics of the park—its remote marine location, lack of inhabitants, tourism potential, and a stakeholder community composed of local and international fishing groups—required a management scheme with a high level, dedicated collaboration between government, non-government, and private sectors. Today, the park is among the few protected areas with a stable source of funds through tourism. Management and protection measures (such as a ban on destructive fishing practices) have greatly improved living coral substrate cover
(White et al. 2000, 2002). The successes in protecting marine areas indicate that multi-sectoral collaborations can succeed and communities can effectively manage marine resources.
Few examples of effective protection and restoration of forests can be found, perhaps because this resource has long been under the control of government and commercial interest groups. Moreover, there is a misguided but long-standing view that wilderness habitats are wastelands and exploitable commodities, instead of vital ecosystems, which leads to their destruction (Primavera 2000). However, policies stemming from the land reform movement that support the devolution of forest resource management are now in place and echo strategies that are being practiced in coastal resource management. Programs such as the CBFM are largely seen to have positive ecological effects and have helped prevent further degradation of forests
(Johnson 1999, Lasco and Pulhin 2006). They are, however, vulnerable to misuse and
12 abuse (Vitug 1993, Malayang 2000) and there is a critical need to evaluate the effectiveness of these reforestation programs for biodiversity conservation.
1.6 Growth in research and knowledge of species
The heightened environmental consciousness in civil society has been paralleled in the academe by a renewed interest in biodiversity research. Field surveys in uncharted and incompletely explored regions of the country have yielded an astonishing haul of species new to science. In addition, an unexpected positive result from this increased field work has been the rediscovery of species that were feared to have become extinct. As early as 1907, ornithologists noted that Cebu, one of the first islands to be settled by European colonizers, had already lost most of its original forest cover
(MacGregor 1907). In 1956, a paper by D. Rabor reported the disappearance of the
Cebu flowerpecker, Dicaeum quadricolor, as well as eight other bird taxa endemic to the island. As D. quadricolor had not been recorded since 1906, it was long considered to be extinct until its rediscovery in 1992 in a small patch of limestone forest (Dutson et al. 1993, Magsalay et al. 1995). Subsequent surveys revealed its presence in two other localities (Collar et al. 1999). More recently, active field surveys uncovered populations of the Philippine bare-backed fruit bat, Dobsonia chapmani, one of the first mammals to be declared extinct in the Philippines. This cave-dwelling fruit bat had not been recorded from its range since 1964 despite intensive searches. But in 2001, three bats were netted in an agricultural clearing on
Cebu (Paguntalan et al. 2004) and two years later, another five were found on nearby
Negros Island from degraded karst habitat (Alcala et al. 2004). Similarly, the
Philippine forest turtle, Siebenrockiella leytensis, was variedly considered for over 80 years to be either rare, on the brink of extinction, or extinct from the island of Leyte.
13 A survey uncovered natural populations on Palawan—an apparent case of workers searching for the species on the wrong island. Along with this turtle, several endemic frogs and reptiles that have not been seen for 15 to 60 years were also rediscovered
(Diesmos et al. 2004, 2005).
As the amount and quality of biodiversity information increases, it is also becoming apparent that a number of species are less extinction-prone than feared. Some are more abundant and widespread than previously thought (e.g., Cryptic flycatcher,
Ficedula crypta, Mindanao gymnure, Podogymnura truei) or maintain good populations even in disturbed habitats (e.g., Philippine tarsier, Tarsius syrichta,
Philippine flying lemur, Cyanocephalus volans) (WCSP 1997, Collar et al. 1999).
Robust data for birds, however, show no consistent pattern to these increases in knowledge of conservation status (Figure 2). The first conservation status assessment of the world’s birds, Collar and Andrew (1988) listed 43 Philippine species as threatened. The second, Collar et al. (1994) listed 86, of which 26 were downlisted from threatened status by the third, BirdLife International (2000). Most of these changes involved new information; only two relate to genuine changes in status
(Butchart et al. 2004)—the increasing threat to Blue-winged racquet-tail, Prioniturus verticalis in the early 1990s and to Philippine duck, Anas luzonica in the late 1990s.
Since then, knowledge of the conservation status of Philippine birds has stabilized, with 69 species considered threatened in the most recent assessment (BirdLife
International 2006, IUCN 2006).
The information gathered from biodiversity studies has drawn attention to previously overlooked biodiversity-rich areas and has led to their inclusion in the protected areas
14 system. Moreover, basic biodiversity information provided by Filipino scientists and their international counterparts has been critical in identifying priority areas and helped fine-tune the national strategy and action plan for biodiversity conservation
(Mallari et al. 2001, Heaney 2001, Ong et al. 2002). Genuine attempts to synthesize available scientific information and arrive at a consensus concerning the status of
Philippine biodiversity has culminated in important publications, such as the
Philippine Red Data Book, Threatened Birds of the Philippines, Key Conservation
Sites of the Philippines, and Philippine Biodiversity Conservation Priorities. The emergence of studies in areas such as biogeography, conservation ecology, resource management, and phylogenetics has greatly contributed to the understanding of diversity in the Philippines. A search of three ISI Web of Knowledge databases
(Biosis Previews, Web of Science, and Zoological Records) for the period 1985 to present reveals an increasing number of publications dealing on biodiversity and conservation (Figure 3).
Numerous scientists, field workers, and students of biodiversity continue to expand the knowledge on Philippine biota, but work is far from complete. A lesson that can be gleaned from these accounts is that there is still much to be learned about
Philippine biodiversity and underscore the importance of gathering empirical data from the field. For instance, the uncritical acceptance of a species’ “extinction” can lead to researchers giving up on them prematurely and, thus, the assumption of their demise may become self-fulfilling (Collar 1998). Urgent targets for potential rediscovery include other “lost” species of birds, amphibians, reptiles, mammals, and vascular plants (Ong et al. 2002, Butchart et al. 2005). Rigorous field surveys are needed to ascertain their status. The state of deforestation in the Philippines means
15 that these species, with their typically small populations, are far from danger of extinction. However, the discovery of these resilient populations has revived research and conservation efforts and offers renewed hope for their survival. Also, the apparent ecological flexibility of some species, rare endemics included, shows that the search should be expanded to include disturbed habitats.
1.7 Networking and synthesis
There are now numerous professional groups that are actively promoting conservation education, research, and advocacy work in the Philippines. Among the pioneers is the
Haribon Foundation, which started out as a bird-watching club in 1972 and now is one of the most active environmental organizations in the country. The Wildlife
Conservation Society of the Philippines is one of the fastest growing societies with diverse membership from the academe, government, NGOs, and people’s organizations (WCSP 1997). Participation in its yearly biodiversity symposium has grown steadily in attendance and membership (Figure 4). Similar progress along this line are being felt by other groups, like the Philippine Association of Marine Science, which holds a well-attended forum on marine biology.
Experience has also been gained through partnerships and alliance-building with international groups and institutions. Prominent species, such as the Philippine eagle,
Pithecophaga jefferyi, have benefited from increased attention and research activities brought about by such collaborations. Active research in the past decade has also amassed critical information on its biology and ecology (e.g., Miranda et al. 2000).
Recent treatises suggest that the species may likely to have a higher population than past estimates, and confirmed recent records from new localities indicate a wider
16 distribution (Collar et al. 1999, BirdLife International 2000). However, known populations remain highly fragmented and are severely threatened from continuing habitat loss and poaching (Bueser et al. 2003). Continued fieldwork and in situ conservation efforts involve an alliance of major local and international conservation groups and government agencies (PEFI 2005).
1.8 Challenges, priorities and future directions
It can be seen that in the past few decades, various sectors of society have responded to the urgent need of conserving the threatened biodiversity of the Philippines. The diverse strategies that have emerged to address the country’s multitude of environmental problems represent progress toward reversing the tide of environmental degradation. However, many challenges remain before the deleterious consequences of past unsustainable practices can be corrected. While the local efforts discussed above are significant developments, they address only part of the larger issues affecting the environment on a national scale. Some of the most pressing problems include finding ways to improve public education, control rampant pollution and the runaway population growth, and change the general lack of political will to pursue meaningful social reforms that favor biodiversity conservation.
The devolution of resource management and the involvement of local communities in conservation initiatives have resulted in promising outcomes for species and habitat protection, as well as sustainable resource use. However, the effectiveness of these local conservation efforts will depend to a large degree on the adequacy of knowledge and capability of communities (Magno 2001). Stakeholders must be further empowered to plan, implement, and monitor their own programs and become
17 financially stable. Continual monitoring and assessment of local programs should be undertaken to ensure that conservation goals are being met. There is a need to invest more in enforcement and implementation of national environmental laws and impose strict sanctions for those found in violation. The importance of curbing ongoing illegal logging and the full realization of an integrated protected parks system cannot be over-emphasized. It will be necessary to pursue alternative and stable sources of funding mechanisms for conservation, since the recurrent lack of funding inhibits the value of many of the larger protected areas.
The recent progress in cataloging the country’s biodiversity is encouraging, but more research is still needed. Basic biological information for many species is still lacking and many areas remain poorly surveyed. There is a need to closely integrate science into conservation efforts through greater involvement of scientists in designing and monitoring programs. As in the rest of South-East Asia, there is a paucity of ecological research in the Philippines, especially on how flora and fauna are affected by habitat disturbance and loss (Sodhi and Liow 2000). Another relevant line of inquiry is assessing the value of degraded areas for species survival and conservation, as well as further exploring the rehabilitation and restoration of such habitats. A greater understanding of the ecological and evolutionary processes that control and maintain biodiversity would help to form the basis of effective conservation.
Fostering collaboration with international organizations and institutions of learning can enrich the capability of local scientists and field workers to conduct research, and should be pursued. Finally, better documentation of research results is essential, as there is a great amount of critical unpublished data, and available information is poorly distributed to the wider community (Brown et al. 2001, Heaney 2001).
18 1.9 Conclusions
The Philippine experience holds valuable lessons on the pitfalls of progress at the expense of the environment, as well as lessons about creating alternative solutions for sustainable and equitable development. With so many ecosystems being pushed to their ecological limit, people have been spurred into action with the realization that human welfare—and their very lives—rests on the state of the environment. The struggle to regain ecological balance now underway in the Philippines serves as an example to other developing tropical countries that face similar challenges today.
The future of the Philippines’ unique biodiversity will hinge on the measures that are currently being taken to prevent the loss of species and habitats. The scale of the tasks is immense and the challenges are multifarious. Effective conservation will require equally enormous amounts of vision, hard work, and dedication to overcome the existing obstacles. What has happened, and is still happening, in the Philippines has broad relevance to many tropical countries. That significant progress has been attained in the Philippines—which some believe to be a “worst case scenario” of ecological ruin in South-East Asia to be written off the conservation agenda (e.g.,
Terborgh 1999)—suggests that grounds for cautious optimism exist for biodiversity conservation in the country.
19 Chapter 2 Effects of land use on forest bird and butterfly communities across a disturbance gradient1
2.1 Introduction
Given the paucity of ecological research from South-East Asia and alarming rates of habitat loss experienced by the region, one of the most important lines of research inquiry is determining the effects of habitat loss and disturbance on forest fauna
(Sodhi and Liow 2000, Sodhi and Brook 2006). Deforested areas are often converted into other land uses such as pasturelands and agricultural plantations, or are developed into cities. Knowledge of which types of disturbance most adversely affect tropical biota and which taxonomic groups are most susceptible to disturbance is generally poor (Dunn 2004). Current knowledge of the impact of anthropogenic activities on forest fauna in the Philippines is very limited. The aim of this study was to provide empirical ecological information on how different land uses affect birds and butterflies in the lowland forest of Subic Bay Watershed Reserve in the Philippines.
Birds and butterflies are well-known indicator taxa because of their sensitivity to environmental perturbations, relevance to ecosystem functioning (e.g., in pollination and seed dispersal), and relative ease in sampling (Brown 1991, Furness et al. 1993,
Blair 1999, Hamann and Curio 1999). In the Philippines, a large proportion of endemics from both groups are dependent on forested habitats (Dickinson et al. 1991,
Settele 1993). Deforestation and forest disturbance are known to have several negative effects on both taxa, which include declines in diversity and abundance
(Hamer et al. 1997, Ghazoul 2002), changes in species assemblages (Johns 1991,
1 Published in Biological Conservation 129:265-270 (2006). 20 Hamer et al. 2003), loss of species guilds (Canaday 1997) and extinction (Magsalay et al. 1995, Castelleta et al. 2000, Brook et al. 2003). However, modified habitats may actually retain some “forest” biodiversity (Hughes et al. 2002, Horner-Devine et al.
2003, Sodhi et al. 2005), but the conservation value of these areas still needs to be assessed.
Most studies on the effects of habitat disturbance on animals have focused on a single taxon and disturbance type (Dunn 2004). By sampling two relatively well-known taxa across a range of land uses, the effects of deforestation and disturbance on
Philippine fauna may be further elucidated. This study aims to provide empirical data on the effects of habitat disturbance on forest birds and butterflies in the Subic Bay
Watershed Reserve and adjacent areas by: (1) comparing species richness and population density among forests and sites with different land uses; (2) determining whether habitat variables (e.g. temperature, humidity, tree density) influence species distribution; and (3) examining if certain ecological traits (e.g., endemicity, body size) influence species vulnerability to disturbance.
2.2 Methodology
2.2.1 Study area
Fieldwork was conducted within the Subic Bay Watershed Reserve and the adjacent
Olongapo City (Figure 5) in west central Luzon, the Philippines. The 9,856-ha reserve lies between 14° 45.0’ to 14° 51.0’ N and 120° 15.5’ to 120° 15.0’ E. Climate is characterized by two distinct seasons, dry from November to April with majority of the rainfall occurring during the wet months of June to September (Coronas 1920).
The reserve contains one of the few semi-evergreen lowland dipterocarp forests
21 remaining in the country and is contiguous with the much larger Bataan Natural Park to the southeast (Mallari et al. 2001). Until 1992, it was part of a US Naval
Reservation and strict security measures were able to prevent much of the natural vegetation from being converted into agricultural land. Military infrastructure was built into parts of the forest and other areas were disturbed by army activities, or were selectively logged. Indigenous Aetas living within the reserve boundaries raise crops in the rural areas and still practice traditional hunting and extraction of non-timber forest products until today. Commercial and industrial activities are concentrated around the waterfront and extend towards nearby Olongapo City.
Five habitat types were chosen a priori based on the definitions of land use classification in the Philippines, as surveyed in the Subic Bay Watershed Reserve in
2000 (Woodward-Clyde 2001) (Figure 5). The habitat types (and approximate areas within the reserve) were defined as follows: (1) closed canopy forest consisting of natural forest where mature dipterocarps or other broadleaf trees cover 40% or more of the area (4,342 ha); (2) open canopy forest that is natural forest with a discontinuous tree layer and coverage of 10–40% (3363 ha); (3) Suburban areas with low density housing developments of 1 to 2-storey detached and semi-detached houses (273 ha); (4) rural areas consisting of mixed-use areas consisting of grasslands, regenerating scrub, small-scale agriculture and reforestation plots (621 ha); and (5) urban areas with commercial and industrial centers with high building and population densities (900 ha).
22 2.2.2 Faunal surveys
Sampling was conducted during the periods of 11 to 25 July, 15 to 25 September and
2 to 13 October in 2003 and 16 to 24 January, 6 to 13 February, 24 to 30 April and 11 to 19 May in 2004. All five habitat types were visited during each sampling period.
Surveys were conducted from 0600 to 1000 h for birds and 1000 to 1400 h for butterflies in good weather (i.e., no heavy rain or strong winds). Both forest and non- forest species were included in the sampling. Forest species were defined as those with ‘forest’, ‘forest edge’ or ‘woodlands’ listed as habitats according to Kennedy et al. (2000) for birds, and Igarashi and Fukuda (1997, 2000) for butterflies. Sampling was conducted by a single observer to reduce observer bias.
