Mutual-information of meteorological-soil and spatial propagation: Agricultural drought assessment based on network science

• A framework for assessing agricultural drought which combines information entropy and network science is proposed. • The vapor pressure deficit played a key to the information on drought propagation. • A highly clustered connectivity structure derived from the drought propagation network is found. • The mutual-information between meteorological and hydrological under high-emission scenarios increased. Agricultural drought causes serious harm to food security and vegetation ecosystems, and challenges the entire food-water-energy-health-environment nexus, complicating, in turn, management strategies. By building a meteorological and soil mutual-information index, this paper developed a novel method for assessing agricultural drought from a network science perspective. The method identified the atmospheric vapor pressure deficit as the key to the information on drought propagation by the network of meteorological and hydrological elements. Results showed that the meteorological and soil mutual-information index was superior to the standardized soil moisture index in monitoring agricultural drought. The spatial propagation of information on drought led to the key source areas for the occurrence of drought. A highly clustered connectivity structure derived from the drought propagation network was demonstrated in China. Under climate scenario SSP1-26, the increase in water vapor pressure deficit was significantly suppressed, thus mitigating agricultural (and/or vegetation) drought. Under high emission scenarios SSP2-45 and SSP5-85, the mutual-information between meteorological and hydrological elements progressively increased, and this increased information to extreme agricultural drought. The information transfer that occurred under a wide range of meteorological and hydrological elements can also be applied to the mutual-information network.