Impacts of Deforestation on Mosquito Community Dynamics

IMPACTS OF DEFORESTATION ON MOSQUITO COMMUNITY DYNAMICS

Hayley Louise Brant

Imperial College London

Centre for Environmental Policy, Faculty of Natural Sciences, Silwood Park Campus, Buckhurst Road, Ascot, SL5 7PY, UK

Thesis submitted for the degree of Doctor of Philosophy

September 2015

Declaration of Originality

I hereby certify that all content of this thesis is my original research and collaborations with other researchers are fully acknowledged. The experimental design, data collection and analysis of Chapter 5 was completed jointly with Borame Dickens, a fellow researcher at the Centre for Environmental Policy. We both contributed equally to this chapter.

Hayley Brant

Names of supervisors

Professor John Mumford

Dr Robert Ewers

The copyright of this thesis rests with the author and is made available under a Creative Commons Attribution Non-Commercial No Derivatives licence. Researchers are free to copy, distribute or transmit the thesis on the condition that they attribute it, that they do not use it for commercial purposes and that they do not alter, transform or build upon it. For any reuse or redistribution, researchers must make clear to others the licence terms of this work.

Abstract

Human-induced land use changes, including deforestation, agricultural encroachment and urbanisation, have caused widespread change in the global distribution of organisms and caused considerable declines in biodiversity through loss of habitat. Oil palm is one of the most rapidly expanding crops in Southeast Asia, but the impact of this crop on mosquito distribution, behaviour and exposure potential has been poorly explored. Understanding these factors is essential for developing, optimising and evaluating novel control measures aimed at reducing disease-transmission. This thesis explored the effect of land use change along an anthropogenic disturbance gradient (primary forest, disturbed forest, highly disturbed forest, oil palm plantations and rural housing estates) in Sabah, Malaysia. The community composition of anthropogenic mosquitoes was separated across land use, with the biggest difference seen between primary forest and oil palm plantations. This was largely driven by medically important mosquitoes attracted to oil palm plantations. Differences in community composition were also seen in areas of rural housing in comparison to primary and disturbed forest sites, due to a high presence of the dengue vector, Stegomyia albopicta, in housing areas. A higher abundance of anopheline vectors were found landing on humans in the disturbed forest and oil palm plantations then primary forest. This thesis found no difference between highly disturbed forest and oil palm plantation sites. This thesis also investigated the host-seeking behaviour of simian malaria vectors, by carrying out human landing catches at ground and canopy level across land use. Results demonstrated the potential ability of one of the vectors, Anopheles balabacensis, to transmit the simian malaria (Plasmodium knowlesi) between canopy-dwelling simian hosts and ground-dwelling humans, and that anthropogenic disturbance increases the abundance of the disease vector. Finally, this thesis investigated the use of different marking methods and the need for an improved dispersal experiment to be carried out.

Acknowledgements

This research was made possible by a Natural Environment Research Council (NERC) funded PhD studentship and financial support from the Stability of Altered Forest Ecosystems (S.A.F.E. project). I am extremely grateful to the Sabah Biodiversity Council (SaBC), the Royal Society Southeast Asia Rainforest Research Project (SEARRP), Maliau Basin Management Committee (MBMC), Danum Valley Management Committee (DVMC), Yayasan Sabah, Benta Wawasan, National Institute of Health (NIH), Institute for Medical Research (IMR), Medical Research and Ethics Committee (MREC) and the Ministry of Health (MOH) in Malaysia for research permits and permissions to carry out fieldwork. I would like to thank Dr Glen Reynolds, Dr Abdul Fatah Amir and Dr Suzan Benedick for their support during the research permit application process. I would like to thank all the S.A.F.E Project staff in Malaysia. I would like to thank the coordinators, MinSheng Khoo, Sarah Watson and Ryan Gray, for sorting out logistical problems, and my main research assistants, Mus, Mai, Harbin and Zinin for their help during mosquito collections. Thanks to Anand Nainar, Dr Terhi Riutta and Steven Hardwick for sending climate data, and to Rajeev Pillay for allowing me to borrow a laser rangefinder in the field. Thanks to the researchers I crossed over with in the field, especially the ones that left food for me after a long evening in the field. A special thanks to Unding for his help with tree climbing, and for training my research assistants. I would like to extend my thanks to the University of Malaya for their hospitality whilst I stayed in Kuala Lumpur. Thanks to Prof Rohela Mahmud for allowing me to work in her faculty, and to Dr Indra Vythilingam for allowing me to work in her laboratory. Thanks to her students for showing me around Kuala Lumpur. I am especially grateful to the Natural History Museum (NHM) for allowing me laboratory space and for helping me with my identifications. I am indebted to Dr Ralph Harbach for his assistance with sample identification, and to Theresa Howard, Dr Erica McAlister, Dr Duncan Sivell and Dr Daniel Whitmore for allowing me to work in the Diptera department. Thanks to the NHM volunteers for the tea breaks. I would like to express my deep gratitude to my supervisors, Prof John Mumford and Dr Robert Ewers, for their support and advice throughout the entirety of the thesis. Also to Dr Indra Vythilingam, Dr Chris Drakeley, Dr Jeffrey Hii, Dr Tilly Collins and Dr Donald Quicke for their valuable advice, for which I am extremely grateful. Lastly, I would like to thank my family and friends for supporting me through the PhD, especially Laurence- for everything.

Contents

Declaration of Originality ...... 3 Names of supervisors ...... 3 Copyright declaration ...... 3 Abstract ...... 4 Acknowledgements ...... 5 Chapter 1 – General Introduction ...... 15 1.1. Overview ...... 15 1.2. Global environmental change and land use change ...... 15 1.2.1. Tropical deforestation and oil palm expansion in Southeast Asia ...... 16 1.3. Human movement, land use change and mosquito-borne diseases ...... 17 1.4. Mosquito biology ...... 21 1.4.1. Host-seeking behaviour and dispersal ...... 22 1.5. Mosquito-borne diseases in Malaysia ...... 23 1.5.1. Dengue ...... 26 1.5.2. Japanese encephalitis, lymphatic filariasis and chikungunya ...... 26 1.5.3. Malaria ...... 27 1.6. Thesis structure and research questions ...... 31 Chapter 2 – The Effects of Land Use Change on Mosquito Community Composition ...... 33 2.1. Abstract ...... 33 2.2. Introduction ...... 34 2.3. Methods ...... 36 2.3.1. Study site ...... 36 2.3.2. Human landing catches ...... 39 2.3.3. Ovitraps ...... 41 2.3.4. Meteorological data ...... 41 2.3.5. Data analysis ...... 41 2.3.6. Ethics ...... 42 2.4. Results ...... 42 2.4.1. Mosquito abundance ...... 42 2.4.2. Effect of land use on mosquito abundance and presence ...... 48 2.4.3. Community composition ...... 52

2.5. Discussion ...... 54 2.6. Conclusions...... 57 Chapter 3: The Effects of Land Use Change on Anopheline Relative Abundance and Human Landing Rates ...... 58 3.1. Abstract ...... 58 3.2. Introduction ...... 59 3.3. Methods ...... 61 3.3.1. Study site ...... 61 3.3.2. Data collection ...... 62 3.3.3. PCR ...... 62 3.3.4. Meteorological data ...... 63 3.3.5. Data analysis ...... 63 3.3.6. Ethics ...... 63 3.4. Results ...... 63 3.5. Discussion ...... 69 3.6. Conclusions...... 72 Chapter 4 – Vertical Stratification of Adult Mosquitoes (Diptera: Culicidae) within a Tropical Rainforest in Sabah, Malaysia ...... 73 4.1. Abstract ...... 73 4.2. Introduction ...... 74 4.3. Methods ...... 76 4.3.1. Study site ...... 76 4.3.2. Data collection ...... 78 4.3.3. Meteorological data ...... 78 4.3.4. Data analysis ...... 78 4.3.5. Ethics ...... 79 4.4. Results ...... 79 4.4.1. Mosquito abundance ...... 79 4.4.2. Effect of height and land use on mosquito abundance ...... 83 4.4.3. Community composition ...... 87 4.5. Discussion ...... 88 4.6. Conclusions...... 90 Chapter 5 – Effects of Marking Methods and Fluorescent Dusts on Stegomyia aegypti (Aedes aegypti) Survival ...... 91

5.1. Abstract ...... 91 5.2. Introduction ...... 92 5.3. Methods ...... 94 5.3.1. Mosquitoes ...... 94 5.3.2. Marking of mosquitoes...... 94 5.3.3. Data analysis ...... 96 5.4. Results ...... 96 5.4.1. Survival analysis of dusted mosquitoes using method controls ...... 99 5.4.2. Survival analysis of dusted mosquitoes using immobilised controls ... 103 5.4.3. Marking efficiency ...... 107 5.5. Discussion ...... 109 5.6. Conclusions...... 112 Chapter 6 – General Discussion ...... 113 6.1. Overview ...... 113 6.2. Land use change and vector-borne diseases ...... 113 6.2.1. Human and simian malaria...... 114 6.2.2. Dengue...... 118 6.2.3. Marking techniques and dispersal ...... 118 6.3. Conclusions ...... 119 References ...... 121 Appendices ...... 143 Appendix A ...... 143 Appendix B ...... 147 Appendix C ...... 148

List of Figures Figure 1.1. Plasmodium malariae/Plasmodium knowlesi cases, by divisions of Sabah, Malaysia, from 2001-2013. Population of divisions in 2010: Interior 424,524; West Coast 1,011,725; Kudat 192,457; Sandakan 702,207; Tawau 819,955. Reproduced from William et al. (2014) ...... 30 Figure 2.1. Map of the Stability of Altered Forest Ecosystems Project, located in Sabah, Malaysia (a) Primary forest sites (b) Continuous twice-logged forest (c) Twice- logged forest and fragmented forest in an oil palm matrix (d) Oil palm plantation sites (e) The fragmentation experiment comprising of six blocks (A-F). Reproduced from Ewers et al. (2011) ...... 38 Figure 2.2. Anopheles landing hourly per person per night during the human landing catch pilot study (17:00-01:00h) in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest (d) Oil palm. Error bars show ± SE of the mean ...... 40 Figure 2.3. Species accumulation curves for human landing catch sampling in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest and (d) Oil palm. Shaded area indicates 95% confidence intervals ...... 46 Figure 2.4. Species accumulation curves for ovitrap sampling in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest and (d) Rural housing. Shaded area indicates 95% confidence intervals ...... 47 Figure 2.5. (a) Anophelines and (b) Culicines landing per person per night across an anthropogenic disturbance gradient: primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP). Error bars show ± SE of the mean ...... 49 Figure 2.6. Mean abundance (total mosquitoes/number of samples) of vector and non- vector species across an anthropogenic disturbance gradient: primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm plantation (OP) and rural housing (RU), using (a) human landing catches (b) ovitraps. Error bars show ± SE of the mean ...... 51 Figure 2.7. Detrended Correspondence Analysis (DCA) plot showing the major axes of variation for (a) Adult mosquitoes collected using human landing catches and in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation sites (OP) and (b) Mosquitoes collected using ovitraps in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and rural housing compounds (RU). The two axes represent linear summaries of the variation in the species numbers and areas ...... 53 Figure 3.1. Number of mosquitoes. (a) Anopheles balabacensis (b) Anopheles macarthuri and An. latens) landing per person per night across an anthropogenic disturbance gradient from primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP). Error bars show ± SE of the mean ...... 66

Figure 3.2. Hourly number of Anopheles balabacensis landing per person per night across an anthropogenic disturbance gradient from (a) Primary forest, (b) Disturbed forest, (c) Highly disturbed forest and (d) Oil palm plantation. Error bars show ± SE of the mean ...... 67 Figure 4.1. Map of the Stability of Altered Forest Ecosystems Project, located in Sabah, Malaysia (a) Primary forest sites (b) Continuous twice-logged forest (c) Twice- logged forest and fragmented forest in an oil palm matrix (d) Oil palm plantation sites (e) The fragmentation experiment comprising of six blocks (A-F). Reproduced from Ewers et al. (2011)...... 77 Figure 4.2. Species accumulation curves for human landing catch sampling in primary forest at (a) Ground and (b) Canopy level,virgin jungle reserve at (c) Ground and (d) Canopy level, and logged forest at (e) Ground and (f) Canopy level. Shaded area indicates 95% confidence intervals ...... 82 Figure 4.3. Effects of collection height on the human landing rate (number of mosquitoes per night per bait) across a forest disturbance gradient: primary forest (PF), lightly logged virgin jungle reserve (VJR), and twice-logged forest (LF). (a) Total abundance of all species combined, (b) Abundance of the most common species, Anopheles balabacensis, alone. Error bars show ± SE of the mean ...... 85 Figure 4.4. Hourly number of Anopheles balabacensis landing per person per night, at ground and canopy level, across an anthropogenic disturbance gradient from (a) Primary forest, (b) Virgin jungle reserve and (c) Logged forest. Error bars show ± SE of the mean ...... 86 Figure 4.5. Detrended Correspondence Analysis (DCA) plot showing the major axes of variation for adult mosquito abundance at ground level and in the canopy of a tropical rainforest. The two axes represent linear summaries of the variation in the species numbers and areas ...... 87 Figure 5.1. Survivorship plot of (a) Male and (b) Female Stegomyia aegypti marked using different methods (bag, bulb duster, dust storm & paint) vs. immobilised controls ...... 98 Figure 5.2. Survivorship plot of male Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. unmarked method controls ...... 101 Figure 5.3. Survivorship plot of female Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. unmarked method controls ...... 102 Figure 5.4. Survivorship plot of male Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. immobilised mosquitoes ...... 105

Figure 5.5. Survivorship plot of female Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. immobilised mosquitoes ...... 106 Figure 5.6. Marking efficiency of methods and colours. Marking efficiency (0–6) of male and female Stegomyia aegypti using different (a) Marking methods and (b) Colours of dust ...... 108 Figure A.1. Ovitrap design ...... 143 Figure A.2. DCA from Figure 2.2a, only showing only species labels. See Table A.1 for full species names ...... 144 Figure A.3. DCA from Figure 2.2b, only showing only species labels. See Table A.1 for full species names ...... 145 Figure C.1. DCA from Figure 4.5, only showing species labels. See Table C.1 for full species names ...... 148

List of Tables

Table 1.1. The key processes involved in disease transmission, following land use change (Martens & Hall 2000; Patz et al. 2000; Norris 2004; Afrane et al. 2005) .... 18 Table 1.2. Medically important mosquito species found in Malaysia (Reid 1968; Rahman, Che’rus & Ahmad 1997; Sallum et al. 2005; Vythilingam et al. 2005; Wiwanitkit 2007; Vythilingam 2012; WRBU 2015) ...... 24 Table 2.1. Mosquitoes collected from primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm (OP) and rural housing (RU) in the district of Tawau, Sabah, Malaysia ...... 44 Table 2.2. Mean species richness and diversity indices (± SE) of mosquito communities, collected using human landing catches, in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantations (OP) ...... 45 Table 2.3. Mean species richness and diversity indices (± SE) of mosquito communities, collected using ovitraps, in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm plantations (OP) and rural housing (RU) ..... 45 Table 2.4. Effects of land use and habitat characteristics on daily mosquito landings of anophelines and culicines in the primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP) ...... 50 Table 2.5. Effects of land use and habitat characteristics on ovitrap and Stegomyia albopicta presence in the primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and rural housing (RU) ...... 50 Table 3.1. Oligonucleotide sequences of PCR primers for detection of malaria parasites ...... 62 Table 3.2. Number of mosquitoes landing per night (18:00-23:00h) per human bait of anophelines collected along an anthropogenic disturbance gradient in the district of Tawau, Sabah, Malaysia (± SE of the mean) ...... 64 Table 3.3. Mosquitoes collected along an anthropogenic disturbance gradient in the district of Tawau, Sabah, Malaysia ...... 64 Table 3.4. Effects of land use and habitat characteristics on Anopheles balabacensis abundance. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 68 Table 3.5. Effects of land use and habitat characteristics on daily mosquito abundance of Anopheles macarthuri and Anopheles latens. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 68 Table 4.1. Mosquitoes collected from different collection sites in the district of Tawau, Sabah, Malaysia ...... 80 Table 4.2. Mean species richness and diversity indices (± SE) of mosquito communities, collected at ground level, using human landing catches, in primary forest (PF), virgin jungle reserve (VJR) and logged forest (LF) ...... 80

Table 4.3. Mean species richness and diversity indices (± SE) of mosquito communities, collected at canopy level, using human landing catches, in primary forest (PF), virgin jungle reserve (VJR) and logged forest (LF) ...... 81 Table 4.4. Effects of height, land use and habitat characteristics on daily mosquito abundance of all species combined, and on Anopheles balabacensis abundance separately. Chi-square (2), degrees of freedom (df), and p-values are given using log likelihood ratio test. Minimum adequate model (Final model) tested against the null model ...... 83 Table 4.5. Effects of height, area and habitat characteristics on daily mosquito abundance of all species combined, and on Anopheles balabacensis abundance separately. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 84 Table 5.1. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster, fan and paint) against immobilised controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given. 97 Table 5.2. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on male Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 100 Table 5.3. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on female Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 100 Table 5.4. Results from cox proportional hazards regression model, testing paint on male and female Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 100 Table 5.5. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on male Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 104 Table 5.6. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on female Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 104 Table 5.7. Results from cox proportional hazards regression model, testing paint on male and female Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given ...... 104 Table A.1. List of full taxonomic names of collected species ...... 146

Table A.2. Effects of land use on sex ratio of hatched larvae, using a generalised linear model with quasipoisson errors ...... 146 Table B.1. List of full taxonomic names of collected species ...... 147 Table C.1. List of full taxonomic names of collected species in Chapter 4 ...... 149

Chapter 1

Chapter 1 – General Introduction

1.1. Overview

Mosquito-borne diseases still thrive in many countries and are an important cause of morbidity and mortality, especially in the tropical and sub-tropical regions of the world. Malaria, the most widespread mosquito-borne disease, caused an estimated 198 million cases resulting in 584,000 deaths in 2013 (WHO 2014b). Anthropogenic modification of ecosystems has often coincided with the emergence and re- emergence of mosquito-borne diseases (Gratz 1999; Patz et al. 2000). Tropical rainforests are currently under threat from anthropogenic modification, particularly in Southeast Asia, which has the highest rate of habitat loss (Sodhi et al. 2004). It is important to understand the effect of deforestation on vector and parasite dynamics to improve our ability to assess health risks in a changing world.

The main aim of the work in this thesis is to investigate the effects of land use change, with a particular focus on deforestation and oil palm plantations, on mosquito ecology and their associated diseases. It also investigates the biting behaviour of simian malaria vectors between canopy and ground level across land use. Finally, it investigates mosquito marking techniques to determine the most cost-effective and least detrimental method for future dispersal experiments on these vectors. This chapter begins by exploring land use change, oil palm expansion and the fundamental aspects of mosquito biology, ecology and associated diseases. It concludes by presenting the main aims and the structure of this thesis.

1.2. Global environmental change and land use change

Human-induced modifications of the global environment have caused widespread change in the global distribution of organisms (Chapin et al. 2000). Land use change is predicted to have the largest global impact on terrestrial ecosystems by the year 2100, followed by climate change, nitrogen deposition, biotic exchange and the changing concentrations of atmospheric carbon dioxide (Sala et al. 2000). These components of change interact and all impose considerable influence on biodiversity

Chapter 1

(Vitousek 1994). Land use and land cover change are predicted to have the largest impact on biodiversity due to loss of habitat availability (Sala et al. 2000). The conversion of natural ecosystems to any other land use, such as agricultural practice, water management and urban areas, inevitably results in habitat loss, degradation and fragmentation and is the major driving force behind worldwide biodiversity loss (Sodhi et al. 2004; Foley et al. 2005). Habitat modification can also play a major role in the emergence or re-emergence of diseases (See Section 1.3).

1.2.1. Tropical deforestation and oil palm expansion in Southeast Asia

Tropical rainforests are exceptionally rich in biodiversity, but are under increasing threat due to conversion to agricultural land (FAO 2010). Since 2000, over 40 million hectares of primary forest have been cleared, and around 13 million ha of forest are converted each year to other uses, such as agriculture (FAO 2010). Southeast Asia includes four biodiversity hotspots (Indo-Burma, Philippines, Sundaland and Wallacea) containing high biodiversity and a large number of endemic species (Myers et al. 2000) but also experiences one of the highest rates of deforestation due to logging, habitat fragmentation, urbanisation, agricultural practice and expansion (Sodhi et al. 2004, 2010b). The expansion of oil palm (Elaeis guineensis) cultivation is a major driver of deforestation, and is now the most rapidly expanding perennial crop within tropical regions due to increasing global demand for a cheap oil source for biofuel and food products (Koh & Wilcove 2007; Fitzherbert et al. 2008; Phalan et al. 2013). Current projections show that Southeast Asia could lose up to three quarters of its original forest and 42% of its biodiversity by 2100 (Sodhi et al. 2004).

Palm oil accounts for nearly 30% of the world’s edible vegetable oil (Carter et al. 2007). It is of huge importance for biofuel production, and in comparison to other edible oils, has high yields and a low production cost (Fitzherbert et al. 2008; Corley 2009). Malaysia and Indonesia are the top producers of palm oil, providing ~85% of global production (FAOSTAT 2015), but also have high rates of deforestation (Wood 1990). Since 1990, the tropical forests in Malaysia have decreased by 1.2 million ha and have been converted into farmlands or agro-forest (Adachi et al. 2011). Sabah, in particular, has experienced a high percentage of forest loss, with 39.5% of its total forest area in 1973 becoming non-forest (oil palm and timber plantations) in 2010 (Gaveau et al.

Chapter 1

2014). Between 1990 and 2005, many oil palm plantations in Malaysia were created by planting on pre-existing cropland, for example rubber, but up to 59% of oil palm followed the logging of secondary forests (Koh & Wilcove 2008; Koh et al. 2011).

Oil palm plantations are structurally less complex than natural forests, with a uniform tree age, lower canopy, a highly variable microclimate, and greater human disturbance (Fitzherbert et al. 2008; Foster et al. 2011). Oil palm plantations are associated with reduced species richness, species diversity and overall abundance across most animal taxa (Koh & Wilcove 2008; Brühl & Eltz 2010; Foster et al. 2011). Oil palm is now the dominant agricultural crop across much of Southeast Asia (FAOSTAT 2015) and is expanding in Central and West Africa, Central America and the Amazon (Butler & Laurance 2009; Foster et al. 2011). The demand for oil palm is likely to increase, and Malaysia’s economy is highly dependent upon this industry (Carter et al. 2007).

1.3. Human movement, land use change and mosquito-borne diseases

The movement of humans through travel and transport has contributed to the spread of mosquito-borne diseases on spatial scales that exceed the limits of natural mosquito dispersal (Wesolowski et al. 2012). People can transport infectious mosquitoes to malaria-free areas, resulting in a resurgence of the disease (Martens & Hall 2000). The malaria risk through human population movement in industrialised countries is mainly by intercontinental travel (Tatem, Hay & Rogers 2006). In developing countries, urbanisation, colonisation, deforestation and agricultural labour migration can increase malaria risk (Martens & Hall 2000). Humans with a low-level immunity to diseases can move into a disease-endemic area and spread the disease (Martens & Hall 2000).

Deforestation, agricultural development, road and dam construction, urbanisation and the increased proximity of people to wildlife all modify the conditions for transmission of infectious disease, which can result in public health and economic harm (Patz et al. 2000; Foley et al. 2005; Coker et al. 2011). Land use changes have been linked to conditions that may increase the transmission of infectious diseases and can lead to outbreaks and emergence episodes, but the effects of land use change on ecosystems and human health are diverse (Gratz 1999; Foley et al. 2005; Yasuoka & Levins 2007). Each environmental change can play an important part in changing the ecological balance within which vectors breed, develop and transmit diseases (Patz et al. 2000;

Chapter 1

Norris 2004). Mosquitoes are the most important vector of human disease and actively transmit a wide range of pathogens. The key processes involved in disease transmission, following land use change, are summarised in Table 1.1.

Table 1.1. The key processes involved in disease transmission, following land use change (Martens & Hall 2000; Patz et al. 2000; Norris 2004; Afrane et al. 2005) Life stage Key processes Breeding sites - Increased breeding sites from social detritus - Changes to microclimate may result in a change in species composition in larval habitats - Water control systems can provide new breeding sites (i.e. irrigation) Larval survival - Increased temperatures can shorten larval developmental times Host finding - Increased human presence may displace other wildlife and humans become a more significant host Fecundity - Land use change can increase fecundity Adult survival - Increased temperatures can shorten larval developmental times Pathogen transmission - Humans may introduce new infections during land use change - Humans with a low-level immunity to disease can move into a disease-endemic area and spread the disease - More humans in one area during urbanisation, means greater opportunity for infections Dispersal - Transport can exceed limits of natural mosquito dispersal

Deforestation is associated with the resurgence of mosquito-borne diseases, such as malaria, Japanese encephalitis and filariasis, even after accounting for the effects of changing human population density (Bunnag et al. 1979; Walsh, Molyneux & Birley 1993; Manga, Toto & Carnevale 1995; Mackenzie, Gubler & Petersen 2004; Foley et al. 2005; Vittor et al. 2009). The resurgence of malaria following deforestation has been shown in Africa (Manga, Toto & Carnevale 1995), Asia (Bunnag et al. 1979) and Latin America (Vittor et al. 2006).

The effects of deforestation on emerging diseases are diverse, and can also reduce the risk of malaria transmission in certain areas (Yasuoka & Levins 2007). In Sarawak, Malaysia, deforestation resulted in reductions of the malaria vector, An. donaldi (Chang et al. 1997). Canopy removal exposed the previously shaded pools to sunlight, rendering them unsuitable for An. donaldi to breed (Chang et al. 1997). In some studies, the abundance of some malaria vectors increased following deforestation, while others decreased (Yasuoka & Levins 2007). The impacts of deforestation on

Chapter 1

mosquito density and malaria incidence are complex, and are influenced by both the agricultural development and the ecological characteristics of the local vector mosquitoes (Yasuoka & Levins 2007).

Land use changes modify temperature and relative humidity, which can affect mosquito survival, density and distribution, resulting in a change in vector abundance, competence and species composition (Patz et al. 2000; Yasuoka & Levins 2007). The temperatures of logged forests and oil palm plantations have a higher fluctuation over a 24 hour period than in primary rainforests (Koh, Levang & Ghazoul 2009; Turner & Foster 2009; Luskin & Potts 2011; Hardwick et al. 2015), which is predicted to have a strong influence on mosquito survival. Provided the temperature is not too extreme, a rising temperature can shorten mosquito development time (Stresman 2010).

As different mosquito species vary in their habitat requirements, a succession of vector species occurs following deforestation (Patz et al. 2000), and usually increases the chance of disease transmission occurring from a vector that was not previously implicated (Norris 2004). Zoonotic disease transmission can occur when anthropogenic changes reduce biodiversity and bring humans closer to wildlife (Pongsiri et al. 2009). The removal of forests and increase of human hosts within localised regions has caused an exponential growth in human-wildlife interaction and conflict, which has resulted in the exposure of newly recognised pathogens to humans, livestock and wildlife (Wolfe et al. 2005).

Agricultural development increases the number of humans working within an area, but can also alter environmental conditions to favour mosquito survival (Norris 2004). Agriculture can cause sedimentation, which can slow or block streams, allowing for water conditions favourable for increased vector development (Zhang & Shi 2001). Rice paddies are an example of agricultural development causing an emerging disease. These shallow bodies of water are an ideal breeding site for mosquitoes and can promote the transmission of Japanese encephalitis (Norris 2004; Coker et al. 2011).

There are also examples of when agricultural development decreased vector populations and malaria prevalence. In Karnataka, India, deforestation for coffee plantations reduced the malaria vector, An. fluviatilis, resulting in malaria elimination in this area (Yasuoka & Levins 2007). However, the development of coffee plantations

Chapter 1

in Southeast Thailand favoured the survival of An. minimus, resulting in hyperendemic malaria (Yasuoka & Levins 2007). The risk of disease transmission in an area depends on the arrival of opportunistic vector species and the adaption of vectors to newly created niches (Pongsiri et al. 2009).

Like deforestation and agricultural development, urbanisation also exposes humans to newly recognised pathogens and vectors (Norris 2004). In urban areas sewage management, runoff, sedimentation and artificial containers can provide many more mosquito breeding sites (Norris 2004). Artificial containers, such as tyres, bottles, buckets, water butts and cups, can provide a large number of mosquito breeding sites, which are particularly suitable for Stegomyia aegypti (=Aedes aegypti, see Reinert, Harbach & Kitching 2004) (Rattanarithikul et al. 2005b). Finally, the construction of roads provides access for new humans and livestock, and can lead to erosion of soil and create ponds by blocking the flow of streams during rainy seasons (Patz et al. 2004).

Water management includes ponds, dams, irrigation systems, paddies, sewage management and storm water management, and overlaps significantly with agricultural development and urbanisation (Norris 2004). Water control systems can provide new vector breeding sites where water was previously limited (e.g. irrigation canals) or by damming water, which is associated with higher malaria and Bancroftian filariasis prevalence (Harb et al. 1993; Alemayehu et al. 1998; Norris 2004; Keiser et al. 2005). Trypanosomiasis and onchocerciasis are also examples of vector-borne diseases affected by changing land use (Patz et al. 2004).

Malaria has been found to be a major health burden within an oil palm plantation in Papua New Guinea (Pluess et al. 2009). To date, there are very few studies looking at malaria incidence in Southeast Asia. Chang et al. (1997) showed a reduction in malaria vectors following deforestation and conversion to oil palm. The abundance of the anopheline vector, An. donaldi, and malaria prevalence remained low during the first two years of oil palm planting and maintenance (Chang et al. 1997). Oil palm plantations have a 25-30 year life cycle, and begin fruiting after 3-5 years (Butler, Koh & Ghazoul 2009). Old plantations (22 years old, ~13 m closed canopy) have shown to have a more buffered microclimate than young plantations (8 years old, ~4 m closed canopy) (Luskin & Potts 2011). There is very little data on vectors and disease rates

Chapter 1

in oil palm plantations, and whether anopheline abundance increases once oil palm plantations have started fruiting and are forming a closed canopy.