To determine bird species richness and abundance, habitat types were surveyed using the point count method, which is a preferred sampling method for assessing birds dense woodland habitats (Bibby et al. 2000). This technique involved identifying all individuals seen or heard within a 25-m radius from a fixed point for a duration of 10 minutes, excluding those that flew over the canopy (e.g. swiftlets and raptors in flight). A ‘rest period’ of 1–2 minutes was allowed to pass after arriving at each point before recording began to allow bird activity to resume. Unidentified birds (<1% of all records) were not included in the analyses. In forests, points were randomly placed to the side of established trails or along newly cleared footpaths. Successive points were spaced at least 200 m apart for independence.
A modification of the line transect walk (Pollard and Yates 1993) was used to determine butterfly richness and abundance. This is a suitable method for surveying butterflies in a wide range of habitats, including tropical rainforests (Walpole and
23 Sheldon 1999, Caldas and Robbins 2003, Koh and Sodhi 2004). In this method, a
100-m transect was slowly traversed at a uniform pace for 10 minutes and all individuals of the families Papilionidae, Pieridae and Nymphalidae that came within an imaginary box, 5 m to either side, above, and in front of the observer were recorded. As in other similar studies, the Lycaenidae and Riodinidae were not included because of the difficulty in identifying them while in flight (Hamer et al.
1997, Ghazoul 2002, Koh et al. 2002). A few congeneric species that could not reliably be identified on the wing were combined into one genus for analyses (e.g.
Ypthima sempera and Y. stellera). Unidentified butterflies (approximately 1% of all records) were not included in analyses. Transects were at least 100 m apart and transects in the forest were located along the same trails as the point counts.
The total number of points and transects are given in Table 1. Different points and transects were used for subsequent counts and no area/trail was visited more than twice. The entry points into the forested areas were chosen to cover as large an area of the study site as possible.
2.2.3 Habitat characterization
Nineteen habitat variables were measured to characterize each point or transect. Light intensity (Topcon IM-2D digital illuminance meter), temperature and relative humidity (Digital min/max thermohygrometer from Forestry Suppliers, Inc.) were recorded before beginning each survey. The circular sample-plot method (James and
Shugart 1970) was used to collect vegetation data. The following variables were measured within a circular plot with a 5-m radius: canopy density (taken with a circular densiometer), percent of shrub and herb cover (visually estimated), diameter
24 at breast height and height to inversion (site of first major branch; Torquebiau 1986) of the nearest 10 trees, number of cultivated and dead trees, number of fruiting and flowering plants, average litter depth and ground cover (estimated in a 0.5-m2 grid at four points in the plot) and vertical vegetation volume (total number of hits on a graduated pole up to 6 m from four points in the plot). The extent of human modification and disturbance was also estimated by measuring the amount of pavement cover (within 5 m), the number of pedestrians and vehicles passing (within
25 m for 5 min), and number of buildings (within 25 m). Plots were centered at the points used in the bird surveys or at the start of the transects. Transects that contained a corresponding point count were considered to have the same vegetation variables.
2.2.4 Statistical analyses
To evaluate the richness of forest bird and butterfly species, a comparison was made of sample-based rarefaction curves rescaled to the number of individuals among the different habitat types. Species richness was computed using a binomial mixture model (Colwell et al. 2004), where any heterogeneity or patchiness in sample data was removed by averaging values over repeated randomizations (Gotelli and Colwell
2001). To compare population density, the number of individuals versus the number of samples were plotted. Additionally, sampling completeness was assessed by generating nonparametric species richness estimators to estimate the total number of species undetected by the surveys. The different estimators (ACE, ICE, Chao1,
Chao2, Jack1, Jack2, Bootstrap) and curve models that compute for asymptotic species richness (MMRuns, MMMeans) have been found to perform differently for different species-abundance distributions and no single method is considered the
“best” (Walther and Morand 1998). Thus, an average of the various estimators was
25 used as a measure of the “true” species richness in each habitat. Values for species richness and the various estimators were generated using EstimateS Version 7
(Colwell 2004).
An indirect gradient analysis was performed in order to relate species distribution with the measured habitat variables. Nonmetric multidimensional scaling (NMS) is an effective ordination method for ecological community data because it does not have the linear constraints that restrict many of the metric ordination methods (e.g. principal components analysis, Kenkel and Orloci 1986). This method allows the biota to “tell their own story” before deducing links to specific environmental variables (Clarke and Ainsworth 1993). It searches for the best arrangement of points in a reduced metric space with k dimensions (axes) that minimizes the stress of the k- dimensional configuration (McCune and Grace 2002). Stress is a measure of the departure from monotonicity in the relationship between the dissimilarity in the original variable space and the reduced k-dimensional space.
Both forest and non-forest birds were included in the ordination to examine the response of the complete community. Habitat variables that were measured as percentage area were arcsine transformed prior to entry into the secondary matrix.
The analysis was run using the settings for ‘slow and thorough’ mode, with a random starting configuration and Sorensen (Bray-Curtis) as the distance measure (McCune and Grace 2002). The two axes representing the highest percent of variance in the data were chosen for the final ordination. A Spearman’s correlation of each of the measured habitat variables was performed against the final axes in the ordination.
Variables that were strongly correlated (r > 0.5) were plotted as vectors, the sizes of
26 which corresponded to the magnitude of the correlation. Species scores generated by weighted averaging of their abundances in each sample were also plotted.
As a further test of differences between the composition of communities in the different habitats, pairwise multi-response permutation procedure analyses were carried out. This nonparametric analysis generates a test statistic, T, to describe the separation between the groups; a p-value to describe the likelihood that the difference is due to chance; and a measure of effect size, A, which describes within-group homogeneity (McCune and Grace 2002). The more negative the T, the greater the separation between groups. Sorensen was used as the distance measure to complement the ordination analysis. The nonmetric multidimensional scaling and multi-response permutation procedure analyses were performed using PC-ORD
Version 4.14 (McCune and Mefford 1999).
2.2.5 Analysis of forest bird species response to canopy cover
Individual responses of forest bird species to canopy cover were examined further through a simple simulation. Binary logistic regressions were performed on each species with their presence/absence as the response to the amount of canopy in point counts in all five habitats (cf. Sodhi et al. 2005). Species that had a significant response (p < 0.05) were used to calculate the proportion of forest species that were present with increasing percentages of canopy cover. Species presence or absence at each increment of canopy cover (e.g., 5%, 10%) was determined by comparison with the null response obtained by taking the natural log of the proportion of point counts wherein the species was present. Results were then plotted to determine species richness at different amounts of canopy cover.
27 2.2.6 Analysis of species vulnerability using ecological traits
To examine if species ecology influences vulnerability to disturbance, information on ecological traits were collated from literature and modeled to the empirical data obtained from the surveys. The main references were Kennedy et al. (2000) and
Robson (2000) for birds and Igarashi and Fukuda (1997, 2000) and Robinson et al.
(2001) for butterflies. For birds, the following traits were considered as predictors of vulnerability: (1) endemicity, (2) presence of sexual dichromatism, (3) body length,
(4) feeding guild, (5) nesting strata and (6) clutch size. For butterflies, the traits were:
(1) endemicity, (2) presence of sexual dichromatism, (3) forewing length and (4) number of larval host plant species. Previous studies have considered these traits as possible factors that affect the differential vulnerability of species to extinction in various taxa, including birds and butterflies (Gaston and Blackburn 1995, Bennett and
Owens 1997, Davies et al. 2000, Owens and Bennett 2000, Purvis et al. 2000, Koh et al. 2004a).
Species traits were analyzed using general estimating equations (GEE), a modification of generalized linear models that allow for correlated data (Hardin and Hilbe 2003).
GEEs are appropriate for this analysis since species cannot be considered independent data points because ecological characteristics are shared by closely-related taxa. A species was defined as vulnerable if there was a significant decrease in its mean abundance outside forested habitats (i.e., abundances were pooled for open and closed canopy forests, as well as for the other 3 habitats), evaluated by Mann-Whitney U tests. Univariate GEEs of each trait as a predictor against vulnerability were fitted with family was used as the clustering variable to control for phylogeny. A binomial
28 distribution was specified since the response was binary (vulnerable or non- vulnerable), and an exchangeable working correlation structure was used in the analyses. Candidate models were then constructed using the significant predictors in different permutations and the final model fit was evaluated using the quasi-likelihood criterion QICu (Pan 2001). QICu is a modified version of the Akaike’s Information
Criterion that can be used with quasi-likelihood models to determine the best subset of covariates for a particular model (Hardin and Hilbe 2003). Smaller values of QICu indicate the best fit among candidate models. The PROC GENMOD procedure in
SAS ver. 8 (SAS Institute, Inc.) was used to fit the generalized estimating equations and compute QICu.
2.3 Results
2.3.1 Community measures for forest species
Most of the sampling curves appeared to be reaching saturation, with the possible exception of the butterfly transects in the rural habitat type (Figures 6 and 7). A comparison of the number of observed forest species with the nonparametric estimators (Table 2), which approximate true asymptotic species richness, indicated that 80–90% of the forest bird species and 68–97% of the butterfly forest species present were detected. For both taxa, the mean number of species detected per sample was highest in the open or closed canopy forest, and decreased in the more disturbed habitats (Table 1).
For birds, a total of 48 forest species and 1576 individuals were detected and 31 forest species and 750 individuals were recorded for butterflies (Table 1). Sixteen bird species (33%) and eight butterfly species (26%) were found to be restricted to either
29 open or closed forest habitats (Table 3). With the exception of the Asian glossy starling, Aplonis panayensis, the five forest species from both taxa that were detected in the urban habitat were present in all other habitat types. Only one was an endemic
(Philippine bulbul, Hypsipetes philippinus), while the others were known forest edge species (e.g. Large-billed crow, Corvus macrorhynchus, and Black-naped oriole,
Oriolus chinensis) or have wide geographical distributions (e.g., Hypolimnas bolina and Papilio alphenor). These species showed increased abundances in modified habitats.
The two taxa showed dissimilar trends for species richness and population densities across the habitats. For birds, rarefaction curves for species richness were highest in the forested sites, followed by the suburban and then rural habitats (Figure 8).
Butterflies showed higher species richness in the rarefaction curve for the rural areas, followed by the forested sites and the suburban habitat (Figure 9). The population density of birds was highest in the closed canopy forest, while the open canopy forest and suburban sites had similar densities (Figure 10). The population density of butterflies was higher in both forest sites than in suburban and rural areas (Figure 11).
The urban sites had the lowest species richness and density of both taxa. The open canopy forest had slightly higher species richness for both taxa than the closed canopy forest at higher sample sizes. Values of the extrapolated “true” species richness calculated from the nonparametric estimators and curve models were similar in ranking order to those from the rarefaction curves (Table 2). For birds, the mean estimated species richness in the open canopy forest was highest with 50.60 + 1.36
(SE), followed by the closed canopy forest with 48.08 + 1.03, suburban habitat with
30.43 + 0.62, rural habitat with 22.44 + 0.83 and urban habitat with 5.52 + 0.14.
30 Butterflies had highest richness in rural habitat with 33.48 + 1.66, followed by open canopy forest with 31.57 + 0.63, closed canopy forest with 27.70 + 0.10, suburban habitat with 20.23 + 0.49 and urban habitat with 6.94 + 0.34.
2.3.2 Indirect gradient analysis
For birds, the final 3-dimensional solution given by the NMS analysis represented
61.4% of the sample variance (Table 4), and the two axes representing the most variance (accounting for 21.1% and 24.4%) were used for the final ordination. Of the variables that were strongly correlated to the axes, tree density, percent canopy cover, average height to inversion, and average ground cover were positively associated with the forest points while light intensity, percent of paved ground and number of buildings were positively associated with the highly modified points located in the urban areas (Table 4). The joint biplot of the point counts (Figure 12) and results of the multi-response permutation procedure (Table 5) showed that the bird species compositions of both forest habitats were similar, having the lowest T statistic, and were distinct from the more disturbed habitats. Suburban and rural habitat had more similar species compositions with each other than with any other site, and urban areas were the dissimilar with all other habitats. The plot of species scores showed most of the forest bird species clustered at the top left quadrant, corresponding to the forest points, with the exception of the species that were detected in the urban areas (Figure
13).
The NMS analysis for butterflies resulted in a final 3-dimensional solution accounting for 44% of the variation in the data (Table 4). The final two axes plotted (accounting for 14.7% and 19.8% of the variance) showed most of the forest transects in the top
31 half of the ordination and transects in the modified habitats on the bottom half (Figure
14). Distances between transects in the same habitat were greater and multiple pairwise comparisons (Table 5) showed that species composition was dissimilar between most habitat pair comparisons. Species composition in the closed and open canopy forest were similar to each other and very dissimilar to the urban habitat.
However, the separation statistic between the open canopy forest and rural habitat was also low, as well as for suburban and other human-modified areas. None of the measured habitat variables were strongly correlated to the ordination, although the strongest relationship found was that of light intensity and average litter depth with r
> 0.45 (not shown in graph). The plot of weighted species scores did not show a clear separation of forest from non-forest species (Figure 15).
2.3.3 Response of forest birds to canopy cover
The ordination showed that canopy cover was strongly associated (r = –0.669 and
0.455 with axes 1 and 2, respectively) with the distribution of forest bird species. The binary logistic regressions showed that twenty-six bird species responded significantly to the percent canopy cover measured in the circular plots (Table 3).
The simulation showed that 24 of these species require canopy cover of 60% or higher (Figure 16). To retain all 26 species, 100% cover was required and none were present when cover was less than 35%.
2.3.4 Ecological traits related to species vulnerability to disturbance
Results of the univariate GEE analysis showed that endemicity and nest location were significant predictors (p < 0.05) of vulnerability for birds, while predictors for butterflies were endemicity and number of larval host plants (Table 6). Out of the
32 candidate models generated (Table 7), the lowest QICu value was given to the model that included only nest location for birds (QICu = 59.27) indicating the best fit among the covariates included in this analysis. Arboreal nesting species were more vulnerable to disturbance, followed by shrub nesting species, while ground nesters were least vulnerable. For butterflies, the model with endemicity alone had the best fit
(QICu = 53.53), with endemic species being more vulnerable than non-endemics.
2.4 Discussion
2.4.1 Faunal communities in forests
The results showed that species richness (Figure 8, Figure 9) and community composition (Figure 12, Table 5) of birds and butterflies were similar between the closed and open canopy forest habitats. This indicates that resources and various microhabitats are still available in the open canopy forest despite disturbances such as past selective timber extraction and harvesting of non-timber products. Studies have shown that low levels of disturbance in tropical forest do not greatly affect levels of diversity and richness of butterflies and birds (Hamer et al. 1997, 2003, Ghazoul
2002, Sodhi et al. 2005). A possible negative effect, however, are the decreased population densities of both taxa. Although the open canopy forest lacks mature dipterocarp cover, a recent floristic inventory in the reserve showed that stocking densities of trees were not necessarily higher, and that trees with large diameters were not restricted to the closed canopy forest (Dalmacio 2001). In places where dipterocarp cover is diminished, other non-dipterocarp species (e.g., Diospyros pilosanthera and Parkia roxburghii) can dominate and grow to similar heights
(Dalmacio 2001), thus providing canopy cover. The presence of these large canopy trees in a forest habitat may be an important structural feature for forest bird species,
33 as canopy cover was identified as a significant habitat characteristic in the ordination.
It also affected the distribution over 50% of the forest species observed (Figure 16) and high values of canopy cover (60% or more) are needed to retain most species.
