Climate Change Arrives in the Extremes, Not the Averages: Implications for Assessing Climate Envelopes of Vector-Borne and Environmental Infectious Diseases

Abstract:

Geohealth researchers wishing to assess the impacts of current and future climate on vector-borne and environmentally-mediated infectious diseases must choose between various forms of climate data for empirical and modeling analyses. For studies with a broad geographical domain, such as those wishing to map the changing spatial extent of diseases or vectors as a result of shifting climate envelopes, there are important implications of selecting average data (e.g., mean monthly temperature) that are more commonly available versus data on extremes (e.g., daily minimum or maximum temperature) that are less commonly available. This situation is especially true for gridded climate data products covering broad areas such as the United States or the entire globe. Yet, the limiting climatic factors for pathogens and vectors and their disease ecologies are more directly associated with changing climate extremes instead, that in turn relate to shifting seasonality, high/low tolerance, threshold behavior (e.g., growing degree days) and more. While changes in extremes can be observed in the shifting climate means, those means are obviously modulated by the averaging process and they may therefore not resolve critical environmental changes that dictate disease behavior. Climate changes can take the form of simple shifts in the distribution such as a change in the mean with no change in the variance, or as more complex combinations of variance change with or without a change in the mean. This paper models a range of permutations for different types of climate changes and their impacts on a representative set of pathogen and vector climatic constraints, identifying important implications for studies assessing spatial and temporal climate envelopes of vector-borne and environmental infectious diseases. The results identify combinations of means and extremes that have greater or lesser reliability in their representation of underlying pathogen or vector response and their associated climate envelopes, and they pinpoint pitfalls and best practices for geohealth climate envelope studies.