1.4. Mosquito biology

Mosquitoes serve as vectors for a variety of pathogens. Many species are significant pests to not only humans, but also domestic animals, with potentially fatal outcomes (Snow 1990). They are found on almost every continent, but the majority are found in the tropics and sub-tropics. The warmer climates allow them to be active all year round, with the ideal conditions being hot and humid with moderate rainfall (Gillett 1971). Although control methods are used, mosquito-borne diseases still thrive in many countries and cause millions of deaths (Tren & Bate 2001).

Mosquitoes are capable of breeding in a variety of environments. Many mosquitoes are generalists and choose a variety of oviposition sites (e.g. artificial containers), whereas others are specialists and choose unique habitats for laying eggs (e.g. bromeliads). The specialist mosquitoes usually disappear after land use changes, but generalists are able to survive in a wide variety of habitats (Rattanarithikul et al. 2005b). There are several types of oviposition sites, which are categorised into ground water (such as rivers, lakes, ground pools), artificial container sites (such as tyres, bottles, cups) or natural container sites (such as fallen leaves, tree holes, tree stumps) (Gillett 1971). Mosquitoes are able to breed in permanent water, semi-permanent water or temporary pools (Rattanarithikul et al. 2005b).

All mosquitoes undergo complete metamorphosis within their lifecycle, with species having very similar patterns of biological development. Eggs are deposited on the surface of an existing or expected water source (Norris 2004). After the eggs hatch, aquatic larvae pass through four larval instars, whilst feeding on detritus, algae and biofilms (Norris 2004). A few non-vector mosquito larvae are predaceous. The larvae take a few days to several weeks to develop, depending on nutrient levels, temperature, competition and water condition (Becker et al. 2010). Pupal stage lasts for around 48 hours, before emergence as adults. Male mosquito adults are short- lived and feed on juices from flower and fruits, but females of most mosquito species require a blood meal to obtain enough nutrients for the development and maturation of eggs (Clements 1992).

Chapter 1

1.4.1. Host-seeking behaviour and dispersal

Activities such as mating, dispersal, oviposition, resting and host choice vary between species and within species in different regions. Host-seeking by female mosquitoes is largely driven by olfactory cues released from a host’s skin and breath (Lehane 1991; Takken & Knols 1999). Hosts emit odour cues (e.g. carbon dioxide) that are diffused into the wind to form an odour plume. Host-seeking mosquitoes follow the behavioural steps of: activation to fly upwind, navigation of the odour plume using olfactory cues, surging and casting, close-range navigation toward skin and finally landing (Takken & Knols 1999). Carbon dioxide, a major component of human breath, is an important kairomone shown to attract a large number of mosquito species (Takken & Knols 1999). Other odours, such as lactic acid and various aldehydes have also shown to be important during the host-seeking behaviour of mosquitoes (Bosch, Geier & Boeckh 2000). Host-seeking nocturnal mosquitoes are usually more active during a full moon than other times of the lunar phase, suggesting light is used during orientation to hosts (Takken & Knols 1999). Physical cues, such as heat and moisture also help host- seeking mosquitoes find a host (Khan, Maibach & Strauss 2007). Gillies & Wilkes (1972) showed the distance at which different species responded to carbon dioxide baits was 30 m or less.

Many flights of mosquitoes are goal-oriented, with flights ending in activities such as nectar-feeding, blood-feeding, resting, mating or oviposition (Silver 2008). Wind plays an important role in the dispersal of mosquitoes (Silver 2008). Transport has also facilitated the movement of disease vectors, as mosquitoes can be transported on aircrafts, trains, vehicles and ships (Tatem, Hay & Rogers 2006; Silver 2008). Understanding dispersal dynamics and flight ranges of mosquito vectors is essential for the mitigation of disease, successful implementation of protection against infection and an important step in understanding the ecology of a vector (Silver 2008). A variety of methods have been used to mark mosquitoes for mark-release-recapture (MRR) experiments, but few studies have addressed the implications of marking efficiency and survivorship on male and female mosquitoes following marking, and even fewer have compared marking methods (Silver 2008). Successful MRR studies require a benign and cost-effective marking technique that adheres to the mosquito for a defined duration.

Chapter 1

1.5. Mosquito-borne diseases in Malaysia

Malaria, dengue fever, chikungunya, Japanese encephalitis and lymphatic filariasis are spread by mosquitoes in Malaysia. There are 434 known species of mosquitoes in Malaysia, representing 20 different genera (Rahman, Che’rus & Ahmad 1997; Wiwanitkit 2007). Stegomyia (=Aedes), Anopheles, Culex and Mansonia are the four main genera containing medically important mosquitoes in Malaysia. Stegomyia species are of concern because they transmit dengue fever and dengue haemorrhagic fever (Lam 1993), Anopheles species transmit malaria and filariasis (Rattanarithikul et al. 2006a), Culex species can transmit Japanese B-encephalitis and filariasis (Rattanarithikul et al. 2005a) and Mansonia species can transmit filariasis (Rattanarithikul et al. 2006b). Anopheles account for 75 of the 434 species, and 9 of these are main vectors of malaria (Table 1.2) (Wiwanitkit 2007).

Chapter 1

Table 1.2. Medically important mosquito species found in Malaysia (Reid 1968; Rahman, Che’rus & Ahmad 1997; Sallum et al. 2005; Vythilingam et al. 2005; Wiwanitkit 2007; Vythilingam 2012; WRBU 2015) Species Disease spread Habitat Typical Resting Biting behaviour behaviour behaviour Stg. aegypti* Dengue, chikungunya virus Clean and clear stagnant water. In artificial Anthropophilic Exophilic/ Exophagic/ (=Aedes aegypti) containers or natural habitats Endophilic Endophagic Stg. albopicta* Dengue, chikungunya virus Clean and clear stagnant water. In artificial Anthropophilic Exophilic/ Exophagic/ (=Aedes albopictus) containers or natural habitats Endophilic Endophagic An. balabacensis* Human malaria, simian Small pools of muddy water, in the forest Anthropophilic Exophilic Exophagic malaria and periphery An. campestris Human malaria, filariasis Still fresh water (marshes, drains, rice Anthropophilic Endophilic Endophagic fields) An. cracens (=An. Human malaria, simian Shaded pools and streams Simio- Exophilic Exophagic dirus B) malaria anthropophilic An. donaldi* Human malaria, filariasis Stagnant pools at forest edge Zoophilic Exophilic Exophagic An. flavirostris* Human malaria Shaded and unshaded stream margins and Anthropophilic/ Exophilic Exophagic/ ground pools Zoophilic Endophagic An. latens (=An. Human malaria, simian Small pools of muddy water, in the forest Simio- Exophilic Exophagic leucosphyrus A) malaria, filariasis and periphery anthropophilic An. letifer Human malaria Stagnant pools, usually in shade Zoophilic Exophilic Exophagic An. maculatus s.l.* Human malaria, filariasis Small pools of muddy water, in the forest Zoophilic Exophilic Exophagic and periphery. Usually sunlit An. epiroticus/ Human malaria Coastal waters Zoophilic Exophilic Exophagic sundaicus s.l.* Cx. gelidus* Possibly Japanese Temporary and semi-permanent ground Anthropophilic Endophilic Endophagic encephalitis pools, puddles and streams. Occasionally artificial containers Cx. Filariasis Stagnant water Anthropophilic Endophilic Endophagic quinquefasciatus* Cx. sitiens* Filariasis and possibly Brackish, salt and fresh groundwater Anthropophilic Endophilic Endophagic Japanese encephalitis habitats. In artificial containers

Chapter 1

Cx. Japanese encephalitis Stagnant water Anthropophilic Endophilic Endophagic tritaeniorhynchus* Cx. vishnui* Japanese encephalitis Stagnant water Anthropophilic Endophilic Endophagic Ma. uniformis* Filariasis, Japanese Open ponds and swamps with floating Zoophilic Exophilic Exophagic encephalitis vegetation *Species found in Sabah, Malaysia

Chapter 1

1.5.1. Dengue

Dengue is mosquito-borne viral infection endemic to tropical and sub-tropical areas, found mainly in urban and suburban areas (Chen et al. 2005; Guzman & Istúriz 2010). There are four known serotypes worldwide (DENV-1, DENV-2, DENV-3 and DENV- 4), all known to cause dengue fever and dengue haemorrhagic fever in Malaysia (Abubakar & Shafee 2002). Dengue is spread by the mosquito genus Stegomyia (=Aedes). Stegomyia aegypti (=Aedes aegypti, see Reinert, Harbach & Kitching 2004), and Stg. albopicta (=Aedes albopictus, see Reinert, Harbach & Kitching 2004) are the main vectors of dengue, and are closely associated with humans and domestic environments (Guzman & Istúriz 2010). The global incidence of dengue has grown rapidly in recent decades (Bhatt et al. 2013; Messina et al. 2014). In Malaysia, there were less than 1,000 cases in 1973, but this has increased to 46,000 cases in 2007, and 108,698 cases in 2014 (Benitez 2009; Lee et al. 2015).

The spread of dengue throughout Malaysia is thought to have followed the spread of Stg. aegypti, which replaced Stg. albopicta as the main carrier of the viruses (Abubakar & Shafee 2002). The accumulation of social detritus and storage containers in peri-urban areas has also contributed to dengue outbreaks (Coker et al. 2011). The distribution of Stg. aegypti and Stg. albopicta overlap in Malaysia and both spread dengue fever and dengue haemorrhagic fever (Chen et al. 2006; Rozilawati, Zairi & Adanan 2007). There is currently no specific treatment for dengue, and a safe vaccine is still in development (WHO 2015a). The current strategy to control or prevent the transmission of dengue virus is to control the mosquito vectors through reducing artificial man-made habitats, applying insecticides to outdoor water storage containers, applying insecticides as space spraying during outbreaks, using repellents and actively monitoring vectors (WHO 2009).

1.5.2. Japanese encephalitis, lymphatic filariasis and chikungunya

Japanese encephalitis is a mosquito-borne viral infection, spread by Culex species (mainly Culex tritaeniorhynchus and Culex vishnui) (Campbell et al. 2011). Japanese encephalitis occurs in East Asia and Southeast Asia, with nearly 68,000 cases every year (WHO 2014a). Japanese encephalitis is endemic in Sarawak, but sporadic cases are reported from all the other Malaysian states (Campbell et al. 2011). The

Chapter 1 transmission cycle occurs in rural and suburban areas, where rice culture and pig farming coexist, with pigs and aquatic birds as the principle vertebrate amplifying hosts (Campbell et al. 2011). Most infections are mild, but 1 in 250 cases can result in severe disease. Approximately 20-30% of severe disease cases are fatal, and 30-50% of survivors suffer permanent intellectual, behavioural or neurological problems (WHO 2014a). There is no cure for the disease, but there are vaccines available to prevent Japanese encephalitis (WHO 2014a).

Lymphatic filariasis is a parasitic infection caused by Wuchereria bancrofti, Brugia malayi, and to a lesser extent, Brugia timori. Wuchereria bancrofi causes 90% of all human lymphatic filarial infections (WHO 2015b). The majority of infections are externally asymptomatic, but filariasis can develop into chronic conditions, such as lymphoedema or elephantiasis of limbs (WHO 2015b). Over 120 million people are infected, with about 40 million of those disfigured by the disease (WHO 2015b). Vectors of filariasis in Southeast Asia belong to species of five mosquito genera; Anopheles, Culex, Downsiomyia, Mansonia and Ochlerotatus (Vythilingam 2012). WHO established the ‘Global Programme to Eliminate Lymphatic Filariasis’ in 2000, to eliminate filariasis by 2020, mainly through the implementation of Mass Drug Administration (MDA) (WHO 2013). Currently, the majority of filariasis cases in Malaysia are introduced into the country from migrant workers (Vythilingam 2012).

Chikungunya is a mosquito-borne viral disease spread by Stegomyia species, which has re-emerged in Africa, southern and Southeast Asia, and the Indian Ocean Islands (Zim et al. 2013). Chikungunya is commonly mistaken for dengue due to similar clinical conditions. An outbreak occurred in Malaysia between December 1998 and February 1999 due to uncollected rubbish and water storage containers (Lam et al. 2001). During the outbreak, public health education, active case detection and house to house fogging was conducted to reduce Stg. aegypti and Stg. albopictus numbers (Lam et al. 2001). Since the outbreak, the disease has become endemic in Malaysia (Lam et al. 2001). There is no vaccine or preventative drug available for chikungunya.

1.5.3. Malaria

Malaria still remains a public health problem within Africa, Southeast Asia, Central America and South America, affecting an estimated 198 million cases, and 584,000 deaths in 2013 (WHO 2014b). It is known that human malaria is caused by four

Chapter 1

Plasmodium species; Plasmodium falciparium, P. malariae, P. ovale and P. vivax, but a fifth Plasmodium species, P. knowlesi, has been recognised to cause symptomatic malaria in humans (Singh et al. 2004; Cox-Singh et al. 2008; White 2008). Since the first two cases of naturally acquired simian malaria to humans (Chin et al. 1965; Fong, Cadigan & Coatney 1971), extensive studies were carried out in the 1960s to determine the vectors and to study simian parasites in non-human primate and humans (Warren et al. 1970). No additional human P. knowlesi cases appeared until 2004, suggesting that naturally-acquired infections were extremely rare (Singh et al. 2004).

Singh et al. (2004) used nested PCR assays to identify 120 cases of naturally acquired P. knowlesi, previously reported by microscopy as P. malariae, in the Kapit Division, Malaysian Borneo. Since then, a large number of P. knowlesi cases have been reported in almost all of the countries in Southeast Asia, with the exception of Laos (Jongwutiwes et al. 2004; Kantele et al. 2008; Luchavez et al. 2008; Cox-Singh et al. 2008; Ng et al. 2008; Van den Eede et al. 2009; Figtree et al. 2010; Jiang et al. 2010; Khim, Siv & Kim 2011). Unlike the other four Plasmodium species, P. knowlesi is largely zoonotic. Monkeys, particularly the long-tailed macaque (Macaca fascicularis) and the pig-tailed macaque (Macaca nemestrina) found in Southeast Asia, are the two main natural hosts of P. knowlesi (Lim & Singh 2013). Before the rapid deforestation in Southeast Asia, the Anopheles (Cellia) mosquitoes responsible for spreading P. knowlesi were only found in the deep forest (Wharton et al. 1964; Chin et al. 1968). Since increased deforestation in many areas, monkeys and their forest dwelling vectors have moved closer to the forest edge and towards human habitation (Vythilingam 2010).

Although P. knowlesi has been extensively studied, there is very little known about the vectors, their distribution and how canopy height and canopy closure affects their biting behaviour. In order to sustain a good control programme, it is important to monitor mosquito populations and how land use change affects their distribution. Malaysia has had a particularly successful malaria control programme, aimed to eliminate malaria by 2020, but P. knowlesi may become a serious threat to malaria elimination (William et al. 2013). From 1994-2011, P. vivax and P. falciparum have decreased 25-fold and 55-fold, respectively within Sabah, Malaysia, but the prevalence of P. knowlesi has increased >10 fold between 2004 and 2011 (Figure 1.1)

Chapter 1

(William et al. 2013, 2014). Plasmodium knowlesi is now the most common cause of malaria in the Malaysian state of Sabah, it accounts for 62% of all malaria incidences in 2013 and presents a threat to malaria elimination (William et al. 2014).

Since current indoor control methods for malaria do not prevent zoonotic transmission, as vectors are generally exophilic and exophagic, P. knowlesi cases are predicted to increase (Kantele & Jokiranta 2011). It is believed the extensive deforestation in Sabah has increased the prevalence of P. knowlesi, resulting in encroachment of humans in previously forested areas, and allowing a higher interaction with vectors and hosts (William et al. 2013). The effect of deforestation and fragmentation on mosquito vectors within Sabah has been poorly studied, but due to the increase in P. knowlesi cases, this area of research has high public health importance.

Chapter 1

Figure 1.1. Plasmodium malariae/Plasmodium knowlesi cases, by divisions of Sabah, Malaysia, from 2001-2013. Population of divisions in 2010: Interior 424,524; West Coast 1,011,725; Kudat 192,457; Sandakan 702,207; Tawau 819,955. Reproduced from William et al. (2014)

Chapter 1

1.6. Thesis structure and research questions

This thesis is centred on the impacts of land use change, particularly deforestation and oil palm expansion, on mosquito community dynamics. It is divided into four main research questions:

1. How does land use change affect the presence of human-landing and container breeding mosquitoes

In Chapter 2, I investigate the ovipositional behaviour and community composition of mosquitoes along an anthropogenic disturbance gradient, from primary rainforest to rural housing estates in oil palm plantations.

2. What are the impacts of oil palm expansion on mosquitoes and their associated diseases

In Chapter 3, I focus on anopheline abundance and human landing rates along an anthropogenic disturbance gradient (primary rainforest to oil palm plantations), and their potential to spread malaria.

3. How do community composition and abundance of mosquitoes differ between ground and canopy

In Chapter 4, I investigate anopheline abundance further by looking at the human host-seeking preference and abundance of mosquitoes at canopy and ground level. In this chapter I mainly focus on the simian malaria, Plasmodium knowlesi, and the potential ability of anophelines to transmit P. knowlesi between canopy- dwelling simian hosts and ground-dwelling humans

4. Can a cost-effective, persistent and non-detrimental marking technique be established for use in dispersal studies

In Chapter 5, I investigate different mosquito marking techniques to determine the most cost-effective, persistent and least detrimental method for use in future dispersal studies.

Chapter 1

Chapter 6 is the concluding chapter in which I discuss the implications of my findings. I also discuss future land use change and malaria control methods. Finally, I outline topics for future study that will extend our knowledge in this research area.

Chapter 2

Chapter 2 – The Effects of Land Use Change on Mosquito Community Composition

Hayley L. Brant1, Robert M. Ewers2, Indra Vythilingam3, Chris Drakeley4, Suzan Benedick5 & John D. Mumford1

1 Centre for Environmental Policy, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, SL5 7PY, UK

2 Department of Life Sciences, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, Berkshire, SL5 7PY, UK

3 Department of Parasitology, Faculty of Medicine, University of Malaya, Kuala Lumpur, 50603, Malaysia

4 Department of Immunology and Infection, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, WC1E 7HT, UK

5 Faculty of Sustainable Agriculture, Universiti Malaysia Sabah, Locked Bag No. 3, 90509, Sandakan, Sabah, Malaysia

2.1. Abstract

Land use changes greatly affect the presence, development and behaviour of vectors, which can influence the transmission of mosquito-borne diseases. Southeast Asia is currently experiencing high rates of deforestation due to oil palm expansion, as well as facing increases in mosquito-borne diseases such as dengue and simian malaria. This study focuses on how land use affects mosquito community composition and the presence of medically important mosquitoes. Mosquito collections were carried out using human landing catches and ovitraps, along an anthropogenic disturbance gradient encompassing primary forest, disturbed forest, highly disturbed forest, oil palm and rural housing estates. Results showed significant differences in the community composition of anthropogenic mosquitoes between primary forest and oil palm sites. This is largely driven by medically important mosquitoes landing in oil palm sites, such as Stegomyia albopicta, Culex gelidus, Cx. sitiens and Cx. quinquefasciatus. Anopheles balabacensis was the most common species landing in

Chapter 2 the disturbed forest, highly disturbed forest and oil palm sites, and is capable of spreading human and simian malaria. Significant differences in community composition were also seen in container breeding mosquitoes, with a high presence of the dengue vector, Stg. albopicta, in areas of rural housing.

2.2. Introduction

Land use changes, including agricultural encroachment and urbanisation, have caused considerable declines in biodiversity through the loss of habitat (Sala et al. 2000; Foley et al. 2005). Southeast Asia contains tropical rainforests exceptionally rich in biodiversity, but these forests are under increasing threat from agricultural expansion and intensification (FAO 2010; Sodhi et al. 2010b). The increase in oil palm (Elaeis guineensis) cultivation is a main driver of deforestation and biodiversity loss in tropical forests, particularly Malaysia and Indonesia (Koh & Wilcove 2007, 2008; Koh et al. 2011). Forest modification and conversion to oil palm plantations has shown to reduce the species richness and abundance of many taxa (Schulze et al. 2004; Fitzherbert et al. 2008; Turner & Foster 2009; Sodhi et al. 2010b; Foster et al. 2011), including birds (Peh et al. 2006; Koh & Wilcove 2008; Edwards et al. 2010), bats (Struebig et al. 2008), small mammals (Bernard, Fjeldså & Mohamed 2009), beetles (Chung et al. 2007; Edwards et al. 2014; Gray et al. 2014), ants (Fayle et al. 2010; Brühl & Eltz 2010) and butterflies (Koh & Wilcove 2008), but only a few studies have focussed on how forest modification will affect mosquito populations in Southeast Asia.

Land use change and biodiversity loss can affect disease transmission through several mechanisms. Fragmentation can result in higher levels of human activity and human- wildlife interactions, which increases the exposure to zoonoses, particularly in biodiversity rich areas containing novel pathogens (Wolfe et al. 2005; Wolfe, Dunavan & Diamond 2007; Keesing et al. 2010). Habitat alteration, deforestation and fragmentation can create new breeding sites for some species and affect microclimates (Walsh, Molyneux & Birley 1993). A warmer microclimate can shorten mosquito and parasite development time, making mosquitoes infectious more quickly (Stresman 2010).

Deforestation, degradation, fragmentation and agricultural development may cause shifts in relative vector abundance (Norris 2004). Deforestation is driven by a variety

Chapter 2 of human activities, including logging, agricultural development and road construction, resulting in higher human host abundance within localised regions and may increase disease risk (Norris 2004; Yasuoka & Levins 2007). Although some agricultural developments can also alter environmental conditions to favour mosquito survival, others can reduce survival. For example, the development of coffee plantations in Southeast Thailand favoured the survival of Anopheles minimus, resulting in hyperendemic malaria (Yasuoka & Levins 2007). In contrast, coffee plantations in Karnataka, India reduced the breeding sites for An. fluviatilis, resulting in the elimination of malaria in this area (Yasuoka & Levins 2007). Agricultural development also results in a higher human host abundance within localised regions, for example in agricultural workers housing areas (Norris 2004). The risk of disease transmission in an area depends on the arrival of opportunistic vectors, adaption of vectors to disrupted or newly created niches and migration of non-immune humans (Pongsiri et al. 2009).

Human migration through travel and transport has contributed to the spread of mosquito-borne diseases on spatial scales that exceed the limits of natural mosquito dispersal (Wesolowski et al. 2012). Migration and urbanisation can transport infectious mosquitoes to malaria-free areas, resulting in a resurgence of the disease, and alter vector habitats and behaviours (Martens & Hall 2000; Norris 2004). Artificial containers, such as tyres, bottles, buckets, water butts and cups, in urban areas can provide a large number of mosquito breeding sites, particularly suitable for the dengue vector Stegomyia aegypti (=Aedes aegypti, see Reinert, Harbach & Kitching 2004) (Rattanarithikul et al. 2005b). In Southeast Asia, the resurgence of dengue has been linked to the accumulation of social detritus and storage containers in peri-urban areas (Coker et al. 2011).

Malaysian Borneo is a hotspot for forest loss and degradation, with nearly 80% of Sabah and Sarawak impacted by logging and clearing operations between 1990 and 2009 (Bryan et al. 2013). The landscape of Sabah and Sarawak are now dominated by degraded forests, timber plantations and oil palm plantations (Bryan et al. 2013). Few studies have looked at mosquito presence and community composition in these areas. Chang et al. (1997) showed a reduction in malaria vector abundance in Sarawak following deforestation and conversion to oil palm, but an increase in dengue vectors. Since land use change is complex, further accumulation of data is needed to

Chapter 2 help develop predictive models to reduce disease transmission. The overall aim of this study was to examine the effect of land use on mosquito abundance, community composition, and the presence and absence of medically important mosquitoes using human landing catches and experimentally-positioned water-filled containers (ovitraps). The present study was aimed to provide preliminary data on the ecology of the mosquito fauna in this region for the development of effective control strategies.

2.3. Methods

2.3.1. Study site

This study was conducted in the Tawau Division of Sabah, Malaysian Borneo. Study sites were selected along an anthropogenic disturbance gradient; primary lowland dipterocarp rainforest (PF), twice-logged disturbed dipterocarp rainforest (DF), twice- logged highly disturbed dipterocarp rainforest (HDF), oil palm plantation (OP) and rural oil palm plantation labour housing compounds estates (RU). Rainforest sites were defined based on canopy closure, measured at each collection site using a spherical densiometer. The average canopy closure values along the anthropogenic disturbance gradient were, 79.9% in the primary forest, 74.9% in disturbed forest, 68.6% in highly disturbed forest, 40.3% in oil palm sites and 0% in rural housing estates.

Primary forest survey points were selected in the Maliau Basin Conservation Area (4°5’N, 116°5’E). Survey points within the primary rainforest were selected in an area that has never been logged commercially, with the exception of a few sites being selectively logged in the 1970s and 1990s to build the Maliau Field Centre. Forest quality is still classed as unaffected by logging and is substantially different from the commercially logged forest (Ewers et al. 2011).

Logged forest and semi-urban survey points were selected within the Benta Wawasan oil palm plantation (4°6’N, 117°5’E). The 45,601 ha area is a mixture of twice-logged rainforest, virgin jungle reserve, acacia and oil palm. Logged forest survey points were in selectively twice-logged forest, logged during the 1970s and 1990s-2000s. Oil palm plantation survey points were in areas of Elaeis guineensis monocultures, planted in 2006 and located 500-700 m from the forest for the majority of the survey points. The

Chapter 2 rest of the survey points were planted in 2000, and were located 1 km from the forest. Semi-urban survey points were located in local housing estates, 500 m from the oil palm plantation sites. Survey points were selected to synchronise with the central sampling design of the ‘Stability of Altered Forest Ecosystems (S.A.F.E.) Project’, a large-scale fragmentation experiment, which is investigating the long-term effects of forest fragmentation (Figure 2.1, Ewers et al. 2011).

Chapter 2

Figure 2.1. Map of the Stability of Altered Forest Ecosystems Project, located in Sabah, Malaysia (a) Primary forest sites (b) Continuous twice- logged forest (c) Twice-logged forest and fragmented forest in an oil palm matrix (d) Oil palm plantation sites (e) The fragmentation experiment comprising of six blocks (A-F). Reproduced from Ewers et al. (2011)

Chapter 2

2.3.2. Human landing catches

Human landing catches were carried out from October 2012 to April 2013 between 18:00-23:00h, for 84 nights. A total of 14 survey points were selected, with two mosquito collectors sampling simultaneously per night. Three survey points, with a minimum separation distance of 600 m, were selected in the primary rainforest and oil palm plantation. Eight survey points were selected in logged forest.

A pilot study was conducted 17:00-01:00h in each sampling area to determine the biting patterns of Anopheles. It demonstrated Anopheles species in each sampling area started biting after 18:00h, with a decrease in abundance by 22:00h (Figure 2.2). One mosquito collector remained constant during the overall sampling period, but the second mosquito collector at each sampling point was rotated every three days to control collector bias. Each survey point was sampled for three nights in a row, with one round of collections from October to December 2012 and the second round from January to April 2013. Each survey point was sampled for a total of six nights.

The two collectors, with the aid of a red torch light, aspirated mosquitoes off their own legs. Collected mosquitoes were placed into cups covered with a net cloth, and a new cup was used during every hour of collection. Mosquitoes were taken back to the field laboratory to be killed and sorted into individual tubes with silica gel. All mosquitoes were identified morphologically using keys (Reid 1968; Rattanarithikul et al. 2005a; b, 2006a; b, 2010; Sallum et al. 2005).

Chapter 2

Figure 2.2. Anopheles landing hourly per person per night during the human landing catch pilot study (17:00-01:00h) in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest (d) Oil palm. Error bars show ± SE of the mean

Chapter 2

2.3.3. Ovitraps

Ovitraps were placed at the survey points from January to March 2013, for a total of 480 trap nights. Eighty-six survey points were sampled in our study: 15 in primary forest, 32 in logged forest of varying forest quality, 18 in oil palm plantation and 15 in the rural housing estates.

The ovitrap consisted of a dark plastic container (700 ml), filled with 500 ml water and an overflow hole near the rim of the container, to avoid overflow of water during heavy rain (Yap & Thiruvengadam 1979; Rozilawati, Zairi & Adanan 2007). A wooden chopstick (8 mm x 20 cm) was added to act as an oviposition substrate (Appendix A, Figure A.1). Wire netting was placed over each trap to prevent large leaves entering the traps. All ovitraps were placed under heavy shade, measured by a densiometer, to ensure mosquitoes were attracted to the traps (Evans & Bevier 1969).

After six days, traps were collected and all ovipositional substrates were dried and placed in individual plastic bags. All water was carried back to the laboratory, for larvae to be counted and reared through to adults for identification. Older, predacious larvae were separated from younger instars. Eggs were counted on the ovipositional substrates and the rim of the ovitraps by using a dissection microscope. After counting, unhatched eggs were placed in water to hatch.

2.3.4. Meteorological data

Rainfall (mm), air temperature (°C) and relative humidity (%) data were obtained from the S.A.F.E. Project and Maliau Basin Field Centre for the duration of the field survey. In addition, lunar illumination (%), cloud cover and unusual climatic events (e.g. strong winds) were recorded every hour by the collectors.

2.3.5. Data analysis

Analyses were performed using R version 3.3.1 (R Core Team 2014). Spatial autocorrelation was tested for using Geary’s C between points in each land use (R package ncf, ‘correlog’: Bjornstad 2013). No significant spatial correlation was detected. Simpson and Shannon diversity indices were calculated in each area using the vegan function ‘diversity’ (R package vegan, 'diversity': Oksanen et al. 2015). Species accumulation curves were calculated in each area using the vegan function

Chapter 2

‘specaccum’ (R package vegan, 'specaccum': Oksanen et al. 2015). The Chao species estimator (Chao 1) and Abundance Coverage Estimator (ACE) were calculated the extrapolated species richness in each area using ‘chao1’ and ‘ACE’ functions in the R package ‘fossil’ (R package fossil, 'chao1', 'ACE': Vavrek 2015).

The presence of larvae in ovitraps was analysed by using a generalised linear model with binomial errors (R package brglm: Kosmidis 2013). The same analysis was performed for the presence of Stg. albopicta in ovitraps. The number of mosquitoes landing per night was used as a measure of relative abundance. The effect of land use change (PF, DF, HDF and OP) on anophelines and culicines was analysed using a generalised linear mixed-effect model (R package lme4: Bates et al. 2015), using day and site as random factors, with Poisson error distribution. A chi-squared test was used to compare the relative abundance of vector and non-vector species in each area.