2.4.2 Faunal communities in modified habitats
Conversion of forests to other land uses results in corresponding changes in community parameters and composition, as species will either avoid or be attracted to disturbance. In human-modified habitats, vegetation attributes such as density and vertical complexity are postulated to remain important for birds (Raman et al. 1998,
Peh et al. 2005). For example, certain species of forest birds could be attracted to suburban areas due to the presence of large trees (Sodhi et al. 1999, Lim and Sodhi
2004). Other species may also be attracted to disturbed areas since they introduce new exploitable resources such as water, ornamental plants, and grasses (Blair and Launer
1997). In this study, suburban habitats had higher species richness and population density of birds than rural habitat. Tree density in suburban habitat was much lower than in forested areas, but large trees are often present in gardens. Because of these trees, forest birds that forage in the canopy, such as the Luzon hornbill. Penelopides manillae, and the Green racquet-tail, Prioniturus luconensis, may be able to extend their ranges to feed in suburban areas. However, they do not go into rural habitats, which has fewer large trees remaining. Species found in the more open rural habitats are perhaps more tolerant of higher disturbance levels and are likely able to utilize planted ornamental or crop trees, or other patches of secondary vegetation present.
For butterflies, community composition was less distinct among the intermediately disturbed habitats. Rural and suburban transects had similar species composition to
34 open canopy forest, as well as each other (Table 5). It is possible that resources are patchily distributed in these areas (e.g., flowering plants) and forest butterflies can actually utilize areas that are usually considered “unsuitable” (Dennis 2004).
However, it is also possible that these individuals represent vagrants and do not form viable populations. The “high” species richness represented by the rarefaction curve for the rural habitat (Figure 9) could be due to the fact that there were only few individuals of each species and a greater evenness of the relative abundance distribution (Gotelli and Colwell 2001). Thus, while there was a high number of species detected, they were present at low densities. No habitat variables that were strongly correlated to the ordination for butterflies. Other factors that were not measured may be more significant in determining their distribution, such as presence of hostplants (Koh et al. 2004a, 2004b). More information on the distribution of adult and larval resources is needed before the extent to which these areas can support forest butterfly populations can be established.
Urbanization was detrimental to forest species in both taxa, as shown by the community measures in the urban sites. Less than 20% of the forest species from both taxa were detected, and this habitat type had the lowest population densities.
Mean abundances for species detected in the urban areas were lower than in other habitats, with the exception of A. panayensis, which seems to have adapted to man- made settings. Some tropical butterflies have been found to survive in urban areas but require green areas with varied vegetation, watercourses and low levels of pollution
(Brown and Freitas 2002, Koh and Sodhi 2004). Such green spaces are lacking in urban environs in the Philippines, which are characterized by high densities of buildings, paved ground and human population. Majority of the native forest species
35 are unable to tolerate the extreme changes to their natural environment, and only more widely-distributed generalists such as the Eurasian tree sparrow, Passer montanus, or
Catopsila pomona are able to exploit urban habitats.
2.4.3 Ecological traits of vulnerable species
In both taxa, endemicity was a predictor of vulnerability to human disturbance. This agrees with the theory that geographically-restricted species are more susceptible to extinction (Jablonski 1991, Smith et al. 1993) since they usually occupy narrow ecological niches. Habitat loss can disproportionately reduce niche availability for such habitat specialists (Norris and Harper 2004). Many Philippine endemics are forest-dependent, and would be most at risk from forest destruction and disturbance
(Dickinson et al. 1991, Settele 1993). Reproductive requirements were also found to be significant in determining vulnerability. Butterfly species with fewer larval hostplants are more at risk than generalists (Koh et al. 2004a), which can utilize a wider range of resources. Specialists would be restricted to areas where they can lay their eggs since they are highly co-evolved with, and dependent on, their hostplants for survival (Ehrlich and Raven 1964, Koh et al. 2004b). The availability of nest sites has been postulated to be an important component of species-habitat relationships and thus can influence the composition of bird species assemblages (Martin 1988a). The distribution of forest species may be limited to intact forests due to the lack of suitable sites in other habitats for nesting. For birds, arboreal nesters were less likely to be more vulnerable since they would be more adversely affected by disturbance that reduce the number of trees and canopy cover. Additionally, changes in vegetation structure outside forests may make nests susceptible to predation because of changes in visibility (Keller et al. 2003).
36 2.5 Conclusions
Forest birds and butterfly communities of the Subic Bay Watershed Reserve show differing responses to various land uses, but showed similar decrease in population densities with increasing disturbance. Although it has been suggested that birds and butterflies may be used as surrogates for each other when assessing biodiversity (Blair
1999), this study found they have dissimilar patterns of response to habitat modification across a disturbance gradient in the Philippines. However, care must be taken in drawing conclusions, since only a subset of the butterfly fauna was surveyed.
Butterflies in the families Lycaenidae and Riodinidae and canopy fauna require methods such as sweep netting or baited traps to sample and thus were excluded. In future studies, efforts should be made to include them, as they can affect measures of community parameters (Dumbrell and Hill 2005). In addition, observed effects of habitat disturbance may depend on the spatial scale used in sampling, and this must be considered when interpreting results. The use of a single spatial scale for different taxa may not be directly comparable if they respond to habitat disturbance in a spatially-dependent manner (Hill and Hamer 2004). Increases in richness were detected in the open canopy forest for both taxa, even with the area surveyed using point counts in this study were approximately twice that of the transect walks. Future studies should examine the scales at which functional loss of diversity occurs for different taxa.
Given how many endemics are dependent on forest habitats, preserving the remaining mature forests should be the primary concern of Philippine conservation efforts.
Some forest species seem to be able to persist in highly modified environments
(Hughes et al. 2002, Brown and Freitas 2002), but whether or not this is due to true
37 resilience to disturbance or to time-lag in extinction is still undetermined (Balmford
1996). It is possible that species were detected in disturbed areas adjacent to the reserve because of their ability to move to and from nearby continuous forest areas
(Johns 1991). It has yet to be determined whether or not these populations can truly persist in degraded areas (Sodhi et al. 2004b, Peh et al. 2005).
38 Chapter 3 Effects of land use on predation of nests and caterpillars across a disturbance gradient 2
3.1 Introduction
Predation is the main mechanism underlying nest mortality in birds (Ricklefs 1969) and larval mortality in butterflies (Feeney et al. 1985), and thereby has a strong influence on habitat selection, community structure and the distribution of species across different habitats (Martin 1988b, Sieving and Willson 1998, Morris 2003,
Shiojiri and Takabayashi 2003). Habitat disturbance may impact species by causing direct mortality, reducing survivorship or decreasing reproductive success (Sih et al.
2000, Schlaepfer et al. 2002, Sodhi et al. 2004b) — ecological processes that are potentially influenced by predation. Patterns of predation can be altered by anthropogenic changes in the landscape (e.g. fragmentation), which often introduce novel predators and facilitate their access to the forest interior (Rodewald and Yahner
2001). Information on how disturbance affects such ecological processes is largely lacking in South-East Asia (Sodhi and Brook 2006). Only a few studies on the effects of habitat disturbance on predation have been conducted in the region (e.g., Cooper and Francis 1998, Wong et al. 1998, Sodhi et al. 2003), and none have been conducted in the Philippines. Knowledge of how disturbance may affect the ecological processes can contribute to the development of strategies to maintain biodiversity in degraded landscapes.
Because actual predation events are rarely observed, previous studies have used artificial models to evaluate predation pressure (e.g., for nests, Wilcove 1985; for
2 Published in Journal of Tropical Ecology 23:27-33 (2007). 39 caterpillars, Koh and Menge 2006). The use of artificial models permits the testing of specific hypotheses (e.g. effects of habitat type), by experimentally controlling for the effects of other factors, such as prey density. However, this method is subject to many inherent biases (Major and Kendall 1996), chief of which is that rates of attack on artificial models may not accurately reflect actual rates of predation (Zanette 2002,
Berry and Lill 2003). Despite the potential limitations, the use of artificial models is a valuable tool for the rapid assessment of relative predation pressure across sites of interest, especially in regions where long-term studies are logistically or economically prohibitive (Loiselle and Hoppes 1983, Koh and Menge 2006).
Predation pressure may vary both between and within habitats (i.e., between microhabitats) due to differences in predator assemblages and density, vegetation structure and complexity, and intensity of human activities (Martin 1993a, Loiselle and Farji-Brener 2002). In this study, artificial models were used to examine the relative levels of predation on bird nests and lepidopteran larvae across three habitats that varied in levels of human disturbance in the Philippines. Predation levels were also compared across three microhabitats of different heights (ground, 1–1.5 m and >
3 m). Based on previous studies, predation pressure was expected to vary for the different habitats and vegetation strata in the following manner: (1) increases with increasing disturbance for both nests and caterpillars; and (2) be highest at ground level for nests and at the heights > 3 m for caterpillars.
40 3.2 Methodology
3.2.1 Study area
We conducted the study at the Subic Bay Watershed Reserve in west central Luzon,
Philippines (Figure 5, see Methodology in Chapter 2 for more details). Three habitat types were chosen to set-up the experiments, based on land-use categories as described in Chapter 2: (1) closed canopy forest consisting of natural forest where mature dipterocarps or other broadleaf trees cover 40% or more of the area; (2) open canopy forest that is natural forest with a discontinuous tree layer and coverage of
10–40% and (3) rural areas consisting of mixed-use areas consisting of grasslands, regenerating scrub, small-scale agriculture and reforestation plots. Experiments were not conducted in the suburban and urban areas due to the limited area of the habitat and lack of available vegetation for nest and egg-laying sites, respectively.
3.2.2 Experimental set-ups
Predation experiments were conducted from late April to early June 2005, a period approximating the breeding season of majority of the birds in the reserve (Dickinson et al. 1991). The artificial nests (10 cm diameter, 6 cm depth) used were woven from stems of climbing ferns (Lygodium sp.) and immersed in boiling water then dried outdoors to reduce odors prior to use. One quail (Coturnix coturnix) egg and one plasticine egg, painted to resemble a real egg, were placed in each nest. Quail eggs were obtained at a local market so that none was more than 1 week old. Artificial caterpillars (35 mm length, 5 mm width) were made from green plasticine and painted to resemble the fifth instar larva of Papilio alphenor, before being glued to a bamboo skewer for ease of attachment in the field.
41 In each habitat, four 750-m transects were randomly established, each with 15 10-m radius plots separated by 50-m intervals. Transects were located at least 150 m away from the forest edge. Within each plot, nests and caterpillars were randomly placed at three heights: ground, 1–1.5 m and > 5 m for nests or > 3 m for caterpillars. One nest and caterpillar placed at each height category. Nests and skewers were secured to stems and branches using thin-gauge wire and a small piece of flagging tape was used to mark the centre of the plot. No attempt was made to conceal the set-ups to avoid any biases related to differences in concealment effort. The set-ups were checked after a 5-day exposure period. Nests were considered depredated if one or both eggs were missing, if the real egg was broken or cracked, or if there were any bite or bill marks on the artificial egg (Söderström et al. 1998). Caterpillars were considered predated if they were missing, or had distinct beak or bite marks. All predated models were collected and examined to determine the predators responsible for the marks made on the plasticine. The percentage of vegetation cover at three vertical strata (0–1 m and
1–2 m cover estimated visually; canopy cover measured with a spherical densiometer) and the number of trees with dbh > 4 cm were recorded within each plot.
3.2.3 Predator identification
A variety of techniques were used to identify potential predators. When checking set- ups, effort was expended to retrieve any eggshell or model remains, and to record any animal traces. Markings on the plasticine models were examined and compared to reference marks made from offering plasticine to live animals, as well as marks from tooth imprints of museum specimens. We also set up 24 camera-monitored nests in the closed and open forest (Trailmaster® Model TM1500 Active Infrared Trail
Monitor and TM35-1 Camera Kit) that were separated from nests at the transects by
42 at least 100 m. Cameras were not set up in the rural habitat for security reasons.
Twenty baited live rodent traps were randomly placed in each transect and deployed overnight after the experiment proper concluded to identify and assess the abundance of potential small-mammal predators.
3.2.4 Statistical analyses
To test whether the probability of predation differed among the set-ups, general linear mixed models (GLMM) were fitted to the data. Mixed models can flexibly represent the covariance structure arising from clustered data (Pinheiro and Bates 2000).
GLMMs explicity model the heterogeneity arising from the non-independence of the data and are subject specific, with coefficients having interpretations for individual subjects (Hardin and Hilbe 2003). Each observation of nests or caterpillars predated vs. not predated was coded as the response variable, height and habitat as fixed effects
(predictor variables), and transects and plots as nested random effects (control variables) to account for the non-independence in spatial location of set-ups. The
GLMMs were fitted using the ‘lme4’ package in R version 2.2.0 (R Development
Core Team; http://www.cran.r-project.org/) by specifying a logit link function and a binomial error structure for the response. Collinearity between the variables was assessed using the ‘perturb’ function in the ‘Perturb’ package, which evaluates changes to the parameter estimates when each predictor variable is randomly re–classified. The effect of each predictor variable was first evaluated separately.
Then, from a maximal model that included both predictors and their interaction term, a set of candidate models was obtained by sequentially removing the variable/interaction term with the least contribution to model adequacy based on the change in log-likelihood value when each variable/interaction term was removed. The
43 minimal adequate model was selected on the basis of parsimony using Akaike’s
Information Criterion (AIC). To determine if vegetation density or predator abundance affected predation levels, Spearman’s rank correlations were performed between the mean predation and measured vegetation variables and small-mammal abundance in each transect.
3.3 Results
3.3.1 Nest predation
Overall predation on nests was lowest in the open canopy forest (16.7%) and highest in the rural areas (58.3%) (Table 8). Predation in the closed canopy forest was intermediate between the other two habitats (32.8%). Single–fixed effect analyses showed that nests in rural areas suffered higher predation than those in closed canopy forest (estimate = 0.26, p = 0.06) (Table 9). Additionally, nests at 1–1.5 m were significantly more likely to be predated than ground nests (estimate = –0.15, p =
0.03), while nests at heights >5 meters did not show a significant difference (Table 9).
The minimal adequate model for nest predation included habitat, height and their interaction term, suggesting that the effect of habitat type on predation differed among different nest heights (Table 10).
3.3.2 Caterpillar predation
Predation on the caterpillars showed an increasing trend with 46.1%, 50.6% and
59.4% of the plasticine models showing beak or bite marks in the closed canopy forest, open canopy forest and rural areas, respectively (Table 8). Single–fixed effect analyses showed that caterpillars in rural areas suffered significantly higher predation than those in closed canopy forests (estimate = 0.13, p = 0.03) (Table 9). The minimal
44 adequate model for caterpillar predation included habitat as the sole predictor (AIC =
802), as the removal of the other terms from the full model did not result in a significant decrease in deviance.
3.3.3 Vegetation variables
Measures of the vegetation variables were significantly different among the three habitats (ANOVA, p < 0.05). Canopy cover and tree density decreased with increasing disturbance, while vegetation at 0–1 m showed the opposite trend.
However, the measured vegetation variables were not significantly correlated with mean predation in the transects for either nests or caterpillars (Spearman’s correlation,
N = 12, p > 0.05).
3.3.4 Potential predators
The infrared cameras were able to capture three separate predation events by long- tailed macaques, Macaca fasicularis. Other large-mammal predators (incisor width
>2 mm) were the Philippine warty pig, Sus philippensis, and the common palm civet,
Paradoxurus hemaphroditus. Small-mammal predators (incisor width <2 mm) identified were rodents, namely the Oriental house rat, Rattus tanezumi, the common brown rat, R. norvegicus, the Polynesian rat R. exulans, and large forest rats, Bullimus sp. The potential avian nest predators that were observed were corvids such as the large-billed crow, Corvus macrorhynchos, and the Crested myna, Acridotheres cristatellus. The majority of the attacks on caterpillars were made by birds and arthropods (e.g. ants, spiders). Marks by arthropod predators were pooled into one category. Out of eighty trap nights in each habitat, overall trap success was 11.3%,
22.5% and 32.5% for closed canopy forest, open canopy forest and rural areas,
45 respectively. There were significantly fewer rodents caught by the live traps in the closed-canopy forest than in the other habitats (n = 12, Kruskal-Wallis H = 8.41, df =
2, p < 0.05). However, the number of rodents was not significantly correlated with predation of caterpillars or nests at ground level (n = 12, Spearman correlation p >
0.05).
