Figure 2.

Schematic diagram of the model and statistical analysis approach in this study. For each county, we selected daily outputs of disease-relevant hydroclimatic variables (related to air temperature T, precipitation P, and runoff R) for the grid cells of each global climate model (GCM) with at least 40% of their area located within each targeted county (35% for Örebro county in The Geophysical Fluid Dynamics Laboratory’s Earth System Model version 4 (GFDL-ESM4) and The Geophysical Fluid Dynamics Laboratory’s climate model version 4 (GFDL-CM4)). We spatially averaged the values of each hydroclimatic output variable of each GCM over these county grid cells to represent a corresponding county-average daily hydroclimatic variable value (air temperature T_c, precipitation P_c, and runoff R_c). Furthermore, for each GCM, we assessed its hydroclimatic projection implications for the future scenario evolution of disease cases in each county, based on a previously established county-relevant statistical disease model that we quantified by preceding-year number of disease cases and relevant GCM-projected county-average hydroclimatic variables (including summer temperature in the preceding year ST_lag, summer precipitation in the same year SP, and standardized relative annual mosquito abundance sRMA, which is in turn related to various river flow Q conditions). Finally, across all GCMs, we calculated the model ensemble mean and inter-model standard deviation (as a GCM uncertainty measure) of scenario-implied possible future evolution of disease cases in each county. SSP1-2.6: Shared Socioeconomic Pathways (SSP) with emissions driven by sustainable practices to produce radiative forcing of 2.6 Wm⁻² by 2100; BCC-CSM2-MR: The Beijing Climate Center Climate System Model, Version 2-MR; MRI-ESM2-0: The Meteorological Research Institute Earth System Model, Version 2.0; NorESM2-MM: The Norwegian Earth System Model, Version 2-MM.