MODIS-derived global land products of shortwave radiation and diffuse and total photosynthetically active radiation at 5 km resolution from 2000

Abstract

Incident shortwave radiation (SW), photosynthetically active radiation (PAR), and diffuse PAR (PAR(dif)) at the land surface drive a multitude of processes related to biosphere-atmosphere interactions and play a critical role in the Earth climate system. Previous global solar radiation products were spatially coarse (> 50-km resolution) or temporally short (a few years), which hindered scaling-up ground based observations of the land surface processes into regional, continental, and global scales across multiple time scales. Here, we report Breathing Earth System Simulator (BESS) SW, PAR, and PAR(dif) products over the global land surface at a 5 km resolution with 4 day intervals between 2000 and 2016. We combined an atmospheric radiative transfer model with an artificial neural network (ANN) to compute SW, PAR, and PAR(dif). A series of MODerate Resolution Imaging Spectroradiometer (MODIS) atmosphere and land products were used as inputs to run the ANN. We test the performance of the products using data from 158 (SW), 77 (PAR), and 22 (PARdif) stations collected in the Baseline Surface Radiation Network (BSRN) and flux tower networks, which covered a range of climatic zones from polar to tropical zones. BESS had strong linear relationships with in-situ SW data (R-2 = 0.95, relative bias = -2.3%), PAR (R-2 = 0.94, relative bias = 1.7%), and PAR hf (R-2 = 0.84, relative bias = 0.2%). BESS captured the interannual variability of SW at both the site (a majority of long-term BSRN sites) and continental levels. Over the study period, global annual SW; PAR, and PAR(dif) values did not show any dimming or brightening trends, although these trends appeared at regional levels, e.g. dimming in India. Mean annual SW over the global land surface was 184.8 W m(-2) (875 ZJ yr(-1), zetta = 10(21)); 46% of SW was partitioned to PAR, which was further split into direct (59%) and diffuse (41%) components. The developed products will be useful in solar energy harvesting research and will improve water, carbon, and energy flux estimates of terrestrial ecosystems from local to the global scales.