Developing and Applying a Unified Weather and Climate Database to Assess Climate Change Impacts on Tropical Infectious Disease Transmission and Burden

Research at the intersection of climate, weather, and health is rapidly expanding and inherently interdisciplinary, requiring integration of information across multiple disciplines. This includes comprehensive, accessible, reliable, and harmonized datasets that combine high-quality observational data with bias-corrected and downscaled climate projections from Global Climate Models (GCMs), such as from the sixth phase of the Coupled Model Intercomparison Project (CMIP6). However, despite the availability of numerous gridded observational datasets and pre-processed projections, individual products vary in strengths, limitations, and representations of fine-scale spatiotemporal patterns, which can substantially affect downstream modelling and projection of current and future health outcomes. Moreover, the operational scale of epidemiological analysis is typically defined by administrative units, rather than by regular grids, and therefore often relies on the inclusion of area-level estimates that are additionally weighted by indicators such as human population. Hence, spatially resolved weather and climate data, typically provided in specialized formats (e.g., NetCDF), generally require substantial preprocessing before they can be used for respective analysis.

To address these challenges, we developed a tailored, quasi-global weather and climate dataset designed to support high-resolution infectious disease transmission modelling in tropical settings. Our dataset comprises (1) high-resolution (0.1°) daily climate projections between 60°N and 60°S, and (2) corresponding spatially averaged (population-weighted) area-level estimates at administrative unit levels 0-2 for over 100 countries. We therefore selected and evaluated multiple global observational datasets, including model- and satellite-based products such as ERA5 and CHIRPS, across heterogeneous, disease-relevant tropical study domains. The observational datasets showing the highest performance in our comparative analysis served as reference climatologies for generating high-resolution, bias-corrected climate projections downscaled from six CMIP6 GCMs, focusing on two scenarios from the Shared Socioeconomic Pathways-Representative Concentration Pathway (SSP-RCP) framework: SSP2-4.5 and SSP5-8.5.

For the first time, we hence provide a robust, open-access resource that combines observational datasets and bias-corrected, downscaled climate projections in a coherent manner and translates them into harmonized, spatially aggregated variables suitable for easy use by non-specialists from various disciplines. As an example application, we present the impact of climate change and the sensitivity of administrative-level vector-borne disease transmission risk in South America to the choice of global climate model and emissions scenario. We focus on yellow fever, a vaccine-preventable zoonotic arbovirus endemic to tropical regions of South America and Africa. We anticipate that our unified weather and climate database will be particularly valuable to infectious disease modelers, epidemiologists, and practitioners conducting climate-sensitive health impact assessments.