Integrating Topographic Effects into Global Downward Shortwave Radiation

Downward shortwave radiation (DSR) is the primary energy source driving Earth’s climate, hydrological, and ecological processes. While mountains occupy approximately 24% of the global land surface and exhibit complex radiative transfer processes, current global climate models and satellite products predominantly rely on plane-parallel assumptions, thereby neglecting topographic effects such as shadowing and terrain-reflected radiation.

To quantify these impacts, we developed a hybrid physical and data-driven method to generate the global, daily DSR product at 0.05° resolution, incorporating topographic effects, spanning from 2003 to 2024. By integrating a mountainous radiative transfer scheme with machine learning, we successfully captured the spatiotemporal heterogeneity of DSR over rugged terrain. Our analysis reveals that ignoring topographic effects results in substantial uncertainties across scales (from daily to annually and grid to global scales). In rough terrain hotspots, such as High Mountain Asia, the annual mean bias exceeds 30 W/m² (>20%). The slope-dependent uncertainties in the original DSR product were substantially reduced in the new DSR product with topographic considerations, i.e., the RMSE decreased globally from 21.7 to 2.2 W/m² in areas with slopes exceeding 25°. The topographically corrected DSR better explains the spatial heterogeneity of land surface temperature variations across the terrains.

These findings suggest that topography acts as a critical modulator of the Earth system's energy flow. The uncertainties of DSR in mountainous areas imply propagated biases in simulations of the cryosphere (snowmelt), carbon cycle (gross primary productivity), and hydrological processes. We underscore the necessity of integrating topographic considerations to improve the understanding of climate mechanisms in vulnerable mountain ecosystems.