Predator type and abundance was found to vary among the three habitat types (Table
8). Out of 194 predated nests, it was not possible to determine predators or recover eggs from 85 (43.8%) of them. Based on the marks on the retrieved models, the closed canopy forest had higher instances of large-mammal predation. Birds and small mammals were responsible for the majority of the nest predation in the open canopy forest and rural areas. A total of 281 caterpillar models was predated upon, 28
(10.0%) of which were not recovered. Bird predation on caterpillars increased with disturbance, while arthropod predation showed the opposite trend. More marks by rodents were also observed in the rural areas, where there was also one incidence of reptile attack.
3.4 Discussion
3.4.1 Effects of disturbance on nest predation
Nest predation experiments have been conducted extensively in temperate habitats
(Söderström 1999), and although knowledge on predation in the tropics has been growing, South-East Asia remains understudied. Predation levels are known to vary with the vertical location of nests (Martin 1993a) and we found that ground nests were significantly more predated than nests at 1–1.5 m, supporting previous findings that ground-dwelling birds are more susceptible to nest mortality in the tropics
46 (Söderström 1999). How pressures differ for tree-nesters has not been elucidated, with only a few studies in temperate areas having examined trends at heights greater than 3 m (i.e., Ortega et al. 1998, Reitsma and Whelan 2000). In this study, nests at heights greater than 5 m had intermediate levels of predation in more disturbed habitats and were mainly attacked by airborne or scansorial predators such as birds and macaques.
In contrast, small terrestrial mammals such as rodents were responsible for most losses at ground level (Gibbs 1991, Wong et al. 1998, Estrada et al. 2002, Sodhi et al.
2003). Nests at 1–1.5 m were the least predated, so they may be somehow less conspicuous. Their intermediate position may make them less accessible to both terrestrial and airborne predators, especially in structurally complex habitats such as tropical forests.
Habitat degradation can lead to higher levels of nest predation (Martin 1993b). Of the three habitats, the open canopy forest had fewer instances of nest predation. This was inconsistent with the expectation that predation would be lowest in the less-disturbed closed canopy forest. One possible reason for the lower predation is that human disturbance may negatively impact the predators in the forest as well (Gibbs 1991).
For instance, predation by large mammals decreased outside of the closed-canopy forest, possibly because M. fasicularis avoid or are present in lower densities in areas with higher disturbance. Rural habitats had a significantly higher number of predated nests than the forest sites. While we did not find a correlation between our measured vegetation variables and nest predation, we observed an increase of attacks by birds, which are visually-oriented predators (Table 8). Previous studies have linked higher rates of predation to habitat openness (Gibbs 1991, Telleria and Diaz 1995, Estrada et al. 2002). In contrast to the closed canopy forest, evidence of attacks by avian
47 predators was found at all three vegetation strata in the rural areas. Small mammals were the other major group of predators in the rural areas, where they also gained access to higher nests. This again suggests that nests may be more conspicuous in the rural areas. Other studies found predation by small mammals to be more common in the tropical forest interior (Telleria and Diaz 1995, Cooper and Francis 1998).
However, we found predation events by rodents increased outside the forest. As they are generalist predators, rodents are abundant near farms and human settlements, which provide novel food sources (Angelstam 1986). Their presence may forest edges adjacent to agriculture unsuitable for nesting, or even possibly ecological traps.
3.4.2 Effects of disturbance on caterpillar predation
The effects of habitat alteration on invertebrate predation rates are even less understood (Koh and Menge 2006) than avian nest predation. Caterpillar predation was found to be higher in rural areas than in closed canopy forest, but predation was not significantly different among the three height strata. Predation by birds increased with disturbance (Table 8), which suggests it is more difficult for them to locate prey in the closed canopy forest—again possibly due to higher visibility in more open areas. Arthropod predation was highest in the closed-canopy site, although the proportion was lower than in other studies (90% or greater, Loiselle and Farji-Brener
2002, Koh and Menge 2006). Arthropod species richness is generally correlated with plant species richness, vegetation height and complexity (Gaston 1992, Haysom and
Coulson 1998). Higher diversity of plants and the presence of higher canopy cover may explain the presence of more invertebrate predators in the forest. In addition, many invertebrates are sensitive to habitat modification (Kremen et al. 1993) and the
48 negative effect of disturbance on arthropods may have caused their decline as the dominant predators of caterpillars in the rural areas.
Aside from affecting the assemblage of predators in a habitat and their ability to locate prey, habitat degradation can cause direct loss of features needed by certain species for reproduction (e.g., reduce the diversity of nest sites) (Martin 1993b).
However, as in other studies in the tropics (e.g., Wong et al. 1998, Koh and Menge
2006), the measured vegetation variables and predator abundance were not found to be significantly correlated with predation levels in the three habitats. Quantifying the variables that directly affect predation rates may continue to be difficult (Sodhi et al.
2003).
3.5 Conclusions
The use of artificial models is well-known to be subject to many biases (Major and
Kendall 1996), and the levels of predation observed in this study cannot be taken to reflect actual losses experienced by real nests and caterpillars. Ideally, natural populations should be monitored to find out if they are being negatively affected by increased predation. However, the differences in the relative predation pressure and predator assemblages among the habitats in this study show that human disturbance can have marked effects on biotic interactions. If altered predation patterns make habitats unsuitable for reproduction, they may become ecological traps (Schlaepfer et al. 2002). To maintain populations of forest fauna, such changes that affect the reproductive success of species must be understood. Future work can evaluate if rural areas, which have been found to retain some biodiversity (Horner-Devine et al. 2003,
Peh et al. 2005, Sodhi et al. 2005), indeed act as ecological traps for forest species.
49 Chapter 4 Correlates of extinction risk for Philippine avifauna
4.1 Introduction
The tropics harbor the majority of the world’s biodiversity (Myers et al. 2000) and it is where human actions are expected to precipitate an extinction crisis as the amount of suitable habitat is diminished and degraded (Pimm and Raven 2000). However, anthropogenic change does not affect all species equally. The non-random extinctions that have occurred in disturbed landscapes suggest that some species may be inherently sensitive (McKinney 1997). While stochastic factors that arise from diminished population size ultimately cause extinction, certain biological factors can predispose species to become extinct. Several traits have been found to lead to higher extinction risk, including small population size, small geographic range, poor dispersal ability, or a combination of characters that lead to a “slow” lifestyle such as slow growth rates, large body size, late sexual maturity, low fecundity (Purvis et al.
2000). The probability of a species’ extinction depends both on these shared characters, as well as on the environment. Different sources of threat can increase the importance of certain traits as extinction correlates (Owens and Bennett 2000, Norris and Harper 2004). For instance, introduced predators impact birds with inherently slow population growth, large birds are more threatened by over-exploitation and habitat specialists are at most risk from habitat loss (Sodhi et al. 2004b).
Knowledge of the extinction-proness of South-East Asian biota is shallow (Sodhi and
Brook 2006). The region is predicted to lose over a third of its biodiversity over the next century given the elevated rates of deforestation (Brook et al. 2003). This study aims to determine which ecological traits contribute to extinction risk among resident
Philippine birds, which are one of the better-known taxa for whom most species are
50 described. Since the current knowledge of distribution, status and biogeography of
Philippine butterflies is limited, and many aspects of their taxonomy and basic biology remain unknown (Settele 1993, Danielsen and Treadaway 2004), data was insufficient to perform the analysis for butterflies. On the other hand, avian phylogeny has been studied in-depth (Sibley and Ahlquist 1990) and threat status of
Philippine bird species has been repeatedly evaluated through Red Lists (IUCN
2006).
Traits that were used for the analysis were chosen based on what is known about extinction biology and the data available for Philippine avifauna. It was expected that range would be an important determinant of extinction risk, since majority of the country’s threatened species are endemics. The greatest threats to Philippine birds include habitat loss and hunting, which affect 97% and 40% of threatened species, respectively (Collar et al. 1999), pointing to body size as a possible risk factor.
Determining the characteristics that contribute to extinction risk can help scientists to better understand the underlying mechanisms behind decline, identify vulnerable species and formulate conservation strategies (McKinney 1997). Results from the analysis can inform conservation actions for Philippine birds, both those that are currently threatened and those that may be at risk of being threatened in the future.
4.2 Methodology
A list of Philippine birds for which traits could be entered accurately was compiled by integrating lists by Dickinson et al. (1991), Kennedy et al. (2000) and BirdLife
International. The analysis was restricted to 396 resident species, after migratory species were excluded from the list. Phylogeny is known to affect extinction risk in
51 birds (Bennett and Owens 1997, Russell et al. 1998). Shared ecological traits among birds belonging to the same family may influence their response and therefore, species cannot be treated as independent of each other. Generalized estimating equations (GEEs) were utilized to analyze the data set. The GEE approach (Liang and
Zeger 1986) uses a generalized linear modeling framework for modeling data with categorical responses that may be correlated. This method has found application in ecological analysis as an alternative to phylogenetically independent contrasts
(Paradis and Claude 2002, Duncan and Blackburn 2004). In the GEE method, the primary interest is the effect of covariates on responses and association between responses is of secondary interest (Fahrmeir and Tutz 1994).
4.2.1 Response variable
Information on the threat status of bird species was obtained from the latest IUCN
Red List (IUCN 2006), which assesses extinction threat consistently for entire taxonomic groups. Species were considered at risk of extinction if they meet the criteria for Critically Endangered, Endangered or Vulnerable under the IUCN Red
List Categories. The response was then coded as binary, ‘Threatened’ or ‘Not threatened’. Sixty of the resident Philippine birds were considered ‘Threatened’ under IUCN criteria.
4.2.2 Clustering variable
Since it has been shown that extinction risk is not randomly distributed among avian families (Bennett and Owens 1997, Russell et al. 1998), family was used as the clustering variable to control for phylogeny in the GEE analyses. Taxonomic classifications were based on BirdLife International, which maintains its own list of
52 all the world’s bird species, reviewed and adopted by the BirdLife Taxonomic
Working Group (BirdLife International 2006).
4.2.3 Predictors
Traits that were considered correlates of extinction risk were determined a priori and information was collated from available literature. The traits used as predictors were defined as follows:
Range — A species’ range can be an important determinant of extinction risk — widely distributed species are usually able to exploit a wider range of habitats than those with narrow distributions and thus may be less prone to extinction (Jablonski
1991, Jones et al. 1991). Species were classified into three categories. To determine if birds restricted to a single island are more at risk, which the first category was defined as ‘Single-Island Endemics’. ‘Multiple-Island Endemics’ were those known from more than one island in the Philippine archipelago, while ‘Widespread’ birds were those whose distributions extend outside the archipelago. Range information was obtained from Kennedy et al. (2000).
Elevation — Elevational range can affect species extinction risk in the same manner as geographical range, in that species that have a wider altitudinal distribution can adapt to a greater variety habitats and thus are less threatened. Elevation was classified into three categories, where ‘Lowland’ species had their median elevational distribution falling between 0 to 600 m.a.s.l.; ‘Montane’ species had median elevational distributions higher than 1200 m.a.s.l. while ‘Widespread’ species had
53 distributions spanning from below 500 m.a.s.l. to higher than 1200 m.a.s.l.
Information was taken from Del Hoyo et al. (1992–2006).
Dichromatism — Presence of dichromatism was included as a trait as species that have a high energetic investment in body coloration may be more susceptible to environmental stresses or be more at risk from predation (Sorci et al. 1998, McLain et al. 1999). Presence or absence dichromatism was coded as a binary categorical variable, using descriptions from Kennedy et al. (2000).
Length — Larger body size has been found to be associated with increases in extinction risk, since large species usually have low population sizes, require bigger ranges and occupy higher tropic niches (Bennett and Owens 1997). Size was given by the total body length in millimeters, as found in Kennedy et al. (2000).
Diet — Certain foraging guilds of tropical birds have been found to be more extinction-prone than others (Sodhi et al. 2004b). Species were classified by their main diet preference, according to Del Hoyo et al. (1992–2006). Categories of diet were ‘Invertebrate’, ‘Vertebrate’, ‘Fruits’, ‘Seeds/Nectar/Plant Material’, and
‘Scavenger/Omnivorous’.
Habitat Breadth — Ecologically specialized species would be at risk from threats that reduce niche availability such as habitat loss (Owens and Bennett 2000). Habitat specialization was coded as a binary categorical variable, with ‘Habitat Restricted’ birds being found in only one habitat type, while ‘Not Habitat Restricted’ birds were
54 found in more than one habitat type. Information was obtained from the Del Hoyo et al. (1992–2006).
Diet Breadth - Diet specialization is another indicator of narrow niche requirements, and was coded as a binary categorical variable, with ‘Diet Restricted’ birds being those that only consume one major food type, while ‘Not Diet Restricted’ birds consume more than one major type of food. Information was obtained from the Del
Hoyo et al. (1992–2006).
Nesting Strata — Nest placement can affect extinction proneness since nests at different heights experience varying levels of predation pressure (Martin 1993a,
Söderstrom et al.1998, Posa et al. 2007). Height of nests were coded as ‘Ground’,
‘Shrub’ when species are known to nest from 1-5 m from the ground or ‘Arboreal’ when nests are found higher than 5 m. Nesting information was taken from Del Hoyo et al. (1992–2006), Kennedy et al. (2000), Robson (2000). Nest heights for species with unknown nests were inferred from congeners.
Clutch Size — Species with low fecundity are often more vulnerable to extinction since they would recover slowly from severe reductions in numbers (Bennett and
Owens 1997). The information on the average number of eggs per clutch was taken from Del Hoyo et al. (1992–2006). Average clutch size values for the family according to Bennett and Owens (1997) were used for species with unknown clutch sizes.
55 4.2.4 Generalized estimating equations.
The effect of each predictor on extinction risk was tested by fitting univariate GEE models, specifying a binomial distribution and an exchangeable working correlation.
Family was used as the clustering variable to control for phylogeny. Significant predictors were then entered into a multivariate analysis to identify a minimum adequate model (MAM) for extinction risk. Starting with a model containing all the explanatory variables, a backward selection process was followed by dropping predictors that resulted in the greatest improvement in model fit. Model fit was assessed using QICu (Pan 2001), which is a modified version of Akaike’s Information
Criterion. QICu makes use of a quasi-likelihood constructed from estimating equations as a model-selection criterion. Removal of predictors was continued until it resulted in a model that had higher values of QICu, indicating poorer fit. Univariate and multivariate GEEs were fit using the ‘repeated’ statement in the PROC
GENMOD procedure, and the SAS QIC macro was used to generate QICu values, in
SAS ver. 9.3.1 (SAS Institute, Inc).
4.3 Results
4.3.1 Univariate analyses
Results of the univariate generalized estimating equation analyses showed that range, elevation, habitat breadth, diet breadth and nesting strata were significant predictors
(p < 0.05) of extinction risk (Table 11). Endemics were found to be significantly more at risk than non-endemic species, with single island endemics being the most threatened. Lowland species were more prone to extinction, as well as both habitat and diet restricted birds. For nesting strata, shrub nesters were significantly less threatened than arboreal nesters.
