Monitoring drought using composite drought indices based on remote sensing

Drought is one of the most frequent disasters occurring in North China and has a great influence on agriculture, ecology and economy. To monitor drought of typical dry areas in North China, Shandong Province, this paper proposed composite drought indices using multivariable linear regression (MCDIs) to integrate Tropical Rainfall Measuring Mission (TRMM) derived precipitation, Global Land Data Assimilation System Version 2.1 (GLDAS-2.1) derived soil moisture, Moderate Resolution Imaging Spectroradiometer (MODIS) derived land surface temperature (LST) and normalized difference vegetation index (NDVI) from 2013 to 2017 (March to September). Pearson correlation analyses were performed between single remote sensing drought indices and in-situ drought indices, standardized precipitation evapotranspiration index (SPEI), in different time scales to assess the capability of single indices over Shandong Province. The multivariable linear regression method was used to established MCDIs, and mediator and moderator variables were introduced to optimize the model. The correlation coefficients (r) between MCDIs and SPEIs was higher than that between each single index and SPEIs. Additionally, when we investigate the correlations of different MCDIs with both standardized precipitation index (SPI) and moisture index (MI), the highest r values with both 1-month SPI and MI were acquired by the MCDI based on 1-month SPEI (MCDI-1). This suggested MCDI-1 was suitable to monitor meteorological drought. Also, the comparison between MCDI based on 9-month SPEI (MCDI-9) and soil moisture showed MCDI-9 was a good indicator for agricultural drought. Therefore, multivariable linear regression and MCDIs were recommended to be an effective method and indices for monitoring drought across Shandong Province and similar areas.