This post is adapted from my final project submission for a class I took in Spring 2020, the term that major COVID disruption began. The class, ESPM 152: Global Change Biology, used a pre-publication version of this textbook. We developed our project proposals in multiple stages over the course of the semester and provided and responded to feedback on each other’s work. The project helped me learn about research funding processes, experimental design, peer review, and quantitative modeling of ecological systems.
Bats are aboundingly diverse, ecologically critical, and deeply threatened. More than 1,400 species of Chiroptera are recognized, and the order is, apart from Rodentia, the most diverse among mammals (York et al. 2019). In many tropical ecosystems, bats are the single most diverse group of mammals (Ghanem and Voigt 2012; Bass et al. 2010; Simmons and Voss 1998). From giant fruit bats to hematophagous vampire bats, Chiroptera display tremendous variety in morphology, roosting preferences, diet, and more. As insectivores, they regulate insect populations; as frugivores, they facilitate seed dispersal; as nectarivores, they provide pollination services. Through these mechanisms and others, bats contribute to all four categories of ecosystem services: provisioning, such as food production via pollination; regulating, such as insect pest control; supporting, such as seed dispersal; and cultural, such as aesthetic and spiritual value (Ghanem and Voigt 2012). Yet these services are not distributed evenly; factors that can affect bat species distributions include: increased use and toxicity of pesticides, which pose myriad problems for bats and their prey; wind turbines, which can precipitate bat mortality via blade action; infectious disease, especially white-nose syndrome, which drains resources and leads bats to prematurely exit hibernation; introduced predators, which prey on bats and their prey; hunting, which reduces population sizes; and habitat fragmentation and destruction, which reduces potential habitat (Ghanem and Voigt 2012, 283, 291, 294). These stressors tend to have decidedly negative impacts on bat populations, but that is not the case for another key stressor: climate change. One might posit that, due to altered weather patterns (which can disrupt physiological behaviors such as hibernation, reshape water regime, and reduce available habitat) and disrupted ecological networks (including changes in prey abundance), climate change would only negatively impact bats. While it’s true that existing literature predicts dramatic range reduction or shifts for certain species, it is also true that it predicts drastic expansion for others (Bilgin, Kesisoglu, and Rebelo 2012; Aguiar et al. 2016). Factors that may account for these differential outcomes include tolerance for arid environments, size of existing range, and specificity of roost preferences, among others (Bilgin, Kesisoglu, and Rebelo 2012).
A number of contemporary conservation efforts could significantly impact bat populations, especially with respect to habitat availability. Nature Needs Half and E.O. Wilson’s Half-Earth Project have called for the protection of 50% of the terrestrial biosphere (Dinerstein et al. 2017). In addition, greenhouse gas offset efforts may include widespread reforestation, and researchers have projected that low-cost carbon pricing could increase reforestation in tropical lands alone by 84.1 million hectares, or 207.8 million acres (Busch et al. 2019). Furthermore, conservationists are increasingly recognizing the problems posed by fragmented landscapes, and accordingly increasing connectivity through projects such as the Mesoamerican Biological Corridor (Vester et al. 2007). Due to this potential for significant near-term developments in conservation planning, and the important, threatened, and varied roles played by bats, it is critical to identify current species ranges and predict future changes.
A common method for predicting ranges is to generate species distribution models (SDMs), which use historical observational data as an input for modeling future scenarios. SDMs are best used for investigations of organisms that have a sizable body of recent observational locality records generated in a consistent method. This data is rare for many groups, such as microorganisms, but is readily available for many mammals. SDMs are classified as “presence-only” when the inputted locality data is solely composed of recorded presences, and not absences. Presence-only SDMs appear to be particularly effective for investigations of bats because they are nocturnal, because they are not easily directly observed, because they have relatively high dispersal capacities, and because it is difficult to definitively determine absence (Razgour et al. 2016). Additionally, bats have slow reproductive rates, and are thus thought to have limited adaptive potential (Aguiar et al. 2016). Therefore, their niches are likely to be conserved, and thus niche modeling is likely to have greater predictive accuracy.
Some projections of bat species distributions have been published, but there are compelling reasons to continue to explore this topic. Progress in computing technology and data digitization efforts has been rapid, so even studies from the recent past may benefit from revision. Bias has been more thoroughly analyzed, and advances have been made in statistical procedures that aim to address bias (Renner et al. 2015; Fisher-Phelps et al. 2017). Additionally, methods for incorporating non-climatic variables have become more common and accessible, resulting in more accurate, insightful results.