56 4.3.2 Minimum adequate model
Removal of diet breadth from the full model (QICu = 252.06) containing the significant predictors resulted in a better fit (QICu = 248.62). The minimum adequate model (Table 12) for extinction risk thus included range, elevation, habitat breadth and nesting strata. More endemic species, lowland species, habitat restricted birds were still more significantly extinction-prone than reference categories, while shrub nesters were also still significantly less extinction-prone than tree nesters.
4.3.3 Species at risk
Based on the results of the minimum adequate model, resident Philippine birds that exhibited the significant correlates of extinction risk that are not currently listed as threatened were identified. These species are listed in Table 13.
4.4 Discussion
The results of the GEE analyses indicate that among resident Philippine birds, species tend to be more extinction-prone if they are endemic, occur at lower altitudes, have specialized diet and habitat requirements (Table 11). Figure 17 shows the proportion of threatened and non-threatened species exhibiting different categories of these traits.
The traits that may predispose species to extinction through other ecological mechanisms were not found to be significant.
Range, as expected, was found to be significant in the univariate analysis, and as a factor in the MAM. Endemic species, especially single-island endemics, were more at risk. Restricted distribution may confer extinction-proneness through several
57 mechanisms. Restricted species tend to have smaller population densities and species with wide distributions may be more adapted to exploit a wider variety of ecological niches (Brown 1984, Jablonski 1991, Smith et al. 1993, Gaston et al. 2000). Island endemics also generally evolve in isolation and thus are more vulnerable to introduced predators, as well as humans (Duncan and Blackburn 2004). Restricted altitudinal range can have similar effect as having a small geographical range, with species having wider ranges more likely to be able to adapt to a greater variety of niches. Elevation is an extrinsic factor that can affect population sizes and threatened species living at higher altitudes have been found to have larger population sizes
(Blackburn and Gaston 2002). However, the significance of elevation as an extinction correlate of Philippine birds may be less due to small population density, and more due to the disproportionate threat to lowland fauna. More lowland areas are inhabited and utilized by humans than less accessible upland areas. Birds with larger elevational ranges can avoid the negative impacts of human action in the lowlands by dispersing to other areas. We found montane species to be less extinction-prone, but research looking at species-area relationships showed that there are more Philippine birds at threat than is predicted by habitat loss in montane areas (Brooks et al. 1999).
This suggests that sources of threat may act through different mechanisms at different elevations.
Widespread habitat destruction such as that which has occurred in the Philippines greatly reduces niche availability. Restricted habitat and diet requirements, which indicate specialized ecological requirements, were related to higher extinction probability, although only habitat breadth was retained in the MAM. This further supports the idea that species that have narrow ecological niches are less adaptable to
58 the changes brought about by habitat loss and degradation (McKinney 1997). This has been also shown in other hotspots of global avian biodiversity. Using data on
Endemic Bird Areas, Norris and Harper (2004) found that extinction risk for habitat specialists increases as habitat loss becomes more severe.
Shrub nesters were found to be significantly less threatened than species that nest at higher strata. The univariate GEE showed ground nesters to be more at risk than arboreal nesters, and while the MAM showed the opposite pattern, neither were significant differences (Table 12). The lack of a clear pattern may be because nest placement can contribute to a species’ vulnerability in several ways. Ground nesting habit can make nests susceptible to higher predation, as shown through artificial nest experiments in the tropics (Söderström 1999, Posa et al. 2007). They would also be heavily impacted by introduced predators (Duncan and Blackburn 2004). However, arboreal nesters are more at risk from disturbances that reduce tree nesting sites, such as selective logging. Shrub nesters are less likely to be directly at risk from both these threats.
High levels of endemism in island birds are usually associated with large body size and low reproductive output, both of which have been found to be correlates of extinction risk. However, neither of these was significant for Philippine birds. Some studies link larger body size to higher risk of extinction (Bennett and Owens 1997,
Purvis et al. 2000, Norris and Harper 2004). Larger bodied animals are also more at risk from hunting pressures and introduced species (Owens and Bennett 2000,
Duncan and Blackburn 2004). Other studies suggest that birds with smaller body size are more at risk from threats that reduce niche availability (Owens and Bennett 2000).
59 Small body size may also mean poor dispersal ability in fragmented landscapes. Thus, it may be difficult to find a direct link between body size and extinction risk, since it is correlated with other traits (McKinney 1997). It is a difficult variable to interpret when comparing between species, although it may explain variation between closely related species (Bennett and Owens 1997). Smaller clutch sizes have been linked to vulnerability since low fecundity leads to low populations and a decreased ability to recover from severe reduction in density (Bennett and Owens 1997). However, the lack of significance in clutch size in this study should be viewed with caution since clutch sizes for many Philippine birds are not known, and were inferred from average values for the families for this analysis.
4.5 Conclusions
The set of factors identified as the most important correlates of extinction risk in
Philippine fauna underscores habitat destruction as the main threat to Philippine fauna. Endemics and ecological specialists, many of which are forest-dependent, are more at risk from such rapid and drastic change where suitable habitats become restricted and the new habitats created are only exploitable by generalist species
(Jones et al. 2001). This is undoubtedly linked to the extent of deforestation in the
Philippines, where roughly 50% of forest was lost in the last century (Liu et al. 1993).
Other intrinsic characteristics such as body size or clutch size were not significant predictors of extinction-proneness. It should be realized, however, that if present deforestation continues, the threat will stop being selective by taxon or trait since no species will be able to withstand complete removal of their habitat (Russell et al.
1998). This reinforces the need to preserve all remaining forest areas in the country,
60 especially in lowland areas where available habitat has dramatically shrunk due to human activities.
61 General Conclusions
The growth in awareness and interest in Philippine biodiversity are encouraging developments for conservation of the country’s highly threatened flora and fauna.
More scientific research should be geared towards providing ecological information that can form an information base for conservation actions. This study sought to provide empirical data on the effects of human action on two different taxa, to understand how communities and natural processes have been impacted by forest disturbance, and identify species that are more at risk. In this way, it can contribute to the knowledge of Philippine birds and butterflies and their conservation.
Results showed that forest species seem to be able to tolerate moderate levels of forest disturbance, as species richness and community composition of forest birds and butterflies were similar between closed canopy and open canopy forest. In addition, there was no significant difference in predation rates on nests and caterpillars between these two habitats, suggesting that their reproductive success is not impacted. On the other hand, conversion of forests into other land uses - agricultural, residential or commercial—has a significant effect on species richness and abundance. While some forest species can persist these modified habitats, important ecological processes such as predation are altered. How these changes affect species survival in disturbed habitats and how valuable mixed-use habitats are for conservation are possible directions for further research. The dissimilar responses of birds and butterflies to habitat modification also indicate that more data is needed on the impacts of disturbance on different taxa, since they cannot be inferred across groups.
62 Many Philippine species, especially endemics, depend on forested habitats. Canopy cover was found to affect the distribution of a number of forest bird species. Traits that were identified as correlates of extinction risk in birds were those that are related to threat from habitat loss. Species that are restricted by geographical range, elevation and habitat are more vulnerable, possibly because they lack the ecological flexibility to exploit the other habitats, especially those created by anthropogenic action.
The results reinforce the need to protect all remaining forests, especially in the lowlands, given the small percentage that remains on the islands. More research into species biology and ecology is urgently needed, especially for endemics with restricted ranges. Work should be done on various key taxonomic groups, since it has been shown that taxa have differing responses to disturbance. Monitoring populations and determining specific habitat requirements of particularly sensitive taxa would be key to their conservation. If management polices are to have solid scientific foundations, more ecological research must be conducted to identify priorities and broaden the poor biological knowledge of Philippine biodiversity.
63 References
Achard F, Eva HD, Stibig H-J, Mayaux P, Gallego J, Richards T, Malingreau JP.
2002. Determination of deforestation rates of the world’s humid tropical
forests. Science 297:999-1002.
Alcala EL, Paalan RB, Averia LT, Alcala AC. 2004. Rediscovery of the Philippine
Bare-Backed Fruit Bat (Dobsonia chapmani Rabor) on southwestern Negros
Island, the Philippines. Silliman Journal 45:123-136
Angelstam P. 1986. Predation on ground-nesting birds’ nests in relation to predator
densities and habitat edge. Oikos 47:365-373.
Balmford A. 1996. Extinction filters and current resilience: the significance of past
selection pressures for conservation biology. Trends in Ecology and
Evolution 11:193-196.
Barcelona JF, Cajano MO, Hadsall AS. 2006. Rafflesia baletei, another new Rafflesia
(Rafflesiaceae) from the Philippines. Kew Bulletin 61:231–237.
Bennett PM, Owens IPF. 1997. Variation in extinction risk among birds: chance or
evolutionary predesposition? Proceedings of the Royal Society of London
Series B 264:401-408.
Berry L, Lill A. 2003. Do predation rates on artificial nests accurately predict
predation rates on natural nests? The effects of nest type, egg type and nest-
site characteristics. Emu 103:207-214.
Bibby CJ, Burgess ND, Hill DA, Mustoe S. 2000. Bird Census Techniques. London
(UK):Academic Press.
BirdLife International. 2000. Threatened Birds of the World. Barcelona (Spain) and
Cambridge (UK): Lynx Edicions and Birdlife International.
64 BirdLife International. 2006. DataZone.
Downloaded on 3 January 2007.
Blackburn TM, Gaston KJ. 2002. Extrinsic factors and population sizes of threatened
birds. Ecology Letters 5:568-576.
Blair RB. 1999. Birds and butterflies along an urban gradient: surrogate taxa for
assessing biodiversity? Ecological Applications 9:164-170.
Blair RB, Launer AE. 1997. Butterfly diversity and human land use: species
assemblages along an urban gradient. Biological Conservation 80:113-125.
Boussekey M. 2000. An integrated approach to conservation on the Philippine or
Red-vented cockatoo Cacatua haematuropygia. International Zoo Yearbook
37:137-145.
Broad R, Cavanagh J. 1993. Plundering Paradise: The Struggle for the Environment
in the Philippines. Berkeley (CA): University of California Press, Ltd.
Brook BW, Sodhi NS, Ng PKL. 2003. Catastrophic extinctions follow deforestation
in Singapore. Nature 424:420-423.
Brooks TM, Pimm S, Kapos V, Ravilious C. 1999. Threat from deforestation to
montane and lowland birds and mammals in insular South-east Asia. Journal
of Animal Ecology 68:1061-1078.
Brown JH. 1984. On the relationship between abundance and distribution of species.
The American Naturalist 124:255-279.
Brown RM, Diesmos AC, Alcala AC. 2001. The state of Philippine herpetology and
the challenges for the next decade. Silliman Journal 42:18-87.
Brown KS Jr. 1991. Conservation of neotropical environments: insects as indicators.
In: Collins NM, Thomas JA, editors. The Conservation of Insects and Their
Habitats. London (UK): Academic Press. p 449-504.
65 Brown KS Jr, Freitas AVL. 2002. Butterfly communities of urban forest fragments in
Campinas, Sao Paolo, Brazil: Structure, instability, environmental correlates,
and conservation. Journal of Insect Conservation 6:217-231.
Bueser GLL, Bueser KG, Afan DS, Salvador DI, Grier JW, Kennedy RS, Miranda
HC Jr. 2003. Distribution and nesting density of the Philippine Eagle
Pithecophaga jefferyi on Mindanao Island, Philippines: What do we know
after 100 years? Ibis 145:130-135.
Butchart SHM, Stattersfield AJ, Bennun, LA, Shutes SM, Akçakaya HR, Baillie JEM,
Stuart SN, Hilton-Taylor C, Mace GM. 2004. Measuring global trends in the
status of biodiversity: red list indices for birds. PLOS Biology 2:2294-2304.
Butchart SHM, Collar NJ, Crosby MJ, Tobias JA. 2005. Lost and poorly known birds:
Top targets for birders in Asia. BirdingASIA 3:41-49.
Caldas A, Robbins RK. 2003. Modified Pollard transects for assessing tropical
butterfly abundance and diversity. Biological Conservation 110:211-219.
Canaday C. 1997. Loss of insectivorous birds along a gradient of human impact in
Amazonia. Biological Conservation 77:63-77.
Carpenter KE, Springer VG. 2005. The center of the center of marine shore fish
biodiversity: The Philippine islands. Environmental Biology of Fishes 72:467-
480.
Castelletta M, Sodhi NS, Subaraj R. 2000. Heavy extinctions of forest avifauna in
Singapore: lessons for biodiversity conservation in Southeast Asia.
Conservation Biology 14:1870-1880.
Clarke KR, Ainsworth M. 1993. A method of linking multivariate community
structure to environmental variables. Marine Ecology Progress Series 92:205-
219.
66 Collar NJ, Andrew P. 1988. Birds to Watch. Cambridge (UK): International Council
for Bird Preservation.
Collar NJ, Crosby MJ, Stattersfield AJ. 1994. Birds to Watch 2. Cambridge (UK):
BirdLife International.
Collar NJ, Mallari NAD, Tabaranza BR Jr. 1999. Threatened Birds of the Philippines.
Makati City (Philippines): Bookmark, Inc.
Collar NJ. 1998. Extinction by assumption; or, the Romeo error on Cebu. Oryx
32:239-244.
Colwell RK. 2004. EstimateS: Statistical Estimation of Species Richness and Shared
Species from Samples, Version 7.0b. User’s guide and application.
Colwell RK, Mao CX, Chang J. 2004. Interpolating, extrapolating, and comparing
incidence-based species accumulation curves. Ecology 85:2717-2727.
Cooper DS, Francis CM. 1998. Nest predation in a Malaysian lowland rainforest.
Biological Conservation 85:199-202.
Coronas J. 1920. The Climate and Weather of the Philippines. Manila (Philippines):
Bureau of Print.
Custodio CC, Molinyawe NM. 2001. The NIPAS Law and the management of
protected areas in the Philippines: Some observations and critique. Silliman
Journal 42:202-228.
Dalmacio MV. 2001. Subic Bay Protected Areas Management Plan Project Resource
Inventory Reports, Volume II – Flora resource inventory report. Unpublished
technical report. Manila (Philippines): Woodward-Clyde.
67 Danielsen F, Treadaway CG. 2004. Priority conservation areas for butterflies
(Lepidoptera: Rhopalocera) in the Philippine islands. Animal Conservation
7:79-92.
Davies KF, Margules CR, Lawrence JF. 2000. Which traits of species predict
population declines in experimental forest fragments? Ecology 81:1450-1461.
Del Hoyo J, Elliot A, Sargatal J, series editors. 1992–2006. Handbook of the Birds of
the World. Volumes 1-11. Barcelona (Spain): Lynx Edicions.
Dennis RLH. 2004. Butterfly habitats, broad-scale biotope affiliations, and structural
exploitation of vegetation at finer scales: the matrix revisited. Ecological
Entomology 29:744-752.
[DENR-PAWB] Department of the Environment and Natural Resources Protected
Areas and Wildlife Bureau. 2003. 2003 Statistics on Philippine Protected
Areas and Wildlife Resources.
September 2006.
Diesmos AC, Diesmos ML, Gee GVA. 2004. 'Lost' frogs and reptiles rediscovered on
Palawan Island, Philippines. Oryx 38:255.
Diesmos AC, Parham JF, Stuart BL, Brown RM. 2005. The phylogenetic position of
the recently rediscovered Philippine forest turtle (Bataguridae: Heosemys
leytensis). Proceedings of the California Academy of Sciences 56:31-41.
Dumbrell AJ, Hill JK. 2005. Impacts of selective logging on canopy and ground
assemblages of tropical butterflies: implications for sampling. Biological
Conservation 125:123-131.
Duncan RP, Blackburn TM. 2004. Extinction and endemism in the New Zealand
avifauna. Global Ecology and Biogeography 13:509-517.
68 Dunn RR. 2004. Managing the tropical landscape: a comparison of the effects of
logging and forest conversion to agriculture on ants, birds, and lepidoptera.