Differences in community composition of human landing catches and ovitraps were explored using Detrended Correspondence Analysis on scaled abundance data (R package vegan, ‘decorana’: Oksanen et al. 2015). We tested for significant differences in community composition using a linear model with the first and second DCA axis as the response variable against land use.

2.3.6. Ethics

This project was approved by the Medical Research and Ethics Committee (NMRR- 12-689-12521), Ministry of Health, Malaysia and Imperial College London Research Ethics Committee (ICREC-12-5-6), UK.

2.4. Results

2.4.1. Mosquito abundance

A total of 2065 adult mosquitoes were collected from 84 human landing catch nights, consisting of 1977 (95.7%) anophelines and 88 (4.3%) culicines (Table 2.1). The primary forest collections consisted of nine species, with An. latens as the predominant species (46.3%). The disturbed forest and highly disturbed forest collections contained

Chapter 2

19 and 15 mosquito species, respectively, with An. balabacensis as the predominant species in both (DF: 93.1%, HDF: 83.1%). Anopheles balabacensis was also predominant in the oil palm plantation site collections (76.4%), consisting of 15 mosquito species in total.

A total of 1145 eggs were counted on the ovipositional substrates and rims of the ovitraps. A total of 2101 larvae were collected from 80 ovitraps, consisting of mostly culicines (Table 2.1). The majority (82%) of the larvae were reared to adults for identification, resulting in a sex ratio of 51.3% ± 1.1 males to 48.7% ± 1.1 females. The majority of larvae identified in the primary forest, highly disturbed forest and disturbed forest were Armigeres jugraensis (PF: 37.4%, DF: 35.6%, HDF: 32.4%) and Culex (Culiciomyia) spp. (PF: 56.8%, DF: 43.3%, HDF: 54.2%). The number of species found in primary forest, disturbed forest and highly disturbed forest were five, seven and six, respectively. The oil palm plantation ovitraps collected zero larvae. The rural housing estate ovitraps collected two species: Culex quinquefasciatus and Stg. albopicta.

For both collection methods, the number of species collected, Shannon index, Simpson index, Chao1 and ACE varied across land use (Table 2.2-3). The species accumulation curves did not reach an asymptote after human landing catch sampling, indicating not all species of mosquitoes had been collected (Figure 2.3). The species accumulation curves of ovitraps reached an asymptote in primary forest and rural housing sites, suggesting sampling was exhaustive, but the disturbed forest and highly disturbed forest did not reach an asymptote (Figure 2.4).

Chapter 2

Table 2.1. Mosquitoes collected from primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm (OP) and rural housing (RU) in the district of Tawau, Sabah, Malaysia Mosquito genera and Human landing catches Ovitraps species PF DF HDF OP PF DF HDF OP RU An. Aitkenii group 1 4 9 ------An. barbirostris - 1 ------An. vanus - 2 - 1 - - - - - An. (Anopheles) spp.* 2 9 8 10 - - - - - An. balabacensis 18 1016 374 359 - - 11 - - An. kochi - 1 ------An. latens 25 6 15 2 - - - - - An. macarthuri 1 26 18 25 - - - - - An. maculatus - 5 2 25 - - - - - An. tessellatus - - - 1 - - - - - An. watsonii 1 5 5 ------Arm. jugraensis 3 1 - - 508 116 70 - - Arm. confuses - - - - - 1 4 - - Arm. flavus 1 ------Col. pseudotaeniatus - - - - 28 - - - - Coq. crassipes - 3 ------Cx. (Culiciomyia) spp.* - - 1 - 772 141 117 - - Cx. nigropunctatus - - - - 1 - - - - Cx. papuensis - 1 - 1 - - - - - Cx. scanloni - - - 1 - - - - - Cx. gelidus - - - 4 - - - - - Cx. mimulus - - 1 ------Cx. quinquefasciatus - - - 12 - - - - 102 Cx. sitiens - - 1 5 - - - - - Cx. vishnui - 3 11 2 - - - - - Cx. (Loph) sp.* - 2 ------Cx. bitaeniorhynchus - 2 ------Do. ganapathi - 1 1 2 - - - - - He. scintillans - - 1 ------Ma. annulata - - 2 ------Orthopodomyia sp.* - 2 ------Pr. ostentatio 2 1 ------Stg. albopicta - - 1 20 - 21 6 - 97 Tripteroides sp.* - - - - - 8 - - - Uranotaenia spp.* - - - - 51 8 - - - Ze. gracilis - - - - - 31 8 - - Total mosquitoes 54 1091 450 470 1360 326 216 0 199 No. of samples 18 24 24 18 15 16 16 18 15 Mosquitoes/samples 3 45.5 18.8 26.1 90.7 20.4 13.5 NA 13.3 * Couldn’t be identified to species level

Chapter 2

Table 2.2. Mean species richness and diversity indices (± SE) of mosquito communities, collected using human landing catches, in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantations (OP) Land Human landing catches use N Species Shannon Simpson Chao1 ACE no. index index PF 18 9 0.36 (0.12) 0.54 (0.09) 13.00 13.63 DF 24 19 0.29 (0.05) 0.14 (0.03) 23.50 21.37 HDF 24 15 0.54 (0.09) 0.28 (0.05) 24.00 17.54 OP 18 15 0.79 (0.09) 0.42 (0.05) 17.67 16.60

Table 2.3. Mean species richness and diversity indices (± SE) of mosquito communities, collected using ovitraps, in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm plantations (OP) and rural housing (RU) Land Ovitraps use N Species Shannon Simpson Chao1 ACE no. index index PF 15 7 0.64 (0.07) 0.40 (0.04) 7.00 7.30 DF 16 7 0.36 (0.09) 0.22 (0.05) 7.00 7.19 HDF 16 7 0.35 (0.09) 0.64 (0.09) 7.00 7.22 RU 15 2 0.02 (0.02) 0.14 (0.09) 2.00 2.00

Chapter 2

Figure 2.3. Species accumulation curves for human landing catch sampling in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest and (d) Oil palm. Shaded area indicates 95% confidence intervals

Chapter 2

Figure 2.4. Species accumulation curves for ovitrap sampling in (a) Primary forest (b) Disturbed forest (c) Highly disturbed forest and (d) Rural housing. Shaded area indicates 95% confidence intervals

Chapter 2

2.4.2. Effect of land use on mosquito abundance and presence

Land use had a significant effect on the number of anopheline landing during the human landing catches (log likelihood test; 2=19.75, df=3, p=0.0002, Figure 2.5a), with a higher abundance of anophelines landing in the disturbed forest, highly disturbed forest and oil palm plantation sites than primary forest. There was also a significantly higher number of anopheline landing in the disturbed forest than highly disturbed forest (p=0.023) and oil palm plantations (p=0.053) (Table 2.4). Land use had a significant effect on the number of culicines landing (log likelihood test; 2=19.47, df=3, p=0.0002, Figure 2.5b), with oil palm plantations having a significantly higher number of culicines landing than in primary forest, disturbed forest and highly disturbed forest (p<0.0001). Rainfall during the collection significantly decreased the number of anopheline landing (2=5.48, df=1, p=0.02), but not culicines (2=1.19, df=1, p=0.275). Moonlight illumination had no effect on the landing of anophelines (2=0.13, df=1, p=0.721) or culicines (2=0.24, df=1, p=0.622). Collector identity also had no effect on the landing of anophelines (2=1.61, df=2, p=0.447) or culicines (2=3.12, df=2, p=0.211).

Land use also had an effect on the presence of mosquito larvae for all species (log likelihood test; 2=65.76, df=4, p<0.0001). There was a significantly lower number of ovitraps with mosquito larvae present in oil palm plantations than the disturbed forest (p<0.001) (Table 2.5). Stegomyia albopicta presence was significantly higher in the rural housing compounds than in other areas (log likelihood test; 2=42.46, df=4, p<0.0001). Rainfall of 1-3 week lags, temperature and humidity during the ovitrap placement were not significant.

Chapter 2

Figure 2.5. (a) Anophelines and (b) Culicines landing per person per night across an anthropogenic disturbance gradient: primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP). Error bars show ± SE of the mean

Chapter 2

Table 2.4. Effects of land use and habitat characteristics on daily mosquito landings of anophelines and culicines in the primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP) Predictor Anophelines Culicines β SE z p β SE z p Intercept 4.166 0.401 10.385 <0.0001*** -0.556 0.305 -1.824 =0.068 Area HDF -1.007 0.443 -2.274 <0.023* 0.159 0.379 0.419 =0.675 Area OP -0.931 0.481 -1.937 =0.053 1.408 0.345 4.083 <0.001 Area PF -3.228 0.527 -6.129 <0.001*** -0.681 0.511 -1.333 =0.183 Rainfall -0.387 0.165 -2.352 =0.019 NA NA NA =0.275

Table 2.5. Effects of land use and habitat characteristics on ovitrap and Stegomyia albopicta presence in the primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and rural housing (RU) Predictor Ovitrap presence Stegomyia albopicta presence β SE z p β SE z p Intercept 3.497 1.480 2.363 <0.01** -0.738 0.534 -1.380 =0.168 Area HDF -3.260 1.563 -2.086 <0.01** -0.284 0.779 -0.365 =0.715 Area OP -7.107 2.087 -3.405 <0.001*** -2.873 1.566 -1.834 =0.067 Area PF -0.063 2.096 -0.030 =0.976 -2.696 1.577 -1.710 =0.087 Area RU -1.810 1.642 -1.103 =0.270 2.424 0.890 2.725 <0.05*

A chi-squared test of the contingency table of vector and non-vector species showed significant differences across land use using ovitraps (2=115.88, df=3, p<0.0001, Figure 2.6b), but no significant differences for the human landing catch method (2=1.27, df=3, p=0.74, Figure 2.6a).

Chapter 2

Figure 2.6. Mean abundance (total mosquitoes/number of samples) of vector and non-vector species across an anthropogenic disturbance gradient: primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF), oil palm plantation (OP) and rural housing (RU), using (a) human landing catches (b) ovitraps. Error bars show ± SE of the mean

Chapter 2

2.4.3. Community composition

The community composition determined from human landing catches and ovitraps was significantly different across land use, separated along the first axis for both methods (Human landing catches: F3,74=30.32, p<0.001, Ovitraps: F4,48=125.5, p<0.001, Figure 2.7). Human landing catches were also separated by land use along the second axis (F3,74=13.09, p<0.001). The first two axes of the human landing catch method accounted for 63.6% of the total variance. Anopheles balabacensis was prevalent in all human landing sampling areas, but the oil palm community included species such as Cx. quinquefasciastus and An. tessellatus, which were not present in the other areas (Appendix A, Figure A.2). Except for one Stg. albopicta collected in the disturbed forest, the rest of the specimens of this dengue vector were found in the oil palm plantation. The first axis of the ovitrap method accounted for 92.5% of the total variance. The separation of the rural housing compound from the forest sites was driven by the presence of Stg. albopicta and Cx. quinquefasciatus (Appendix A, Figure A.3).

Chapter 2

Figure 2.7. Detrended Correspondence Analysis (DCA) plot showing the major axes of variation for (a) Adult mosquitoes collected using human landing catches and in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation sites (OP) and (b) Mosquitoes collected using ovitraps in primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and rural housing compounds (RU). The two axes represent linear summaries of the variation in the species numbers and areas

Chapter 2

2.5. Discussion

Agricultural expansion and intensification are the main drivers of habitat change in tropical regions (Laurance, Sayer & Cassman 2014), with the rapid oil palm expansion across Southeast Asia threatening many species, but there are few studies focusing on the mosquito populations present in different land use areas. The overall objective of this study was to examine the effect of land use on mosquito community composition and presence. Our data show that there is a different community composition of mosquitoes across different land uses. Oil palm plantations and rural housing compounds differed significantly from rainforest sites, due to the presence of Stg. albopicta. The species richness also differed across land use, with the disturbed forest containing the highest number of species and species richness, estimated by Chao1 and ACE.

The compositional differences between the human landing catches in oil palm and forest were very evident amongst the medically important mosquitoes, such as Culex gelidus, Cx. sitiens, Cx. quinquefasciatus and Stg. albopicta. Medically important species also dominated the rural sites in the ovitrap data. Other studies have suggested that vector abundance can increase following environmental change and modifications in larval habitat availability (Vittor et al. 2006; Johnson, Gómez & Pinedo- Vasquez 2008). These results support those from prior work in Sarawak, where Stg. albopicta was more abundant after the clearing of forests and cultivation of oil palm (Chang et al. 1997). Our results further showed that Stg. albopicta was only breeding in residential areas, not in the oil palm itself. As Stg. albopicta was collected using human landing catches in the oil palm sites, the absence in the ovitraps suggests their host-searching diversion from more suitable habitats. Since Stg. albopicta has been shown to have a limited host-searching range, it is unlikely to be far from the oil palm sites.

As Stg. albopicta is predominately a daytime biter, bias may have been introduced to the human landing catches by only sampling between 18:00h and 23:00h, but this was consistent in each sampling area. These hours provided information on malaria vectors landing during the evening when humans are most likely to be bitten. Because of logistical problems, 24 hour collections were not able to be conducted, but these

Chapter 2 would have given an insight into the composition of mosquitoes biting throughout the day and night.

The human landing catches showed anopheline mosquitoes were landing at a significantly higher rate in logged and converted lands than in primary forest. The number of culicines landing was significantly higher in oil palm than in other areas due to the presence of Stg. albopicta in oil palm. Although the twice-logged disturbed forest has had more than 15 years of regeneration, the numbers of mosquitoes are still significantly higher than in original primary forest. In Sabah and Sarawak, there are few areas of primary forest (Bryan et al. 2013) and efforts are being made to preserve these areas to maintain biodiversity and prevent the further increase in the number of medially important mosquitoes.

It is unknown why some species of mosquitoes were not collected by the ovitraps in the oil palm plantation, despite their presence in the rural housing compounds. The temperature differences were very minor. In this study, the housing areas were a distance of 440-1000 m from the nearest oil palm plantation site. Perhaps the rural housing compounds provided suitable breeding habitats (e.g. large water containers) for Stg. albopicta. These mosquitoes have been shown to fly 400-800 m and may fly into the oil palm plantation in search of hosts (Niebylski & Craig 1994; Reiter et al. 1995; Honório et al. 2003).

In this study, rainfall during the collection period did not have an effect on the presence of mosquitoes in ovitraps but it did reduce the number of anophelines landing. Rainfall is an important environmental factor for the survival of mosquitoes. It provides a breeding site for the aquatic stage of the mosquito lifecycle, and can extend the lifespan of adult mosquitoes (Martens et al. 1995). Increased rainfall is often associated with mass egg hatching and an increase in the number of mosquitoes (Ndiaye et al. 2006). The study area has been described as aseasonal, with no dry season, but subject to the occasional drought (Marsh & Greer 1992; Walsh & Newbery 1999). Although this study did not look at the seasonal pattern associated with temperature and rainfall, it is predicted that both variables would influence the abundance of vectors in this area during strong climatic changes (e.g. El Niño effects), but the differences will still remain between land uses.

Chapter 2

Oil palm plantations have been shown to have a higher fluctuation in temperature and humidity over a 24 hour period than forests (Koh, Levang & Ghazoul 2009; Turner & Foster 2009; Luskin & Potts 2011; Hardwick et al. 2015). In this study, mean daytime temperature across all the study areas varied from 29°C in primary forest to 34.9°C in oil palm. By 18:00h, all areas cooled to 24°C, with high humidity. A temperature above 34°C will generally decrease the survival rate of mosquito vectors and their parasites (Rueda et al. 1990). No differences were seen in the regional climate data, but the microclimate could have affected the compositional differences seen in this experiment for ovitraps. Moonlight illumination was not a significant factor affecting the abundance of anophelines or culicines landing in this study. Some studies have shown that moonlight increases relative abundance (Bidlingmayer 1964; Charlwood et al. 1986; Birley & Charlwood 1989; Chadee 1992; Kampango, Cuamba & Charlwood 2011) whereas others have shown a decrease (Miller et al. 1970; Davies 1975; Rubio- Palis 1992; Souza et al. 2005) or no effect at all (Singh et al. 1996).

A higher number of larvae were collected in the ovitraps than eggs on the ovipositional substrate. This difference is due to the mosquito ovipositing behaviours of different genera. Some mosquito genera lay eggs on the surface of the water, either singly (Anopheles and Orthopodomyia), or in batches (Culex, Uranataenia, Mansonia and Coquillettidia) (Gillett 1971; Snow 1990). The genus Stegomyia lays eggs singly, but attaches eggs at the water’s edge (Gillett 1971). During the egg counts on the ovipositional substrates and the rim of the artificial containers, mosquitoes from the tribe Aedini were most likely to be counted. As these eggs couldn’t be identified to species, larval presence was used in the analysis.

This study focused on the container breeding mosquitoes using ovitraps, and human landing catches for anthropogenic host-seeking mosquitoes. As mosquito species differ in choice of oviposition sites and host-seeking preference, a range of trapping techniques need to be used. Human landing catches are the standard method for collecting malaria vectors in Malaysia, as Anopheles mosquitoes are not attracted to CDC light traps (Vythilingam, Chiang & Chan 1992). Searching for container breeding mosquitoes, such as Stg. albopicta and Cx. quinquefasciatus, can be time-consuming but ovitraps set in specific locations have proved to be a useful tool in detecting the presence and for estimating adult population sizes (Silver 2008). Ovitrap surveillance is the most common sampling method used for detecting the presence of dengue

Chapter 2 vectors in Malaysia (Lau et al. 2013). Ovitraps can provide presence data, but this does not mean an outbreak is going to happen.

Dengue has a profound impact in Malaysia, with the number of cases increasing exponentially in the last two decades (Sam et al. 2013). The resurgence of dengue in Southeast Asia has been linked to urbanisation and the accumulation of social detritus and storage containers in peri-urban area (Coker et al. 2011). Controlling the Stegomyia vectors is the main tool for the management of dengue, as there is currently no vaccine or specific treatment (WHO 2015a). Human behaviour, such as reducing man-made containers and tyres, contributes to controlling the breeding grounds of many dengue vectors (Chandren, Wong & AbuBakar 2015). Although dengue is usually described as an urban disease, it is increasingly being found in rural areas (Azami et al. 2011). The effects of deforestation, agricultural expansion and urbanisation appear to be complex. There is a need for further long term surveillance of mosquitoes in fragmented forest due to the rising cases of simian malaria, and in rural areas due to dengue risks. Deforestation, agricultural expansion, and urbanisation is continuing in Sabah, and the composition of mosquito communities may continue to change.

2.6. Conclusions

This study has given an overview of container breeding and anthropophilic mosquito species found in the Tawau Division. There were significant differences in the mosquito community composition between land uses, indicated by medically important mosquitoes landing in oil palm and rural housing compound sites. Anopheles balabacensis was the most common species landing in the disturbed forest, highly disturbed forest and oil palm sites, and is capable of spreading human and simian malaria. The dengue vector, Stegomyia albopicta, was found breeding in rural housing compounds, and landing in oil palm plantation sites. As dengue is increasingly being found in rural areas, control techniques in oil palm plantations should target potential breeding sites of vectors, such as water storage containers. This study provided preliminary data on the ecology and presence of the mosquito fauna in this region.

Chapter 3

Chapter 3: The Effects of Land Use Change on Anopheline Relative Abundance and Human Landing Rates

Hayley L. Brant1, Robert M. Ewers2, Indra Vythilingam3, Chris Drakeley4, Suzan Benedick5 & John D. Mumford1

1 Centre for Environmental Policy, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, SL5 7PY, UK

2 Department of Life Sciences, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, Berkshire, SL5 7PY, UK

3 Department of Parasitology, Faculty of Medicine, University of Malaya, Kuala Lumpur, 50603, Malaysia

4 Department of Immunology and Infection, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, WC1E 7HT, UK

5 Faculty of Sustainable Agriculture, Universiti Malaysia Sabah, Locked Bag No. 3, 90509, Sandakan, Sabah, Malaysia

3.1. Abstract

Southeast Asia is currently experiencing high rates of deforestation due to logging, oil palm expansion, and urbanisation. Deforestation and agricultural expansion are associated with increased vector-borne diseases. The objective of this study was to evaluate how land use change affects the anopheline fauna in Sabah, Malaysia, with a particular focus on young oil palm plantations (7-13 years). Mosquito collections were carried out using human landing catches along an anthropogenic disturbance gradient encompassing primary forest, disturbed forest, highly disturbed forest and oil palm plantation between 18:00- 23:00h. Results show that Anopheles balabacensis was the predominant species in the disturbed forest, highly disturbed forest and oil palm sites. Anopheles latens was the predominant species in the primary forest. In disturbed forest there was a greater abundance of An. balabacensis landing compared to the primary forest (p<0.0001), highly disturbed forest (p=0.026) and oil palm (p<0.020). Anopheles balabacensis is the main vector of malaria in Sabah, with the

Chapter 3 potential to transmit Plasmodium knowlesi, which is increasing in Sabah. One hypothesis is that this increase is associated with deforestation. Although the abundance of anophelines decreased in this study from disturbed forest through to oil palm expansion, mosquito landing rates were still occurring at relatively high numbers and have the potential of transmitting malaria.

3.2. Introduction

Anthropogenic land use changes, such as deforestation, agricultural encroachment and urbanisation, are associated with emerging and re-emerging infectious disease transmission in wildlife, domestic animals and humans (Gratz 1999; Patz et al. 2000). Environmental change can have a significant effect on habitat quality and microclimatic conditions, which in turn influence the abundance and survival of vectors and their parasites (Patz et al. 2000). For example, the clearing of forest is associated with the emergence of malaria in Africa, South America and Southeast Asia (Patz et al. 2004; Guerra, Snow & Hay 2006). Understanding the effects of land use on vector distribution, abundance and behaviour is important for predicting disease risks and minimising disease outbreaks.

Mosquitoes are among the forest vectors most sensitive to environmental change (Yasuoka & Levins 2007). Environmental alterations, such as temperature, humidity, water quality and the availability of suitable breeding sites may lead to significant changes in anopheline fauna. This may also lead to increased malaria risks within a small area or a large region (Dorvillé 1996; Afrane et al. 2005; Stresman 2010). The removal of intact forest may cause shifts in relative vector abundance, and increases the chance of disease transmission occurring from a vector that was not previously implicated (Norris 2004). There are also higher levels of human activity and human- wildlife interactions following habitat destruction, increasing the exposure to zoonoses (Wolfe et al. 2005; Wolfe, Dunavan & Diamond 2007)

Southeast Asia contains four biodiversity ‘hotspots’ containing high biodiversity and a large number of endemic species, but also has one of the highest rates of deforestation in any tropical region (Myers et al. 2000; Sodhi et al. 2004, 2010a). The expansion of oil palm (Elaeis guineensis) cultivation is the main driver of deforestation in Southeast Asia, with Malaysia and Indonesia being the top producers (Koh & Wilcove 2007; Koh

Chapter 3 et al. 2011). Oil palm plantations are associated with reduced species richness and overall abundance across most animal taxa (Fitzherbert et al. 2008; Foster et al. 2011).

Malaria, caused by the blood protozoan parasite of the genus Plasmodium, remains a public health problem in tropical and sub-tropical regions, affecting an estimated 198 million cases and 584,000 deaths in 2013 (WHO 2014b). Traditionally four Plasmodium species; Plasmodium falciparum, P. malariae, P. ovale, and P. vivax, were known to cause symptomatic malaria in humans, but a fifth Plasmodium species, P. knowlesi, is now responsible for human malaria (Singh et al. 2004; Cox-Singh et al. 2008; White 2008). Malaysia has been successful in reducing Plasmodium falciparum and P. vixax malaria cases over recent decades and aims to eliminate malaria by 2020 (William et al. 2014). However, there has been a recent increase in zoonotic human malaria cases caused by P. knowlesi (Rajahram et al. 2012; William et al. 2013, 2014). Plasmodium knowlesi accounted for 62% of all malaria incidences in 2013, and is thought to have increased due to deforestation (William et al. 2013), presenting a threat to malaria elimination. The current control measures in Malaysia include IRS, ITN distribution, artemisinin-based combination anti-malarial drugs, larviciding, environmental management measures and personal protection (Manguin 2013). Despite large reductions in malaria vectors, there are still significant challenges remaining such as insecticide resistance, and the zoonotic nature and increase of P. knowlesi (William et al. 2013; Manguin 2013).

Malaria has been shown to be a major health problem within an oil palm plantation in Papua New Guinea (Pluess et al. 2009), but to date there has only been one study on mosquito abundance in an oil palm plantation within Southeast Asia (Chang et al. 1997). Chang et al. (1997) showed a reduction in vector abundance following deforestation and conversion to oil palm, but it is unknown whether malaria vector abundance remains low as the oil palm matures. In this study, we investigated how anopheline distribution and landing rates change in relation to land use reflecting a gradient of anthropogenic land disturbance in Sabah, Malaysia, and whether anopheline abundance increases once oil palm plantations have started fruiting, and are forming a closed canopy.

Chapter 3

3.3. Methods

3.3.1. Study site

The study was conducted in the Tawau Division of Sabah, Malaysia. Four areas were selected along an anthropogenic disturbance gradient; primary lowland dipterocarp rainforest (PF), twice-logged disturbed dipterocarp rainforest (DF), twice-logged highly disturbed dipterocarp rainforest (HDF) and oil palm plantation (OP). Rainforest sites were defined based on canopy closure, measured at each collection site using a spherical densiometer. The average canopy closure values along the anthropogenic disturbance gradient were, 87.3% in the primary forest, 70.2% in disturbed forest, 61.8% in highly disturbed forest and 40.7% in oil palm sites. All data collection was carried out from October 2012 to April 2013.

Primary forest survey points were selected in the Maliau Basin Conservation Area (4°5’N, 116°5’E). The primary forest survey points were selected in an area that has never been logged commercially, with the exception of one site being selectively logged in the 1970s and 1990s to build the Maliau Field Centre. Forest quality is still classed as unaffected by logging and is substantially different from the commercially logged forest (Ewers et al. 2011). Logged forest and oil palm survey points were selected within the Benta Wawasan oil palm plantation. The 45,601 ha area is a mixture of twice-logged rainforest, virgin jungle reserve, acacia and oil palm. Logged forest survey points were in selectively twice-logged forest, logged during the 1970s and 1990s-2000s. Oil palm plantation survey points were in areas of Elaeis guineensis monocultures, planted in 2000 for one survey point, and 2006 for the rest. Further details of the project area are given by Ewers et al. (2011).

Survey points were selected to synchronise with the central sampling design of the ‘Stability of Altered Forest Ecosystems (S.A.F.E.) Project’, a large-scale fragmentation experiment, which is investigating the long-term effects of forest fragmentation (Figure 2.1, Ewers et al. 2011). We selected survey points separated by maximum distance (≥600 m) in each area. Three survey points were selected in the primary forest and oil palm plantation. Four survey points were selected in the disturbed forest and highly disturbed forest.

Chapter 3

3.3.2. Data collection

Mosquitoes were collected using paired human landing catches between 18:00- 23:00h. Human landing catches are the standard method for collecting malaria vectors in Malaysia, as Anopheles mosquitoes are not attracted to CDC light traps (Vythilingam, Chiang & Chan 1992). A pilot study demonstrated that Anopheles species in each sampling area started biting at 18:00h, with a decrease in abundance by 22:00h (Chapter 2, Figure 2.2). One mosquito collector remained constant during the overall sampling period, but the second mosquito collector at each sampling point was rotated every three days to control collector bias. Each survey point was sampled for a total of six nights.

3.3.3. PCR

Genomic data was extracted from the guts and salivary glands of a subset of collected Anopheles balabacensis using the DNeasy tissue kit (Qiagen, Germany) according to the manufacturer’s protocol. Sporozoite detection was carried out using a Plasmodium genus-specific PCR and primers (Table 3.1), as described by (Irene 2011). The PCR cycles used were 95°C for 4.5 minutes, followed by 44 cycles of 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds. The final extension cycle was 72°C for 2 minutes.

Table 3.1. Oligonucleotide sequences of PCR primers for detection of malaria parasites Primer Sequence Tm (°C) Expected product name size (bp) PlasF 5'- AGTGTGTATCAATCGAGTTTCT -3' 44.9 188 PlasR 5’- CTTGTCACTACCTCTCTTCTTTAGA -3’ 48.2

Chapter 3

3.3.4. Meteorological data

Rainfall (mm), air temperature (°C) and relative humidity (%) data were obtained from the SAFE Project and Maliau Basin Field Centre for the duration of the field survey (October 2012 to April 2013). In addition, lunar illumination (%), cloud cover and unusual climatic events (e.g. strong winds) were recorded every hour by the collectors.

3.3.5. Data analysis

Analyses were performed using R version 3.1.1 (R Core Team 2014). The number of mosquitoes landing per night was used as a measure of relative abundance. The effect of land use change (PF, DF, HDF and OP) on An. balabacensis, An. macarthuri, and An. latens was analysed using a generalised linear mixed-effect model (R package lme4: Bates et al. 2015), using day and site as random factors, with Poisson error distribution.

3.3.6. Ethics

This project was approved by the Medical Research and Ethics Committee (NMRR- 12-689-12521), Ministry of Health, Malaysia and Imperial College London Research Ethics Committee (ICREC-12-5-6), UK.

3.4. Results

A total of 2065 mosquitoes were collected from 84 human landing catch nights, consisting of 1977 (95.7%) anophelines and 88 (4.3%) culicines. Anopheles balabacensis was the predominant species in the disturbed forest (94.5%), highly disturbed forest (86.8%) and oil palm plantation sites (84.9%). A total of five (5%) An. balabacensis out of 108 were found positive for Plasmodium infection. Anopheles latens was the predominant species in the primary forest sites (37.5%). Human landing rates for each Anopheles species are given in Table 3.2, and total counts given in Table 3.3. The number of species collected, Shannon index, Simpson index and Chao1 varied across land use (Chapter 2, Table 2.2). The species accumulation curves did not reach an asymptote after human landing catch sampling, indicating not all species of mosquitoes had been collected (Chapter 2, Figure 2.3).

