Expansion of Coccidioidomycosis Endemic Regions in the United States in Response to Climate Change

Abstract

Coccidioidomycosis (Valley fever) is a fungal disease endemic to the southwestern United States. Across this region, temperature and precipitation influence the extent of the endemic region and number of Valley fever cases. Climate projections for the western United States indicate that temperatures will increase and precipitation patterns will shift, which may alter disease dynamics. We estimated the area potentially endemic to Valley fever using a climate niche model derived from contemporary climate and disease incidence data. We then used our model with projections of climate from Earth system models to assess how endemic areas will change during the 21st century. By 2100 in a high warming scenario, our model predicts that the area of climate-limited endemicity will more than double, the number of affected states will increase from 12 to 17, and the number of Valley fever cases will increase by 50%. The Valley fever endemic region will expand north into dry western states, including Idaho, Wyoming, Montana, Nebraska, South Dakota, and North Dakota. Precipitation will limit the disease from spreading into states farther east and along the central and northern Pacific coast. This is the first quantitative estimate of how climate change may influence Valley fever in the United States. Our predictive model of Valley fever endemicity may provide guidance to public health officials to establish disease surveillance programs and design mitigation efforts to limit the impacts of this disease.

Keywords: Coccidioidomycosis; Valley fever; human health; infectious diseases; mycoses; niche model.

Figure 1

Valley fever incidence for counties in the southwestern United States (n = 152) as a function of (a) mean annual temperature and (b) mean annual precipitation. All counties that have endemic levels of Valley fever incidence (defined as meeting or exceeding 10 or more cases per 100,000 population per year during 2000–2015; n = 23) have a mean annual temperature greater than or equal to 10.7°C and a mean annual precipitation level less than or equal to 600 mm/year. (c) Counties with higher levels of mean annual Valley fever incidence are concurrently hotter and drier. We adapted panels a and b of this figure from Gorris et al. (2018) and added the gray lines to indicate the position of the climate thresholds we used to build our climate‐constrained niche model.

Figure 2

Representative concentration pathway (RCP) 8.5 climate projections indicate warming throughout the contiguous United States with the highest levels occurring in northern states (a–c). Changes in precipitation will vary by region (d–f). RCP8.5 projections indicate drying in the southwestern United States and south‐central Great Plains and wetting across the Pacific Northwest and eastern United States. The difference panels (c and f) are the difference between the 2095 and 2007 maps for each climate variable.

Figure 3

(a) Counties our climate‐constrained niche model identify as endemic (with a mean annual temperature greater than or equal to 10.7°C and a mean annual precipitation level less than or equal to 600 mm/year) are colored in magenta. (b) There is reasonable agreement between this set of counties and the endemic region identified by the CDC. Counties shown in red in panel a have a mean annual temperature greater than or equal to 10.7°C but unsuitable mean annual precipitation (greater than 600 mm/year). Counties shown in blue have a mean annual precipitation level less than or equal to 600 mm/year but unsuitable mean annual temperature (less than 10.7°C). Counties in white our model defines as unsuitable according to both thresholds.

Figure 4

For the representative concentration pathway (RCP) 8.5 climate change scenario, areas where climate will permit Valley fever endemicity are shown for the years (a) 2035, (b) 2065, and (c) 2095. Areas where mean annual temperature will permit endemicity are shown in red, areas where mean annual precipitation will permit endemicity are shown in blue, and areas where both temperature and precipitation will permit endemicity are shown in magenta, following the color scheme used in Figure 3. The area endemic to Valley fever will extend farther north in future decades, especially in the rain shadows of the Sierra Nevada and Rocky Mountain Ranges. Precipitation will play a key role in determining which areas become endemic through time, as greater rainfall and moisture availability will limit the eastward extent of Valley fever as well as its presence in the Pacific Northwest and in western counties at higher elevations.

Figure 5

Time series of change in (a) the total area potentially endemic to Valley fever, (b) the number of endemic states, (c) the number of endemic counties, (d) the number of people living within endemic regions, and (e) the estimated number of annual cases from 2007 to 2095 for both representative concentration pathway (RCP) 8.5 and RCP4.5 climate scenarios. The shaded areas are the standard deviation describing variation among the 30 Coupled Model Intercomparison Project Phase 5 (CMIP5) Earth system models used in our analyses.

Figure 6

There is strong model agreement throughout the majority of the area we estimate as endemic to Valley fever for the RCP8.5 climate scenario in years (a) 2035, (b) 2065, and (c) 2095. The model agreement shows a measure of uncertainty for the counties along the edge of the endemic area. Some models predict that the endemic range in 2095 will expand into counties as far east as western Minnesota. Percent model agreement was calculated as the number of individual CMIP5 models that predict the county will have a climate that permits endemicity, divided by the total number of models (n = 30), as projected by the climate‐constrained niche model.

Figure 7

We estimated an upper bound of future Valley fever incidence using a 90th percentile regression model for (a) our 2007 baseline period, (b) 2035, (c) 2065, and (d) 2095 for representative concentration pathway (RCP) 8.5. Over time, our model predicts Valley fever incidence will increase throughout the extreme southwestern US and the southern Great Plains. Incidence will also increase throughout the Central Valley of California and in the northwestern United States.