Copula-Based Joint Drought Index Using Precipitation, NDVI, and Runoff and Its Application in the Yangtze River Basin, China

Abstract

Drought monitoring ensures the Yangtze River Basin’s social economy and agricultural production. Developing a comprehensive index with high monitoring precision is essential to enhance the accuracy of drought management strategies. This study proposes the standardized comprehensive drought index (SCDI) using a novel approach that utilizes the joint distribution of C-vine copula to effectively combine three critical drought factors: precipitation, NDVI, and runoff. The study analyzes the reliability and effectiveness of the SCDI in detecting drought events through quantitative indicators and assesses its applicability in the Yangtze River Basin. The findings are as follows: (1) The SCDI is a highly reliable and applicable drought index. Compared to traditional indices like the SPI, VCI, and SRI, it has a consistency rate of over 67% and can detect drought events in more sensitive months by over 51%. It has a low false negative rate of only 2% and a false positive rate of 0%, making it highly accurate. The SCDI is also applicable to all the third-level sub-basins of the Yangtze River Basin, making it a valuable tool for regional drought monitoring. (2) The time lag effect of the NDVI can affect the sensitivity of the SCDI. When the NDVI time series data are shifted forward by one month, the sensitivity of the SCDI in detecting agricultural drought improves from 47.8% to 53%. (3) The SDCI can assist in monitoring drought patterns in the Yangtze River Basin. From 2001 to 2018, the basin saw fluctuations in drought intensity, with the worst in December 2008. The western region had less frequent but more intense and prolonged droughts, while the eastern part had more frequent yet less severe droughts.

Evaluating Performance of Multiple Machine Learning Models for Drought Monitoring: A Case Study of Typical Grassland in Inner Mongolia

“…Furthermore, mutual feedback and time lag between vegetation and dry conditions are crucial for the construction of a drought index. Vegetation changes reflect the wet and dry conditions of the region, as well as the relationship between soil, atmosphere, and water [5,[26][27][28][29]. The Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), precipitation (PRE), evapotranspiration (ET), and soil moisture (SM) were selected as independent variables to construct a remote sensing drought monitoring model.…”