Notably, bat species distribution modeling has disproportionately represented Europe, with Central America among the least studied areas; out of 89 bat SDM studies published from 2001 to January 2016, 39 focused on Europe (44%) while 8 focused on Central America (9%) (Razgour et al. 2016).
I hope to focus my investigation on Central America. Despite its underrepresentation in published literature, Central America, composed of Belize, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and Panama, is considered a biodiversity conservation hotspot, meaning that it is both highly diverse and highly threatened (Chiabai 2015). It has been estimated that 1,159 vertebrates are endemic to Central America, representing 4.2% of global vertebrates, among the highest of any region worldwide (Myers et al. 2000, 2). Yet only 20% of the area of primary vegetation remains (ibid.). This is a result of various global change drivers, but especially habitat fragmentation and deforestation caused by agricultural expansion. This disruption continues in modern times; in fact, the average annual forest surface loss in Central America from 2000 to 2005 was around 1.3% (Chiabai 2015, 39).
There are 144 bat species known to exist in southeast Mexico and Central America (F. A. Reid 2009, 72), and 123 species are known to exist in Costa Rica and Nicaragua alone (York et al. 2019, 1). Nine families of Chiroptera are found in Central America, and they have striking differences; diets range from nectar to blood (F. A. Reid 2009, 72). Thus, modeling the future ranges of these species is no small task. However, it is likely that models would not ultimately be generated for each species that occurs in the region. Observational presence records are scarce for many species, and insufficient high-quality data could disqualify many species from modeling analysis. (Missing parameters, high spatial autocorrelation, and inconsistent methodology could all reduce the quality of data).
There’s another reason that the modeled species would be a subset of the total species. It isn’t necessarily insightful to conduct species distribution modeling within geopolitically-defined spatial parameters, because geopolitical boundaries don’t always overlap with ecological boundaries (Razgour et al. 2016, 4). Many Central American bat species (e.g., Ectophylla alba, Myotis cobanensis) are found exclusively in Central America, while others (e.g., Diclidurus albus, Lasiurus intermedius) have much broader ranges (F. Reid et al. 2016; Miller and C 2016; Rodriguez and Miller 2016; “The IUCN Red List” n.d.). This investigation would focus on species that are endemic to Central America, because those with broader ranges tend to be more generalist species, and thus more resilient in the face of global change. Identifying species endemic to the study area would therefore be a critical component of this investigation.
This investigation will aim to answer the question: how will near-term anthropogenic climate change impact the spatial distribution of bats endemic to Central America? The focus is on the near-term, as opposed to the long-term, because of the urgency of conservation planning and because there is greater certainty with near-term climate projections. (Large-scale climate action could dramatically alter long-term outcomes.) The emphasis on anthropogenic climate change, as opposed to other factors that may determine spatial distributions, is because of the variable outcomes reported in the existing literature and because of the strong likelihood of significant near-term change. The choice of spatial distribution was made because knowledge of organismal ranges can inform decisions related to land management, conservation efforts, agriculture, and more. The selection of Central America was made because of its underrepresentation in published SDM literature, its abundant biodiversity, and its threatened status.
In order to respond to this question, I will gather observed bat occurrence and climate data and use them to construct presence-only, point-process species distribution models. I will use these models and existing projected climate scenarios to generate possible bat species distributions. When possible, I will include relevant non-climatic variables in order to refine these possible ranges. In essence, by evaluating altered environmental conditions (the explanatory variable), I can predict changes in bats’ realized niches (the response variable).
I hypothesize that modeling will indicate that change in niche area will differ across species, both in sign and in magnitude, because, in very general terms, the expansion of hotter, drier environments will lead to range expansions and the decline in cooler, wetter environments will lead to range contractions. This picture is, of course, complicated by the many non-climatic variables that affect species distributions, but it is nonetheless a key trend. Furthermore, I anticipate an overall trend of a decrease in species richness because of the low adaptability and high niche specificity of many of the target species.