Chapter 3

Table 3.2. Number of mosquitoes landing per night (18:00-23:00h) per human bait of anophelines collected along an anthropogenic disturbance gradient in the district of Tawau, Sabah, Malaysia (± SE of the mean) Species Primary Disturbed Highly Oil palm forest (PF) forest (DF) disturbed plantation (OP) forest (HDF) An. Aitkenii group 0.03 (0.03) 0.08 (0.04) 0.19 (0.08) 0 (0) An. barbirostris 0 (0) 0.02 (0.02) 0 (0) 0 (0) An. vanus 0 (0) 0.04 (0.04) 0 (0) 0.03 (0.03) An. spp.* 0.06 (0.04) 0.19 (0.09) 0.17 (0.09) 0.03 (0.17) An. balabacensis 0.50 (0.14) 21.17 (2.96) 7.79 (1.47) 9.97 (3.63) An. kochi 0 (0) 0.02 (0.02) 0 (0) 0 (0) An. latens 0.69 (0.42) 0.13 (0.05) 0.31 (0.13) 0.06 (0.04) An. macarthuri 0.03 (0.07) 0.54 (0.18) 0.38 (0.13) 0.69 (0.32) An. maculatus 0 (0) 0.10 (0.10) 0.04 (0.04) 0.69 (0.29) An. tessellatus 0 (0) 0 (0) 0 (0) 0.03 (0.03) An. watsonii 0.03 (0.03) 0.10 (0.10) 0.1 (0.05) 0 (0) Total (An.) 1.33 (0.47) 22.4 (3.03) 8.8 (1.59) 12 (3.83) * Couldn’t be identified to species level

Table 3.3. Mosquitoes collected along an anthropogenic disturbance gradient in the district of Tawau, Sabah, Malaysia Species Primary Disturbed Highly Oil palm forest (PF) forest (DF) disturbed plantation forest (HDF) (OP) An. Aitkenii group 1 4 9 0 An. barbirostris 0 1 0 0 An. vanus 0 2 0 1 An. spp.* 2 9 8 10 An. balabacensis 18 1016 374 359 An. kochi 0 1 0 0 An. latens 25 6 15 2 An. macarthuri 1 26 18 25 An. maculatus 0 5 2 25 An. tessellatus 0 0 0 1 An. watsonii 1 5 5 0 Total 48 1075 431 423 * Couldn’t be identified to species level

Anopheles balabacensis abundance in the primary forest was significantly lower than disturbed forest, highly disturbed forest and oil palm plantation sites (log likelihood test; 2=21.10, df=3, p<0.0001, Figure 3.1a, Table 3.4). This was also seen in An. macarthuri abundance (log likelihood test; 2=7.86, df=3, p=0.049, Figure 3.1b, Table 3.5). Anopheles latens was the most abundant species in the primary forest, but no significant differences were seen in this species across land use (log likelihood test; 2=5.233, df=3, p=0.156, Figure 3.1, Table 3.5). Significant decreases were seen in

Chapter 3 the abundance of An. balabacensis from disturbed forest to highly disturbed forest (p=0.026) and oil palm plantations (p=0.020, Figure 3.1, Table 3.4).

Collector identity had no effect on An. balabacensis (2=3.3, df=2, p=0.193), An. macarthuri (2=0.93, df=2, p=0.629) or An. latens (2=3.75, df=2, p=0.15). Moonlight had no significant effect on the abundance of An. balabacensis (2=0.13, df=1, p=0.715), An. macarthuri (2=0.98, df=1, p=0.322) or An. latens (2=0.01, df=1, p=0.960). Peak biting of An. balabacensis during the collection period was observed between 18:30 and 20:00h in the disturbed forest, highly disturbed forest and oil palm sites (Figure 3.2). Rainfall slightly decreased mosquito abundance but not significantly for each species (An. balabacensis: 2=3.81, df=1, p=0.051, An. macarthuri: 2=3.41, df=1, p=0.06, An. latens: 2=2.94, df=1, p=0.09). Temperature during the collection had no effect on An. balabacensis (2=0.13, df=1, p=0.721), An. macarthuri (2=0.27, df=1, p=0.606) or An. latens (2=0.1, df=1, p=0.976).

Chapter 3

Figure 3.1. Number of mosquitoes. (a) Anopheles balabacensis (b) Anopheles macarthuri and An. latens) landing per person per night across an anthropogenic disturbance gradient from primary forest (PF), disturbed forest (DF), highly disturbed forest (HDF) and oil palm plantation (OP). Error bars show ± SE of the mean

Chapter 3

Table 3.4. Effects of land use and habitat characteristics on Anopheles balabacensis abundance. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Predictor Anopheles balabacensis Β SE Z p Intercept 3.512 0.349 10.06 <0.0001*** Area HDF -1.110 0.497 -2.234 =0.026* Area OP -1.257 0.541 -2.323 =0.020* Area PF -3.946 0.606 -6.510 <0.0001***

Table 3.5. Effects of land use and habitat characteristics on daily mosquito abundance of Anopheles macarthuri and Anopheles latens. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Predictor Anopheles macarthuri Anopheles latens Β SE Z p β SE z p Intercept -1.211 0.620 -1.955 =0.051 -2.272 0.645 -3.522 <0.0001*** Area HDF -0.337 0.722 -0.466 =0.641 0.849 0.738 1.150 =0.250 Area OP 0.163 0.784 -0.208 =0.835 -0.793 1.011 -0.785 =0.433 Area PF -2.963 1.291 -2.295 =0.022* 1.267 0.760 1.666 =0.096

Chapter 3

3.5. Discussion

Deforestation, agricultural expansion and intensification are the main drivers of habitat change in tropical regions (Laurance, Sayer & Cassman 2014). The rapid expansion of oil palm plantations threatens many species, however, less is understood about how vector abundance, distribution, exposure to humans and biting rates vary relative to land use change. Our data show, compared to primary rainforest, disturbed forest has a higher abundance of Anopheles balabacensis and An. macarthuri. There was a decrease in An. balabacensis abundance from disturbed forest to oil palm plantations.

Chang et al. (1997) sampled anophelines in Sarawak for five days yearly from the forest before development, through to oil palm maintenance, and found an 82% reduction in adult mosquitoes landing. During the four years of surveys, reductions in human-biting rates, entomological inoculation rates (EIR) and adult survival of An. donaldi and An. letifer decreased the risk of malaria transmission by 90% (Chang et al. 1997), but this percentage doesn’t account for the large increase in human population density that accompany plantation development. In this study, oil palm had a lower abundance of An. balabacensis than disturbed forest, but also had a higher human population density. Although vector abundance decreased from disturbed forest to oil palm, there is an increase in exposure potential due to the human population density and activity in the oil palm area.

In this study, we sampled in young oil palm plantations, ranging from trees planted 7 to 13 years ago (4-8 m high). Oil palm plantations have a 25-30 year life cycle, and begin to fruit after 3-5 years (Butler, Koh & Ghazoul 2009). Old plantations (22 years old, ~13 m closed canopy) have shown to have a more buffered microclimate than young plantations (8 years old, ~4 m closed canopy) (Luskin & Potts 2011). Chang et al. (1997) showed huge reductions in human-biting rates of An. donaldi from logged forest through to oil palm planting and maintenance (1 year old, 2-3m high). The reduction of An. donaldi may be explained by the microclimate and canopy cover of oil palm plantations. Recently planted oil palms have a highly variable microclimate and an open canopy (Luskin & Potts 2011). Our study found there was no significant difference between the abundance of An. balabacensis in highly disturbed forests and oil palm plantations. This may be due to a more buffered microclimate and closed canopy than the previous study, resulting in no difference in mosquito abundance

Chapter 3 between oil palm plantations and highly disturbed forests. The disturbed forest had a significantly higher abundance of An. balabacensis than highly disturbed forests and oil palm plantations, but also had a higher percentage of canopy cover. If the high abundance of An. balabacensis is due to increased canopy cover and a buffered microclimate, there is the potential for this species to increase as oil palm matures and the canopy eventually closes.

Mosquito survival and reproduction rates are strongly affected by fluctuations in temperature, precipitation and humidity (Clements 1992). Oil palm plantations have been shown to have a higher fluctuation in temperature and humidity over a 24 hour period than forests (Koh, Levang & Ghazoul 2009; Turner & Foster 2009; Luskin & Potts 2011; Hardwick et al. 2015), which is predicted to strongly influence mosquito survival, especially larval survival. Temperature affects mosquitoes at every stage of their life cycle, as well as parasite survival. Provided the temperature is not too extreme, a higher temperature can shorten mosquito and parasite development time (Stresman 2010), increasing the likelihood of that mosquitoes are able to transmit disease.

Rainfall, which has a direct effect on humidity, affects the expected lifespan of adult mosquitoes. Rainfall also provides a breeding site for the aquatic stage of the mosquito life cycle. Whilst vectors need rain to breed, excessive rainfall may lead to flushing of breeding sites and thus reducing survival rates (Martens et al. 1995). There were no collections occurring during high winds or heavy rainfall, but light rain during the collection slightly decreased mosquito catch.

In this study, although the mean daytime maximum temperature varied from 29°C in primary forest to 34.9°C in oil palm, by 18:00h, all areas cooled to 24°C and had high humidity. The temperature and humidity remained similar in each area during the sampling night. Moonlight was not a significant factor in this study. Some studies have shown that moonlight increases relative abundance (Bidlingmayer 1964; Charlwood et al. 1986; Birley & Charlwood 1989; Chadee 1992; Kampango, Cuamba & Charlwood 2011) whereas others have shown a decrease (Miller et al. 1970; Davies 1975; Rubio-Palis 1992; Souza et al. 2005) or no effect at all (Singh et al. 1996).

Anopheles balabacensis was the most predominant anopheline collected in the disturbed forest, highly disturbed forest and oil palm plantation sites, accounting for

Chapter 3

94.5% of anophelines collected in the disturbed forest. Anopheles balabacensis is considered the most important vector of human malaria parasites in Sabah, Malaysia (Hii & Vun 1985; Khoon 1985; Hii et al. 1988; Sallum et al. 2005). In earlier studies in Sabah, Anopheles balabacensis was the predominant species in the Kinabatangan, followed by An. donaldi and An. maculatus (Hii et al. 1987). A later study between 2001 and 2003, showed An. donaldi replaced An. balabacensis as the predominant species and primary vector (Vythilingam et al. 2005). Anopheles balabacensis has now recolonised and become the predominant species and primary vector in Sabah (Wong et al. 2015).

Anopheles balabacensis is also a member of the Leucosphyrus group, and able to transmit P. knowlesi. Many members of the Leucosphyrus group are found within forested areas, they feed primarily on monkeys, and are capable of transmitting various Plasmodium species (Sallum et al. 2005). The immature stages can be found in shaded temporary pools of fresh water, including ground puddles, animal footprints, wheel tracks, ditches and rock pools (Manguin 2013). In this study, Anopheles balabacensis larvae were observed in wheel tracks occurring in the oil palm plantation. A pilot run of this study showed An. balabacensis was biting as early as 18:00h, with a high rate of biting between 18:30 and 20:00h. Studies during previous decades showed this species was mainly a late night biter (21:00-22:00h), but now appears to be biting during the early evenings (Hii & Vun 1985; Rohani et al. 1999; Vythilingam et al. 2005). The effects of insecticide-treated bednets (ITNs) on the time of biting and host choice has been reported in multiple studies, with mosquitoes biting earlier in the evening, before mosquito hosts go to bed (Mbogo et al. 1996; Takken 2002).

Anopheles macarthuri and An. latens also belong to the Leucosphyrus group (Sallum et al. 2005) and were the second and third most abundant species in this study. In previous studies, An. macarthuri was found to mainly be attracted to monkey-baited traps, but is not a confirmed vector of P. knowlesi (Vythilingam et al. 2006; Baird 2009; Vythilingam 2010). Anopheles latens has been incriminated as a P. knowlesi vector, and is attracted to both humans and macaque hosts (Vythilingam et al. 2006; Tan et al. 2008; Vythilingam 2010). Current vector control methods in Malaysia, including insecticide treated bednets and indoor residual spraying, are not sufficient in controlling P. knowlesi due to the exophagic and exophilic behaviour of vectors (Manguin 2013). Long-lasting insecticidal hammocks (LLIH) within forested areas may

Chapter 3 prove to be an effective additional tool in malaria reduction in Malaysia (Thang et al. 2009; Sochantha et al. 2010; Manguin 2013).

Malaysia is in the pre-elimination phase of malaria elimination (WHO 2014b), and has successfully reduced P. falciparum and P. vivax in Sabah (William et al. 2014). Plasmodium knowlesi, currently the most common cause for human malaria in Sabah, continues to threaten the achievement of malaria control goals (William et al. 2014). Family clusters are occurring in P. knowlesi cases, suggesting malaria transmission is occurring peri-domestically (Barber et al. 2012). Family clusters were previously considered rare because transmission was mainly occurring in densely forested areas. This change in transmission is thought to be linked to deforestation and land use change (Barber et al. 2012). Deforestation is still occurring at high rates in Malaysia, and although human-vector-human transmissions are still considered unlikely, P. knowlesi may increase and become a significant public health problem (Imai et al. 2014).

3.6. Conclusions

This study has given an overview of anophelines found along an anthropogenic disturbance gradient in the Tawau Division of Sabah, Malaysia. Anopheles balabacensis, the main vector of malaria in Sabah, was the predominant species found in the disturbed forest, highly disturbed forest and oil palm sites. Anopheles latens was the predominant species in the primary forest. Although many studies have shown deforestation is associated with increased vector abundance, this study has shown that there is a decrease in An. balabacensis abundance from disturbed forest to oil palm plantations. Despite a decrease in vector abundance, human exposure increases from disturbed forest to oil palm plantation, due to higher human population and activity in these areas. Since land use change is complex, further accumulation of data on biting rates is needed to help develop predictive models to reduce disease transmission.

Chapter 4

Chapter 4 – Vertical Stratification of Adult Mosquitoes (Diptera: Culicidae) within a Tropical Rainforest in Sabah, Malaysia

Hayley L. Brant1, Robert M. Ewers2, Indra Vythilingam3, Chris Drakeley4, Suzan Benedick5 & John D. Mumford1

1 Centre for Environmental Policy, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, SL5 7PY, UK

2 Department of Life Sciences, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, Berkshire, SL5 7PY, UK

3 Department of Parasitology, Faculty of Medicine, University of Malaya, Kuala Lumpur, 50603, Malaysia

4 Department of Immunology and Infection, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, WC1E 7HT, UK

5 Faculty of Sustainable Agriculture, Universiti Malaysia Sabah, Locked Bag No. 3, 90509, Sandakan, Sabah, Malaysia

4.1. Abstract

Malaria cases caused by Plasmodium knowlesi, a simian parasite naturally found in long-tailed and pig-tailed macaques, are increasing rapidly in Sabah, Malaysia. One hypothesis is that this increase is associated with changes in land use. A study was carried out to identify the anopheline vectors present in different forest types and to observe the human landing behaviour of mosquitoes. Mosquito collections were carried out using human landing catches at ground and canopy levels in the Tawau Division of Sabah. Collections were conducted along an anthropogenic disturbance gradient (primary forest, lightly logged virgin jungle reserve and salvage logged forest) between 18:00-22:00h. Results showed Anopheles balabacensis, a vector of P. knowlesi, was the predominant species in all collection areas, accounting for 70% of the total catch, with a peak landing time of 18:30-20:00h. Anopheles balabacensis had a preference for landing on humans at ground level compared to the canopy (p<0.0001). A greater abundance of mosquitoes were landing in the logged and lightly

Chapter 4 logged forest compared to the primary forest (p<0.0001). Our results demonstrate the potential ability of An. balabacensis to transmit P. knowlesi between canopy-dwelling simian hosts and ground-dwelling humans, and that forest disturbance increases the abundance of this disease vector. These results, in combination with regional patterns of land use change, may partly explain the rapid rise in P. knowlesi cases in Sabah. This study provides essential data on anthropophily for the principal vector of P. knowlesi which is important for the planning of vector control strategies.

4.2. Introduction

Malaria still remains a public health problem throughout tropical and sub-tropical regions of the world, with an estimated 198 million cases causing 584,000 deaths in 2013 (WHO 2014b). Four Plasmodium species are recognised as causing human malaria; Plasmodium falciparum, P. malariae, P. ovale and P. vivax, but recently a fifth species, P. knowlesi, has been recognised as causing symptomatic malaria in humans (Singh et al. 2004; Cox-Singh et al. 2008; White 2008). Plasmodium knowlesi, transmitted by the forest-dwelling Anopheles from the Leucosphyrus group, is an emerging cause for zoonotic human malaria in Southeast Asia (Cox-Singh et al. 2008; Kantele & Jokiranta 2011; Antinori et al. 2013; Millar & Cox-Singh 2015). Malaysia has had a successful malaria control program, aimed to eliminate malaria by 2020, with marked reductions in reported cases of P. falciparum and P. vivax, but there has been a recent increase in P. knowlesi cases (Rajahram et al. 2012; William et al. 2013, 2014; Yusof et al. 2014). Plasmodium knowlesi is now the most common cause of malaria in the Malaysian state of Sabah, accounting for 62% of all malaria incidences in 2013 and presenting a threat to malaria elimination (William et al. 2014).

It has been proposed that land use change, including deforestation, forest fragmentation and agricultural practices, has increased the prevalence of P. knowlesi by increasing the encroachment of humans into previously forested areas, allowing a higher interaction between vectors and human and macaque hosts (Vythilingam et al. 2008; William et al. 2013). The increase in P. knowlesi cases may also be underestimated due to misdiagnosis during microscopic examination (Barber et al. 2013b; Singh & Daneshvar 2013). Microscopy of stained blood smears allows

Chapter 4 differentiation between species, but frequent misdiagnosis occurs in areas containing P. falciparum, P. vivax and P. knowlesi (Barber et al. 2013b).

Monkeys, particularly the long-tailed macaque (Macaca fascicularis) and the pig-tailed macaque (Macaca nemestrina) found in Southeast Asia, are the two main natural hosts of P. knowlesi (Garnham 1966). A study in Sabah showed nearly all patients with P. knowlesi malaria had a recent history of forest or forest-edge exposure, and had seen a macaque in the preceding month (Barber et al. 2013a), supporting the premise that P. knowlesi is primarily zoonotic. The existence of a zoonotic reservoir of P. knowlesi malaria poses a challenge to malaria elimination, because even if infections are eliminated from humans, there is the risk of future spill-over from macaque hosts.

Most members of the Leucosphyrus group, from the genus Anopheles, feed primarily on monkeys in the canopy and are capable of transmitting various Plasmodium species (Sallum, Peyton & Wilkerson 2005). Previously incriminated mosquito vectors of P. knowlesi, from the Leucosphyrus group, bite and rest outdoors, which poses additional challenges to malaria elimination, as current control methods (ITNs, IRS) will not be effective (Tan et al. 2008; Jiram et al. 2012). Anopheles balabacensis is the predominant vector of human malaria in Sabah (Hii & Vun 1985; Khoon 1985), and has also been incriminated as a P. knowlesi vector (Collins, Contacos & Guinn 1967; Vythilingam 2010).

Most primates are arboreal. Although some species of chimpanzees, baboons and macaques rest and feed at ground level during the day, primates almost always sleep in the canopy during the night (Anderson 2000). As Anopheles species generally bite between 6pm and 6am, primate roosting sites are potentially a key location for disease transmission between primate hosts. It is hypothesised that vectors are biting humans at ground level, but if given the opportunity, will bite at canopy level. We also expect that key vector species should be present in the disturbed forest habitats where people come into contact with monkeys, but it’s unknown what should happen in primary forest.

Since transmission is increasing in Sabah, it is important to identify the vectors present and understand their biting behaviour. Despite ongoing studies in the Interior, West Coast, Kudat and Sandakan Sabah Divisions, Tawau Division is less studied. This

Chapter 4 study was conducted to determine the vertical distribution of mosquitoes and their biting preference in Sabah, Malaysia.

4.3. Methods

4.3.1. Study site

The study was conducted in the Tawau Division of Sabah, Malaysia. Three areas were selected along a forest disturbance gradient; primary lowland dipterocarp rainforest (PF), virgin jungle reserve (VJR) and twice-logged disturbed dipterocarp rainforest (LF). All data collection was carried out from April to July 2014.

Primary forest survey points were selected in the vicinity of Danum Valley Field Centre (4°58’N, 117°42’E), located within the Danum Valley Conservation Area. This area consists of 43,800 ha of protected dipterocarp rainforest (Marsh & Greer 1992). Virgin jungle reserve (4°40’N, 117°32’E) and logged forest survey points (4°41’N, 117°34’E) were selected within the Benta Wawasan oil palm plantation. The 45,601 ha area is a mixture of twice-logged rainforest, virgin jungle reserve, acacia and oil palm. The VJR, of 2200 ha, has been logged around the edge, but never logged in the steep interior (Ewers et al. 2011). Survey points were selected 500-1000 m from the VJR edge in locations that had undergone light logging. Logged forest survey points were in selectively twice-logged forest, logged during 1970s, 1990s-2000s and currently disturbed by logging activity in surrounding areas. Further details of the project area are given by Ewers et al. (2011).

Survey points were selected to synchronise with the central sampling design of the ‘Stability of Altered Forest Ecosystems (S.A.F.E.) Project’, a large-scale fragmentation experiment, which is investigating the long-term effects of forest fragmentation (Ewers et al. 2011, Figure 4.1). Three survey points, with a minimum separation distance of 500 m, were selected in each area. One tree was selected at each point based on its accessibility into the canopy, low density of epiphytes and height. Visual tree assessments (VTA) were carried out to make sure every tree was safe to climb. The trees selected ranged from a height of 15 m in the logged forest to 30 m in the virgin jungle reserve and primary forest.

Chapter 4

Figure 4.1. Map of the Stability of Altered Forest Ecosystems Project, located in Sabah, Malaysia (a) Primary forest sites (b) Continuous twice- logged forest (c) Twice-logged forest and fragmented forest in an oil palm matrix (d) Oil palm plantation sites (e) The fragmentation experiment comprising of six blocks (A-F). Reproduced from Ewers et al. (2011).

Chapter 4

4.3.2. Data collection

Mosquitoes were collected using human landing catches at ground and canopy height between 18:00-22:00h. Four nights of collections were carried out in PF and VJR, and five nights in LF, using a rotation of collectors. Access was gained into the canopy using line insertion to high branches (Ellwood & Foster 2001), followed by the double rope climbing technique taught by Canopy Access Limited (http://www.canopyaccess.co.uk). The average canopy height surrounding each selected tree was calculated using a laser rangefinder. Canopy samples were collected at a height of two-thirds the average canopy height at that location (10-20 m). Ground and canopy collections were conducted simultaneously.

The collectors, with the aid of a red torch light, aspirated mosquitoes off their own legs. Collected mosquitoes were placed into cups covered with a net cloth, and a new cup was used during every half an hour of collection. Mosquitoes were taken back to the field laboratory to be killed and sorted into individual tubes with silica gel. All mosquitoes were identified morphologically using keys (Reid 1968; Rattanarithikul et al. 2005a; b, 2006a; b, 2010; Sallum et al. 2005).

4.3.3. Meteorological data

Rainfall (mm), air temperature (°C) and relative humidity (%) data were obtained from the S.A.F.E. Project and Danum Valley Field Centre for the duration of the field survey (March to July 2014). In addition, lunar illumination (%), cloud cover and unusual climatic events (e.g. strong winds) were recorded every half an hour by the collectors.

4.3.4. Data analysis

Analyses were performed using R version 3.1.1 (R Core Team 2014). Simpson and Shannon diversity indices were calculated in each area using the vegan function ‘diversity’ (R package vegan, 'diversity': Oksanen et al. 2015). Species accumulation curves were calculated in each area using the vegan function ‘specaccum’ (R package vegan, 'specaccum': Oksanen et al. 2015). The Chao species estimator (Chao 1) and Abundance Coverage Estimator (ACE) were calculated the extrapolated species richness in each area using ‘chao1’ and ‘ACE’ functions in the R package ‘fossil’ (R package fossil, 'chao1', 'ACE': Vavrek 2015).

Chapter 4

The effect of canopy height and forest disturbance (PF, VJR & LF) on mosquito abundance was analysed using a generalised linear mixed-effect model (R package lme4: Bates et al. 2015), using day and site as random factors, with Poisson error distribution. Chi-squared tests were used to compare the relative abundance of vector and non-vector species in each area, and between ground and canopy level. Differences in community composition at ground and canopy height were explored using Detrended Correspondence Analysis (R package vegan, ‘decorana’: Oksanen et al. 2015). The number of mosquitoes landing per night was used as a measure of relative abundance. We tested for significant differences in community composition using a linear model with the first DCA axis as the response variable against canopy height and forest disturbance.

4.3.5. Ethics

This project was approved by the Medical Research and Ethics Committee (NMRR- 12-689-12521), Ministry of Health, Malaysia and Imperial College London Research Ethics Committee (ICREC-12-5-6), UK.

4.4. Results

4.4.1. Mosquito abundance

A total of 807 mosquitoes were collected from 39 human landing catch nights, consisting of 743 (92.1%) anophelines, and 64 (7.9%) culicines. A total of 555 (68.8%) mosquitoes from 21 species were found at ground level in comparison to 252 (31.2%) mosquitoes from 10 species at canopy level. Anopheles balabacensis was the predominant species at ground and canopy level at each collection site. A full list of species is given in Table 4.1.

The number of species collected, Shannon index, Simpson index, Chao1 and ACE varied across land use and between ground and canopy level (Table 4.2-3). The species accumulation curves did not reach an asymptote at ground and canopy level, indicating not all species of mosquitoes had been collected (Figure 4.2).

Chapter 4

Table 4.1. Mosquitoes collected from different collection sites in the district of Tawau, Sabah, Malaysia Mosquito genera and Number collected at: species Primary forest Virgin jungle reserve Logged forest Ground Canopy Ground Canopy Ground Canopy Am. orbitae 0 0 0 0 1 0 An. Aitkenii group 0 0 21 0 18 0 An. barbirostris 0 0 1 0 3 0 An. sp.* 0 0 1 0 10 3 An. balabacensis 21 12 83 23 296 130 An. latens 0 1 1 5 1 1 An. macarthuri 0 1 10 4 16 5 An. maculatus 1 0 7 2 12 4 An. watsonii 1 6 3 15 14 11 Arm. confusus 0 0 1 0 0 0 Arm. jugraensis 0 0 0 0 5 0 Arm. sp.* 0 0 1 0 0 0 Col. pseudotaeniatus 0 0 1 0 0 0 Coq. crassipes 0 0 0 1 0 1 Cx. sitiens 0 0 0 0 5 1 Cx. vishnui 0 0 0 0 2 0 Cx. (Lophoceraomyia) 0 0 0 0 1 0 Do. ganapathi 1 2 0 3 4 16 Pr. ostentatio 0 0 1 0 0 0 Ph. prominens 0 0 0 0 1 0 Stg. albopicta 0 0 0 0 4 0 Stg. sp.* 0 0 1 0 4 2 Ve. sp.* 0 0 1 0 1 3 Total mosquitoes 24 22 133 53 398 177 No. of collection nights 12 12 12 12 15 15 Mosquitoes/nights 2 1.8 11.1 4.42 26.5 11.8 * Couldn’t be identified to species level

Table 4.2. Mean species richness and diversity indices (± SE) of mosquito communities, collected at ground level, using human landing catches, in primary forest (PF), virgin jungle reserve (VJR) and logged forest (LF) Land use Human landing catches N Species Shannon Simpson Chao1 ACE no. index index PF 12 4 0.06 (0.04) 0.45 (0.14) 7 7 VJR 12 14 0.70 (0.10) 0.40 (0.05) 50 19.4 LF 15 18 0.88 (0.12) 0.43 (0.06) 30.5 19.8

Chapter 4

Table 4.3. Mean species richness and diversity indices (± SE) of mosquito communities, collected at canopy level, using human landing catches, in primary forest (PF), virgin jungle reserve (VJR) and logged forest (LF) Land use Human landing catches N Species Shannon Simpson Chao1 ACE no. index index PF 12 5 0.24 (0.13) 0.65 (0.12) 7 6 VJR 12 7 0.71 (0.15) 0.50 (0.09) 7.5 7.36 LF 15 11 0.56 (0.11) 0.32 (0.06) 15.5 12.41

Chapter 4

Figure 4.2. Species accumulation curves for human landing catch sampling in primary forest at (a) Ground and (b) Canopy level,virgin jungle reserve at (c) Ground and (d) Canopy level, and logged forest at (e) Ground and (f) Canopy level. Shaded area indicates 95% confidence intervals

Chapter 4

4.4.2. Effect of height and land use on mosquito abundance

Mosquito abundance in the canopy was significantly lower than at ground level (Table 4.4-5, Figure 4.3a). Logged forest had a significantly higher abundance than virgin jungle reserve or primary forest (Figure 4.3a). Similar patterns were seen with abundance of An. balabacensis (Table 4.4-5, Figure 4.3b). Rainfall and a higher evening temperature significantly decreased mosquito abundance, whereas moonlight increased the abundance (Table 4.4-5). Collector identity had no effect on mosquito abundance (2=3.871, df=2, p=0.144). Peak biting of An. balabacensis during the collection period was observed between 19:00 and 20:00h in logged forest and virgin jungle reserve (Figure 4.4). A chi-squared test of the contingency table of vector and non-vector species showed significant differences between ground and canopy level (2=3.77, df=1, p=0.05), but no differences across land use (2=1.14, df=2, p=0.565).

