Modeling the effects of precipitation and temperature patterns on agricultural drought in China from 1949 to 2015

Agricultural drought is one of the most frequent and widespread natural disasters occurring in China. Drought is associated with hydrological and meteorological conditions that lead to water-deficient vegetation, which has a negative effect on agricultural activities. The monitoring of droughts, as well as early-warning and timely information, is significant for crop production and food security. However, the spatial and temporal patterns of precipitation and temperature have rarely been reported when monitoring the agricultural drought loss rate on a national scale. In this study, we analyzed the spatial and temporal patterns of drought based on model simulation. An artificial neural network (ANN) model for drought warning was developed using monthly temperature and precipitation data from 1949 to 2015. Our results demonstrated that the agricultural drought loss rate can be simulated in most agricultural areas of China. Our ANN model simulation revealed that the areal percentages of precipitation and temperature are strongly correlated with agricultural drought, with the agricultural drought loss rate exhibiting greater sensitivity to precipitation than temperature. We suggest that the spatial and temporal patterns of precipitation are useful for capturing drought warning signals. The precipitation thresholds play an important role in detecting agricultural drought in critical months or seasons of crop growth in different regions. This study presents a framework and reference for drought monitoring in the regions and countries facing frequent agricultural drought.