In Haiti’s Cities, People, Heat and Greenery Shape Which Mosquitoes Thrive

In the densely populated Ouest Department of Haiti, the mosquitoes that carry dengue, malaria, and lymphatic filariasis do not distribute themselves randomly across the landscape. A new study published in the journal Parasites & Vectors has mapped the biodiversity of mosquito communities in one of the country’s most urbanized regions and found that human population density, temperature, and vegetation indices are among the strongest predictors of how diverse—and how dangerous—local mosquito assemblages become. The research, led by Ian A. Pshea-Smith of the University of Florida together with collaborators from Haiti’s Laboratoire National de la Sante Publique, the Uniformed Services University, and several U.S. Navy and Army research units, offers one of the most detailed ecological portraits to date of the vector communities circulating in a nation where mosquito-borne disease remains a persistent public health challenge.

The scale of the field effort was considerable. Between August 2018 and September 2019, the research team deployed three complementary trap types—Biogents Sentinel traps, CDC Gravid traps, and CDC Light Traps—at 19 study locations across the Ouest Department. Over the thirteen-month sampling window, they captured a total of 22,504 mosquitoes. Two species dominated the catch: Culex quinquefasciatus, which accounted for 53.4 percent of all specimens, and Aedes aegypti, which made up 16.9 percent. These are not incidental species. Cx. quinquefasciatus is the principal vector of lymphatic filariasis in Haiti, while Ae. aegypti transmits the arboviruses responsible for dengue, Zika, chikungunya, and yellow fever. Their numerical dominance underscores how heavily the region’s disease burden is concentrated in a small number of highly adapted urban vectors.

Alongside the two dominant species, the collections included Anopheles albimanus, the country’s key malaria vector, and a broader cast of secondary species, including Psorophora columbiae. All mosquitoes were identified morphologically, and the team then quantified biodiversity using a standard ecological toolkit: species richness measures of alpha diversity, Shannon’s index, and Simpson’s index. Each of these metrics captures a slightly different aspect of community structure—raw species counts, the balance between richness and evenness, and the probability that two randomly drawn individuals belong to the same species, respectively. Applying all three allowed the researchers to distinguish sites that merely hosted many species from those that supported genuinely balanced, multi-species communities.

A crucial methodological finding emerged from this diversity assessment: the metrics varied significantly both by trap type and by month. No single trapping method painted a complete picture of the mosquito community. Biogents Sentinel traps, which target host-seeking container-breeding Aedes, CDC Gravid traps, which attract egg-laying females enriched for Culex, and CDC Light Traps, which sample nocturnal flying insects broadly, each bias collections toward particular behaviors and species. The implication for surveillance programs is direct—reliance on a single trap design risks systematically undercounting the very species that matter most for a given disease system, and the authors emphasize the need for diversified trapping methods in future monitoring efforts.

To move beyond description and identify the environmental forces structuring these communities, the team turned to boosted regression trees, a machine-learning technique increasingly favored in disease ecology for its ability to model nonlinear relationships and interactions among many predictors simultaneously. The models incorporated both static and temporally varying covariates, including temperature, precipitation, vegetation indices measured through the normalized difference vegetation index (NDVI), and human population density. By relating the diversity indices computed for each collection event to these environmental layers, the BRTs could rank the relative influence of each factor and reveal the shapes of their effects—something classical linear models often obscure.

The results were striking in their consistency. Human population density, temperature, and vegetation indices emerged as the key predictors of mosquito diversity across the study area. This triad tells an ecologically coherent story. Dense human settlement creates the artificial containers, storm drains, and water-holding debris that container-breeding Aedes require, while also supplying the blood meals that sustain high vector populations. Temperature governs mosquito development rates, viral replication within vectors, and seasonal activity windows. Vegetation, as indexed by NDVI, shapes the availability of shaded resting sites, organic-rich larval habitats, and the microclimates that favor some species over others. In a highly urbanized tropical setting, these three forces interact to determine which mosquito species coexist at any given site and in any given month.

Perhaps the most novel component of the study was its use of species co-occurrence networks to describe the relationships among mosquito species. The researchers constructed two types of networks—one based on a cross-product approach and one on a probabilistic framework—to test whether pairs of species occurred together more or less often than expected by chance. The networks revealed significant positive associations between Aedes species and Cx. quinquefasciatus, suggesting that the vectors of arboviruses and lymphatic filariasis frequently share the same habitats and may respond similarly to urban environmental conditions. For control programs, this is a double-edged finding: interventions targeting shared breeding habitats could suppress multiple disease vectors simultaneously, but conditions that favor one dominant species may also be quietly building populations of the others.

Not all relationships in the networks were positive. Psorophora columbiae exhibited negative associations with multiple other species, standing out as the community’s apparent contrarian. Negative co-occurrence patterns can arise from several mechanisms, including competition for larval resources, differing habitat preferences, or predation, although the study’s observational design means the underlying cause cannot be pinned down with certainty. What the pattern does establish is that inter-species dynamics are a real and measurable feature of Haiti’s mosquito communities, not a statistical artifact, and that single-species control strategies may produce ecological ripple effects that are currently poorly understood.

The public health stakes of this work are substantial. Haiti continues to confront endemic malaria transmitted by An. albimanus, recurring dengue outbreaks driven by Ae. aegypti, and lymphatic filariasis sustained by Cx. quinquefasciatus. Ae. albopictus, the Asian tiger mosquito, is also present as an arbovirus vector. By identifying the environmental conditions under which these vectors co-occur, the study provides a foundation for multi-species control measures—interventions designed not against one target insect but against the ecological communities that collectively transmit the country’s major vector-borne diseases. The finding that human population density is a central driver also highlights how urbanization itself functions as a disease ecology variable, concentrating vectors and humans in the same environmental envelope.

The authors are careful to frame their conclusions as a beginning rather than an endpoint. They call for further research on inter-species dynamics to build a more unified understanding of vector ecology in Haiti, and their study design—combining year-round field surveillance, multiple trap types, rigorous biodiversity metrics, machine-learning environmental modeling, and network analysis—offers a template that other resource-limited settings could adapt. The work was funded by the Armed Forces Health Surveillance Branch’s Global Emerging Infections Surveillance section and through a Battelle Memorial Institute contract supporting Navy force health protection efforts, reflecting the strategic importance placed on vector surveillance in the Caribbean region. As climate change alters temperature regimes and urbanization accelerates across the tropics, the environmental levers identified in this study—population density, heat, and vegetation—will only grow in importance for predicting where mosquito biodiversity, and the diseases it carries, will flourish next.

Subject of Research: Environmental and demographic drivers of mosquito biodiversity and species co-occurrence in urban Haiti

Article Title: Vegetation, temperature, and human population density influence mosquito biodiversity in Haiti

Article References: Pshea-Smith, I. A., Okech, B. A., Boncy, J., Existe, A., So, J., Hamilton, T., Blanton, J., Matulis, G. A., Metrailer, M. C., Cleary, N. G., Dunford, J. C., Koehler, J. W., Blackburn, J. K., & von Fricken, M. E. (2026). Vegetation, temperature, and human population density influence mosquito biodiversity in Haiti. Parasites & Vectors. https://doi.org/10.1186/s13071-026-07724-3

Image Credits: AI Generated

DOI: 10.1186/s13071-026-07724-3

Keywords: mosquito biodiversity, Haiti, Aedes aegypti, Culex quinquefasciatus, boosted regression trees, species co-occurrence networks, vector ecology, NDVI, urban ecology, dengue, malaria, lymphatic filariasis