Table 4.4. Effects of height, land use and habitat characteristics on daily mosquito abundance of all species combined, and on Anopheles balabacensis abundance separately. Chi-square (2), degrees of freedom (df), and p-values are given using log likelihood ratio test. Minimum adequate model (Final model) tested against the null model Predictor All species Anopheles balabacensis 2 Df p 2 df p Height 81.89 1 <0.0001*** 75.429 1 <0.0001*** Land use 10.939 2 =0.004** 7.150 2 =0.028* Temperature 34.920 1 <0.0001*** 24.166 1 <0.0001*** Rainfall 14.555 1 =0.0001*** 12.480 1 =0.0004*** Moonlight 4.489 1 =0.034* 5.500 1 =0.019* Final model 157.65 6 <0.0001*** 139.41 6 <0.0001***

Chapter 4

Table 4.5. Effects of height, area and habitat characteristics on daily mosquito abundance of all species combined, and on Anopheles balabacensis abundance separately. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Predictor All species Anopheles balabacensis β SE z p β SE z p Intercept 10.367 1.241 8.356 <0.0001*** 10.233 1.423 7.190 <0.0001*** Height -0.050 0.006 -8.804 <0.0001*** -0.060 0.007 -8.348 <0.0001*** Area PF -2.268 0.545 -4.165 <0.0001*** -2.147 0.676 -3.176 =0.001** Area VJR -0.307 0.518 -0.593 =0.554 -0.633 0.657 -0.965 =0.335 Temperature -0.295 0.047 -6.330 <0.0001*** -0.307 0.053 -5.788 <0.0001*** Rainfall -0.427 0.100 -4.306 <0.0001*** -0.448 0.115 -3.883 <0.0001*** Moonlight 0.006 0.003 2.129 =0.033* 0.007 0.003 2.507 =0.012*

Chapter 4

Figure 4.3. Effects of collection height on the human landing rate (number of mosquitoes per night per bait) across a forest disturbance gradient: primary forest (PF), lightly logged virgin jungle reserve (VJR), and twice-logged forest (LF). (a) Total abundance of all species combined, (b) Abundance of the most common species, Anopheles balabacensis, alone. Error bars show ± SE of the mean

Chapter 4

Figure 4.4. Hourly number of Anopheles balabacensis landing per person per night, at ground and canopy level, across an anthropogenic disturbance gradient from (a) Primary forest, (b) Virgin jungle reserve and (c) Logged forest. Error bars show ± SE of the mean

Chapter 4

4.4.3. Community composition

The DCA plot indicated that the community composition of the canopy was significantly different to ground level collection, and mainly separated along the first axis

(F1,60=24.72, p<0.0001, Figure 4.5). The difference of composition across land use was not separated along the first axis (F2,60=0.92, p=0.4), but the interaction between height and land use was significant (F2,60=8.37, p<0.001). The first two axes accounted for 66.9 % of the total variance. Anopheles balabacensis was prevalent at both ground and canopy level, but the ground level community also included species such as An. barbirostris, Arm. jugraensis, Cx. vishnui and the An. Aitkenii group, which were not present in the canopy (Appendix C, Figure C.1).

Figure 4.5. Detrended Correspondence Analysis (DCA) plot showing the major axes of variation for adult mosquito abundance at ground level and in the canopy of a tropical rainforest. The two axes represent linear summaries of the variation in the species numbers and areas

Chapter 4

4.5. Discussion

Studying the host preference of simian malaria vectors in the canopy is essential because the hosts are primarily arboreal. In order to develop, sustain or adapt a good control programme, it is important to monitor mosquito populations as well as their hosts and host-seeking preference, distribution and behaviour. Although previous studies in Southeast Asia have used monkey-baited traps at different canopy heights (Tan et al. 2008; Vythilingam et al. 2008; Jiram et al. 2012), this is the first study to attempt human landing catches, using this method, in the canopy. This study found that there was a higher abundance and human landing rate of mosquitoes at ground level, where people tend to be, than in the canopy where the simian hosts reside. This trend was driven by An. balabacensis, a key malaria vector in Sabah, and highlights the potential importance of this species in transmitting Plasmodium species from simian to human hosts.

Anopheles balabacensis was the most abundant mosquito in all sampled areas, accounting for 70% of all collected species. Anopheles balabacensis is considered the most important vector of human malaria parasites on Banggi Island and Sabah, Malaysia (Hii & Vun 1985; Khoon 1985; Hii et al. 1988; Sallum et al. 2005) . In Sabah, An. balabacensis was found to be mainly exophagic, but could also be endophagic and exophilic (Sallum et al. 2005; Sinka et al. 2011). There were also two distinct subpopulations, one more zoophilic and one more anthropophilic (Hii 1985; Hii, Birley & Sang 1990; Hii et al. 1991). Anopheles balabacensis occurs in forested areas, and readily bites human and monkey hosts, making it an ideal vector of simian malaria (Eyles et al. 1964; Vythilingam 2010).

Anopheles balabacensis has shown to bite as early as 18:00h in the Tawau Division (Chapter 3). Studies have shown the species bites as early as 19:00h in recent years in comparison to late night biters in previous decades (Hii 1985; Rohani et al. 1999; Vythilingam et al. 2005). Currently insecticide treated betnets (ITN) and indoor residual spraying (IRS) are the two main control methods in Malaysia (Rundi 2011). Given that An. balabacensis is early evening biting, highly anthropophilic, exophagic and exophilic, current control methods (ITN, IRS) are not sufficient to break the transmission cycle of P. knowlesi (Manguin 2013). In Vietnam and Cambodia, long- lasting insecticidal hammocks (LLIH) were shown to reduce malaria incidence and

Chapter 4 prevalence in forested areas, and may prove to be an additional effective tool in reduction of malaria in Malaysia (Thang et al. 2009; Sochantha et al. 2010; Manguin 2013). The use of repellents have been used for malaria control, but need to be tested in forest and plantations areas.

Different mosquito species have a particular flight distribution, with certain species flying and feeding close to ground, some species showing a preference for higher canopy heights, while others show a random distribution (Bates 1944; Clements 1999). The percentage biting at different canopy heights can be affected by microclimate conditions, such as relative humidity, temperature, wind speed and rainfall (Bates 1945; Clements 1992), but may also change according to time of day (Silver 2008). This study found a different community composition of mosquitoes in the canopy in comparison to ground level. The species turnover that causes the difference in community composition between ground and canopy was driven by species that have limited reported medical importance.

Moonlight appeared to have a significant impact on mosquito activity, with human landing rates increasing on bright nights. Although some studies have shown moonlight increases relative abundance of biting vectors (Bidlingmayer 1964; Charlwood et al. 1986; Birley & Charlwood 1989; Chadee 1992; Kampango, Cuamba & Charlwood 2011), others have shown a decrease (Miller et al. 1970; Davies 1975; Rubio-Palis 1992; Souza et al. 2005) or no effect at all (Singh et al. 1996). Collection bias was reduced in this study by collecting in each area under different phases of the moon.

This study also showed how forest disturbance affected mosquito abundance, species richness and human landing rates. Vector abundance increased in the virgin jungle reserve and logged forest, but remained low in the primary forest. These results may be explained by the availability of larval breeding sites. Logged forests and virgin jungle reserves have increased levels of solar radiation reaching breeding habitats, in comparison to primary forests, which decreases mosquito development time (Leisnham et al. 2004). Wheel tracks in logged areas due to logging activities can also provide breeding sites for a range of mosquito species, whereas wheel tracks are not present within primary forests or virgin jungle reserves (Rattanarithikul et al. 2005b). Species richness, estimated by the Chao1 index and ACE, differed across land use

Chapter 4 and height, with logged forest and ground level having a higher species richness than primary forest and canopy.

Malaysian Borneo is a hotspot for forest loss and degradation due to timber and oil palm demands, resulting in nearly 80% of land area affected by logging and clearing operations from 1990 to 2009 (Bryan et al. 2013). Deforestation, logging and forest fragmentation increase human-wildlife interactions, which can expose humans, livestock and wildlife to newly recognised pathogens (Wolfe et al. 2000). Deforestation in many areas has caused monkeys, humans and forest dwelling vectors to interact more frequently (Vythilingam 2010), which may be associated with the increase in prevalence of P. knowlesi in Sabah (William et al. 2013). Moreover, our results demonstrate that the disturbed forest habitats in this particular setting have higher abundances of malaria vectors, further potential for increasing the probability of simian-human disease transmission at these transition zones between people and forest.

4.6. Conclusions

This study has given an overview of mosquito species found in the Tawau Division, including human host-seeking preference at canopy and ground levels. Anopheles balabacensis was the predominant species found in primary forest, virgin jungle reserve and logged forest with a preference for landing on humans at ground level. As An. balabacensis is a vector of human and simian malaria, these findings will be useful for the planning of control strategies of malaria vectors.

Chapter 5

Chapter 5 – Effects of Marking Methods and Fluorescent Dusts on Stegomyia aegypti (Aedes aegypti) Survival

Borame L. Dickens1 & Hayley L. Brant2

1 Centre for Environmental Policy, Faculty of Natural Sciences, Imperial College London, South Kensington Campus, London SW7 1NA, UK

2 Centre for Environmental Policy, Faculty of Natural Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, SL5 7PY, UK

Published in Parasites & Vectors: Dickens, B.L. & Brant, H.L. (2014) Effects of marking methods and fluorescent dusts on Aedes aegypti survival, Parasites & Vectors, 7, 65.

5.1. Abstract

Tracking the movement of mosquitoes and understanding dispersal dynamics is essential for the control and prevention of vector-borne diseases. A variety of marking techniques have been used, including dusts and dyes. In this study, Stegomyia aegypti were marked using fluorescent dusts (‘DayGlo’: A-19 Horizon Blue & A-13-N Rocket Red; ‘Brian Clegg’: pink, blue & red), fluorescent paints (‘Brian Clegg’: blue, red & yellow) and metallic gold dust (‘Brian Clegg’). Dusting methods were those previously used in mark-release-recapture experiments, including application with a bulb duster, creation of a dust storm or shaking in a bag. Results showed marking mosquitoes using a dust storm allowed relatively high survival, compared to unmarked method controls, and high marking efficiency. The remaining methods gave varying results, with different colours significantly reducing mosquito survival. Mosquitoes marked with blue dusts had the most significant reduction in survivorship in comparison to controls. Mosquitoes marked with pinks or reds showed both reasonable performance in marking and impact on overall survival across males and females. This study showed that marking technique and colour can have a significant impact on the survival and marking coverage of a mosquito.

Chapter 5

5.2. Introduction

Dispersal is an essential component of understanding insect biology, behaviour, life history and population dynamics (Hagler & Jackson 2001). Understanding dispersal dynamics and flight ranges of mosquito vectors is essential for the mitigation of disease, successful implementation of protection against infection and an important step in understanding the ecology of a vector (Silver 2008). Mark-release-recapture (MRR) techniques can be used to estimate mosquito population densities, feeding behaviour, duration of gonotropic cycles and their dispersal behaviour (Silver 2008). An ideal marking technique should be cost-effective, easily applied, visible, non-toxic and should not affect the behaviour, development, longevity or reproduction of mosquitoes (Southwood & Henderson 2000; Hagler & Jackson 2001). Marking techniques have the potential to affect the mortality and dispersal rates of marked individuals in a way that could bias the results arising from MRR experiments, so preliminary experiments on marking methodologies are required prior to carrying out MRR experiments (Silver 2008).

A variety of methods have been used to mark mosquitoes, including dusts (Darling 1925; Sheppard et al. 1969; Reisen, Mahmood & Parveen 1979; Muir & Kay 1998; Russell et al. 2005; Williams et al. 2012), dyes (Welch et al. 2006; Tsuda et al. 2008), paints (Conway, Trpis & McClelland 1974; Cho et al. 2002), trace elements (Wilkins et al. 2007) and radioactive isotopes (Lindquist et al. 1967; Hamer et al. 2012) (see reviews in Southwood & Henderson 2000; Hagler & Jackson 2001; Silver 2008). One of the most common methods of marking mosquitoes externally is to apply micronized particles of dust (also known as powder or pigments), particularly fluorescent dust, to a large number of mosquitoes (Hagler & Jackson 2001). Dusts are cost-effective, available in a range of colours, easily applied and very detectable.

Several types of fluorescent dusts have been used to mark insects, including ‘Helecon’ and Radiant dusts, but the ‘DayGlo’ series A and AX are now the most commonly used as no adhesive is needed for mosquitoes to retain marks (Silver 2008). Dusts can be applied using a syringe (Muir & Kay 1998) or bulb duster (Jones et al. 2012), putting them in a bag containing dust and shaking them gently (Ikeshoji & Yap 1990), or by creating a dust storm within a cage (Renshaw, Service & Birley 1994). Many shaking methods can cause high mortality by applying too much dust to mosquitoes. This can

Chapter 5 increase mortality, decrease mobility and affect the sensory organs (Hagler & Jackson 2001), giving biased results in MRR studies. There is also the constraint of dusts not persisting long enough for long-term studies and transference of dusts to unmarked individuals (Fryer & Meek 1989; Hagler & Jackson 2001).

Paints can be applied individually (Conway, Trpis & McClelland 1974) or to large groups of individuals (Cho et al. 2002; Tsuda et al. 2008). Applying paints individually can be time-consuming, but mass marking by spraying paints can be easy and quick. Usually paints are diluted with acetone or alcohol before being sprayed from hand atomizers or spray guns (Hagler & Jackson 2001). Myles and Grace (1991) experimented with spray paints as an adhesive for borate dusts on termites, claiming they were non-toxic. Fluorescent paints adhere to body parts, can achieve 100% coverage and are readily identifiable under UV light (Forschler 1994). Marking with paints is usually believed to cause little mortality, but it is likely that applying paint spots to mosquito wings affects behaviour and possibly survival (Silver 2008). Droplet size and visibility needs to be balanced in order to ensure that the mosquito is unaffected and able to be seen.

The effects of marking adults with powders or stains, or any other substance, should always be carefully evaluated by comparing mortalities of marked and unmarked mosquitoes of the same species, sex and if possible same age over the lifetime of the insect. Unfortunately it appears from the literature that in many experiments where marking was carried out, the effects of the marking were not evaluated statistically, or inadequately so (Silver 2008). In order to improve marking efficiency for future dispersal and population studies, methods should be rigorously compared. This study was conducted to determine the best method and colour for marking mosquitoes for MRR experiments by comparing the mortalities of marked and unmarked individuals, whether immobilising the mosquitoes had adverse effects on survival and calculating how efficient each method is in marking.

Chapter 5

5.3. Methods

5.3.1. Mosquitoes

Stegomyia aegypti (=Aedes aegypti, see Reinert, Harbach & Kitching 2004) (Linnaeus) were used for all laboratory experiments. For the rearing of mosquitoes, second instar larvae were placed in plastic trays and fed daily with fish food (TetraMinBaby©). Pupae were transferred to a cup of water in an insect rearing cage. Larvae and adults were maintained under insectary conditions (27°C, 70% RH and a photoperiod of 12:12 [L:D] h) and provided with 10% sucrose solution.

5.3.2. Marking of mosquitoes

Two or three day old Stg. aegypti mosquitoes were randomly aspirated into 90 mm plastic containers (1 L tumblers) with gauze tops, until there were 15-18 females and 15-18 males in each container. After applying the marking technique, all experimental containers had cotton wool soaked in 10% sucrose solution placed on top of the gauze, which was refreshed daily. A plastic lid was placed over the top of the container to keep the humidity high. Dead mosquitoes were recorded daily and removed.

Mosquitoes were immobilised prior to the marking methods to allow a better coverage, and to prevent accidental release. To check that survival was not affected by immobilising them, an experiment was set up to compare briefly immobilised mosquitoes to control mosquitoes. Two containers were placed in a freezer (-18°C) for one minute, and then changed over to a container held at room temperature (22°C) until the mosquitoes had recovered. Two containers of immobilised mosquitoes and two containers of control mosquitoes were then placed back in the insect rearing room (27°C) until all mosquitoes died.

Dusts from two companies were used in this experiment; A-19 Horizon Blue and A- 13-N Rocket Red ‘DayGlo’ series A fluorescent pigments (DayGlo Color Corp, Cleveland, OH, USA), which will be referred to as ‘D Blue’ and ‘D Red’, gold metallic dust (Brian Clegg, UK), pink, blue and red fluorescent dusts (Brian Clegg, UK), which will be referred to as ‘BC Gold’, ‘BC Pink’, ‘BC Blue’ and ‘BC Red’. Yellow, blue and red fluorescent paints (Brian Clegg, UK) were also used in this experiment. DayGlo dusts are manufactured from organic dyes, incorporated into a melamine

Chapter 5 formaldehyde resin and grounded into a fine powder (Silver 2008). Brian Clegg dusts are composed of calcium/magnesium carbonate, for use as powder paints.

The marking methods used in this experiment were; placing mosquitoes in a bag with dust at the bottom and gently shaking (hereafter known as ‘bag’ method), using a bulb duster to create a small puff of dust (hereafter known as ‘bulb duster’ method), using a fan to create a small dust storm within a cage (hereafter known as ‘dust storm’ method) and lightly spraying paint in small droplets (hereafter known as ‘paint’ method). The paint solution was made by mixing 2 g dust, 200 ml paint of the same colour and 200 ml distilled water. This was repeated for each paint colour. The paint control was made up of 200 ml distilled water. After the mosquitoes had been immobilised, they were transferred to a tray and the paint solution was finely sprayed, using an atomiser, three times over the mosquitoes. Mosquitoes were then transferred to a container to recover, and placed in the insect rearing room. All DayGlo and Brian Clegg dusts were used for the bulb duster, bag and dust storm method. The bulb duster was loaded with 0.3 g dust per container; mosquitoes were transferred to a tray and sprayed four times. The bag method had 0.3 g dust placed in the bottom of the bag; immobilised mosquitoes were placed in the bag and gently shaken in the bag. Mosquitoes for the dust storm method were also placed in a bag with 0.3 g dust, a fan was used to create a dust storm within the bag. After all dusting methods, immobilised mosquitoes were gently placed back in their original container. Each method had a control, where mosquitoes were handled similarly to marked mosquitoes, but dusts or paints were not added. Each method, control and colour had three repeats, each containing at least 30 mosquitoes. The survival experiment continued until all mosquitoes died.

‘Marking efficiency’ scores were given to each colour and method used. Scores were calculated by whether dust could be seen from 20 cm away, and if dust was present on their head, thorax, abdomen, legs or wings when observed using a dissection microscope. A score of ‘1’ was given if present, or ‘0’ if not present, for each category. The sum of these categories gave the overall marking efficiency score, out of a maximum of six.

Chapter 5

5.3.3. Data analysis

The survival data was analysed using a Cox proportional hazards regression model (R package survival, ‘coxph’: Therneau 2015). Across all experimental treatments there was a clear difference between male (N = 1260), and female (N = 1235) survival (2=1080.80, df=1, p<0.0001), therefore analysis was performed separately for each sex. Immobilised and mobile mosquitoes were compared prior to testing the effect of dust and marking methods on mosquito survival.

Immobilised mosquitoes were compared to unmarked method controls using a Cox proportional hazards regression model. Marking methods (bag, bulb duster, dust storm and paint) were compared separately, to see what effect colour had on each method. Each method was compared to an unmarked control, which had been handled in a similar way. Finally, marked mosquitoes were compared to immobilised mosquitoes.

A generalised linear model (GLM) with quasi-poisson errors was used to test marking efficiency in relation to marking method and colour. A GLM with poisson errors showed that the data was over-dispersed. Interactions were tested for, but were not present. Mosquitoes that died early (days 1-20) and late (days 21-61) within the experiment were compared separately for marking efficiency. Males and females were combined within the model because their marking efficiencies were not significantly different (df=1, p=0.897). All models were plotted to see how well the model fitted the data. All graphs were plotted using R version 3.1.1 (R Core Team 2014).

5.4. Results

The experiment ran for 61 days, at which point all mosquitoes had died. Median longevity was 12 and 28 days for males and females, respectively. Overall, 716 (58%) of females survived beyond 30 days, but only 21 (2%) male individuals survived beyond 30 days. There were no significant differences in longevity between immobilised mosquitoes and their controls for males (2=1.46, df=1, p=0.227) or females (2=2.10, df=1, p=0.148). The bag and dust methods had significantly higher longevity in males than the control (Table 5.1, Figure 5.1a). In females, the bag, bulb duster, dust storm and paint methods had significantly higher longevity than the control (Table 5.1, Figure 5.1b).

Chapter 5

Table 5.1. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster, fan and paint) against immobilised controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given. Method Males Females β SE z p β SE z p Bag -0.759 0.237 -3.207 <0.01** -0.551 0.226 -2.443 <0.05* Bulb duster -0.517 0.229 -2.264 <0.05* -1.335 0.238 -5.612 <0.0001*** Dust storm -0.136 0.228 -0.598 =0.550 -0.930 0.235 -3.952 <0.0001*** Paint -0.053 0.229 -0.231 =0.817 -0.795 0.225 -3.535 <0.0001***

Chapter 5

Figure 5.1. Survivorship plot of (a) Male and (b) Female Stegomyia aegypti marked using different methods (bag, bulb duster, dust storm & paint) vs. immobilised controls

Chapter 5

5.4.1. Survival analysis of dusted mosquitoes using method controls

With the exception of BC Blue, the survival of male mosquitoes was significantly lower using the bag method, in comparison to the method control (Table 5.2, Figure 5.2a). The use of Dayglo dusts with the bag method, particularly D Blue significantly lowered male survival. The use of D Red and BC Blue and bag method significantly reduced female survival, in comparison to the method control, but the remaining colours did not (Table 5.3, Figure 5.3a). There was no significant difference between the survival of male mosquitoes, marked using a bulb duster, and the method controls, except for BC Gold, which significantly increased male longevity (Table 5.2, Figure 5.2b). Except for BC Blue, all dust colours significantly reduced female survival whilst using the bulb duster (Table 5.3, Figure 5.3b).

Dusts applied to male mosquitoes using the dust storm method were not significantly different to method controls, with the exception of D Blue, which has a significantly decreased longevity (Table 5.2, Figure 5.2c). BC Red significantly reduced the longevity of female mosquitoes marked using the dust storm method, but the other colours had no impact (Table 5.3, Figure 5.3c). Applying paint had no significant impact on the survival and longevity of male mosquitoes (Table 5.4, Figure 5.2d), but the blue paint significantly reduced the survival and longevity of female mosquitoes (Table 5.4, Figure 5.3d). Yellow and red paint did not significantly reduce the survival of females marked with paint (Table 5.4, Figure 5.3d).

Chapter 5

Table 5.2. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on male Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Bag Bulb duster Dust storm β SE z p β SE z p β SE z p D Blue 2.160 0.226 9.566 <0.0001 0.005 0.204 0.024 =0.980 0.453 0.207 2.187 =0.029 D Red 1.179 0.220 5.363 <0.0001 0.164 0.204 0.803 =0.422 0.042 0.207 0.202 =0.840 BC Gold 0.572 0.220 2.604 =0.009 -0.581 0.215 -2.699 =0.007 -0.228 0.209 -1.093 =0.274 BC Blue 0.316 0.215 1.467 =0.142 0.089 0.207 0.428 =0.668 -0.270 0.210 -1.288 =0.198 BC Pink 1.321 0.222 5.950 <0.0001 0.337 0.207 1.633 =0.103 0.098 0.203 0.484 =0.628 BC Red 0.898 0.218 4.129 <0.0001 0.009 0.204 0.046 =0.963 -0.389 0.207 -1.881 =0.056

Table 5.3. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on female Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Bag Bulb duster Dust storm β SE z p β SE z p β SE z p D Blue -0.147 0.209 -0.704 =0.481 2.918 0.248 11.752 <0.0001 0.339 0.210 1.615 =0.106 D Red 1.387 0.212 6.536 <0.0001 1.847 0.231 7.996 <0.0001 0.353 0.209 1.687 =0.092 BC Gold 0.099 0.213 0.463 =0.643 1.241 0.219 5.675 <0.0001 0.101 0.210 0.478 =0.632 BC Blue 0.591 0.208 2.837 =0.005 0.202 0.212 0.951 =0.342 0.309 0.211 1.464 =0.143 BC Pink 0.077 0.207 0.372 =0.710 2.415 0.238 10.139 <0.0001 0.292 0.208 1.405 =0.160 BC Red -0.305 0.211 -1.447 =0.148 1.608 0.232 6.935 <0.0001 0.721 0.212 3.395 <0.0001

Table 5.4. Results from cox proportional hazards regression model, testing paint on male and female Stegomyia aegypti against unmarked method controls. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Males Females β SE z p β SE z p Blue -0.259 0.209 -1.240 =0.215 1.219 0.214 5.702 <0.0001 Red -0.188 0.204 -0.926 =0.355 0.385 0.202 1.909 =0.06 Yellow -0.250 0.207 -1.206 =0.228 -0.249 0.204 -1.218 =0.223

100

Chapter 5

Figure 5.2. Survivorship plot of male Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. unmarked method controls

101

Chapter 5

Figure 5.3. Survivorship plot of female Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. unmarked method controls

102

Chapter 5

5.4.2. Survival analysis of dusted mosquitoes using immobilised controls

In comparison to unmarked immobilised mosquitoes, D Blue, D Red and BC Pink, applied using the bag method, significantly reduced male survival, but BC Blue increased survival (Table 5.5, Figure 5.4a). Female mosquitoes responded differently, with D Red significantly reducing survival, but D Blue, BC Gold, BC Pink and BC Red increasing survival (Table 5.6, Figure 5.5a). Using the bulb duster method, all colours increased the survival of male mosquitoes (Table 5.5, Figure 5.4b), but only BC Blue increased the survival of female mosquitoes (Table 5.6, Figure 5.5b). BC Gold and BC Red had no impact on the survival of female mosquitoes, but D Blue, D Red and BC Pink reduced survival (Table 5.6, Figure 5.5b).

There was no significant difference between immobilised mosquitoes and male mosquitoes marked, using the dust storm method, with D Blue, D Red and BC Pink but the remaining colours increased survival (Table 5.5, Figure 5.4c). With the exception of BC Red marked using the dust storm method, which was not significantly different to immobilised mosquitoes, the remaining colours increased female survival (Table 5.6, Figure 5.5c). The paint method significantly increased the survival of males (Table 5.7, Figure 5.4d), and females marked with red and yellow paint (Table 5.7, Figure 5.5d). Blue paint significantly reduced the longevity of female mosquitoes (Table 5.7, Figure 5.5d).

103

Chapter 5

Table 5.5. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on male Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Bag Bulb duster Dust storm β SE z p β SE z p β SE z p D Blue 1.542 0.204 7.549 <0.0001 -0.723 0.192 -3.769 <0.0001 0.143 0.189 0.757 =0.449 D Red 0.496 0.192 2.585 0.009 -0.562 0.190 -2.961 =0.003 -0.294 0.193 -1.524 =0.127 BC Gold -0.149 0.194 -0.766 0.443 -1.289 0.208 -6.203 <0.0001 -0.538 0.193 -2.787 =0.005 BC Blue -0.411 0.196 -2.100 0.036 -0.633 0.193 -3.282 =0.001 -0.582 0.196 -2.969 =0.003 BC Pink 0.646 0.193 3.345 <0.0001 -0.387 0.191 -2.033 =0.042 -0.212 0.185 -1.147 =0.251 BC Red 0.176 0.194 0.909 0.363 -0.717 0.194 -3.693 <0.0001 -0.705 0.192 -3.668 <0.0001

Table 5.6. Results from cox proportional hazards regression model, testing different methods (bag, bulb duster and dust storm) on female Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Bag Bulb duster Dust storm β SE z p β SE z p β SE z p D Blue -0.729 0.198 -3.687 <0.0001 1.716 0.205 8.378 <0.0001 -0.447 0.196 -2.276 =0.023 D Red 0.775 0.194 3.988 <0.0001 0.613 0.195 3.146 =0.002 -0.448 0.195 -2.304 =0.021 BC Gold -0.477 0.201 -2.366 =0.018 0.022 0.192 0.115 =0.909 -0.681 0.197 -3.450 <0.0001 BC Blue 0.002 0.195 0.012 =0.990 -0.965 0.215 -4.485 <0.0001 -0.476 0.197 -2.414 =0.016 BC Pink -0.519 0.195 -2.661 =0.008 1.189 0.195 6.087 <0.0001 -0.499 0.194 -2.572 =0.010 BC Red -0.866 0.200 -4.322 <0.0001 0.365 0.194 1.879 =0.06 -0.076 0.193 -0.394 =0.694

Table 5.7. Results from cox proportional hazards regression model, testing paint on male and female Stegomyia aegypti against immobilised mosquitoes. Coefficient estimates (β), standard errors, associated Wald’s z-score, and p-values are given Colour Males Females β SE z p β SE z p Blue -0.506 0.194 -2.606 =0.009 0.408 0.193 2.117 =0.034 Red -0.427 0.188 -2.271 =0.024 -0.347 0.195 -1.784 =0.074 Yellow -0.502 0.192 -2.613 =0.009 -0.956 0.202 -4.744 <0.0001

104

Chapter 5

Figure 5.4. Survivorship plot of male Stegomyia aegypti marked with DayGlo colour dust (D Blue & D Red) and Brian Clegg coloured dust (BC Blue, BC Pink, BC Red, BC Gold), using different marking methods: (a) Bag (b) Bulb duster (c) Dust storm and (d) Paint vs. immobilised mosquitoes

105

Chapter 5

106

Chapter 5

5.4.3. Marking efficiency

For marking efficiency, no statistical difference was observed between early (days 1- 20) and late (days 21-61) survival, indicating that those individuals who showed signs of being marked in the beginning of the experiment remained so for the rest of the experiment (df=1, p<0.0001). All methods of application and colours were visible (p≤0.002 for all dusting combinations), but some more visible than others (Figure 5.6). The paint method was the least visible, followed by the bulb duster (Figure 5.6a). D Blue, D Red, BC Red and BC Pink showed the greatest marking efficiency (Figure 5.6b). DayGlo dusts had a higher marking efficiency than Brian Clegg dusts.

107

Chapter 5

Figure 5.6. Marking efficiency of methods and colours. Marking efficiency (0–6) of male and female Stegomyia aegypti using different (a) Marking methods and (b) Colours of dust

108

Chapter 5

5.5. Discussion

The effect of dusts on insect survival and behaviour varies with species (Narisu, Lockwood & Schell 1999) thus the amount of dust used and application technique is likely to have an effect. Although some studies have found no significant differences between the survival rates of marked insects to unmarked controls (Crumpacker 1974; Moth & Barker 1975; Sempala 1981; Lysyk & Axtell 1986; Oloumi-Sadeghi & Levine 1990; Chiang et al. 1991; Takken et al. 1998; Narisu, Lockwood & Schell 1999; Watson, Saul & Kay 2000; Cameron et al. 2002; Weldon 2005; Nakata 2008; Hoddle et al. 2011; Johnson, Spitzauer & Ritchie 2012; Liu et al. 2012), others have documented adverse dusting effects (Moffitt & Albano 1972; LaBrecque et al. 1975; Williams, LaBrecque & Patterson 1977; Naranjo 1990; Dye, Davies & Lainson 1991; Coviella et al. 2006; Reid & Reid 2008; de Guzman, Frake & Rinderer 2012), with factors including brand and colour often associated with dramatic differences between marked and unmarked controls.