Forest Ecology and Management 191:215-224.
Dutson GCL, Magsalay PM, Timmins RJ. 1993. The rediscovery of the Cebu
Flowerpecker Dicaeum quadricolor, with notes on other forest birds on Cebu,
Philippines. Bird Conservation International 3:235-243.
Ehrlich PR, Raven PH. 1964. Butterflies and plants: a study in coevolution.
Evolution 18:586-608.
Estrada A, Rivera A, Coates-Estrada R. 2002. Predation of artificial nests in a
fragmented landscape in the tropical region of Los Tuxtlas, Mexico.
Biological Conservation 106:199-209.
Fahrmeir L, Tutz G. 1994. Multivariate Statistical Modelling Based on Generalized
Linear Models. New York (NY):Springer-Verlag.
Feeny P, Blau WS, Kareiva PM. Larval growth and survivorship of the black
swallow tail butterfly in central New York. Ecological Monographs 55:167-
187.
[FAO] Food and Agriculture Organization. 2000. Forest Resources Assessment:
2000 Main Report. Food and Agriculture Organization Forestry Paper 140.
Rome: Forestry Department, Food and Agricultural Organization of the
United Nations.
[FAO] Food and Agriculture Organization of the United Nations. 2005. Global
Forest Resources Assessment 2005. Country Report 202: Philippines. Rome:
Forestry Department, Food and Agricultural Organization of the United
Nations.
69 Furness RW, Greenwood JJD, Jarvis PJ. 1993. Can birds be used to monitor the
environment? In: Furness RW, Greenwood JJD, editors. Birds as Monitors of
Environmental Change. London (UK):Chapman and Hall. p 1-41.
Gaston KJ, Blackburn TM, Greenwood JJD, Gregory RD, Quinn RM, Lawton JH.
2000. Abundance-occupancy relationships. Journal of Applied Ecology 37:39-
59.
Gaston KJ. 1992. Regional numbers of insect and plant species. Functional Ecology
6:243-247.
Gaston KJ, Blackburn TM. 1995. Birds, body size and the threat of extinction.
Philosophical Transactions of the Royal Society of London Series B 347:205-
212.
Ghazoul J. 2002. Impact of logging on the richness and diversity of forest butterflies
in a tropical dry forest in Thailand. Biodiversity and Conservation 11:521-
541.
Gibbs JP. 1991. Avian nest predation in tropical wet forest: an experimental study.
Oikos 60:155-161.
Gomez ED, Aliño PM, Yap HT, Licuanan WY. 1994. A review of the status of
Philippine reefs. Marine Pollution Bulletin 29:62-68.
Gotelli NJ, Colwell RK. 2001. Quantifying biodiversity: procedures and pitfalls in
the measurement and comparison of species richness. Ecology Letters 4:379-
391.
Hamann A, Curio E. 1999. Interactions among frugivores and fleshy fruit trees in a
Philippine submontane rainforest. Conservation Biology 13:766-773.
Hamer KC, Hill JK, Benedick S, Mustaffa N, Sherratt TN, Maryati M, Chey VK.
2003. Ecology of butterflies in natural and selectively logged forests of
70 northern Borneo: the importance of habitat heterogeneity. Journal of Applied
Ecology 40:150-162.
Hamer KC, Hill JK, Lace LA, Langan AM. 1997. Ecological and biogeographical
effects of forest disturbance on tropical butterflies of Sumba, Indonesia.
Journal of Biogeography 24:67-75.
Hardin JW, Hilbe JM. 2003. Generalized Estimating Equations. Florida: Chapman
and Hall.
Haysom KA, Coulson JC. 1998. The Lepidoptera fauna associated with Calluna
vulgaris: effects of plant architecture on abundance and diversity. Ecological
Entomology 23:377-385.
Heaney LR. 1985. Zoogeographic evidence for middle and late Pleistocene land
bridges to the Philippines. Modern Quaternary Research of South-East Asia
9:127-143.
Heaney LR. 2001. A decade of research on Philippine mammals: progress and
challenges. Silliman Journal 42:88-108.
Heaney LR, Regalado JC. 1998. Vanishing Treasures of the Philippine Rainforest.
Chicago (IL): The Field Museum.
Hill JK, Hamer KC. 2004. Determining impacts of habitat modification on diversity
of tropical forest fauna: the importance of spatial scale. Journal of Applied
Ecology 41:744-754.
Horner-Devine MC, Daily GC, Ehrlich PR, Boggs CL. 2003. Countryside
biogeography of tropical butterflies. Conservation Biology 17:168-177.
Hughes JB, Daily GC, Ehrlich PR. 2002. Conservation of tropical forest birds in
countryside habitats. Ecology Letters 5:121-129.
71 Igarashi S, Fukuda H. 1997. The Life Histories of Asian Butterflies, Volume I.
Tokyo (Japan):Tokai University Press.
Igarashi S, Fukuda H. 2000. The Life Histories of Asian Butterflies, Volume II.
Tokyo (Japan):Tokai University Press.
[IUCN] International Union for Conservation of Nature and Natural Resources. 2006.
2006 IUCN Red List of Threatened Species.
Downloaded on 4 May 2006.
Jablonski D. 1991. Extinctions: A paleontological perspective. Science 253:754-757.
James FC, Shugart HH Jr. 1970. A quantitative method of habitat description.
Audubon Field Notes 24:727-736.
Johns AD. 1991. Responses of Amazonian rain forest birds to habitat modification.
Journal of Tropical Ecology 7:417-437.
Johnson TR. 1999. Community-based forest management in the Philippines. Journal
of Forestry 97:26-30.
Jones MJ, Sullivan MS, Mardsen SJ, Linsley MD. 2001. Correlates of extinction risk
of birds from two Indonesian islands. Biological Journal of the Linnean
Society 73:65-79.
Keller JK, Richmond ME, Smith CR. 2003. An explanation of patterns of breeding
bird species richness and density following clearcutting in northeastern USA
forests. Forest Ecology and Management 174:541-564.
Kenkel NC, Orloci L. 1986. Applying metric and nonmetric multidimensional scaling
to ecological studies: some new results. Ecology 67:919-928.
Kennedy RS, Gonzales PC, Dickinson EC, Miranda HC Jr, Fisher TH. 2000. A Guide
to the Birds of the Philippines. New York (NY):Oxford University Press.
72 Koh LP, Menge DNL. 2006. Rapid assessment of Lepidoptera predation rates in
neotropical forest fragments. Biotropica 38:1-3.
Koh LP, Sodhi NS. 2004. Importance of reserves, fragments and parks for butterfly
conservation in a tropical urban landscape. Ecological Applications 14:1695-
1708.
Koh LP, Sodhi NS, Brook BW. 2004a. Ecological correlates of extinction proneness
in tropical butterflies. Conservation Biology 18:1571-1578.
Koh LP, Sodhi NS, Brook BW. 2004b. Co-extinctions of tropical butterflies and their
hostplants. Biotropica 36:272-274.
Koh LP, Sodhi NS, Tan THT, Peh KS-H. 2002. Factors affecting the distribution of
vascular plants, springtails, butterflies and birds on small tropical islands.
Journal of Biogeography 29:93-108.
Kremen C, Colwell RK, Erwin TL, Murphy DD, Noss RR, Sanjayan MA. 1993.
Terrestrial arthropod assemblages: their use in conservation planning.
Conservation Biology 7:796-808.
Kummer DM. 1992. Deforestation in the Postwar Philippines. Chicago (IL):Chicago
University Press.
Lasco RD, Puhlin JM. 2006. Environmental impacts of community-based forest
management in the Philippines. International Journal of Environment and
Sustainable Development 5:46–56.
Liang K-Y, Zeger SL. 1986. Longitudinal data analysis using generalized linear
models. Biometrika 73:13-22.
Lim HC, Sodhi NS. 2004. Responses of avian guilds to urbanization in a tropical
city. Landscape and Urban Planning 66:99-215.
73 Liu DS, Iverson LR, Brown SB. 1993. Rates and patterns of deforestation in the
Philippines: application of geographic information system analysis. Forest
Ecology and Management 57:1-16.
Loiselle BA, Hoppes WG. 1983. Nest predation in insular and mainland lowland
rainforest in Panama. Condor 85:93-95.
Loiselle BA, Farji-Brener AG. 2002. What’s up? An experimental comparison of
predation levels between canopy and understory in a tropical wet forest.
Biotropica 34:327-330.
Magno F. 2001. Forest devolution and social capital: State-civil society relations in
the Philippines. Environmental History 6:264-286.
Magsalay P, Brooks T, Dutson G, Timmins R. 1995. Extinction and conservation on
Cebu. Nature 373:294.
Major RE, Kendall CE. 1996. The contribution of artificial nest experiments to
understanding avian reproductive success: a review of methods and
conclusions. Ibis 138:298-307.
Malayang BS. 2000. The changing role of government in forest protection. In Utting
P, editor. Forest Policy and Politics in the Philippines. Quezon City
(Philippines): Ateneo de Manila University Press. p 40-66.
Mallari NAD, Tabaranza BR Jr, Crosby MJ. 2001. Key Conservation Sites in the
Philippines. Makati City (Philippines): Bookmark, Inc.
Martin TE. 1988a. Habitat and area effects on forest bird assemblages: is nest
predation an influence? Ecology 69:74-84.
Martin TE. 1988b. Processes organizing open-nesting bird assemblages: competition
or nest predation? Evolutionary Ecology 2: 7-50.
74 Martin TE. 1993a. Nest predation among vegetation layers and habitat types:
revisiting the dogmas. The American Naturalist 141:897-913.
Martin TE. 1993b. Nest predation and nest sites. BioScience 43:523-532.
McCune B, Grace JB. 2002. Analysis of Ecological Communities. Oregon:MjM
Software.
McCune B, Mefford MJ. 1999. PC-ORD Multivariate Analysis of Ecological Data,
Version 4.14. Oregon: MjM Software.
McGregor RC. 1907. Notes on birds collected in Cebu. Philippine Journal of Science
2:298-309.
McKinney ML. 1997. Extinction vulnerability and selectivity: combining ecological
and paleontological views. Annual Review of Ecology and Systematics
28:495-516.
McLain DK, Moulton MP, Sanderson JG. 1999. Sexual selection and extinction: the
fate of plumage-dimorphic and plumage-monomorphic birds introduced onto
islands. Evolutionary Ecology Research 1:549-565.
Miranda HC Jr, Salvador DI, Ibañez J, Ibañez GB. 2000. Summary of Philippine
Eagle reproductive success, 1978-1998. Journal of Raptor Research 34:37-41.
Morris DW. 2003. Shadows of predation: habitat-selecting consumers eclipse
competition between coexisting prey. Evolutionary Ecology 17:393-422.
Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J. 2000.
Biodiversity hotspots for conservation priorities. Nature 403:853-858.
Myers N. 1988. Environmental degradation and some economic consequences in the
Philippines. Environmental Conservation 15:205-214.
75 Norris K, Harper N. 2004. Extinction processes in hot spots of avian biodiversity and
the targeting of pre-emptive conservation action. Proceedings of the Royal
Society of London Series B 271:123-130.
Ong PS, Afuang LE, Rosell-Ambal RG, editors. 2002. Philippine Biodiversity
Conservation Priorities: A second iteration of the national biodiversity strategy
and action plan. Final Report. Quezon City (Philippines):Department of
Environment and Natural Resources-Protected Areas Management Bureau,
Conservation International Philippines, Biodiversity Conservation Program-
University of the Philippines Center for Integrative and Development Studies,
and Foundation for the Philippine Environment.
Ortega CP, Ortega JC, Rapp CA, Backensto SA. 1998. Validating the use of
artificial nests in predation experiments. Journal of Wildlife Management
62:925-932.
Owens IPF, Bennett PM. 2000. Ecological basis of extinction risk in birds: habitat
loss versus human persecution and introduced predators. Proceedings of the
National Academy of Science 97:12144-48.
Paguntalan LMJ, Pedregosa MDG, Gadiana MJC. 2004. The Philippine Bare-backed
Fruit Bat Dobsonia chapmani Rabor, 1952: Rediscovery and conservation
status on Cebu Island. Silliman Journal 45:113-122.
Pan W. 2001. Akaike’s information criterion in generalized estimating equations.
Biometrics 57:120-125.
Paradis E, Claude J. 2002. Analysis of comparative data using generalized estimating
equations. Journal of Theoretical Biology 218:178-185.
Peh KS-H, de Jong J, Sodhi NS, Lim SL-H, Yap CAM. 2005. Lowland rainforest
avifauna and human disturbance: persistence of primary forest birds in
76 selectively logged forests and mixed-rural habitats of southern Peninsular
Malaysia. Biological Conservation 123:489-505.
[PEFI] Philippine Eagle Foundation Incorporated. 2005. Annual Report.
Pimm SL and Raven P. 2000. Extinction by numbers. Nature 403:843-845.
Pinheiro JC, Bates DM. 2000. Mixed effects models in S and S-PLUS. Springer,
New York.
Pollard E, Yates TJ. 1993. Monitoring Butterflies for Ecology and Conservation.
London (UK): Chapman and Hall.
Posa, MRC and Sodhi NS. 2006. Effects of anthropogenic land use on forest birds and
butterflies in Subic Bay, Philippines. Biological Conservation 129:256-270.
Posa MRC, Sodhi NS and Koh LP. 2007. Predation on artificial nests and caterpillar
models across a disturbance gradient in Subic Bay, Philippines. Journal of
Tropical Ecology 23:27-33.
Primavera JH. 2000. Development and conservation of Philippine mangroves:
Institutional issues. Ecological Economics 35:91–106.
Purvis A, Gittleman JL, Cowlishaw G, Mace GM. 2000. Predicting extinction risk in
declining species. Proceedings of the Royal Society of London Series
267:1947-1952.
Rabor DS. 1959. The impact of deforestation on the birds of Cebu, Philippines, with
new records for that island. The Auk 76:37-43.
Raman TRS, Rawat GS, Johnsingh AJT. 1998. Recovery of tropical forest avifauna
in relation to vegetation succession following shifting cultivation in Mizoram,
Northeast India. Journal of Applied Ecology 35:214-231.
77 Reitsma LR, Whelan CJ. 2000. Does vertical partitioning of nest sites decrease nest
predation? The Auk 117:409-415.
Ricklefs RE. 1969. An analysis of nesting mortality in birds. Smithsonian
Contributions to Zoology 9:1-48.
Roberts CM, McClean CJ, Veron JEN, Hawkins JP, Allen GR, McAllister DE,
Mittermeier CG, Schueler FW, Spalding M, Wells F, Vynne C, Werner TB.
2002. Marine biodiversity hotspots and conservation priorities for tropical
reefs. Science 295:1280-1284.
Robinson GS, Ackery PR, Kitching IJ, Beccaloni GW, Hernandez LM. 2001.
Hostplants of the Moth and Butterfly Caterpillars of the Oriental Region.
Kuala Lumpur (Malaysia):United Selangor Press.
Robson C. 2000. A Field Guide to the Birds of South-east Asia. United
Kingdom:New Holland Publishers.
Rodewald AD, Yahner RH. 2001. Avian nesting success in forested landscapes:
influence of landscape composition, stand and nest-patch microhabitat, and
biotic interactions. The Auk 118:1018-1028.
Russ GR, Alcala AC. 1999. Management histories of Sumilon and Apo Marine
Reserves, Philippines, and their influence on national marine resource policy.
Coral Reefs 18:307-319.
Russell GJ, Brooks TM, McKinney MM, Anderson CG. 1998. Present and future
taxonomic selectivity in bird and mammal extinctions. Conservation Biology
12:1365-1376.
Schlaepfer MA, Runge MC, Sherman PW. 2004. Ecological and evolutionary traps.
Trends in Ecology and Evolution 17:474-480.