In order to generate a species distribution model, one must assemble data of current and historical bat presence records and climate metrics, as well as data for predicted future climate conditions. A potential source for climate data would be worldclim.org, which includes data for historical climate from the late 20th century onwards and projected climates based on models from the Intergovernmental Panel on Climate Change, or the IPCC (UN FAO n.d.). The projections of the IPCC provide a reasonable foundation for exploratory analysis because they reflect the views of a broad coalition of experts, rather than a lone viewpoint. One source for open-access bat locality data would be gbif.org, the Global Biodiversity Information Facility. The existing bat observational data has the advantage of being open access, but its scope is limited, so it would be advantageous to supplement it. Additional data could be procured from museum records and published literature. To access museum records, I would contact staff members at institutions with non-digitized collections known to hold specimens from the study area and request relevant materials. I would compile this data in a format compatible with the open-access data. To review published literature, I would utilize targeted queries on search engines such as Web of Science, record a list of relevant results, contact authors for raw data as necessary, and, after determining a consistent method, compile the observational data in a format compatible with the other sources. A final source of data would be field work, which would likely involve both ultrasound detection and visual identification (Ahlén and Baagøe 1999). In order to best focus this more expensive and labor-intensive stage, I would first exhaust the above options and identify areas with limited data, whether phylogenetically, spatially, or otherwise. For these areas with limited data, I would review available literature to determine best practices for field work, evaluate constraints (financial, temporal, legal), and develop a more comprehensive plan. This would allow me to more effectively allocate resources.
In addition to this extensive effort to assemble both climatic data and presence-only observational data, I would also seek to identify non-climatic variables that may determine the ecological niches of the study organisms, such as roost availability. This would primarily be accomplished through a literature review of the various species but could also be evaluated based on consultation with experts, study of existing field notes, or original field research. Having created this list of additional variables to consider, I would then seek out potential sources of spatially explicit data for these variables in the study area across a corresponding temporal scale. Where applicable, I would include them in the models. Even if it did not result in the inclusion of additional variables, researching non-climatic variables could help to determine functionally relevant climatic variables. (E.g., presence of adequate roosting sites may be linked to rainfall.)
Given the wide range of possible variables and the limitations in available data, I would consider, in addition to or instead of evaluating change for individual variables, evaluating change in entire ecoregions, as determined by the World Wildlife Foundation and more than 1,000 experts, available at https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world (Olson et al. 2001). These projected changes in ecoregions could be used as a proxy for a large number of ecological variables. In this scenario, I would use the climatic range of each ecoregion to project changes in its size and location. Ultimately, I would compare the projected changes in ecoregions to the projected changes in bat occurrences, and thus refine my predictions.
After gathering this data, I would perform some curation using R software. I would eliminate species with insufficient observations for model generation, possibly those with fewer than 25 presence records, and also species whose observed ranges fall substantially outside of the target region of Central America (using a numeric threshold to determine substantiality, perhaps 5% of observations). I would also evaluate the data for spatial sampling bias using a bias grid. With the data set narrowed down, I would select portions of the historical data to train models (determined randomly) and use the rest of the data to test these models. This would be accomplished using open-source MaxEnt (Maximum Entropy) software available through the American Museum of Natural History at https://biodiversityinformatics.amnh.org/open_source/maxent/, along with tutorials and best-practice guidelines. Specifically, I would seek to use point-process modeling, as discussed by Renner et al (2015). At that point, it would probably be necessary to eliminate some variables in order to minimize co-linearity and also to average multiple models in order to generate a less noisy result. Jackknife or bootstrap resampling methods could also be applied. Model accuracy could be evaluated by calculating the area under a receiver operating characteristic curve and comparing the values to those of null models as controls. After making these adjustments, I would apply these models to selected dates from the future climate condition data (e.g., 2050, 2080) in order to ascertain potential future distributions, then analyze the implications of my findings.
This investigation would advance understanding of spatiotemporal distributions of bats under predicted climate regimes and thus inform bat-related conservation decisions, such as selection of protected areas. Ultimately, Central American bat conservation is critical because these bats are critical to forest ecosystems, agriculture, and human health and culture.
Chiroptera are critical to the health of forests. Bats are recognized for providing pollination services, distributing seeds, and redistributing nutrients via guano. These services facilitate growth of new trees, which is important for long-term forest health in general but especially crucial for forest resiliency in the face of accelerating global change. Elevated seed dispersal distances, for example, can help forests shift ranges in response to changing temperatures. Consequences of diminished forests include, broadly, the deterioration of water quality, elevated soil erosion, more extreme drought and flood regimes, diminished carbon sequestration, and the elimination of provisioning, recreational, and cultural services (Chiabai 2015, 39–40).
