Fine-scale wetland characteristics modulate vector-specific and climate-dependent macroscale patterns in mosquito-borne disease incidence

Background/Question/Methods

Identifying linkages between fine and macroscale patterns and processes can help facilitate the development of a more predictive and mechanistic understanding of ecological systems. Using West Nile virus (WNV) as a model system, we evaluate the potential links between scales to better understand the determinants of seemingly erratic human outbreak patterns. WNV is the most prevalent mosquito-borne disease in North America and is an ideal model system due to its extensive geographic range, ubiquitous yet variable temporal impact, and well-studied ecology at multiple scales. We quantify and evaluate the role of three key fine-scale wetland characteristics – wetland size, connectivity, and inundation regime – in driving macroscale human infection trends over 11 years and 1,600,000 km². To accomplish this, we have derived new macroscale wetland metrics based on the fine-scale wetland characteristics discussed above, and a new climate metric describing inter-year drought conditions. We then tested for spatial and temporal associations between our drought metric, wetland metrics (average wetland size, the proportional area of connected wetlands, and the proportional area of semi-permanent wetlands), and annual human WNV incidence.

Results/Conclusions

We found that, in regions dominated by the Culex tarsalis WNV vector, counties with a small average wetland size had more than 100% higher human WNV incidence than counties with a large average wetland size. Additionally, in regions dominated by the Culex pipiens vector, counties with a low proportional area of connected wetland had at least 50% higher human WNV incidence than counties with a high proportional area of connected wetland. Finally, Cx. tarsalis-dominated counties with a high proportional area of semi-permanent wetland and experiencing drought conditions had over 300% higher annual WNV incidence than drought-affected counties with a low proportional area of semi-permanent wetland. These counties also had more than 150% higher incidence than counties with a high proportional area of semi-permanent wetland but not experiencing drought conditions. Overall, our results provide convincing evidence that local processes aggregate to influence macroscale phenomena, but also demonstrate that the linkage between scales is buttressed by the complex interplay of tangential ecological factors like species-specific traits and climate.