Agricultural drought dynamics with satellite soil moisture and impacts on crop loss in China during 2003-2022

Agricultural drought poses a significant risk to national food security and regional stability. Studying on agricultural drought enhances the understanding of drought dynamics and facilitates more effective drought risk management. However, cropland is typically managed at a fine scale, while the data resolution used in previous studies has been relatively coarse, making it insufficient to meet the requirements of agricultural drought analysis. Therefore, it is essential to utilize a higher resolution dataset to capture drought on a larger scale. Firstly, this study calculated the Standardized Soil Moisture Index (SSMI) using a 1 km resolution soil moisture dataset covering China from 2003 to 2022. Secondly, methods such as MannKendall and Spatiotemporal Drought Clustering were employed to conduct a spatiotemporal analysis of agricultural drought. At the national scale, a correlation analysis was conducted between SPEI-based drought intensity, SSMI-based drought intensity, and crop loss. At the provincial scale, a correlation analysis was conducted between SSMI-based drought intensity and the affected area, disaster area, and crop failure area. Results demonstrate that: 1) From 2003 to 2022, agricultural drought in China exhibited an overall decreasing trend; 2) The centroids of agricultural drought events showed a southwestward shift, with a declining drought trend in Northeast China; 3) The frequency distribution of different drought severity levels varied, with moderate droughts occurring more frequently in Northeast and Central China, severe droughts primarily in Central China, and extreme droughts concentrated in southern and southwestern China; 4) The correlation analysis between SSMI-based drought intensity and agricultural crop loss yielded correlation coefficient (R) values of 0.84, 0.77, 0.78 and 0.75 for the affected area, disaster area, crop failure area and food loss respectively. 5) From 2003 to 2020, the correlations between SPEI-based drought intensity and crop loss were 0.63, 0.69, 0.61, and 0.63, respectively, all of which are lower than the corresponding correlations with SSMI-based drought intensity during the same period. 6) At the provincial scale, the R values for the correlation between SSMI-based drought intensity and the affected area, disaster area, and crop failure area are 0.58, 0.53, 0.39 respectively. Future research is encouraged to integrate SAR data with crop growth descriptors to capture drought and explore the coupling between agricultural droughts and hydrological, ecological, and environmental processes.