This study found that males and female Stg. aegypti were adversely affected by the bag and the bulb duster methods. It is possible that the bag method was more damaging to males due to their fragility and that physical shaking damaged their extremities. Shaking procedures have shown to be detrimental to delicate insects because they place too much dust on the insects, and cause high mortality immediately after marking (Meyerdirk, Hart & Burnside 1979; Hagler & Jackson 2001). It is less certain as to why the bulb duster method gave greater mortality for females but it may also be due to an excess of dust that female mosquitoes were unable to groom off. Too much dust can decrease mobility, interfere with sensory organs and increase mortality (Davey 2009). Excessive moisture whilst marking can cause insects to become ‘gummed up’ with dust (Crumpacker 1974), but this was not an issue in this study as relative humidity was constant and there were no observations of any water droplet formation or gumming of dust.

The commonly used dust storm method had less impact on survival and thus is better than the other two dusting methods. This is perhaps due to dust storms atomising the dust better than the bag and bulb dusting methods. As for marking efficiency for each method, the dust storm and bag provided the greatest coverage of mosquitoes. Spraying mosquitoes with fluorescent paint had no marked effect on survival, possibly

109

Chapter 5 due to the small droplet use in the study, which was designed to balance survival and coverage. The paint method showed a low marking efficiency in comparison to the dry marking methods. This could be increased by extra sprays during marking, but too much moisture can affect behaviour and survival (Silver 2008). Whilst relative humidity prolongs mosquito survival, excess wetting mosquitoes can be a source of mortality (Crumpacker 1974), so this method must be used with caution. It is unknown why the treated mosquitoes outlived the controls in Figure 5.1. There was no difference between the longevity of immobilised mosquitoes and mobilised mosquitoes, but perhaps the addition of dust, and moving them around during the marking methods increased their survival due to dust being held at room temperature.

Colour choice appears to be important for MRR studies. The reds and pinks used had intermediate values for both survival and marking efficiency. Although D Blue could be seen using the naked eye, the significant negative effect on survival makes it unsuitable for studies where an assumption is made that the marked individuals are of equal fitness to the unmarked. Other shades of blue which have a different chemical make-up and/or different concentrations of agents may not create this effect. This was observed in Reticulitermes which had a significant mortality rate over 15 days for Sudan Red 7B but not Neutral Red (Su, Ban & Scheffrahn 1991). This may contribute to the success and failure of any MRR studies which use different shades of D Blue or other colours, although many other factors should be considered such as the scale, location, season, duration of the study and marking method. The low impact of BC Blue on survival could be used as evidence to argue this but owing to its poor marking efficiency, it is possible that the dust was unable to adhere to individuals and was thereby spuriously associated with high survivorship. A colour change occurred in BC Gold tests with blue webbing across appendages at 15 days and it additionally had greater longevity for those marked with this colour for reasons unknown.

Mosquitoes marked with the two particular shades of DayGlo dust had reduced survival, but further testing with more colours, shades and chemical compositions is required to conclude whether one brand has lower mortality. Other studies have also observed reduced survival rates that are dependent on the colour of dust used. Coviella et al. (2006) observed reduced survival in cicadellids, but only for one of the six colours of DayGlo dusts tested. One of 14 colours of powders tested by Toepfer et al. (2005) reduced the survival in corn rootworm adults.

110

Chapter 5

Two concerns in scaling the results from this laboratory study up to a semi-field or field trial are establishing a practical immobilisation method and preventing contact-transfer among marked and unmarked individuals. Firstly, using a freezer to immobilise mosquitoes in the field would prove difficult in certain situations, but diethyl ether is a suitable replacement. Even though freezing was chosen here over a chemical agent, provided the knock-out time is similar and the chemical has no lasting effects, it is likely to have no significant effect in the survival of the mosquito (Silver 2008). As it is well established that many species and strains vary in environmental resilience and excessive immobilising methods have the potential to cause significant mortality, immobilisation and marking technique should be tested in the laboratory and suitable concentrations established before field experiments. It should also be noted that marking under a laboratory setting without the uncontrolled weather and environmental conditions of a field study are likely to mean that marking efficiency will be high for a longer duration than can be expected in the field.

Secondly, contact-transfer can bias results when using methods which heavily rely on dust coverage, potentially marking 1-3% of the unmarked mosquitoes after 24 hours if exposed to DayGlo mosquitoes in a confined space (Fryer & Meek 1989). Dust particles can potentially be transferred to unmarked insects in traps and sweep nets used for sampling (Miller 1993), where individuals are forced to come into contact with each other, but is not an issue with sticky traps. It should be noted that same-sex transfer is likely to be of little concern and most transfer would occur during mating. Crumpacker et al. (1974) observed no dust transfer during mating of Drosophila pseudoobscura or following the crowding of heavily marked and unmarked individuals after they were allowed to clean themselves before mixing. The implication of this for field studies is that they should mark individuals then give them time to clean themselves in relatively uncrowded conditions before releasing them.

We expect it is likely that the effects we observed here of dusting methods on survival of Stg. aegypti will be similar to the effects on Anopheles species, although other experiments have shown little difference in survival of marked and unmarked An. punctulatus, An. maculatus, An. sinensis, An. subpictus and An. stephensi (Reisen, Mahmood & Parveen 1979; Reisen & Aslamkhan 1979; Charlwood, Graves & Birley 1986; Chiang et al. 1991; Miller 1993; Liu et al. 2012). The survival of a mosquito is dependant on many variables, including its size and resistance to environmental stress

111

Chapter 5

(Haramis 1983; Kitthawee, Edman & Upatham 1992; Ameneshewa & Service 1996). The effects of marking represents an additional stress on the body of the insect, making it necessary to carry out this experiment on other medically-important mosquito subfamilies, including the malaria vector species in Anopheles, where a range of body sizes and survival rates have been reported (Petrarca et al. 1998; Silver 2008; Castillo, Brown & Strand 2011).

We found that application method had a significant effect on mosquito survival, and that the dust storm method caused the least mortality. Of the colours, the two shades of blue we tested should be avoided. For increased survival and marking efficiency, BC Pink or BC Red appear to be the most viable options. Care should be taken before using new colours (e.g. BC Gold) and when assuming there are no significant effects on mosquito survival with a chosen method, as our results suggest they may strongly bias the results of a MRR study. Choosing the best technique for use in the field will be essential to the success of field-based studies in mosquito dispersal.

5.6. Conclusions

Few studies have addressed the implications of marking efficiency and survivorship on male and female mosquitoes following marking, and even fewer have compared marking methods. Stg. aegypti is the primary vector of dengue, and its spatial movement is of interest to many policymakers and modellers. With no specific treatment, and the increased global incidence of dengue, it is becoming paramount to understand the vector’s dispersal through studies such as MRR. Successful MRR studies require a benign technique that adheres to the mosquito for a defined duration. This study showed that treatments not only affected males and females differently, but also particular colours and methods were significantly different to controls. Males dusted with D Blue and females with D Red had the most significant reduction in survivorship in comparison to the control. Dusting using BC Red or BC Pink showed both reasonable performance in marking and impact on overall survival across males and females. Overall, the dust storm method provided the best trade-off between survival and marking efficiency.

112

Chapter 6

Chapter 6 – General Discussion

6.1. Overview

Land use changes have caused considerable declines in biodiversity due to loss of habitat availability, but are also a main driver in a range of infectious disease outbreaks and emergence events (Patz et al. 2004). The tropical rainforests of Southeast Asia are currently under threat from anthropogenic modification, due to the expansion of oil palm plantations (Fitzherbert et al. 2008; Phalan et al. 2013). It is important to understand the effect of land use change on vector and parasite dynamics to improve our ability to assess health risks in a changing world. The objective of this thesis was to investigate the impacts of land use change, particularly deforestation and oil palm expansion, on mosquito community dynamics. In this chapter, I will review the results of my research, discuss the implications of my findings within a wider context, and outline topics for future study.

6.2. Land use change and vector-borne diseases

The results presented in this thesis show land use change can increase some species, but decrease others. Land use change disrupts, eliminates or creates new ecological niches, which can either increase or decrease vector species (Walsh, Molyneux & Birley 1993; Patz et al. 2000, 2004; Leisnham et al. 2007). This thesis showed a significant increase of the landing rate of Anopheles vectors on human collectors from primary forest to logged forest (Chapter 2-3). This was also seen from primary forest to virgin jungle reserve (Chapter 4). In Sabah and Sarawak, there are few areas of primary forest remaining (Bryan et al. 2013), and efforts are being made to preserve areas to maintain biodiversity. By preserving these areas, it will prevent further increases in the number of medically important mosquitoes, such as Anopheles balabacensis.

The differences in the community composition of mosquitoes collected between primary forest and oil palm (Chapter 2) may be explained through the water bodies present in each area. Although it was not studied in this thesis, water bodies can differ in water quality (e.g. turbidity, pH) across land use (Patz et al. 2000). The amount of shade in each area from canopy cover may have contributed to differences seen between land use. In this thesis, canopy cover ranged from 37.5% in oil palm to 87.3%

113

Chapter 6 in primary forest (Chapter 3). Anopheles balabacensis is a forest-dwelling species, and deposits eggs in the shaded pools (Chapter 1, Table 1.1). This may explain why there was a significant difference seen between the abundance of An. balabacensis between disturbed forest, and highly disturbed forest (Chapter 3).

Deforestation and agricultural expansion can increase human risk from disease due to change in habitats and altered behaviour of humans. For example, forest clearing can result in humans with a low-level immunity to diseases moving into a disease- endemic area and spreading the disease (Martens & Hall 2000). Workers involved in logging and agricultural activities can be at risk from vector-borne diseases. For example, increased contact between humans and forested areas has increased other vector-borne diseases, such as Leishmaniasis (Gratz 1999) Forest clearing may deplete mosquito species initially, but this may be followed by colonisation by a more efficient vector species and an increase in disease transmission (Guerra, Snow & Hay 2006).

Overall, there are unanswered questions about which environmental factors are affecting the community composition across land use. It is clear from our results in Chapter 2 that land use change affected malaria and dengue vectors differently. For this reason, the implications of land use change on malaria and dengue vectors are discussed separately in Sections 6.3.1-2.

6.2.1. Human and simian malaria

This thesis investigated the preference of mosquitoes landing at ground and canopy level (Chapter 4). Chapter 4 provided essential data on the anthropophily of Anopheles species in the Tawau Division of Sabah, Malaysia, for the planning of vector control strategies. Anopheles balabacensis had a preference for landing on humans at ground level within the logged and lightly logged forest. For studies looking at sampling anthropogenic mosquitoes in Malaysian forests, ground level human landing catches would be more beneficial than canopy catches.

Previously monkey-baited traps at different canopy heights (0 m, 3 m and 6 m) have been used to study Plasmodium knowlesi vectors, alongside human landing catches at ground level (Tan et al. 2008; Jiram et al. 2012). These studies found An. latens was positive for sporozoites, and was attracted to both human and monkey hosts.

114

Chapter 6

Anopheles latens preferred to bite long-tailed macaques at six meters rather than ground level or three meters. In this thesis, Anopheles latens was found landing on humans at ground and canopy level, but numbers were too low to determine preference. In Chapter 3, An. latens was the predominant species in the primary forest (Maliau Basin Conservation Area), but with a relatively low number of bites per human bait, per night (18:00-23:00h). Low numbers of An. latens were also seen in Chapter 4, where only one An. latens was collected in the primary forest (Danum Valley Conservation Area). Anopheles latens has been incriminated as a P. knowlesi vector in Sarawak (Vythilingam et al. 2006; Tan et al. 2008), but based on the low number of An. latens catches in this thesis, it appears unlikely that this vector is causing the increase of P. knowlesi in Sabah.

Anopheles balabacensis was the predominant species collected in highly disturbed forest, disturbed forest and oil palm plantations in this thesis (Chapter 2-4). It is one of the main Anopheles vector capable of spreading human malaria, and has also been incriminated as a simian malaria vector (Collins, Contacos & Guinn 1967; Vythilingam 2010; Wong et al. 2015). In Palawan Island, An. balabacensis was more attracted to monkey-baited traps than human baited catches (Miyagi 1973; Manguin 2013). It is unknown if An. balabacensis has the same biting preference as An. latens by biting macaques in the canopy, but the species is known to bite humans and macaques (Manguin 2013), which means it has the potential ability to transmit P. knowlesi between simian hosts and ground-dwelling humans.

Between 2001 and 2003, An. donaldi appeared to have replaced An. balabacensis as the main vector in Kinabatangan, Sabah (Vythilingam et al. 2005). This displacement was thought to have occurred from deforestation and malaria control activities, and resulted in a reduction of human malaria cases (Vythilingam et al. 2005; Vythilingam 2012). It appears in recent years An. balabacensis has re-established as the primary vector of malaria in Sabah (Wong et al. 2015), which may pose a threat to malaria elimination targets of 2020 in Malaysia. Chapter 3 showed the number of An. balabacensis landing in young oil palm plantations (7-13 years old, 4-8 m high) was not significantly different to highly disturbed forest sites. Chang et al. (1997) showed reductions in human-biting rates of Anopheles from logged forest through to oil palm planting and maintenance (1 year old, 2-3 m high). Oil palm plantations have a 25-30 year life cycle, and begin to fruit after 3-5 years (Butler, Koh & Ghazoul 2009). As oil

115

Chapter 6 palm matures, the microclimate becomes more buffered, leaf litter increases and the canopy eventually closes (Luskin & Potts 2011). This thesis showed mosquito abundance appears to recover as oil palm mature. Further surveys should be conducted in older plantations, once the canopy fully closes.

This thesis found Anopheles balabacensis started landing on humans from 18:00h, and biting within close proximity to housing estates within oil palm plantations (Chapter 2-4). Studies from previous decades showed this species was mainly a late night biter (21:00-22:00h), but they now appear to be biting during the early evenings before adult hosts are under insecticide-treated bed nets (ITNs) (Hii 1985; Rohani et al. 1999; Vythilingam et al. 2005). Other studies have reported earlier mosquito biting times following the use of ITNs, due to selective pressure placed on the vector population (Mbogo et al. 1996; Takken 2002). Current vector control methods in Malaysia include ITNs and indoor residual spraying (IRS), but as P. knowlesi vectors are highly anthropophilic, exophagic and exophilic, current control methods are not sufficient to break the transmission cycle of P. knowlesi from primate to human or human to human (Manguin 2013).

Results from this thesis have highlighted the need for a new control method within Malaysia to reduce the number of P. knowlesi vectors, especially within oil palm plantations. The use of long-lasting insecticidal hammocks (LLIH) have shown to reduce malaria incidence and prevalence in forested areas of Vietnam and Cambodia (Thang et al. 2009; Sochantha et al. 2010). The use of LLIH should be encouraged in areas of ecotourism, where people stay in or near forests overnight (Manguin 2013). Oil palm workers should be made aware of the risks of visiting forests and forest- fringes during the evenings, and encouraged to use ITNs, repellents and LLIH.

This thesis also found An. macarthuri was the second most abundant mosquito in the disturbed forests (Chapter 3-4). This species also belongs to the Leucosphyrus group, and although it hasn’t been incriminated as a P. knowlesi vector, all members of the Leucosphyrus group are capable of transmitting P. knowlesi experimentally (Baird 2009). Tan et al. (2008) found the second most abundant species to their monkey trap was An. macarthuri, meaning it readily bites humans and macaques. It is unknown how actively this species is involved in the transmission of P. knowlesi, but it could potentially have a role in transmitting the parasite from primate to primate, and primate to human. Further entomological studies need to be carried out to understand the

116

Chapter 6 primate-vector-human interactions occurring in P. knowlesi infected areas, and to predict exposure risk in response to environmental change.

Plasmodium knowlesi cases have increased during the last decade, and is now the most common cause of malaria in Sabah (William et al. 2013, 2014). Malaysia is currently in the pre-elimination phase of malaria elimination (WHO 2014b), and the rise in cases threatens the achievement of malaria control goals (William et al. 2014). Family clusters of P. knowlesi cases were considered rare because transmission was mainly occurring in densely forested areas, but these cluster of cases suggest human- vector-human transmission is occurring in peridomestic areas (Barber et al. 2012). This change in transmission is thought to be linked to deforestation and land use change (Barber et al. 2012; Fornace et al. 2016). Although humans have always shared habitats with non-human primates, the destruction of primate habitats has changed the dynamics of human-primate interactions (Chapman, Gillespie & Goldberg 2005). Deforestation, logging and forest fragmentation increases human-wildlife interactions, and in many areas has caused monkeys, humans and forest dwelling vectors to interact more frequently (Wolfe et al. 2005; Vythilingam 2010). This increases the potential for zoonotic malaria transmission.

Previously, cases of P. knowlesi have appeared in patients with a recent history of forest or forest-edge exposure, and had seen a macaque in the preceding month (Barber et al. 2013a). Given the current rate of deforestation in Malaysia, the number of human-wildlife interactions are likely to increase. This thesis found a high abundance of P. knowlesi vectors within oil palm plantations. The increase in human- wildlife interactions and the high number of P. knowlesi vectors may increase the risk of future human to human transmission of simian malaria.

There are important issues regarding P. knowlesi that remain unanswered, particularly information on the ecology of simian vectors. Further studies are required to address these knowledge gaps, such as risk factors for the acquisition of the disease, the extent of human-vector-human transmission occurring, identifying the mosquito vectors in Sabah and monitoring how land use change affects the disease risk as ecological changes continue. With continued oil palm expansion and fragmentation of remaining forests, there is the potential for vectors to switch to hosts, and must be actively monitored.

117

Chapter 6

6.2.2. Dengue

This thesis investigated the presence and absence of container breeding mosquitoes across an anthropogenic gradient. Chapter 2 found the dengue vector, Stg. albopicta, breeding in rural housing compounds. Human landing catches could not be carried out in the rural housing compounds due to resident’s interest in the study, which would have resulted in a biased collection. The oil palm plantation sites, which were 440- 1000 m from the rural housing compounds, did not collect any container breeding mosquitoes. As Stg. albopicta was collected using human landing catches in the oil palm sites, the absence in the ovitraps suggests their host-searching diversion from more suitable habitats. Since Stg. albopicta has been shown to have a limited host- searching range, it is unlikely to be far from the oil palm sites. Stegomyia albopicta has been shown to fly 400-800 m in dispersal studies, and may fly into the oil palm plantation in search of hosts (Niebylski & Craig 1994; Reiter et al. 1995; Honório et al. 2003).

The incidence of dengue in Malaysia has grown rapidly in recent decades, from 46,000 cases in 2007 to 108,698 cases in 2014 (Benitez 2009; Lee et al. 2015). Controlling the Stegomyia vectors is the main tool for the management of dengue as there is no vaccine or specific treatment (WHO 2015a). This study observed high levels of social detritus within each rural housing compound, with a high presence of Stg. albopicta larvae. The removal of these potential breeding sites is important, because dengue is increasingly being reported in urban and rural areas (Azami et al. 2011).

Ovitrap surveillance is the most common method used to detect the presence of dengue vectors (Lau et al. 2013), which could result in a disease outbreak. There are important issues regarding dengue that remain unanswered, for example: Will further urbanisation in Malaysia increase the number of dengue cases? What is the best way to reduce artificial containers in rural areas? How far do the dengue vectors fly to seek hosts (for use in dengue risk models)?

6.2.3. Marking techniques and dispersal

Chapter 5 investigated the effect of different mosquito marking methods and colour on Stg. aegypti, to find the best technique for use in mark-release-recapture (MRR) studies in the field. This chapter found that marking technique is important for the

118

Chapter 6 survival of Stg. aegypti before releasing them. The mosquito marking methods compared in this study were; applying the dust using a bulb duster, placing mosquitoes in a bag containing dust and gently shaking or by creating a dust storm within a cage. The best marking method to optimise mosquito survival, whilst maintaining visibility, was the dust storm in a cage. This method provides a cost-effective, persistent and non-detrimental marking technique.

MMR techniques are a powerful tool for the estimation of mosquito population densities, feeding behaviour, duration of gonotropic cycles and their dispersal behaviour, which are important for understanding mosquito-borne disease transmission (Silver 2008; Guerra et al. 2014). Chapter 5 showed marking techniques have the potential to affect the mortality of marking individuals in a way that could bias the results arising from MRR experiments. Marking methods and dust colour should be rigorously compared before use in the field (Silver 2008), as the effects of dusts on insect survival and behaviour vary with species (Narisu, Lockwood & Schell 1999).

There is a scarcity of MRR information for vector species in Borneo (Guerra et al. 2014). MRR studies in Sabah on Stg. aegypti and Stg. albopicta could provide essential dispersal and host-searching range in oil palm plantations and surrounding rural and urban areas for use in future dengue risk models. The use of MRR studies in studying simian malaria could provide information on how far simian vectors travel from the forest to rural areas. This would be particularly useful in oil palm plantations and fragmented forest to assess risk of simian malaria. This data could then be used for the planning of future housing areas, and how far away from the forest they should be to decrease simian malaria.

6.3. Conclusions

Deforestation, agricultural expansion and intensification are the main drivers of habitat change in tropical regions (Laurance, Sayer & Cassman 2014), but also a range of infectious disease outbreaks and emergence events (Patz et al. 2004). The rapid expansion of oil palm plantations in Southeast Asia threatens many species, however, less is understood about how vector abundance, distribution, exposure to humans and biting rates vary relative to land use change. The results of this thesis show the community composition of anthropogenic mosquitoes was separated along land use, which was driven by medically important mosquitoes. Differences in community

119

Chapter 6 composition were also seen in areas of rural housing in comparison to disturbed forest and primary forest, due to the presence of Stg. albopicta in rural housing compounds. Anopheles balabacensis was the predominant species collected in highly disturbed forest, disturbed forest and oil palm plantations in this thesis. This species is the main malaria vector in Sabah, and is capable of spreading simian malaria. It had a preference for feeding on humans at ground level, when given a choice between ground and canopy level. This thesis highlighted the need for future MMR studies to be conducted across land use, to estimate dengue and simian malaria vector population densities, feeding and dispersal behaviour. It is important to understand the effect of land use change on vector and parasite dynamics to improve our ability to assess health risks in a changing world. Further data collection is critical for a better understanding of complex interactions between land use and infectious diseases.

120

Bibliography

References

Abubakar, S. & Shafee, N. (2002) Outlook of dengue in Malaysia: A century later. Malaysian Journal of Pathology, 24, 23–27.

Adachi, M., Ito, A., Ishida, A., Kadir, W.R., Ladpala, P. & Yamagata, Y. (2011) Carbon budget of tropical forests in Southeast Asia and the effects of deforestation: An approach using a process-based model and field measurements. Biogeosciences Discussions, 8, 3051–3079.

Afrane, Y., Lawson, B., Githeko, A. & Yan, G. (2005) Effects of microclimatic changes caused by land use and land cover on duration of gonotrophic cycles of Anopheles gambiae (Diptera: Culicidae) in Western Kenya Highlands. Journal of Medical Entomology, 42, 974–980.

Alemayehu, T., Ye-ebiyo, Y., Ghebreysesus, T.A., Witten, K.H., Bosman, A. & Teklehaimanot, A. (1998) Malaria, schistosomiasis, and intestinal helminthes in relation to microdams in Tigray, Northern Ethiopia. Parassitologia, 40, 259–267.

Ameneshewa, B. & Service, M.W. (1996) Resting habits of Anopheles arabiensis in the Awash river valley of Ethiopia. Annals of Tropical Medicine and Parasitology, 90, 515–521.

Anderson, J.R. (2000) Sleep-related behavioural adaptations in free-ranging anthropoid primates. Sleep Medicine Reviews, 4, 355–373.

Antinori, S., Galimberti, L., Milazzo, L. & Corbellino, M. (2013) Plasmodium knowlesi: The emerging zoonotic malaria parasite. Acta Tropica, 125, 191–201.

Azami, N.A.M., Salleh, S.A., Neoh, H.-M., Zakaria, S.Z.S. & Jamal, R. (2011) Dengue epidemic in Malaysia: Not a predominantly urban disease anymore. BMC research notes, 4, 216.

Baird, J.K. (2009) Malaria zoonoses. Travel Medicine and Infectious Disease, 7, 269– 77.

Barber, B.E., William, T., Dhararaj, P., Anderios, F., Grigg, M.J., Yeo, T.W. & Anstey, N.M. (2012) Epidemiology of Plasmodium knowlesi malaria in north-east Sabah, Malaysia: Family clusters and wide age distribution. Malaria Journal, 11, 401.

Barber, B.E., William, T., Grigg, M.J., Menon, J., Auburn, S., Marfurt, J., Anstey, N.M. & Yeo, T.W. (2013a) A prospective comparative study of knowlesi, falciparum, and vivax malaria in Sabah, Malaysia: high proportion with severe disease from Plasmodium knowlesi and Plasmodium vivax but no mortality with early referral and artesunate therapy. Clinical Infectious Diseases, 56, 383–397.

Barber, B.E., William, T., Grigg, M.J., Yeo, T.W. & Anstey, N.M. (2013b) Limitations of microscopy to differentiate Plasmodium species in a region co-endemic for

121

Bibliography

Plasmodium falciparum, Plasmodium vivax and Plasmodium knowlesi. Malaria Journal, 12, 10–1186.

Bates, M. (1944) Observations on the distribution of diurnal mosquitoes in a tropical forest. Ecology, 25, 159–170.

Bates, M. (1945) Observations on climate and seasonal distribution of mosquitoes in Eastern Colombia. Journal of Animal Ecology, 14, 17–25.

Bates, D., Maechler, M., Bolker, B. & Walker, S. (2015) lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-7. Available from http://CRAN.Rproject.org/package=lme4.

Becker, N., Petrić, D., Zgomba, M., Boase, C., Madon, M., Dahl, C. & Kaiser, A. (2010) Mosquitoes and Their Control, 2nd ed. Springer, Heidelberg.

Benitez, M.A. (2009) Climate change could affect mosquito-borne diseases in Asia. The Lancet, 373, 1070.

Bernard, H., Fjeldså, J. & Mohamed, M. (2009) A case study on the effects of disturbance and conversion of tropical lowland rain forest on the non-volant small mammals in North Borneo: Management implications. Mammal Society of Japan, 34, 85–96.

Bhatt, S., Gething, P.W., Brady, O.J., Messina, J.P., Farlow, A.W., Moyes, C.L., Drake, J.M., Brownstein, J.S., Hoen, A.G., Sankoh, O., Myers, M.F., George, D.B., Jaenisch, T., Wint, G.R.W., Simmons, C.P., Scott, T.W., Farrar, J.J. & Hay, S.I. (2013) The global distribution and burden of dengue. Nature, 496, 504–507.

Bidlingmayer, W.L. (1964) The effect of moonlight on the flight activity of mosquitoes. Ecology, 45, 87–94.

Birley, M.H. & Charlwood, J.D. (1989) The effect of moonlight and other factors on the ovipositon cycle of malaria vectors in Madang, Papua New Guinea. Annuals of Tropical Medicine and Parasitology, 83, 415–422.

Bjornstad, O.N. (2013) ncf: Spatial nonparametric covariance functions. R package version 1.1-5. Available from http://onb.ent.psu.edu/onb1/R.

Bosch, O.J., Geier, M. & Boeckh, J. (2000) Contribution of fatty acids to olfactory host finding of female Aedes aegypti. Chemical Senses, 25, 323–330.

Brühl, C.A. & Eltz, T. (2010) Fuelling the biodiversity crisis: Species loss of ground- dwelling forest ants in oil palm plantations in Sabah, Malaysia (Borneo). Biodiversity and Conservation, 19, 519–529.

Bryan, J.E., Shearman, P.L., Asner, G.P., Knapp, D.E., Aoro, G. & Lokes, B. (2013) Extreme differences in forest degradation in Borneo: Comparing practices in Sarawak, Sabah, and Brunei. PLoS One, 8, e69679.

122

Bibliography

Bunnag, T., Sornmani, S., Pinithpongse, S. & Harinasuta, C. (1979) Surveillance of water-borne parasitic infections and studies on the impact of ecological changes on vector mosquitoes of malaria after dam construction. Southeast Asian Journal of Tropical Medicine and Public Health, 10, 656–660.

Butler, R.A., Koh, L.P. & Ghazoul, J. (2009) REDD in the red: Palm oil could undermine carbon payment schemes. Conservation Letters, 2, 67–73.

Butler, R.A. & Laurance, W.F. (2009) Is oil palm the next emerging threat to the Amazon? Tropical Conservation Science, 2, 1–10.

Cameron, P.J., Walker, G.P., Penny, G.M. & Wigley, P.J. (2002) Movement of potato tuberworm (Lepidoptera: Gelechiidae) within and between crops, and some comparisons with diamondback moth (Lepidoptera: Plutellidae). Environmental Entomology, 31, 65–75.

Campbell, G.L., Hills, S.L., Fischer, M., Jacobson, J.A., Hoke, C.H., Hombach, J.M., Marfin, A.A., Solomon, T., Tsai, T.F., Tsu, V.D. & Ginsburg, A.S. (2011) Estimated global incidence of Japanese encephalitis: A systematic review. Bulletin of the World Health Organization, 89, 766–774.

Carter, C., Finley, W., Fry, J., Jackson, D. & Willis, L. (2007) Palm oil markets and future supply. European Journal of Lipid Science and Technology, 109, 307–314.

Castillo, J., Brown, M.R. & Strand, M.R. (2011) Blood feeding and insulin-like peptide 3 stimulate proliferation of hemocytes in the mosquito Aedes aegypti. PLoS Pathogens, 7, e1002274.

Chadee, D.D. (1992) Indoor and outdoor host-seeking rhythms of Anopheles bellator (Diptera: Culicidae) in Trinidad, West Indies. Journal of Medical Entomology, 29, 567–569.

Chandren, J.R., Wong, L.P. & AbuBakar, S. (2015) Practices of dengue fever prevention and the associated factors among the Orang Asli in Peninsular Malaysia. PLOS Neglected Tropical Diseases, 9, e0003954.

Chang, M.S., Hii, J., Buttner, P. & Mansoor, F. (1997) Changes in abundance and behaviour of vector mosquitoes induced by land use during the development of an oil palm plantation in Sarawak. Transactions of the Royal Society of Tropical Medicine and Hygiene, 91, 382–386.

Chapin, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L., Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C. & Díaz, S. (2000) Consequences of changing biodiversity. Nature, 405, 234–242.