78 Settele J. 1993. Lepidopterological research in the Philippines – a short survey.
Nachrichten des Entomologischen Vereins Apollo Supplement 12:12-24.
Shiojiri K, Takabayashi J. 2003. Effects of specialist parasitoids on oviposition
preference of phytophagous insects: encounter-dilution effects in a tritrophic
interaction. Ecological Entomology 28:573-578.
Sibley CG, Ahlquist AE. 1990. Phylogeny and Classification of Birds: A Study in
Molecular Evolution. New Haven (CT):Yale University Press.
Sieving KE, Willson MF. 1998. Nest predation and avian species diversity in
northwestern forest understory. Ecology 79:2391-2402.
Sih A, Jonsson BG, Luikart G. 2000. Habitat loss: ecological, evolutionary and
genetic consequences. Trends in Ecology & Evolution 15:132-134.
Smith FDM, May RM, Pellew R, Johnson TH, Walter KR. 1993. How much do we
know about the current extinction rate? Trends in Ecology and Evolution
8:375-378.
Söderström B. 1999. Artificial nest predation rates in tropical and temperate forests: a
review of the effects of edge and nest site. Ecography 22:455-463.
Söderström B, Part T, Ryden J. 1998. Different nest predator faunas and nest
predation risk on ground and shrub nests at forest ecotones: an experiment and
a review. Oecologia 117:108-118.
Sodhi NS, Brook BW. 2006. Southeast Asian Biodiversity In Crisis. Cambridge
(UK): Cambridge University Press.
Sodhi NS, Liow LH. 2000. Improving conservation biology research in Southeast
Asia. Conservation Biology 14:1211-1212
79 Sodhi NS, Briffett C, Kong LLL, Yuen BKP. 1999. Bird use of linear areas of a
tropical city: implications for park connector design and management.
Landscape and Urban Planning 45:123-130.
Sodhi NS, Peh KS-H, Lee TM, Turner IM, Tan HTW, Prawiradilaga DM, Darjono.
2003. Artificial nest and seed predation experiments on tropical southeast
Asian islands. Biodiversity and Conservation 12:2415-2433.
Sodhi NS, Koh LP, Brook, BW, Ng PKL. 2004a. Southeast Asian biodiversity: the
impending disaster. Trends in Ecology and Evolution 19:654-660
Sodhi NS, Liow LH, Bazzaz FA. 2004b. Avian extinctions from tropical and
subtropical forests. Annual Review of Ecology, Evolution and Systematics
35:323-345.
Sodhi NS, Koh LP, Prawiradilaga DM, Darjono, Tinulele I, Putra DD, Tan THT.
2005. Land use and conservation value for forest birds in Central Sulawesi
(Indonesia). Biological Conservation 122:547-558.
Sodhi NS, Brooks TM, Koh LP, Acciaioli G, Erb M, Tan AK-J, Curran LM, Brosius
P, Lee TM, Patlis JM, Gumal M, Lee RJ. 2006. Biodiversity and human
livelihood crises in the Malay Archipelago. Conservation Biology 20:1811-
1813.
Sorci G, Moller AP, Clobert JC. 1998. Plumage dichromatism of birds predicts
introduction success in New Zealand. Journal of Animal Ecology 67:263-269.
Stibig H-J, Achard F, Fritz S. 2004. A new forest cover map of continental Southeast
Asia derived from SPOT-VEGETATION satellite imagery. Applied
Vegetation Science 7:153-162.
Telleria JL, Diaz M. 1995. Avian nest predation in a large natural gap of the
Amazonian rainforest. Journal of Field Ornithology 66:343-351.
80 Terborgh J. 1999. Requiem For Nature. Washington (DC): Island Press. van der Ploeg J, van Weerd M. 2004. Devolution of natural resource management and
crocodile conservation: The case of San Mariano, Isabela. Philippine Studies
3:345-382.
Vitousek PM, Mooney HA, Lubchenco J, Melillo JM. 1997. Human domination of
earth's ecosystems. Science 277:494-499.
Vitug MD. 1993. The Politics of Logging: Power From the Forest. Manila
(Philippines): Philippine Center for Investigative Journalism.
Walpole MJ, Sheldon IR. 1999. Sampling butterflies in tropical rainforest: an
evaluation of a transect walk method. Biological Conservation 87:85-91.
Walters BB. 2003. People and mangroves in the Philippines: fifty years of
environmental change. Environmental Conservation 30:293-303.
Walther RA, Morand S. 1998. Comparative performance of species richness
estimation methods. Parasitology 116:395-405.
White AT, Courtney CA, Salamanca A. 2002. Experience with marine protected area
planning and management in the Philippines. Coastal Management 30:1–6.
White AT, Vogt HP and Arin T. 2000. Philippine coral reefs under threat: economic
losses caused by reef destruction. Marine Pollution Bulletin 40:598-605.
White AT, Vogt HP. 2000. Philippine coral reefs under threat: Lessons learned after
25 years of community-based reef conservation. Marine Pollution Bulletin
40:537–550.
Widmann P, Widmann IDL, Diaz SH, van den Beukel DV, Cruz R. 2006. Potentials
and limitations of community-based parrot conservation projects – the
example of the Philippine Cockatoo Conservation Program. Proceedings of
the 6th International Parrot Convention, Tenerife, Spain.
81 Wilcove DS. 1985. Nest predation in forest tracts and the decline of migratory
songbirds. Ecology 66:1211-1214.
[WCSP] Wildlife Conservation Society of the Philippines. 1997. Philippine Red Data
Book. Makati City (Philippines): Bookmark, Inc.
Wong TCM, Sodhi NS, Turner IM. 1998. Artificial nest and seed predation
experiments in tropical lowland rainforest remnants of Singapore. Biological
Conservation 85: 97-104.
Woodward-Clyde, 2001. Subic Bay Protected Areas Management Plan Project,
Volume I – Introduction and area ecological profile. Unpublished technical
report. Manila (Philippines): Woodward-Clyde.
[WRI] World Resources Institute. 2003. Earthtrends
Downloaded on 23 August 2006.
Zanette L. 2002. What do artificial nests tell us about nest predation? Biological
Conservation 103:323-329.
82 TABLES
83 Table 1. Number of samples, forest species and individuals observed in the five habitat types in the Subic Bay Watershed Reserve. Values in parentheses represent the percentage of total number of forest species recorded. Means are given with standard error.
Closed Open Suburban Rural Urban Canopy Canopy No. of Point Counts 95 90 72 78 82 Forest Bird Species 43 (89.58) 43 (89.58) 26 (54.17) 18 (37.50) 5 (10.42) Forest Bird Individuals 558 410 347 212 49 Mean No. Spp./Point 4.60 ± 0.26 3.59 ± 0.24 2.79 ± 0.21 1.80 ± 0.17 0.33 ± 0.07 Mean No. Indiv./Point 5.87 ± 0.36 4.56 ± 0.32 4.82 ± 0.45 2.72 ± 0.29 0.60 ± 0.17
No. of Transect Walks 94 90 74 76 79 Forest Butterfly Species 27 (87.10) 28 (90.32) 17 (54.84) 23 (74.19) 5 (16.13) Forest Butterfly Indiv. 276 267 105 89 13 Mean No. Spp./Transect 2.16 ± 0.18 2.31 ± 0.21 1.07 ± 0.14 0.88 ± 0.14 0.15 ± 0.06 Mean No. Indiv./Point 2.84 ± 0.25 2.97 ± 0.30 1.42 ± 0.21 1.17 ± 0.20 0.16 ± 0.06 Indiv./Transect
84 Table 2. Nonparametric species richness estimators and curve models for asymptotic species richness for the five habitat types in the Subic Bay Watershed Reserve. Mean species richness given with standard error. Closed Open Canopy Canopy Suburban Rural Urban Forest Birds ACE 46.47 47.65 31.39 23.97 5.00 ICE 47.97 48.01 30.19 22.19 5.55 Chao1 45.58 50.25 31.33 23.63 5.00 Chao2 48.76 55.50 29.12 23.63 5.13 Jack1 50.92 51.90 31.92 22.94 5.99 Jack2 54.87 58.77 33.92 26.85 6.00 Bootstrap 46.74 46.90 28.87 20.03 5.50 MMRuns 45.70 48.06 28.62 19.28 5.51 MMMean 45.75 48.36 28.52 19.41 5.97
Mean Species Richness 48.08 ± 1.03 50.6 ± 1.36 30.43 ± 0.62 22.44 ± 0.83 5.52 ± 0.14 Proportion detected by surveys 89.43 84.98 85.44 80.21 90.57
Forest butterflies ACE 27.38 30.25 19.23 30.42 6.95 ICE 27.42 30.79 18.9 30.2 6.75 Chao1 27.01 29.63 19.22 35.5 5.75 Chao2 27.01 31.61 19.22 30.38 5.75 Jack1 27.99 32.94 20.95 31.88 6.97 Jack2 21.22 35.9 22.92 36.8 7.96 Bootstrap 28.53 30.32 18.83 27.13 5.92 MMRuns 31.39 31.25 21.62 43.87 7.99 MMMean 31.32 31.43 21.22 35.1 8.43
Mean Species Richness 27.70 ± 0.99 31.57 ± 0.63 20.23 ± 0.49 33.48 ± 1.66 6.94 ± 0.34 Proportion detected by surveys 97.47 88.69 84.03 68.70 72.05
85 Table 3. Mean abundance of forest bird and butterfly species detected in the five habitat types the Subic Bay Watershed Reserve. Endemic species are in boldface, * indicate bird species that were significantly affected by canopy loss.
Mean Abundance Spp. Closed Open No. Forest Bird Species Canopy Canopy Suburban Rural Urban 1 Gallus gallus* 0.053 0.133 0 0 0 2 Turnix ocellata 0.011 0 0 0 0 3 Dendrocopus maculatus 0.032 0.022 0.111 0.128 0 4 Dryocopus javensis* 0.095 0.033 0.014 0 0 5 Chrysocopates lucidus* 0.137 0.100 0.028 0 0 6 Mulleripicus funebris 0.021 0.022 0 0 0 7 Megalaima haemacephala 0.074 0.033 0.222 0 0 8 Penelopides manillae* 0.084 0.044 0.014 0 0 9 Hierococcyx fugax 0 0.011 0 0 0 10 Surniculus lugubris 0.011 0.011 0 0 0 11 Eudynamys scolopacea 0.011 0.022 0 0 0 12 Phaenicophaeus superciliosus* 0.137 0.078 0.069 0 0 13 Phaenicophaeus cumingii* 0.042 0.056 0.014 0 0 14 Centropus viridis* 0.505 0.278 0.097 0.423 0 15 Centropus unirufus* 0.084 0.133 0 0 0 16 Bolbopsittacus lunulatus* 0.358 0.233 0.208 0.038 0 17 Prioniturus luconensis* 0.032 0.056 0.014 0 0 18 Tanygnathus lucionensis 0.095 0.078 0.181 0 0 19 Loriculus philippensis 0.084 0.089 0.194 0.218 0 20 Hemiprocne comata 0.011 0.011 0 0 0 21 Chalcophaps indica* 0.021 0.044 0 0.013 0 22 Phapitreron leucotis 0.179 0.078 0.042 0.192 0 23 Treron vernans 0.042 0 0.014 0 0 24 Treron pompadora* 0.011 0.011 0 0 0 25 Ducula aenea* 0.168 0.089 0 0 0 26 Amaurornis phoenicurus 0 0.011 0 0.064 0 27 Haliastur Indus 0.011 0 0 0 0 28 Microhierax erythrogenys 0 0.011 0 0 0 29 Pitta erythrogaster* 0.021 0.011 0 0 0 30 Corvus macrorhynchos 0.105 0.133 1.194 0.179 0.232 31 Artamus leucorhynchus 0.032 0.044 0.444 0.064 0.085 32 Oriolus chinensis 0.274 0.189 0.458 0.231 0.037 33 Coracina striata 0.253 0.111 0.250 0 0 34 Coracina coerulescens* 0.347 0.167 0 0 0 35 Rhipidura cyaniceps* 0.063 0.056 0 0 0 36 Dicrurus balicassius* 0.400 0.444 0.139 0 0 37 Hypothymis azurea* 0.063 0.033 0 0.013 0 38 Cyornis rufigastra 0 0.011 0 0.013 0 39 Copsychus luzoniensis* 0.337 0.178 0.014 0.026 0 40 Aplonis panayensis 0 0 0.056 0 0.220
86 Mean Abundance Spp. Closed Open No. Forest Bird Species Canopy Canopy Suburban Rural Urban 41 Sarcops calvus* 0.189 0.144 0.181 0.013 0 42 Parus elegans* 0.032 0.044 0 0 0 43 Ixos philippinus* 1.105 0.867 0.694 0.641 0.024 44 Zoosterops meyeni 0.021 0 0 0.346 0 45 Orthotomus derbianus* 0.232 0.200 0.056 0.103 0 46 Rhabdornis mysticalis 0.021 0.067 0.083 0 0 47 Dicaeum hypoleucum* 0.042 0.044 0 0.013 0 48 Dicaeum pygmaeum* 0.032 0.122 0.028 0 0 Forest Butterfly Species 1 Papilio alphenor 0.053 0.256 0.081 0.039 0.025 2 Achillides daedalus 0.021 0.022 0 0.013 0 3 Menelaides rumanzovia 0.011 0.011 0 0 0 4 Graphium agamemnon 0.032 0.022 0.014 0.013 0 5 Lamproptera meges 0.043 0.056 0.014 0.013 0 6 Gandaca harina 0.745 0.233 0.041 0.039 0 7 Leptosia nina 0 0.022 0.108 0.039 0.076 8 Cepora aspasia 0.053 0.011 0 0.026 0 9 Appias nero 0.096 0.067 0 0.013 0 10 Appias nephele 0.128 0.211 0.081 0.053 0 11 Pareronia boebera 0.032 0.067 0.122 0.053 0 12 Hebomoia glaucippe 0.106 0.178 0.419 0.039 0 13 Cupha arias 0 0.044 0 0.013 0 14 Vindula dejone 0.021 0.033 0 0 0 15 Junonia hedonia 0.085 0.067 0.095 0.013 0.013 16 Hypolimnas bolina 0.021 0.011 0.081 0.039 0.038 17 Cyrestis maenalis 0.021 0 0 0 0 18 Bassarona piratica 0 0.011 0 0 0 19 Dophla evelina 0 0.033 0 0 0 20 Faunis phaon 0.053 0.111 0 0 0 21 Melanitis boisduvalia 0.064 0.033 0 0 0 22 Elymnias melias 0.074 0.078 0.027 0.013 0 23 Zethera pimplea 0.255 0.433 0.014 0.026 0 24 Ptychandra leucogyne 0.138 0.256 0 0.013 0 25 Mycalesis ita 0.128 0.067 0 0.013 0 26 Mycalesis tagala 0.021 0 0 0 0 27 Mycalesis spp.(mineus/igoleta) 0.074 0.211 0.054 0.474 0 28 Ypthima spp. (stellera/sempera) 0.202 0.111 0.014 0.105 0 29 Ideopsis juventa 0.021 0 0.027 0.039 0 30 Euploea mulciber 0.266 0.156 0.162 0.039 0.013 31 Phaedyma columella 0.170 0.122 0.054 0.039 0
87 Table 4. Variance in bird and butterfly community composition represented by the final 3 ordination axes in nonmetric multidimensional scaling analysis and Spearman correlation coefficients for the most strongly correlated (r > 0.5 for birds, r > 0.4 for butterflies) habitat variables.
