The Aedes aegypti mosquito is the vector for four infectious diseases of global concern – Yellow Fever, Dengue, Chikungunya, and Zikavirus. Previous attempts to model the expansion of the vector habitat due to global climate change have rarely included characteristics related to the human populations on which this mosquito is dependent. The purpose of this research was to determine whether the inclusion of human population density improves model performance while creating risk maps that can be used to determine where humans are most likely to be exposed to the vector in the future. The resulting model demonstrated that the inclusion of human population density improves the predictive power for A. aegypti and should be considered during model development. Maps produced by the model were also suitable for identifying regions where human populations are most likely to experience increased risk. Finally, two areas at risk of expansion were highlighted as a case study in pairing risk mapping with evidence-based intervention strategies to identify sites that would benefit from mosquito-control efforts. In this case, a low-cost program of insecticide-treated covers for water storage containers would likely work well in both Minas Gerais, Brazil and Northwestern Province, Zambia to mitigate mosquito risk. This research demonstrates that human population characteristic not only improve model fit but also increase the extent to which risk maps are actionable by aiding in targeting interventions.

The Important Bird Areas (IBAs) network of BirdLife International aims to identify sites that are essential for the long-term conservation of the world’s avifauna. A number of global change events have the potential to negatively affect, either directly or indirectly, most bird species, biodiversity in general and associated ecological processes in these areas identified as IBAs. To assist conservation decisions, I assessed a suite of ten landscape scale anthropogenic pressures to 115 Important Bird Areas (IBAs) in South Africa, both those currently placing pressures on IBAs and those that constitute likely future vulnerability to transformation. These threats are combined with irreplaceability, a frequently used measure of conservation importance, to identify the suite of IBAs which are high priority sites for conservation interventions: those with high irreplaceability and are highly vulnerable to anthropogenic threats. A total of 22 (19%) of the South African IBAs are highly irreplaceable and are highly vulnerable to at least some of the pressures assessed. Afforestation, current and potential future patterns of alien plant invasions affect the largest number of highly irreplaceable IBAs. Only 9% of the area of highly irreplaceable IBAs is formally protected. A total of 81 IBAs (71%) are less than 5% degraded or transformed. This result, together with seven highly irreplaceable IBAs found outside of formally protected areas with lower human densities than expected by chance provides an ideal opportunity for conservation interventions. However, all the pressures assessed vary geographically, with no discernible systematic pattern that might assist conservation managers to design effective regional interventions. Furthermore, I used the newly emerging technique of ensemble forecasting to assess the impact of climate change on endemic birds in relation to the IBAs network. I used 50 endemic species, eight bioclimatic envelope models, four climate change models and two methods of transformation to presence or absence, which essentially creates 2400 projections for the years 2070-2100. The consensual projection shows that climate change impacts are very likely to be severe. The majority of species (62%) lose climatically suitable space and 99% of grid cells show species turnover. Five species lose at least 85% of climatically suitable space. The current locations of the South African Important Bird Areas network is very likely ineffective to conserve endemic birds under climate change along a “business a usual” emissions scenario. Many IBAs show species loss (41%; 47 IBAs) and species turnover (77%; 95 IBAs). However, an irreplaceability analysis identified mountainous regions in South Africa as irreplaceable refugia for endemic species, and some of these regions are existing IBAs. These IBAs should receive renewed conservation attention, as they have the potential to substantially contribute to a flexible conservation network under realistic scenarios of climate change. Considering all the global change threats assessed in this study, the Amersfoort-Bethal-Carolina District and the Grassland Biosphere Reserve (IBA codes: SA018; SA020) are the key IBAs in South Africa for conservation prioritisation. / Dissertation (MSc)--University of Pretoria, 2008. / Zoology and Entomology / unrestricted

Introduction: The Aedes aegypti mosquito is responsible for the transmission of Yellow Fever, Dengue, Chikungunya and Zikavirus, making it a deadly vector and global public health threat. Zikavirus and Chikungunya, which were previously restricted to smaller geographic areas, have both appeared in the Western Hemisphere in the past three years and spread to areas where A. aegypti are present. This means that the pathogens have now entered areas in which the population has no previous immunity, which can lead to extensive outbreaks and epidemics. As the effects of global climate change become apparent, the areas of the globe that are suitable for inhabitance by A. aegypti may change. Additionally, this vector prefers human hosts for blood meals and requires standing water to breed, which is often created by water storage containers. This means that increasing urbanization and human population density are likely to put populations at higher risk of exposure to this vector. Methods: To create maps of the future risk of exposure to Aedes aegypti globally, species occurrence data for the vector and the Maxent modeling approach were used. Current and projected climate data were downloaded from WorldClim.org for the four representative concentration pathways (RCPs) used to model future climate change. Human population density, projected to 2050, the same timeframe as the future climate data, were used to model changes in human populations. To identify areas at high risk for future presence of A. aegypti populations, current and future models were compared across areas with at least a 50% probability of increased risk. These results where then used to create maps displaying high risk areas. Results: The AUC, an indicator of model fit, signaled that the models had high predictive power. However, high omission rates indicated that the trade-off of risk mapping may be a need to decrease probability thresholds below 50% to capture the full at-risk population. Future high-risk areas were most often those surrounding current cities, which supports the idea that the combination of urbanization and increasing human population density will work synergistically to increase the disease burden within and around urban centers. Additionally, expansion at the current geographic margins of this species shows that incursion into currently non-endemic areas is possible. Conclusions: Urban and peri-urban populations are likely to be at higher risk of exposure compared to rural areas due to global climate change and changes in population density. Attempts to model expansion of vector habitats should consider how these human population characteristics will change the risk to populations and how to best identify the areas at highest risk. Thresholds for the probability of a population being at risk of exposure to a vector may need to be different from those required to determine whether or not a habitat is suitable for a species. Appropriately determining which areas are high-risk results in maps and models can then be used to identify areas where climate change mitigation and vector control efforts are likely to have the highest impacts.

The Aedes aegypti mosquito is responsible for transmission of four vector-borne diseases that cause considerable global morbidity and mortality. Projections of the future effects of global climate change indicate that expansion of this species due to changing habitats is possible. Furthermore, since A. aegypti is highly dependent on human populations for feeding and egg-laying sites, changing human population characteristics are likely to alter the risk of exposure for humans based on geographic location. This study aims to create future potential risk maps for human exposure to A. aegypti using human population density as a predictor. Using current population density data and future growth trajectories, high-resolution human population density forecasts were created for 2050, then included as variables in ecological niche models developed using Maxent. Species occurrence data and high resolution climate data for current and future conditions (best and worst case scenarios) were included in the model, as well. Model fit indices and variable contributions indicated that the inclusion of human population density improves model accuracy for A. aegypti. Risk maps created by these models showed that areas currently adjacent to large cities within endemic regions, such as central Africa and western Brazil, are likely to see the greatest increase in risk to human populations. This corroborates current projections on increasing urbanization in the future and suggests that these models can be used to target interventions in high risk areas.