Chapman, C.A., Gillespie, T.R. & Goldberg, T.L. (2005) Primates and the ecology of their infectious diseases: How will anthropogenic change affect host-parasite interactions? Evolutionary Anthropology, 14, 134–144.

123

Bibliography

Charlwood, J.D., Graves, P.M. & Birley, M.H. (1986) Capture-recapture studies with mosquitoes of the group of Anopheles punctulatus Dönitz (Diptera: Culicidae) from Papua New Guinea. Bulletin of Entomological Research, 76, 211–227.

Charlwood, J.D., Paru, R., Dagaro, H. & Lagog, M. (1986) Influence of moonlight and gonotrophic age on biting activity of Anopheles farauti (Diptera, Culicidae) from Papua New Guinea. Bulletin of Entomological Research, 23, 132–135.

Chen, C.D., Benjamin, S., Saranum, M.M., Chiang, Y.F., Lee, H.L., Nazni, W.A. & Sofian-Azirun, M. (2005) Dengue vector surveillance in urban residential and settlement areas in Selangor, Malaysia. Tropical Biomedicine, 22, 39–43.

Chen, C.D., Seleena, B., Nazni, W.A., Lee, H.L., Masri, S.M., Chiang, Y.F. & Sofian- Azirun, M. (2006) Dengue vectors surveillance in endemic areas in Kuala Lumpur city centre and Selangor State, Malaysia. Dengue Bulletin, 30, 197–203.

Chiang, G.L., Loong, K.P., Chan, S.T., Eng, K.L. & Yap, H.H. (1991) Capture- recapture studies with Anopheles maculatus Theobald (Diptera: Culicidae) the vector of malaria in peninsular Malaysia. Southeast Asian Journal of Tropical Medicine in Public Health, 22, 643–647.

Chin, W., Contacos, P.G., Coatney, G.R. & Kimball, H.R. (1965) A naturally acquired quotidian-type malaria in man transferable to monkeys. Science, 149, 865.

Chin, W., Contacos, P.G., Collins, W.E., Jeter, M.H. & Alpert, E. (1968) Experimental mosquito-transmission of Plasmodium knowlesi to man and monkey. American Journal of Tropical Medicine and Hygiene, 17, 355–358.

Cho, S.-H., Lee, H.-W., Shin, E.-H., Lee, H.-I., Lee, W.-G., Kim, C.-H., Kim, J.-T., Lee, J.-S., Lee, W.-J., Jung, G.-G. & Kim, T.-S. (2002) A mark-release-recapture experiment with Anopheles sinensis in the northern part of Gyeonggi-do, Korea. The Korean Journal of Parasitology, 40, 139–148.

Chung, A.Y.C., Eggleton, P., Speight, M.R., Hammond, P.M. & Chey, V.K. (2007) The diversity of beetle assemblages in different habitat types in Sabah, Malaysia. Bulletin of Entomological Research, 90, 475–496.

Clements, A.N. (1992) The Biology of Mosquitoes, Vol. 1. Development, Nutrition and Reproduction. CABI Publishing, Wallingford.

Clements, A.N. (1999) The Biology of Mosquitoes. Vol. 2. Sensory Reception and Behaviour. CABI Publishing, Wallingford, UK.

Coker, R.J., Hunter, B.M., Rudge, J.W., Liverani, M. & Hanvoravongchai, P. (2011) Emerging infectious diseases in Southeast Asia: Regional challenges to control. Lancet, 377, 599–609.

Collins, W.E., Contacos, P.G. & Guinn, E.G. (1967) Studies on the transmission of simian malarias II. Transmission of the H strain of Plasmodium knowlesi by Anopheles balabacensis balabacensis. The Journal of Parasitology, 53, 841–844.

124

Bibliography

Conway, G.R., Trpis, M. & McClelland, G.A.H. (1974) Population parameters of the mosquito Aedes aegypti (L.) estimated by mark-release-recapture in a suburban habitat in Tanzania. The Journal of Animal Ecology, 43, 289–304.

Corley, R.H.V. (2009) How much palm oil do we need? Environmental Science & Policy, 12, 134–139.

Coviella, C.E., Garcia, J.F., Jeske, D.R., Redak, R.A. & Luck, R.F. (2006) Feasibility of tracking within-field movements of Homalodisca coagulata (Hemiptera: Cicadellidae) and estimating its densities using fluorescent dusts in mark-release- recapture experiments. Journal of Economic Entomology, 99, 1051–1057.

Cox-Singh, J., Davis, T.M.E., Lee, K.-S., Shamsul, S.S.G., Matusop, A., Ratnam, S., Rahman, H.A., Conway, D.J. & Singh, B. (2008) Plasmodium knowlesi malaria in humans is widely distributed and potentially life threatening. Clinical Infectious Diseases, 46, 165–171.

Crumpacker, D.W. (1974) The use of micronized fluorescent dusts to mark adult Drosophila pseudoobscura. American Midland Naturalist, 91, 118–129.

Darling, S.T. (1925) Entomological research in malaria. Southern Medical Journal, 18, 446–449.

Davey, J.T. (2009) A method of marking isolated adult locusts in large numbers as an aid to the study of their seasonal migrations. Bulletin of Entomological Research, 46, 797–802.

Davies, J.B. (1975) Moonlight and the biting activity of Culex (Melanoconion) portesi Senevet & Abonnenc and C. (M.) taeniopus D. & K. (Diptera, Culicidae) in Trinidad forests. Bulletin of Entomological Research, 65, 81–96.

De Guzman, L.I., Frake, A.M. & Rinderer, T.E. (2012) Marking small hive beetles with thoracic notching: Effects on longevity, flight ability and fecundity. Apidologie, 43, 425–431.

Dorvillé, L.F.M. (1996) Mosquitoes as Bioindicators of Forest Degradation in Southeastern Brazil, a Statistical Evaluation of Published Data in the Literature. Studies on Neotropical Fauna and Environment, 31, 68–78.

Dye, C., Davies, C.R. & Lainson, R. (1991) Communication among phlebotomine sandflies: A field study of domesticated Lutzomyia longipalpis populations in Amazonian Brazil. Animal Behaviour, 42, 183–192.

Edwards, F.A., Edwards, D.P., Larsen, T.H., Hsu, W.W., Benedick, S., Chung, A., Vun Khen, C., Wilcove, D.S. & Hamer, K.C. (2014) Does logging and forest conversion to oil palm agriculture alter functional diversity in a biodiversity hotspot? Animal Conservation, 17, 163–173.

125

Bibliography

Edwards, D.P., Hodgson, J.A., Hamer, K.C., Mitchell, S.L., Ahmad, A.H., Cornell, S.J. & Wilcove, D.S. (2010) Wildlife-friendly oil palm plantations fail to protect biodiversity effectively. Conservation Letters, 3, 236–242.

Ellwood, M.D.F. & Foster, W.A. (2001) Line insertion techniques for the study of high forest canopies. Selbyana, 22, 97–102.

Evans, B.R. & Bevier, G.A. (1969) Measurement of field populations of Aedes aegypti with the ovitrap in 1968. Mosquito News, 29, 347–353.

Ewers, R.M., Didham, R.K., Fahrig, L., Ferraz, G., Hector, A., Holt, R.D., Kapos, V., Reynolds, G., Sinun, W., Snaddon, J.L. & Turner, E.C. (2011) A large-scale forest fragmentation experiment: The Stability of Altered Forest Ecosystems Project. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 366, 3292–3302.

Eyles, D.E., Wharton, R.H., Cheong, W.H. & Warren, M. (1964) Studies on malaria and Anopheles balabacensis in Cambodia. Bulletin of the World Health Organization, 30, 7–21.

FAO. (2010) Global Forest Resources Assessment 2010. Main Report. Rome.

FAOSTAT. (2015) Online statistical service, http://faostat.fao.org

Fayle, T.M., Turner, E.C., Snaddona, J.L., Chey, V.K., Chung, A.Y.C., Eggleton, P. & Foster, W.A. (2010) Oil palm expansion into rain forest greatly reduces ant biodiversity in canopy, epiphytes and leaf-litter. Basic and Applied Ecology, 11, 337–345.

Figtree, M., Lee, R., Bain, L., Kennedy, T., Mackertich, S., Urban, M., Cheng, Q. & Hudson, B.J. (2010) Plasmodium knowlesi in human, Indonesian Borneo. Emerging Infectious Diseases, 16, 672–674.

Fitzherbert, E.B., Struebig, M.J., Morel, A., Danielsen, F., Brühl, C.A., Donald, P.F. & Phalan, B. (2008) How will oil palm expansion affect biodiversity? Trends in Ecology & Evolution, 23, 538–545.

Foley, J.A., Defries, R., Asner, G.P., Barford, C., Bonan, G., Carpenter, S.R., Chapin, F.S., Coe, M.T., Daily, G.C., Gibbs, H.K., Helkowski, J.H., Holloway, T., Howard, E.A., Kucharik, C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Ramankutty, N. & Snyder, P.K. (2005) Global consequences of land use. Science, 309, 570–574.

Fong, Y.L., Cadigan, F.C. & Coatney, G.R. (1971) A presumptive case of naturally occurring Plasmodium knowlesi malaria in man in Malaysia. Transactions of the Royal Society of Tropical Medicine and Hygiene, 65, 839–840.

Fornace, K.M., Abidin, T.R., Alexander, N., Brock, P., Grigg, M.J., Murphy, A., William, T., Menon, J., Drakeley, C.J. & Cox, J. (2016) Association between landscape factors and spatial patterns of Plasmodium knowlesi infections in Sabah, Malaysia. Emerging Infectious Disease, 22, 3–10.

126

Bibliography

Forschler, B.T. (1994) Fluorescent spray paint as a topical marker on subterranean termites (Isoptera: Rhinotermitidae). Sociobiology, 24, 27–38.

Foster, W.A., Snaddon, J.L., Turner, E.C., Fayle, T.M., Cockerill, T.D., Ellwood, M.D.F., Broad, G.R., Chung, A.Y.C., Eggleton, P., Khen, C.V. & Yusah, K.M. (2011) Establishing the evidence base for maintaining biodiversity and ecosystem function in the oil palm landscapes of South East Asia. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 366, 3277–3291.

Fryer, J.C. & Meek, C.L. (1989) Further studies on marking an adult mosquito, Psorophora columbiae, in situ, using fluorescent pigments. Southwestern Entomologist, 14, 409–418.

Garnham, P.C.C. (1966) Malaria Parasites and Other Haemosporidia. Blackwell Scientific Publications Ltd., Oxford, UK.

Gaveau, D.L.A., Sloan, S., Molidena, E., Yaen, H., Sheil, D., Abram, N.K., Ancrenaz, M., Nasi, R., Quinones, M., Wielaard, N. & Meijaard, E. (2014) Four decades of forest persistence, clearance and logging on Borneo. PLoS One, 9, 1–11.

Gillett, J.D. (1971) Mosquitoes. Richard Clay Ltd, Suffolk.

Gillies, M.T. & Wilkes, T.J. (1972) The range of attraction of animal baits and carbon dioxide for mosquitoes. Studies in a freshwater area of West Africa. Bulletin of Entomological Research, 61, 389–404.

Gratz, N.G. (1999) Emerging and resurging vector-borne diseases. Annual Review of Entomology, 44, 51–75.

Gray, C.L., Slade, E.M., Mann, D.J. & Lewis, O.T. (2014) Do riparian reserves support dung beetle biodiversity and ecosystem services in oil palm-dominated tropical landscapes? Ecology and Evolution, 4, 1049–1060.

Guerra, C.A., Reiner, R.C., Perkins, T.A., Lindsay, S.W., Midega, J.T., Brady, O.J., Barker, C.M., Reisen, W.K., Harrington, L.C., Takken, W., Kitron, U., Lloyd, A.L., Hay, S.I., Scott, T.W. & Smith, D.L. (2014) A global assembly of adult female mosquito mark-release-recapture data to inform the control of mosquito-borne pathogens. Parasites & Vectors, 7, 276.

Guerra, C.A., Snow, R.W. & Hay, S.I. (2006) A global assessment of closed forests, deforestation and malaria risk. Annals of Tropical Medicine and Parasitology, 100, 189–204.

Guzman, A. & Istúriz, R.E. (2010) Update on the global spread of dengue. International Journal of Antimicrobial Agents, 36, S40–S42.

Hagler, J.R. & Jackson, C.G. (2001) Methods for marking insects: Current techniques and future prospects. Annual Review of Entomology, 46, 511–543.

127

Bibliography

Hamer, G.L., Donovan, D.J., Hood-Nowotny, R., Kaufman, M.G., Goldberg, T.L. & Walker, E.D. (2012) Evaluation of a stable isotope method to mark naturally- breeding larval mosquitoes for adult dispersal studies. Journal of Medical Entomology, 49, 61–70.

Haramis, L.D. (1983) Increased adult size correlated with parity in Aedes triseriatus. Mosquito News, 43, 77–79.

Harb, M., Faris, R., Gad, A.M., Hafez, O.N., Ramzy, R. & Buck, A.A. (1993) The resurgence of lymphatic filariasis in the Nile delta. Bulletin of the World Health Organization, 71, 49–54.

Hardwick, S.R., Toumi, R., Pfeifer, M., Turner, E.C., Nilus, R. & Ewers, R.M. (2015) The relationship between leaf area index and microclimate in tropical forest and oil palm plantation: Forest disturbance drives changes in microclimate. Agricultural and Forest Meteorology, 201, 187–195.

Hii, J.L.K. (1985) Evidence for the existence of genetic variability in the tendency of Anopheles balabacensis to rest in houses and to bite man. The Southeast Asian Journal of Tropical Medicine and Public Health, 16, 173–182.

Hii, J.L.K., Birley, M.H. & Sang, V.Y. (1990) Estimation of survival rate and oviposition interval of Anopheles balabacensis mosquitoes from mark-recapture experiments in Sabah, Malaysia. Medical and Veterinary Entomology, 4, 135–140.

Hii, J.L.K., Chew, M., Sang, V.Y., Munstermann, L.E., Tan, S.G., Panyim, S. & Yasothornsrikul, S. (1991) Population genetic analysis of host seeking and resting behaviors in the malaria vector, Anopheles balabacensis (Diptera: Culicidae). Journal of Medical Entomology, 28, 675–684.

Hii, J.L.K., Kan, S., Vun, Y.S., Chin, K.F., Tambakau, S., Chan, M.K., Lye, M.S., Mak, J.W. & Cochrane, A.H. (1988) Transmission dynamics and estimates of malaria vectorial capacity for Anopheles balabacensis and An. flavirostris (Diptera: Culicidae) on Banggi island, Sabah, Malaysia. Annals of Tropical Medicine & Parasitology, 82, 91–101.

Hii, J.L.K. & Vun, Y.S. (1985) A study of dispersal, survival and adult population estimates of the malaria vector, Anopheles balabacensis Baisas (Diptera: Culicidae) in Sabah, Malaysia. Tropical Biomedicine, 2, 121–131.

Hii, J.L.K., Vun, Y.S., Chin, K.F., Chua, R., Tambakau, S., Binisol, E.S., Fernandez, E., Singh, N. & Chan, M.K.C. (1987) The influence of permethrin-impregnated bednets and mass drug administration on the incidence of Plasmodium falciparum malaria in children in Sabah, Malaysia. Medical and Veterinary Entomology, 1, 397–407.

Hoddle, M.S., Millar, J.G., Hoddle, C.D., Zou, Y., McElfresh, J.S. & Lesch, S.M. (2011) Field optimization of the sex pheromone of Stenoma catenifer (Lepidoptera: Elachistidae): Evaluation of lure types, trap height, male flight distances, and

128

Bibliography

number of traps needed per avocado orchard for detection. Bulletin of Entomological Research, 101, 145–152.

Honório, N.A., Silva, W.D.C., Leite, P.J., Gonçalves, J.M., Lounibos, L.P. & Lourenço- de-Oliveira, R. (2003) Dispersal of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) in an urban endemic dengue area in the state of Rio de Janeiro, Brazil. Memórias do Instituto Oswaldo Cruz, 98, 191–198.

Ikeshoji, T. & Yap, H.H. (1990) Impact of the insecticide-treated sound traps on an Aedes albopictus population. Japanese Journal of Sanitary Zoology, 41, 213– 217.

Imai, N., White, M.T., Ghani, A.C. & Drakeley, C.J. (2014) Transmission and control of Plasmodium knowlesi: A mathematical modelling study. PLoS neglected tropical diseases, 8, e2978.

Irene, M.L. (2011) Identification and Molecular Characterisation of Simian Malaria Parasites in Wild Monkeys of Singapore. National University of Singapore.

Jiang, N., Chang, Q., Sun, X., Lu, H., Yin, J., Zhang, Z., Wahlgren, M. & Chen, Q. (2010) Co-infections with Plasmodium knowlesi and other malaria parasite, Myanmar. Emerging Infectious Diseases, 16, 1476–1478.

Jiram, A.I., Vythilingam, I., NoorAzian, Y.M., Yusof, Y.M., Azahari, A.H. & Fong, M.-Y. (2012) Entomologic investigation of Plasmodium knowlesi vectors in Kuala Lipis, Pahang, Malaysia. Malaria Journal, 11, 213.

Johnson, M.F., Gómez, A. & Pinedo-Vasquez, M. (2008) Land use and mosquito diversity in the Peruvian Amazon. Journal of Medical Entomology.

Johnson, P.H., Spitzauer, V. & Ritchie, S.A. (2012) Field sampling rate of BG-Sentinel traps for Aedes aegypti (Diptera: Culicidae) in suburban Cairns, Australia. Journal of Medical Entomology, 49, 29–34.

Jones, C.E., Lounibos, L.P., Marra, P.P. & Kilpatrick, A.M. (2012) Rainfall influences survival of Culex pipiens (Diptera: Culicidae) in a residential neighborhood in the mid-Atlantic USA. Journal of Medical Entomology, 49, 467–473.

Jongwutiwes, S., Putaporntip, C., Iwasaki, T., Sata, T. & Kanbara, H. (2004) Naturally acquired Plasmodium knowlesi malaria in human, Thailand. Emerging Infectious Diseases, 10, 2211–2213.

Kampango, A., Cuamba, N. & Charlwood, J.D. (2011) Does moonlight influence the biting behaviour of Anopheles funestus? Medical and Veterinary Entomology, 25, 240–246.

Kantele, A. & Jokiranta, T.S. (2011) Review of cases with the emerging fifth human malaria parasite, Plasmodium knowlesi. Clinical Infectious Diseases, 52, 1356– 1362.

129

Bibliography

Kantele, A., Marti, H., Felger, I., Müller, D. & Jokiranta, T.S. (2008) Monkey malaria in a European traveler returning from Malaysia. Emerging Infectious Diseases, 14, 1434–1436.

Keesing, F., Belden, L.K., Daszak, P., Dobson, A., Harvell, C.D., Holt, R.D., Hudson, P., Jolles, A., Jones, K.E., Mitchell, C.E., Myers, S.S., Bogich, T. & Ostfeld, R.S. (2010) Impacts of biodiversity on the emergence and transmission of infectious diseases. Nature, 468, 647–652.

Keiser, J., De Castro, M.C., Maltese, M.F., Bos, R., Tanner, M., Singer, B.H. & Utzinger, J. (2005) Effect of irrigation and large dams on the burden of malaria on a global and regional scale. American Journal of Tropical Medicine and Hygiene, 72, 392–406.

Khan, A.A., Maibach, H.I. & Strauss, W.G. (2007) The role of convection currents in mosquito attraction to human skin. British Journal of Psychiatry, 28, 462–464.

Khim, N., Siv, S. & Kim, S. (2011) Plasmodium knowlesi infection in humans, Cambodia, 2007–2010. Emerging infectious …, 17, 1900–1902.

Khoon, C.C. (1985) Status of malaria vectors in Malaysia. The Southeast Asian Journal of Tropical Medicine in Public Health, 16, 133–138.

Kitthawee, S., Edman, J.D. & Upatham, E.S. (1992) Relationship between female Anopheles dirus (Diptera: Culicidae) body size and parity in a biting population. Journal of Medical Entomology, 29, 921–926.

Koh, L.P., Levang, P. & Ghazoul, J. (2009) Designer landscapes for sustainable biofuels. Trends in Ecology and Evolution, 24, 431–438.

Koh, L.P., Miettinen, J., Liew, S.C. & Ghazoul, J. (2011) Remotely sensed evidence of tropical peatland conversion to oil palm. Proceedings of the National Academy of Sciences of the United States of America, 108, 5127–5132.

Koh, L.P. & Wilcove, D.S. (2007) Cashing in palm oil for conservation. Nature, 448, 993–994.

Koh, L.P. & Wilcove, D.S. (2008) Is oil palm agriculture really destroying tropical biodiversity? Conservation Letters, 1, 60–64.

Kosmidis, I. (2013) brglm: Bias reduction in binomial-response generalized linear models. R package version 0.5-9. Available from http://www.ucl.ac.uk/~ucakiko/index.html.

LaBrecque, G.C., Bailey, D.L., Meifert, D.W. & Weidhaas, D.E. (1975) Density estimates and daily mortality rate evaluations of stable fly Stomoxys calcitrans (Diptera: Muscidae) populations in field cages. The Canadian Entomologist, 107, 597–600.

130

Bibliography

Lam, S.K. (1993) Two decades of dengue in Malaysia. Tropical Medicine, 35, 195– 200.

Lam, S.K., Chua, K.B., Hooi, P.S., Rahimah, M.A., Kumari, S., Tharmaratnam, M., Chuah, S.K., Smith, D.W. & Sampson, I.A. (2001) Chikungunya infection- an emerging disease in Malaysia. The Southeast Asian Journal of Tropical Medicine and Public Health, 32, 447–451.

Lau, K.W., Chen, C.D., Lee, H.L., Izzul, A.A., Asri-Isa, M., Zulfadli, M. & Sofian-Azirun, M. (2013) Vertical distribution of Aedes mosquitoes in multiple storey buildings in Selangor and Kuala Lumpur, Malaysia. Tropical Biomedicine, 30, 36–45.

Laurance, W.F., Sayer, J. & Cassman, K.G. (2014) Agricultural expansion and its impacts on tropical nature. Trends in Ecology and Evolution, 29, 107–116.

Lee, H.L., Rohani, A., Khadri, M.S., Nazni, W.A., Rozilawati, H., Nurulhusna, A.H., Nor Afizah, A.H., Roziah, A., Rosilawati, R. & Teh, C.H. (2015) Dengue vector control in Malaysia- challenges and recent advances. International Medical Journal Malaysia, 14, 11–16.

Lehane, M.J. (1991) Biology of Blood-Sucking Insects. HarperCollins Academic, London.

Leisnham, P.T., Lester, P.J., Slaney, D.P. & Weinstein, P. (2004) Anthropogenic Landscape Change and Vectors in New Zealand: Effects of Shade and Nutrient Levels on Mosquito Productivity. EcoHealth, 1, 306–316.

Leisnham, P.T., Slaney, D.P., Lester, P.J., Weinstein, P. & Heath, A.C.G. (2007) Mosquito density, macroinvertebrate diversity, and water chemistry in water-filled containers: Relationships to land use. New Zealand Journal of Zoology, 34, 203– 218.

Lim, K.G. & Singh, B. (2013) Zoonotic Malaria in Malaysia. Medical Journal of Malaysia, 68, 4–5.

Lindquist, A.W., Ikeshoji, T., Grab, B., de Meillon, B. & Khan, Z.H. (1967) Dispersion studies of Culex pipiens fatigans tagged with 32P in the Kemmendine area of Rangoon, Burma. Bulletin of the World Health Organization, 36, 21–37.

Liu, Q., Liu, X., Zhou, G., Jiang, J., Guo, Y., Ren, D., Zheng, C., Wu, H., Yang, S., Liu, J., Li, H., Li, H., Li, Q., Yang, W. & Chu, C. (2012) Dispersal range of Anopheles sinensis in Yongcheng City, China by mark-release-recapture methods. PloS One, 7, e51209.

Luchavez, J., Espino, F., Curameng, P., Espina, R., Bell, D., Chiodini, P., Nolder, D., Sutherland, C., Lee, K.-S. & Singh, B. (2008) Human infections with Plasmodium knowlesi, the Philippines. Emerging Infectious Diseases, 14, 811–813.

Luskin, M.S. & Potts, M.D. (2011) Microclimate and habitat heterogeneity through the oil palm lifecycle. Basic and Applied Ecology, 12, 540–551.

131

Bibliography

Lysyk, T.J. & Axtell, R.C. (1986) Estimating numbers and survival of house flies (Diptera: Muscidae) with mark/recapture methods. Journal of Economic Entomology, 79, 1016–1022.

Mackenzie, J.S., Gubler, D.J. & Petersen, L.R. (2004) Emerging flaviviruses: The spread and resurgence of Japanese encephalitis, West Nile and dengue viruses. Nature Medicine, 10, S98–109.

Manga, L., Toto, J.C. & Carnevale, P. (1995) Malaria vectors and transmission in an area deforested for a new international airport in Southern Cameroon. Annales- societe Belge de Medecine Tropicale, 75, 43–49.

Manguin, S. (2013) Anopheles Mosquitoes - New Insights into Malaria Vectors. InTech, Rijeka, Croatia.

Marsh, C.W. & Greer, A.G. (1992) Forest land-use in Sabah, Malaysia: An introduction to Danum Valley. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 335, 331–339.

Martens, P. & Hall, L. (2000) Malaria on the move: Human population movement and malaria transmission. Emerging Infectious Diseases, 6, 103–109.

Martens, W.J.M., Niessen, L.W., Rotmans, J., Jetten, T.H. & McMichael, A.J. (1995) Potential impact of global climate change on malaria risk. Environmental Health Perspectives, 103, 458–464.

Mbogo, C.N.M., Baya, N.M., Ofulla, A.V.O., Githure, J.I. & Snow, R.W. (1996) The impact of permethrin-impregnated bednets on malaria vectors of the Kenyan coast. Medical and Veterinary Entomology, 10, 251–259.

Messina, J.P., Brady, O.J., Scott, T.W., Zou, C., Pigott, D.M., Duda, K.A., Bhatt, S., Katzelnick, L., Howes, R.E., Battle, K.E., Simmons, C.P. & Hay, S.I. (2014) Global spread of dengue virus types: Mapping the 70 year history. Trends in Microbiology, 22, 138–146.

Meyerdirk, D.E., Hart, W.G. & Burnside, J. (1979) Marking and dispersal study of adults of the citrus blackfly, Aleurocanthus woglumi. Southwestern Entomologist, 4, 325–329.

Millar, S.B. & Cox-Singh, J. (2015) Human infections with Plasmodium knowlesi- zoonotic malaria. Clinical Microbiology and Infection, 21, 640–648.

Miller, L.R. (1993) Fluorescent dyes as markers in studies of foraging biology of termite colonies (Isoptera). Sociobiology, 23, 127–134.

Miller, T.A., Stryker, R.G., Wilkinson, R.N. & Esah, S. (1970) The influence of moonlight and other environmental factors on the abundance of certain mosquito species in light-trap collections in Thailand. Journal of Medical Entomology, 7, 555–561.

132

Bibliography

Moffitt, H.R. & Albano, D.J. (1972) Codling moths: Fluorescent powders as markers. Environmental Entomology, 1, 750–753.

Moth, J.J. & Barker, J.S.F. (1975) Micronized fluorescent dusts for marking Drosophila adults. Journal of Natural History, 9, 393–396.

Muir, L.E. & Kay, B.H. (1998) Aedes aegypti survival and dispersal estimated by mark- release-recapture in northern Australia. The American Journal of Tropical Medicine and Hygiene, 58, 277–282.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. & Kent, J. (2000) Biodiversity hotspots for conservation priorities. Nature, 403, 853–858.

Myles, T.G. & Grace, J.K. (1991) Behavioural ecology of the eastern subterranean termite in Ontario as a basis for control. Proceedings: Technology transfer conference, Ontario Ministry of the environment, 2, 547–554.

Nakata, T. (2008) Effectiveness of micronized fluorescent powder for marking citrus psyllid, Diaphorina citri. Applied Entomology and Zoology, 43, 33–36.

Naranjo, S.E. (1990) Influence of two mass-marking techniques on survival and flight behavior of Diabrotica virgifera virgifera (Coleoptera: Chrysomelidae). Journal of Economic Entomology, 83, 1360–1364.

Narisu, Lockwood, J.A. & Schell, S.P. (1999) A novel mark-recapture technique and its application to monitoring the direction and distance of local movements of rangeland grasshoppers (Orthoptera: Acrididae) in the context of pest management. Journal of Applied Ecology, 36, 604–617.

Ndiaye, P.I., Bicout, D.J., Mondet, B. & Sabatier, P. (2006) Rainfall triggered dynamics of Aedes mosquito aggressiveness. Journal of Theoretical Biology, 243, 222–229.

Ng, O.T., Ooi, E.E., Lee, C.C., Lee, P.J., Ng, L.C., Pei, S.W., Tu, T.M., Loh, J.P. & Leo, Y.S. (2008) Naturally acquired human Plasmodium knowlesi infection, Singapore. Emerging Infectious Diseases, 14, 814–6.

Niebylski, M.L. & Craig, G.B. (1994) Dispersal and survival of Aedes albopictus at a scrap tire yard in Missouri. Journal of the American Mosquito Control Association, 10, 339–343.

Norris, D.E. (2004) Mosquito-borne diseases as a consequence of land use change. EcoHealth, 1, 19–24.

Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H. & Wagner, H. (2015) vegan: Community Ecology Package. R package version 2.2-1. Available from http://CRAN.R-project.org/package=vegan.

133

Bibliography

Oloumi-Sadeghi, H. & Levine, E. (1990) A simple, effective, and low-cost method for mass marking adult western corn rootworms (Coleoptera: Chrysomelidae). Journal of Entomological Science, 25, 170–175.

Patz, J.A., Daszak, P., Tabor, G.M., Aguirre, A.A., Pearl, M., Epstein, J., Wolfe, N.D., Kilpatrick, A.M., Foufopoulos, J., Molyneux, D. & Bradley, D.J. (2004) Unhealthy landscapes: Policy recommendations on land use change and infectious disease emergence. Environmental Health Perspectives, 112, 1092–1098.

Patz, J.A., Graczyk, T.K., Geller, N. & Vittor, A.Y. (2000) Effects of environmental change on emerging parasitic diseases. International Journal for Parasitology, 30, 1395–1405.