Axis 1 Axis 2 Axis 3 Birds Variance explained (r2) 0.211 0.244 0.159
Correlation with axis (r) Light Intensity 0.608 –0.387 % Canopy Cover –0.669 0.455 Tree Density –0.663 0.420 Height to Inversion –0.586 0.342 % Ground Cover –0.467 0.538 % Paved Ground 0.403 0.640 Number of Buildings 0.356 –0.531
Butterflies Variance explained (r2) 0.095 0.147 0.198
Correlation with axis (r) Light Intensity –0.251 –0.489 Litter Depth 0.045 0.488
88 Table 5. Results of pairwise comparisons between habitats using multi-response permutation procedures. All statistics were p < 0.001 except for * which was p < 0.05. More negative values of T shows greater separation between groups. The A statistic describes within-group homogeneity.
Habitat Pair Birds Butterflies T A T A Closed canopy vs. Open canopy –1.969 0.008 –8.432 0.033 * Closed canopy vs. Suburban –43.942 0.201 –25.065 0.116 Closed canopy vs. Rural –54.769 0.232 –19.514 0.086 Closed canopy vs. Urban –101.143 0.432 –32.373 0.138 Open canopy vs. Suburban –32.647 0.152 –15.639 0.076 Open canopy vs. Rural –44.426 0.190 –9.193 0.043 Open canopy vs. Urban –91.281 0.372 –24.569 0.112 Suburban vs. Rural –25.167 0.120 –6.333 0.037 Suburban vs. Urban –50.811 0.213 –7.604 0.047 Rural vs. Urban –75.775 0.339 –12.319 0.071
89 Table 6. Parameter estimates from univariate general estimating equations on ecological traits used to predict species vulnerability to disturbance. Smaller values of the estimate indicate traits of species with a higher probability of being vulnerable to disturbance. Z scores test significant difference of parameter estimates from the category with an estimate of 0.
Trait Category Estimate z p Birds (n = 68) Endemicity Endemic 0 Non-Endemic 2.065 2.38 0.017
Dichromatism Yes 0 No 0.553 1.28 0.202
Body Length - –0.004 –1.58 0.114
Feeding Guild Frugivores 0.481 0.62 0.537 Insectivores 0.191 0.24 0.808 Frugivore/Insectivore 1.164 1.41 0.158 Carnivore/Omnivore 0
Nest Location Ground 1.888 2.11 0.035 Shrub 0.954 2.19 0.028 Arboreal 0
Clutch Size - 0.142 0.75 0.454
Butterflies (n = 44) Endemicity Endemic 0 Non-endemic 2.071 3.55 0.0004
Dichromatism Yes 0 No 0.580 1.79 0.073
Forewing Length - –0.006 -0.28 0.780
No. Host Plants - 0.153 2.68 0.007
90 Table 7. QICu values of candidate models for species vulnerability using significant ecological traits as predictors. Lower values of QICu indicate better fit.
Covariates QICu Birds (n = 68) Endemicity + Nest Location 93.131 Endemicity 83.857 Nest Location 59.269
Butterflies (n = 44) Endemicity + No. Host Plants 70.295 Endemicity 53.530 No. Host Plants 69.514
91 Table 8. Numbers of artificial nests and caterpillar models predated at different habitats and height categories. Predator type was determined from nest remains and bite marks on plasticine (incisor width > 2 mm for large mammals and < 2 mm for small mammals). Set-ups where neither eggs nor caterpillar models were recovered were considered missing/not determined.
Artificial nests Caterpillar models Closed Open Rural Closed Open Rural canopy canopy canopy canopy Ground (total) 19 14 43 22 30 43 Missing/not determined 9 3 6 4 2 3 Large mammals 4 2 - - - - Small mammals 6 7 23 2 - 8 Birds - 2 11 7 24 31 Reptiles - - 2 - - 1 Arthropods - - - 9 4 -
1-1.5 m (total) 16 6 27 27 28 35 Missing/not determined 6 1 13 2 4 1 Large mammals 9 2 - - - - Small mammals 1 1 2 1 1 2 Birds - 2 11 6 17 29 Reptiles - - 1 - - - Arthropods - - - 18 6 3
>5 and > 3 m (total) 24 10 35 34 33 29 Missing/not determined 12 4 31 3 6 3 Large mammals 10 1 - 1 - - Small mammals - 1 1 - - - Birds 2 4 2 6 16 23 Reptiles ------Arthropods - - - 24 11 3
92 Table 9. Single-fixed effect models of probability of nest and caterpillar predation (df for both models = 537) with habitat and height as predictors and the inclusion of transect and plot as nested clustering variables to control for spatial autocorrelation.
Fixed effect Estimate t p Nest predation Habitat Closed-canopy forest (reference) - - - Open-canopy forest –0.16 –1.33 0.184 Rural areas 0.26 1.90 0.058
Height Ground (reference) - - - 1-1.5 m –0.15 –2.20 0.028 > 5 m –0.04 –0.82 0.413
Caterpillar predation Habitat Closed-canopy forest (reference) - - - Open-canopy forest 0.04 0.40 0.689 Rural areas 0.13 2.21 0.027
Height Ground (reference) - - - 1-1.5 m –0.03 –0.37 0.713 > 3 m 0.01 0.07 0.943
93 Table 10. Minimal adequate model of nest predation probability (loglikelihood = –326; df = 531, AIC = 707). The full model included both predictors and their interaction term and transect and plot as nested clustering variables to control for spatial autocorrelation. Model simplification was based on Akaike's Information Criterion.
Fixed effect Estimate t p Nest predation Intercept 1.32 12.0 <0.001
Habitat Closed-canopy forest (reference) - - - Open-canopy forest –0.08 –0.56 0.576 Rural areas 0.40 2.52 0.012
Height Ground (reference) - - - 1-1.5 m –0.05 –0.49 0.624 > 5 m 0.08 0.82 0.414
Habitat x height Open-canopy forest x 1-1.5 m –0.08 –0.52 0.603 Rural areas x 1-1.5 m –0.22 –1.35 0.177 Open-canopy forest x > 5 m –0.15 –0.94 0.350 Rural areas x > 5 m –0.22 –1.35 0.177
94 Table 11. Parameter estimates from univariate generalized estimating equations using traits to predict extinction risk for resident Philippine bird species with family as included as a clustering variable (n = 396). For each trait, categories with more positive parameter estimates have a higher risk of extinction relative to categories with a parameter estimate of 0. Z scores test whether the parameter estimate for a category differs significantly from category with parameter of zero.
Trait Category Estimate z p Range Single-Island Endemic 3.69 5.13 <0.0001 Multiple-Island Endemic 3.41 4.88 <0.0001 Widespread 0
Elevation Lowland 1.47 4.95 <0.0001 Montane 0.68 1.24 0.213 Widespread 0
Dichromatism Absent 0.03 0.12 0.905 Present 0
Length 0.002 1.10 0.271
Diet Invertebrate –0.51 0.20 0.610 Vertebrate –0.02 0.61 0.984 Fruit 0.07 0.95 0.947 Seed/Nectar/Plant Material –0.39 0.70 0.699 Scavenger/Omnivorous 0
Habitat Breadth Habitat Restricted 1.55 4.12 <0.0001 Habitat Not Restricted 0
Diet Breadth Diet Restricted 1.22 1.96 0.050 Diet Not Restricted 0
Nesting Strata Ground –0.45 –1.27 0.202 Shrub –0.69 –2.09 0.036 Arboreal 0
Clutch Size 0.03 0.23 0.816
95 Table 12. Minimum adequate model of extinction risk in Philippine birds (df=388, QICu = 248.62) using significant ecological traits as predictors and family as the clustering variable. For each trait, categories with more positive parameter estimates have a higher risk of extinction relative to categories with a parameter estimate of 0. Z scores test whether the parameter estimate for a category differs significantly from category with parameter of zero.
Trait Category Estimate z p Intercept –4.98 –6.37 <0.0001
Range Single-Island Endemic 3.92 5.30 <0.0001 Multiple-Island Endemic 3.37 4.48 <0.0001 Widespread 0
Elevation Lowland 0.88 2.66 0.008 Montane –0.95 –1.68 0.093 Widespread 0
Habitat Breadth Habitat Restricted 0.88 2.56 0.010 Habitat Not Restricted 0
Nesting Strata Ground 0.76 1.46 0.144 Shrub –0.71 –2.33 0.020 Arboreal 0
96 Table 13. Resident Philippine birds that possess traits identified as correlates of extinction risk and are not currently listed as threatened.
Species Common name Level of endemicity Nest site Spilornis holospilus Philippine serpent-eagle Multiple islands Arboreal Microheirax erythrogenys Philippine falconet Multiple islands Arboreal Mearnsia picina Philippine needletail Multiple islands Arboreal Penelopides samarensis Samar hornbill Multiple islands Arboreal Penelopides affinis Mindanao hornbill Multiple islands Arboreal Aceros leucocephalus Mindanao wrinkled hornbillMultiple islands Arboreal Coracina coerulescens Blackish cuckooshrike Multiple islands Arboreal Rhipidura cyaniceps Blue-headed fantail Multiple islands Arboreal Rhabdornis grandis Long-billed rhabdornis Single island Arboreal Rhabdornis inornatus Stripe-breasted rhabdornis Multiple islands Arboreal Oriolus albiloris White-lored oriole Single island Arboreal Dicrurus balicassius Balicassiao Multiple islands Arboreal
97 FIGURES
98 Figure 1. Map of the Philippine archipelago showing approximate percentages (pie charts) and distribution of forest cover (including secondary forest and plantations) remaining on the major islands. Map modified from Stibig et al. (2004).
99 Figure 2. Numbers and stability of Philippine bird species considered threatened in four global conservation assessments for the IUCN Red List (Collar and Andrew 1988, Collar et al. 1994, BirdLife International 2000, BirdLife International 2006). Bars indicate numbers of species considered threatened in a given assessment, with shading showing if they are also considered threatened in the preceding and following assessments (solid grey), not considered threatened in the preceding assessment but considered threatened in the subsequent one (left-rising stripes), also considered threatened in the preceding assessment but not in the subsequent one (right-rising stripes), or considered threatened in neither the preceding nor the subsequent assessment (vertical stripes).
100 120
100
80
60
40
20 No. of Publications
0 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Year
Figure 3. Number of publications on Philippine biodiversity and conservation obtained from searching three ISI Web of Knowledge databases (Biosis Previews, Web of Science, Zoological Records) using the search words TS=((Philippin*) AND (Biodiversity OR Biological Diversity OR Conservation)).
101 180
160
140
120
100
80
60
40
20
Number of individuals in attendance 0
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Year
Figure 4. Growth in attendance at the annual symposium on biodiversity by the Wildlife Conservation Society of the Philippines.
102 Figure 5. Map of the study area showing the five habitat types surveyed in the Subic Bay Watershed Reserve. Inset: approximate location of Subic Bay in Luzon island. Map adapted from Subic Bay Protected Areas Management Plan Project reports (Woodward- Clyde 2001)
103 50 Open Canopy
40 Closed Canopy
30 Suburban
20 Rural Number of species
10
Urban
0 0 20 40 60 80 100 Number of points
Figure 6. Species accumulation curves for forest birds in the five habitat types.
104 30 Open Canopy
25 Closed Canopy Rural
20
Suburban 15
Number of species 10
5 Urban
0 0 20 40 60 80 100 Number of transects
Figure 7. Species accumulation curves for forest butterflies in the five habitat types.
105 50 Open Canopy
40 Closed Canopy
30 Suburban
20 Rural Number of species
10
Urban
0 0 100 200 300 400 500 600 Number of individuals
Figure 8. Rarefaction curves of forest bird species richness in the five habitat types.
106 30 Open Canopy
25 Rural Closed Canopy
20
Suburan 15
Number of species 10
Urban 5
0 0 50 100 150 200 250 300 Number of individuals
Figure 9. Rarefaction curves of forest butterfly species richness in the five habitat types.
107 600 Closed Canopy
500
400 Suburban
Open Canopy 300
Rural 200 Number of individuals
100 Urban
0 0 20 40 60 80 100 Number of points
Figure 10. Population densities of forest birds in the five habitat types.
108 300
Open Canopy 250
Closed canopy
200
150
Suburban 100 Number of individuals Rural
50
Urban 0 0 20 40 60 80 100 Number of transects
Figure 11. Population densities of forest birds in the five habitat types.
109 1.0
0.5
Den Lit Can GCR Inv
0.0 NMS Axis 2 Lux Bld
Pav -0.5
-1.0
-1.0 -0.5 0.0 0.5 1.0 NMS Axis 1
Figure 12. Nonmetric multidimensional scaling ordination joint biplot of sample scores for the entire bird community with an overlay of strongly correlated (r > 0.5) habitat variables. BLD = building density; CAN = canopy cover; DEN = tree density (DBH >15 cm); GCR = ground cover; INV = height to inversion; LIT = litter depth; PAV = paved ground. Shaded circle = closed canopy forest; Open circle = open canopy forest; Shaded triangle = suburban habitat; Open triangle = rural habitat; Cross = Urban habitat; Shaded diamonds = forest species; Open Diamonds = non- forest species. See Table 3 for species codes.
110 1.0
28
20 23 2 43 0.5 38 42 Lit 5 Den GCR 21 19 37 Can 4 46 9 13 12 7 34 48 3 25 45 Inv 22 11 47 33 39 1 15 16 18 30 8 17 14 41 26 0.0 35 6 32 31 36 10 44 24
NMS Axis 2 29 Lux Bld
-0.5 Pav
40 -1.0
-1.0 -0.5 0.0 0.5 1.0 NMS Axis 1
Figure 13. Nonmetric multidimensional scaling ordination joint biplot of species scores for the entire bird community with an overlay of strongly correlated (r > 0.5) habitat variables. BLD = building density; CAN = canopy cover; DEN = tree density (DBH >15 cm); GCR = ground cover; INV = height to inversion; LIT = litter depth; PAV = paved ground. Shaded circle = closed canopy forest; Open circle = open canopy forest; Shaded triangle = suburban habitat; Open triangle = rural habitat; Cross = Urban habitat; Shaded diamonds = forest species; Open Diamonds = non-forest species. See Table 3 for species codes.
111 1.0
0.5
0.0 NMS Axis 3
-0.5
-1.0
-1.0 -0.5 0.0 0.5 1.0 NMS Axis 2
Figure 14. Nonmetric multidimensional scaling ordination of sample scores for the entire butterfly community. Shaded circle = closed canopy forest; Open circle = open canopy forest; Shaded triangle = suburban habitat; Open triangle = rural habitat; Cross = Urban habitat; Shaded diamonds = forest species; Open Diamonds = non-forest species. See Table 3 for species codes.
112 1.0
6 9 26 25 24 21 0.5 28 8 31 3
14 13 22 29 4 30 2 27 0.0 1 12 10 15 5 11 NMS Axis 3 29 18 17 20
-0.5 7 16
-1.0
-1.0 -0.5 0.0 0.5 1.0 NMS Axis 2
Figure 15. Nonmetric multidimensional scaling of species scores for the entire butterfly community. Shaded circle = closed canopy forest; Open circle = open canopy forest; Shaded triangle = suburban habitat; Open triangle = rural habitat; Cross = Urban habitat; Shaded diamonds = forest species; Open Diamonds = non-forest species. See Table 3 for species codes.
113 30
25
20
15
10
5
0
0 20 40 60 80 100 Canopy cover (%)
Figure 16. Results of simulations showing number of forest bird species present versus amount of canopy cover.
114 a.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0% Range Elevation Habitat Diet Nesting Breadth Breadth Strata b.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0% Range Elevation Habitat Diet Nesting Breadth Breadth Strata
Figure 17. Proportion of threatened (a) and nonthreatened (b) resident Philippine bird species with ecological traits that were significant correlates of extinction risk. Categories labels are as follows: SE = Single island endemics; ME = Multiple island endemics; LO = Lowland; MN = Montane; R = Restricted; NR = Not restricted; G = ground; S = shrub; A = Arboreal.
115