Bats are key to agricultural productivity. Agriculture is reliant on steady water supplies, and thus reliant on forests, and thus reliant on bats. In addition, bats play crucial roles in suppressing populations of insects that can pose problems in agricultural settings, including Lepidoptera and Coleoptera (Kunz et al. 2011, 8). Furthermore, bats aid in dispersal for economically important crops such as cashews and coffee and pollination for crops such as agave and durian (Kunz et al. 2011, 16–19).
Bats are also important to human health and culture. Reliable sources of clean water are crucial to human health, and, as noted above, water supplies and forest health are deeply intertwined. Similarly, reliable sources of food are necessary for human survival, and bats have notable impacts on agriculture. Bats also reduce populations of insects that can be harmful to human health, including mosquitoes, which are vectors of ailments such as malaria, dengue, and West Nile Virus. In fact, +1 million people die from mosquito-borne diseases annually (Cuervo-Parra, Cortés, and Ramirez-Lepe 2016). As far as medicinal and cultural services, bats have been used in folk medicine since ancient times, and today the anticoagulant found in the saliva of Desmodus rotundus, the common vampire bat, is viewed as a potential treatment for strokes (Kunz et al. 2011, 23). Bats figure prominently in cultures both contemporary (Batman, Dracula, Halloween) and ancient (the Mayan bat god, Chinese bowl carvings, Egyptian wall paintings), and are central to cave visits, nocturnal tours, and educational programs (Kunz et al. 2011, 23). In these ways and many more, bats are critical to human health and wellness.
Aguiar, Ludmilla M. S., Enrico Bernard, Vivian Ribeiro, Ricardo B. Machado, and Gareth Jones. 2016. “Should I Stay or Should I Go? Climate Change Effects on the Future of Neotropical Savannah Bats.” Global Ecology and Conservation 5 (January): 22–33. https://doi.org/10.1016/j.gecco.2015.11.011.
Ahlén, Ingemar, and Hans J Baagøe. 1999. “Use of Ultrasound Detectors for Bat Studies in Europe: Experiences from Field Identification, Surveys, and Monitoring,” 137–50.
Bass, Margot S., Matt Finer, Clinton N. Jenkins, Holger Kreft, Diego F. Cisneros-Heredia, Shawn F. McCracken, Nigel C. A. Pitman, et al. 2010. “Global Conservation Significance of Ecuador’s Yasuní National Park.” PLOS ONE 5 (1): e8767. https://doi.org/10.1371/journal.pone.0008767.
Bilgin, Rasit, Ari Kesisoglu, and Hugo Rebelo. 2012. “Distribution Patterns of Bats in the Eastern Mediterranean Region Through a Climate Change Perspective.” ResearchGate. 2012. https://dx.doi.org/10.3161/150811012X661747.
Busch, Jonah, Jens Engelmann, Susan C. Cook-Patton, Bronson W. Griscom, Timm Kroeger, Hugh Possingham, and Priya Shyamsundar. 2019. “Potential for Low-Cost Carbon Dioxide Removal through Tropical Reforestation.” Nature Climate Change 9 (6): 463–66. https://doi.org/10.1038/s41558-019-0485-x.
Chiabai, Aline. 2015. Climate Change Impacts on Tropical Forests in Central America: An Ecosystem Service Perspective. First edition. Routledge. https://doi.org/10.4324/9781315866703.
Cuervo-Parra, Jaime A., Teresa Romero Cortés, and Mario Ramirez-Lepe. 2016. “Mosquito-Borne Diseases, Pesticides Used for Mosquito Control, and Development of Resistance to Insecticides.” In Insecticides Resistance, edited by Stanislav Trdan. InTech. https://doi.org/10.5772/61510.
Dinerstein, Eric, David Olson, Anup Joshi, Carly Vynne, Neil D. Burgess, Eric Wikramanayake, Nathan Hahn, et al. 2017. “An Ecoregion-Based Approach to Protecting Half the Terrestrial Realm.” BioScience 67 (6): 534–45. https://doi.org/10.1093/biosci/bix014.
Fisher-Phelps, Marina, Guofeng Cao, Rebecca M. Wilson, and Tigga Kingston. 2017. “Protecting Bias: Across Time and Ecology, Open-Source Bat Locality Data Are Heavily Biased by Distance to Protected Area.” Ecological Informatics 40: 22–34. https://doi.org/10.1016/j.ecoinf.2017.05.003.