Peh, K.S.H., Sodhi, N.S., De Jong, J., Sekercioglu, C.H., Yap, C.A.M. & Lim, S.L.H. (2006) Conservation value of degraded habitats for forest birds in southern Peninsular Malaysia. Diversity and Distributions, 12, 572–581.

Petrarca, V., Sabatinelli, G., Touré, Y.T. & Di Deco, M.A. (1998) Morphometric multivariate analysis of field samples of adult Anopheles arabiensis and An. gambiae s.s. (Diptera: Culicidae). Journal of Medical Entomology, 35, 16–25.

Phalan, B., Bertzky, M., Butchart, S.H.M., Donald, P.F., Scharlemann, J.P.W., Stattersfield, A.J. & Balmford, A. (2013) Crop expansion and conservation priorities in tropical countries. PLOS One, 8, e51759.

Pluess, B., Mueller, I., Levi, D., King, G., Smith, T.A. & Lengeler, C. (2009) Malaria- a major health problem within an oil palm plantation around Popondetta, Papua New Guinea. Malaria Journal, 8, 56.

Pongsiri, M.J., Roman, J., Ezenwa, V.O., Goldberg, T.L., Koren, H.S., Newbold, S.C., Ostfeld, R.S., Pattanayak, S.K. & Salkeld, D.J. (2009) Biodiversity loss affects global disease ecology. BioScience, 59, 945–954.

R Core Team (2014) R: A language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, URL http:www.R- project.org/

Rahman, W.A., Che’rus, A. & Ahmad, A.H. (1997) Malaria and Anopheles mosquitos in Malaysia. The Southeast Asian Journal of Tropical Medicine and Public Health, 28, 599–605.

Rajahram, G.S., Barber, B.E., William, T., Menon, J., Anstey, N.M. & Yeo, T.W. (2012) Deaths due to Plasmodium knowlesi malaria in Sabah, Malaysia: Association with reporting as Plasmodium malariae and delayed parenteral artesunate. Malaria Journal, 11, 284.

Rattanarithikul, R., Harbach, R.E., Harrison, B.A., Panthusiri, P., Coleman, R.E. & Richardson, J.H. (2010) Illustrated keys to the mosquitoes of Thailand VI. Tribe Aedini. The Southeast Asian Journal of Tropical Medicine and Public Health, 41, 1–225.

134

Bibliography

Rattanarithikul, R., Harbach, R.E., Harrison, B.A., Panthusiri, P., Jones, J.W. & Coleman, R.E. (2005a) Illustrated keys to the mosquitoes of Thailand II. Genera Culex and Lutzia. The Southeast Asian Journal of Tropical Medicine and Public Health, 36, 1–97.

Rattanarithikul, R., Harrison, B.A., Harbach, R.E., Panthusiri, P. & Coleman, R.E. (2006a) Illustrated keys to the mosquitoes of Thailand IV. Anopheles. The Southeast Asian Journal of Tropical Medicine and Public Health, 37, 1–128.

Rattanarithikul, R., Harrison, B.A., Panthusiri, P. & Coleman, R.E. (2005b) Illustrated keys to the mosquitoes of Thailand I. Background; geographic distribution; lists of genera, subgenera, and species; and a key to the genera. The Southeast Asian Journal of Tropical Medicine and Public Health, 36, 1–80.

Rattanarithikul, R., Harrison, B.A., Panthusiri, P., Peyton, E.L. & Coleman, R.E. (2006b) Illustrated keys to the mosquitoes of Thailand III. Genera Aedeomyia, Ficalbia, Mimomyia, Hodgesia, Coquillettidia, Mansonia, and Uranotaenia. Southeast Asian Journal of Tropical Medicine in Public Health, 37, 1–85.

Reid, J.A. (1968) Anopheline mosquitoes of Malaya and Borneo. Studies from the Institute for Medical Research Malaysia, 31, 1–520.

Reid, T.G. & Reid, M.L. (2008) Fluorescent powder marking reduces condition but not survivorship in adult mountain pine beetles. The Canadian Entomologist, 140, 582–588.

Reinert, J.F., Harbach, R.E. & Kitching, I.J. (2004) Phylogeny and classification of Aedini (Diptera: Culicidae), based on morphological characters of all life stages. Zoological Journal of the Linnean Society, 142, 289–368.

Reisen, W.K. & Aslamkhan, M. (1979) A release-recapture experiment with the malaria vector, Anopheles stephensi Liston, with observations on dispersal, survivorship, population size, gonotrophic rhythm and mating behaviour. Annals of Tropical Medicine and Parasitology, 73, 251–269.

Reisen, W.K., Mahmood, F. & Parveen, T. (1979) Anopheles subpictus Grassi: Observations on survivorship and population size using mark-release-recapture and dissection methods. Researches on Population Ecology, 21, 12–29.

Reiter, P., Amador, M.A., Anderson, R.A. & Clark, G.G. (1995) Short report: Dispersal of Aedes aegypti in an urban area after blood feeding as demonstrated by rubidium-marked eggs. American Journal of Tropical Medicine and Hygiene, 52, 177–179.

Renshaw, M., Service, M.W. & Birley, M.H. (1994) Host finding, feeding patterns and evidence for a memorized home range of the mosquito Aedes cantans. Medical and Veterinary Entomology, 8, 187–193.

135

Bibliography

Rohani, A., Lokman Hakim, S., Hassan, A.R., Chan, S.T., Ong, Y.F., Abdullah, A.G. & Lee, H.L. (1999) Bionomics of Anopheles balabacensis Baisas, the principal malaria vector in Ranau, Sabah. Tropical Biomedicine, 16, 31–38.

Rozilawati, H., Zairi, J. & Adanan, C.R. (2007) Seasonal abundance of Aedes albopictus in selected urban and suburban areas in Penang, Malaysia. Tropical Biomedicine, 24, 83–94.

Rubio-Palis, Y. (1992) Influence of moonlight on light trap catches of the malaria vector Anopheles nuneztovari in Venezuela. Journal of America Mosquito Control Association, 8, 178–180.

Rueda, L.M., Patel, K.J., Axtell, R.C. & Stinner, R.E. (1990) Temperature-dependent development and survival rates of Culex quinquefasciatus and Aedes aegypti (Diptera: Culicidae). Journal of Medical Entomology, 27, 892–898.

Rundi, C. (2011) Malaria Elimination in Malaysia. Third annual APMEN technical and business meeting, 9–12 May 2011; Kota Kinabalu, Malaysia. Available: http://apmen.org/apmen-iii-meeting-proceedings/. Accessed 3 February 2015.

Russell, R.C., Webb, C.E., Williams, C.R. & Ritchie, S.A. (2005) Mark-release- recapture study to measure dispersal of the mosquito Aedes aegypti in Cairns, Queensland, Australia. Medical and Veterinary Entomology, 19, 451–457.

Sala, O.E., Chapin, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber- Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M. & Wall, D.H. (2000) Global biodiversity scenarios for the year 2100. Science, 287, 1770–1774.

Sallum, M.A.M., Peyton, E.L., Harrison, B.A. & Wilkerson, R.C. (2005) Revision of the Leucosphyrus group of Anopheles (Cellia) (Diptera, Culicidae). Revista Brasileira de Entomologica, 49, 1–152.

Sallum, M.A.M., Peyton, E.L. & Wilkerson, R.C. (2005) Six new species of the Anopheles Leucosphyrus group, reinterpretation of An. elegans and vector implications. Medical and Veterinary Entomology, 19, 158–199.

Sam, S.S., Omar, S.F.S., Teoh, B.T., Abd-Jamil, J. & AbuBakar, S. (2013) Review of dengue hemorrhagic fever fatal cases seen among adults: A retrospective study. PLoS Neglected Tropical Diseases, 7, 1–7.

Schulze, C.H., Waltert, M., Kessler, P.J.A., Pitopang, R., Shahabuddin, Veddeler, D., Mühlenberg, M., Gradstein, S.R., Leuschner, C., Steffan-Dewenter, I. & Tscharntke, T. (2004) Biodiversity indicator groups of tropical land-use systems: comparing plants, birds, and insects. Ecological ResearchApplications, 14, 1321– 1333.

136

Bibliography

Sempala, S.D.K. (1981) The ecology of Aedes (Stegomyia) africanus (Theobald) in a tropical forest in Uganda: mark-release-recapture studies on a female adult population. International Journal of Tropical Insect Science, 1, 211–224.

Sheppard, P.M., Macdonald, W.W., Tonn, R.J. & Grab, B. (1969) The dynamics of an adult population of Aedes aegypti in relation to dengue haemorrhagic fever in Bangkok. Journal of Animal Ecology, 38, 661–702.

Silver, J.B. (2008) Mosquito Ecology: Field Sampling Methods, 3rd ed. Springer, London.

Singh, B. & Daneshvar, C. (2013) Human infections and detection of Plasmodium knowlesi. Clinical Microbiology Reviews, 26, 165–184.

Singh, B., Kim Sung, L., Matusop, A., Radhakrishnan, A., Shamsul, S.S.G., Cox- Singh, J., Thomas, A. & Conway, D.J. (2004) A large focus of naturally acquired Plasmodium knowlesi infections in human beings. Lancet, 363, 1017–1024.

Singh, N., Mishra, A.K., Curtis, C.F. & Sharma, V.P. (1996) Influence of moonlight on light trap catches of the malaria vector Anopheles culicifacies (Diptera: Culicidae) in India. Bulletin of Entomological Research, 86, 475–479.

Sinka, M.E., Bangs, M.J., Manguin, S., Chareonviriyaphap, T., Patil, A.P., Temperley, W.H., Gething, P.W., Elyazar, I.R.F., Kabaria, C.W., Harbach, R.E. & Hay, S.I. (2011) The dominant Anopheles vectors of human malaria in the Asia-Pacific region: occurrence data, distribution maps and bionomic précis. Parasites & Vectors, 4, 89.

Snow, K.R. (1990) Mosquitoes. Naturalist’s Handbooks Series. Richmond Publishing Co. Ltd, London.

Sochantha, T., Van Bortel, W., Savonnaroth, S., Marcotty, T., Speybroeck, N. & Coosemans, M. (2010) Personal protection by long-lasting insecticidal hammocks against the bites of forest malaria vectors. Tropical Medicine & International Health, 15, 336–341.

Sodhi, N.S., Koh, L.P., Brook, B.W. & Ng, P.K.L. (2004) Southeast Asian biodiversity: An impending disaster. Trends in Ecology & Evolution, 19, 654–660.

Sodhi, N.S., Koh, L.P., Clements, R., Wanger, T.C., Hill, J.K., Hamer, K.C., Clough, Y., Tscharntke, T., Posa, M.R.C. & Lee, T.M. (2010a) Conserving Southeast Asian forest biodiversity in human-modified landscapes. Biological Conservation, 143, 2375–2384.

Sodhi, N.S., Posa, M.R.C., Lee, T.M., Bickford, D., Koh, L.P. & Brook, B.W. (2010b) The state and conservation of Southeast Asian biodiversity. Biodiversity and Conservation, 19, 317–328.

Southwood, T.R.E. & Henderson, P.A. (2000) Ecological Methods, 3rd ed. Blackwell Science, Oxford.

137

Bibliography

Souza, N.A., Andrade-Coelho, C.A., Silva, V.C., Peixoto, A.A. & Rangel, E.F. (2005) Moonlight and blood-feeding behaviour of Lutzomyia intermedia and Lutzomyia whitmani (Diptera: Psycododae: Phlebotominae) vectors of American cutaneous leishmaniasis in Brazil. Memorias do Instituto Oswaldo Cruz, 100, 39–42.

Stresman, G.H. (2010) Beyond temperature and precipitation: Ecological risk factors that modify malaria transmission. Acta Tropica, 116, 167–172.

Struebig, M.J., Kingston, T., Zubaid, A., Mohd-Adnan, A. & Rossiter, S.J. (2008) Conservation value of forest fragments to Palaeotropical bats. Biological Conservation, 141, 2112–2126.

Su, N.-Y., Ban, P.M. & Scheffrahn, R.H. (1991) Evaluation of twelve dye markers for population studies of the eastern and Formosan subterranean termite (Isoptera: Rhinotermitidae). Sociobiology, 19, 349–362.

Takken, W. (2002) Do insecticide treated bednets have an effect on malaria vectors? Tropical Medicine & International Health, 7, 1022–1030.

Takken, W., Charlwood, J.D., Billingsley, P.F. & Gort, G. (1998) Dispersal and survival of Anopheles funestus and An. gambiae s.l. (Diptera: Culicidae) during the rainy season in southeast Tanzania. Bulletin of Entomological Research, 88, 561–566.

Takken, W. & Knols, B.G.J. (1999) Odor-mediated behavior of Afrotropical malaria mosquitoes. Annual Review of Entomology, 44, 131–157.

Tan, C.H., Vythilingam, I., Matusop, A., Chan, S.T. & Singh, B. (2008) Bionomics of Anopheles latens in Kapit, Sarawak, Malaysian Borneo in relation to the transmission of zoonotic simian malaria parasite Plasmodium knowlesi. Malaria Journal, 7, 52.

Tatem, A.J., Hay, S.I. & Rogers, D.J. (2006) Global traffic and disease vector dispersal. Proceedings of the National Academy of Sciences of the United States of America, 103, 6242–6247.

Thang, N.D., Erhart, A., Speybroeck, N., Xa, N.X., Thanh, N.N., Ky, P. Van, Hung, L.X., Thuan, L.K., Coosemans, M. & D’Alessandro, U. (2009) Long-lasting insecticidal hammocks for controlling forest malaria: A community-based trial in a rural area of central Vietnam. PloS One, 4, e7369.

Therneau, T.M. (2015) survival: Survival Analysis. R package version 2.38-3. Available from https://cran.r-project.org/web/packages/survival/survival.pdf.

Toepfer, S., Levay, N. & Kiss, J. (2005) Suitability of different fluorescent powders for mass-marking the Chrysomelid, Diabrotica virgifera virgifera LeConte. Journal of Applied Entomology, 129, 456–464.

Tren, R. & Bate, R. (2001) Malaria and the DDT Story. The Institute of Economic Affairs, London.

138

Bibliography

Tsuda, Y., Komagata, O., Kasai, S., Hayashi, T., Nihei, N., Saito, K., Mizutani, M., Kunida, M., Yoshida, M. & Kobayashi, M. (2008) A mark-release-recapture study on dispersal and flight distance of Culex pipiens pallens in an urban area of Japan. Journal of the American Mosquito Control Association, 24, 339–343.

Turner, E.C. & Foster, W.A. (2009) The impact of forest conversion to oil palm on arthropod abundance and biomass in Sabah, Malaysia. Journal of Tropical Ecology, 25, 23–30.

Van den Eede, P., Van, H.N., Overmeir, C.V., Vythilingam, I., Duc, T.N., Hung, L.X., Manh, H.N., Anné, J., D’Alessandro, U. & Erhart, A. (2009) Human Plasmodium knowlesi infections in young children in central Vietnam. Malaria Journal, 8, 249.

Vavrek, M.J. (2015) fossil: Palaeoecological and Palaeogeographical Analysis Tools. R package version 0.3.7 . Available from https://cran.r- project.org/web/packages/fossil/fossil.pdf.

Vitousek, P.M. (1994) Beyond global warming: Ecology and global change. Ecology, 75, 1861–1876.

Vittor, A.Y., Gilman, R.H., Tielsch, J., Glass, G., Shields, T., Lozano, W.S., Pinedo- Cancino, V. & Patz, J.A. (2006) The effect of deforestation on the human-biting rate of Anopheles darlingi, the primary vector of falciparum malaria in the Peruvian Amazon. The American Journal of Tropical Medicine and Hygiene, 74, 3–11.

Vittor, A.Y., Pan, W., Gilman, R.H., Tielsch, J., Glass, G., Shields, T., Sánchez- Lozano, W., Pinedo, V. V, Salas-Cobos, E., Flores, S. & Patz, J.A. (2009) Linking deforestation to malaria in the Amazon: Characterization of the breeding habitat of the principal malaria vector, Anopheles darlingi. The American Journal of Tropical Medicine and Hygiene, 81, 5–12.

Vythilingam, I. (2010) Plasmodium knowlesi in humans: A review on the role of its vectors in Malaysia. Tropical Biomedicine, 27, 1–12.

Vythilingam, I. (2012) Plasmodium knowlesi and Wuchereria bancrofti: Their vectors and challenges for the future. Frontiers in Physiology, 3, 1–9.

Vythilingam, I., Chan, S.T., Shanmugratnam, C., Tanrang, H. & Chooi, K.H. (2005) The impact of development and malaria control activities on its vectors in the Kinabatangan area of Sabah, East Malaysia. Acta Tropica, 96, 24–30.

Vythilingam, I., Chiang, G.L. & Chan, S.T. (1992) Evaluation of carbon dioxide and 1- octen-3-ol as mosquito attractants. The Southeast Asian Journal of Tropical Medicine and Public Health, 23, 328–331.

Vythilingam, I., Noorazian, Y.M., Huat, T.C., Jiram, A.I., Yusri, Y.M., Azahari, A.H., NorParina, I., NoorRain, A. & LokmanHakim, S. (2008) Plasmodium knowlesi in humans, macaques and mosquitoes in peninsular Malaysia. Parasites & Vectors, 1, 26.

139

Bibliography

Vythilingam, I., Tan, C.H., Asmad, M., Chan, S.T., Lee, K.S. & Singh, B. (2006) Natural transmission of Plasmodium knowlesi to humans by Anopheles latens in Sarawak, Malaysia. Transactions of the Royal Society of Tropical Medicine and Hygiene, 100, 1087–1088.

Walsh, J.F., Molyneux, D.H. & Birley, M.H. (1993) Deforestation: Effects on vector- borne disease. Parasitology, 106, S55–S75.

Walsh, R.P. & Newbery, D.M. (1999) The ecoclimatology of Danum, Sabah, in the context of the world’s rainforest regions, with particular reference to dry periods and their impact. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 354, 1869–1883.

Walter Reed Biosystematics Unit (2015) Medically Important Mosquitoes, http://www.wrbu.org/pacom_MQ.html [accessed 10/09/2015]

Warren, M., Cheong, W.H., Fredericks, H.K. & Coatney, G.R. (1970) Cycles of jungle malaria in West Malaysia. American Journal of Tropical Medicine and Hygiene, 19, 383–393.

Watson, T.M., Saul, A. & Kay, B.H. (2000) Aedes notoscriptus (Diptera: Culicidae) survival and dispersal estimated by mark-release-recapture in Brisbane, Queensland, Australia. Journal of Medical Entomology, 37, 380–384.

Welch, C.H., Kline, D.L., Allan, S.A. & Barnard, D.R. (2006) Laboratory evaluation of a dyed food marking technique for Culex quinquefasciatus (Diptera: Culicidae). Journal of the American Mosquito Control Association, 22, 626–628.

Weldon, C.W. (2005) Marking Queensland fruit fly, Bactrocera tryoni (Froggatt) (Diptera: Tephritidae) with fluorescent pigments: Pupal emergence, adult mortality, and visibility and persistence of marks. General and Applied Entomology, 34, 99–108.

Wesolowski, A., Eagle, N., Tatem, A.J.J., Smith, D.L.L., Noor, A.M.M., Snow, R.W. & Buckee, C.O. (2012) Quantifying the impact of human mobility on malaria. Science, 338, 267–270.

Wharton, R.H., Eyles, D.E., Warren, M.W. & Cheong, W.H. (1964) Studies to determine the vectors of monkey malaria in Malaya. Annals of tropical medicine and parasitology, 58, 56–77.

White, N.J. (2008) Plasmodium knowlesi: The fifth human malaria parasite. Clinical Infectious Diseases, 46, 172–173.

WHO. (2009) Dengue Guidelines for Diagnosis, Treatment, Prevention and Control. World Health Organization, Geneva, Switzerland.

WHO. (2013) Sustaining the Drive to Overcome the Global Impact of Neglected Tropical Diseases. Geneva, Switzerland.

140

Bibliography

WHO. (2014a) Japanese encephalitis. Fact sheet No 386, http://www.who.int/mediacentre/factsheets/fs386/en/

WHO. (2014b) World Malaria Report 2014. Geneva, Switzerland.

WHO. (2015a) Dengue and dengue haemorrhagic fever. Fact sheet No 117, http://www.who.int/mediacentre/factsheets/fs117/en/

WHO. (2015b) Lymphatic filariasis. Fact sheet No 102, http://www.who.int/mediacentre/factsheets/fs102/en/

Wilkins, E.E., Smith, S.C., Roberts, J.M. & Benedict, M. (2007) Rubidium marking of Anopheles mosquitoes detectable by field-capable X-ray spectrometry. Medical and Veterinary Entomology, 21, 196–203.

William, T., Jelip, J., Menon, J., Anderios, F., Mohammad, R., Mohammad, T.A.A., Grigg, M.J., Yeo, T.W., Anstey, N.M. & Barber, B.E. (2014) Changing epidemiology of malaria in Sabah, Malaysia: increasing Incidence of Plasmodium knowlesi. Malaria Journal, 13, 390.

William, T., Rahman, H.A., Jelip, J., Ibrahim, M.Y., Menon, J., Grigg, M.J., Yeo, T.W., Anstey, N.M. & Barber, B.E. (2013) Increasing incidence of Plasmodium knowlesi malaria following control of P. falciparum and P. vivax Malaria in Sabah, Malaysia. PLoS Neglected Tropical Diseases, 7, e2026.

Williams, C.R., Bader, C.A., Williams, S.R. & Whelan, P.I. (2012) Adult mosquito trap sensitivity for detecting exotic mosquito incursions and eradication: a study using EVS traps and the Australian southern saltmarsh mosquito, Aedes camptorhynchus. Journal of Vector Ecology, 37, 110–116.

Williams, D.F., LaBrecque, G.C. & Patterson, R.S. (1977) Effect of gamma rays and/or fluorescent pigments on sterility and survival of the stable fly. The Florida Entomologist, 60, 297–299.

Wiwanitkit, V. (2007) Malaria Research in Southeast Asia. Nova Science Publishing, New York.

Wolfe, N.D., Daszak, P., Kilpatrick, A.M. & Burke, D.S. (2005) Bushmeat hunting, deforestation, and prediction of zoonotic disease emergence. Emerging Infectious Diseases, 11, 1822–1827.

Wolfe, N.D., Dunavan, C.P. & Diamond, J. (2007) Origins of major human infectious diseases. Nature, 447, 279–283.

Wolfe, N.D., Eitel, M.N., Gochowski, J., Muchaal, P.K., Nolte, C., Prosser, A.T., Torimiro, J.N., Weise, S.F. & Burke, D.S. (2000) Deforestation, hunting and the ecology of microbial emergence. Global Change and Human Health, 1, 10–25.

Wong, M.L., Chua, T.H., Leong, C.S., Khaw, L.T., Fornace, K., Wan-Sulaiman, W.-Y., William, T., Drakeley, C., Ferguson, H.M. & Vythilingam, I. (2015) Seasonal and

141

Bibliography

Spatial Dynamics of the Primary Vector of Plasmodium knowlesi within a Major Transmission Focus in Sabah, Malaysia. PLOS Neglected Tropical Diseases, 9, e0004135.

Wood, W.B. (1990) Tropical deforestation concerns. Global Environmental Change, 1, 23–41.

Yap, H.H. & Thiruvengadam, V. (1979) Relative abundance of Aedes aegypti (Linnaeus) and Aedes albopictus (Skuse) in different habitat- an ovitrap survey conducted in Georgetown, Penang, Malaysia. Medical Journal of Malaysia, 34, 76–79.

Yasuoka, J. & Levins, R. (2007) Impact of deforestation and agricultural development on anopheline ecology and malaria epidemiology. The American Journal of Tropical Medicine and Hygiene, 76, 450–460.

Yusof, R., Lau, Y.L., Mahmud, R., Fong, M.Y., Jelip, J., Ngian, H.U., Mustakim, S., Hussin, H.M., Marzuki, N. & Mohd Ali, M. (2014) High proportion of knowlesi malaria in recent malaria cases in Malaysia. Malaria Journal, 13, 168.

Zhang, D.D. & Shi, C. (2001) Sedimentary causes and management of two principal environmental problems in the lower Yellow River. Environmental Management, 28, 749–760.

Zim, M.A.M., Sam, I.-C., Omar, S.F.S., Chan, Y.F., AbuBakar, S. & Kamarulzaman, A. (2013) Chikungunya infection in Malaysia: Comparison with dengue infection in adults and predictors of persistent arthralgia. Journal of Clinical Virology, 56, 141–145.

142

Appendix A

Appendices

Appendix A

Chapter 2 supplementary figures and tables

Figure A.1. Ovitrap design

143

Appendix A

Figure A.2. DCA from Figure 2.2a, only showing only species labels. See Table A.1 for full species names

144

Appendix A

Figure A.3. DCA from Figure 2.2b, only showing only species labels. See Table A.1 for full species names

145

Appendix A

Table A.1. List of full taxonomic names of collected species Species Mosquito genera and species abbreviation An.Ait Anopheles (Anopheles) Aitkenii group (Reid & Knight 1961) An.bar Anopheles (Anopheles) barbirostris (van der Wulp 1884) An.van Anopheles (Anopheles) vanus (Walker 1860) An.sp Anopheles (Anopheles) spp. An.bal Anopheles (Cellia) balabacensis (Baisas 1936) An.koc Anopheles (Cellia) kochi (Dönitz 1901) An.lat Anopheles (Cellia) latens (Sallum & Peyton 2005) An.maca Anopheles (Cellia) macarthuri (Colless 1956) An.macu Anopheles (Cellia) maculatus (Theobald 1901) An.tes Anopheles (Cellia) tessellatus (Theobald 1901) An.wat Anopheles (Cellia) watsonii (Leicester 1908) Arm.jug Armigeres (Armigeres) jugraensis (Leicester 1908) Arm.con Armigeres (Armigeres) confusus (Edwards 1915) Arm.fl Armigeres (Armigeres) flavus (Leicester 1908) Coll.pse Collessius (Alloemyia) pseudotaeniatus (Giles 1901) Coq.cr Coquillettidia (Coquillettidia) crassipes (van der Wulp 1881) Cx.Cul Culex (Culiciomyia) spp. Cx.nig Culex (Culiciomyia) nigropunctatus (Edwards 1926) Cx.pap Culex (Culiciomyia) papuensis (Taylor 1914) Cx.sca Culex (Culiciomyia) scanloni (Bram 1967) Cx.gel Culex (Culex) gelidus (Theobald 1901) Cx.mim Culex (Culex) mimulus (Edwards 1915) Cx.qui Culex (Culex) quinquefasciatus (Say 1823) Cx.sit Culex (Culex) sitiens (Wiedemann 1828) Cx.vis Culex (Culex) vishnui (Theobald 1901) Cx.Lop Culex (Lophoceraomyia) sp. Cx.bit Culex (Oculeomyia) bitaeniorhynchus (Giles 1901) Do.gan Downsiomyia ganapathi (Colless 1958) He.sci Heizmannia (Heizmannia) scintillans (Ludlow 1905) Ma.ann Mansonia (Mansonioides) annulata (Leicester 1908) Or.sp Orthopodomyia sp. (Theobald 1904) Pr.ost Paraedes ostentatio (Leicester 1908) Stg.al Stegomyia albopicta (Skuse 1895) Tri.sp Tripteroides sp. (Giles 1904) Ur.sp Uranotaenia spp. (Lynch Arribalzaga 1891) Ze.gra Zeugnomyia gracilis (Leicester 1908)

Table A.2. Effects of land use on sex ratio of hatched larvae, using a generalised linear model with quasipoisson errors Predictor β SE z p Intercept 0.074 0.111 0.664 =0.506 Area HDF -0.102 0.176 -0.578 =0.563 Area PF -0.068 0.123 -0.549 =0.583 Area SU -0.053 0.181 -0.294 =0.768 Area OP NA NA NA NA

146

Appendix B

Chapter 3 supplementary figures and tables

Table B.1. List of full taxonomic names of collected species Mosquito genera and species Anopheles (Anopheles) Aitkenii group (Reid & Knight 1961) Anopheles (Anopheles) barbirostris (van der Wulp 1884) Anopheles (Anopheles) vanus (Walker 1860) Anopheles (Anopheles) spp. Anopheles (Cellia) balabacensis (Baisas 1936) Anopheles (Cellia) kochi (Dönitz 1901) Anopheles (Cellia) latens (Sallum & Peyton 2005) Anopheles (Cellia) macarthuri (Colless 1956) Anopheles (Cellia) maculatus (Theobald 1901) Anopheles (Cellia) tessellatus (Theobald 1901) Anopheles (Cellia) watsonii (Leicester 1908)

147

Appendix C

Chapter 4 supplementary figures and tables

Figure C.1. DCA from Figure 4.5, only showing species labels. See Table C.1 for full species names

148

Appendix C

Table C.1. List of full taxonomic names of collected species in Chapter 4 Species Mosquito genera and species abbreviation Am.orb Aedimorphus orbitae (Edwards 1922) An.Ait Anopheles (Anopheles) Aitkenii group (Reid & Knight 1961) An.bar Anopheles (Anopheles) barbirostris (van der Wulp 1884) An.sp Anopheles (Anopheles) sp. An.bal Anopheles (Cellia) balabacensis (Baisas 1936) An.lat Anopheles (Cellia) latens (Sallum & Peyton 2005) An.maca Anopheles (Cellia) macarthuri (Colless 1956) An.macu Anopheles (Cellia) maculatus (Theobald 1901) An.wat Anopheles (Cellia) watsonii (Leicester 1908) Arm.con Armigeres (Armigeres) confusus (Edwards 1915) Arm.jug Armigeres (Armigeres) jugraensis (Leicester 1908) Arm.sp Armigeres (Armigeres) sp. Coll.pse Collessius (Alloemyia) pseudotaeniatus (Giles 1901) Coq.cr Coquillettidia (Coquillettidia) crassipes (van der Wulp 1881) Cx.sit Culex (Culex) sitiens (Wiedemann 1828) Cx.vis Culex (Culex) vishnui (Theobald 1901) Cx.Lop Culex (Lophoceraomyia) sp. Do.gan Downsiomyia ganapathi (Colless 1958) Ph.pro Phagomyia prominens (Barraud 1923) Pr.ost Paraedes ostentatio (Leicester 1908) Stg.al Stegomyia albopicta (Skuse 1895) Stg.sp. Stegomyia sp. Ve,sp Verrallina sp.

149