Ghanem, Simon J., and Christian C. Voigt. 2012. “Increasing Awareness of Ecosystem Services Provided by Bats.” In Advances in the Study of Behavior, 279–302. Elsevier. https://doi.org/10.1016/b978-0-12-394288-3.00007-1.
Kunz, Thomas H., Elizabeth Braun De Torrez, Dana Bauer, Tatyana Lobova, and Theodore H. Fleming. 2011. “Ecosystem Services Provided by Bats.” Annals of the New York Academy of Sciences 1223 (1): 1–38. https://doi.org/10.1111/j.1749-6632.2011.06004.x.
Miller, Bruce W., and Jose O. Cajas C. 2016. “IUCN Red List of Threatened Species: Myotis cobanensis.” IUCN Red List of Threatened Species. June 30, 2016. https://www.iucnredlist.org/species/14154/22058031.
Myers, Norman, Russell A. Mittermeier, Cristina G. Mittermeier, Gustavo A. B. Da Fonseca, and Jennifer Kent. 2000. “Biodiversity Hotspots for Conservation Priorities.” Nature 403 (6772): 853–58. https://doi.org/10.1038/35002501.
Olson, David M., Eric Dinerstein, Eric D. Wikramanayake, Neil D. Burgess, George V. N. Powell, Emma C. Underwood, Jennifer A. D’amico, et al. 2001. “Terrestrial Ecoregions of the World: A New Map of Life on Earth.” BioScience 51 (11): 933. https://doi.org/10.1641/0006-3568(2001)051[0933:TEOTWA]2.0.CO;2.
Razgour, Orly, Hugo Rebelo, Mirko Di Febbraro, and Danilo Russo. 2016. “Painting Maps with Bats: Species Distribution Modelling in Bat Research and Conservation.” Hystrix, the Italian Journal of Mammalogy 27 (1). https://doi.org/10.4404/hystrix-27.1-11753.
Reid, Fiona A. 2009. A Field Guide to the Mammals of Central America and Southeast Mexico. Second edition. Oxford ; New York: Oxford University Press.
Reid, Fiona, Bruce W. Miller, Burton Lim, A. D. Cuarón, P. de Grammont, and Joaquin Arroyo-Cabrales. 2016. “IUCN Red List of Threatened Species: Diclidurus albus.” IUCN Red List of Threatened Species. July 1, 2016. https://www.iucnredlist.org/species/6561/21986615.
Renner, Ian W., Jane Elith, Adrian Baddeley, William Fithian, Trevor Hastie, Steven J. Phillips, Gordana Popovic, and David I. Warton. 2015. “Point Process Models for Presence-Only Analysis.” Methods in Ecology and Evolution 6 (4): 366–79. https://doi.org/10.1111/2041-210X.12352.
Rodriguez, Bernal, and Bruce W. Miller. 2016. “IUCN Red List of Threatened Species: Northern Yellow Bat.” IUCN Red List of Threatened Species. August 8, 2016. https://www.iucnredlist.org/species/11352/115101697.
Simmons, Nancy B., and Robert S. Voss. 1998. “The Mammals of Paracou, French Guiana, a Neotropical Lowland Rainforest Fauna. Part 1, Bats. Bulletin of the AMNH ; No. 237.” https://digitallibrary.amnh.org/handle/2246/1634.
“The IUCN Red List.” n.d. IUCN Red List of Threatened Species. Accessed March 26, 2020. https://www.iucnredlist.org/en.
UN FAO. n.d. “WorldClim - Global Climate Data (WORLDCLIM).” Accessed March 24, 2020. https://www.fao.org/land-water/land/land-governance/land-resources-planning-toolbox/category/details/en/c/1043064/.
Vester, Henricus F. M., Deborah Lawrence, J. Ronald Eastman, B. L. Turner, Sophie Calmé, Rebecca Dickson, Carmen Pozo, and Florencia Sangermano. 2007. “LAND CHANGE IN THE SOUTHERN YUCATÁN AND CALAKMUL BIOSPHERE RESERVE: EFFECTS ON HABITAT AND BIODIVERSITY.” Ecological Applications 17 (4): 989–1003. https://doi.org/10.1890/05-1106.
York, Heather A., Bernal Rodríguez-Herrera, Richard K. Laval, Robert M. Timm, and Kaitlin E. Lindsay. 2019. “Field Key to the Bats of Costa Rica and Nicaragua.” Journal of Mammalogy 100 (6): 1726–49. https://doi.org/10.1093/jmammal/gyz150.