Investigating the Propagation From Meteorological to Hydrological Drought by Introducing the Nonlinear Dependence With Directed Information Transfer Index

Abstract

Meteorological drought is the fuse of hydrological drought, and understanding the propagation mechanisms from meteorological to hydrological drought is of great significance to the monitoring and prevention of hydrological drought. Besides the linear dependence, this study thoroughly investigated the propagation from meteorological to hydrological drought by introducing the nonlinear dependence with directed information transfer index (DITI) for the first time. In this study, the standardized precipitation index and the standardized runoff index were used to represent meteorological drought and hydrological drought, respectively. A new drought response time (DRT) evaluation system was constructed based on the maximum Pearson correlation coefficient and DITI, simultaneously considering the linear and nonlinear relationships between meteorological drought and hydrological drought. Moreover, the relationships of drought characteristics (duration and severity) between these two types of drought were established by using run theory and mathematical function, and the trigger thresholds from meteorological to hydrological drought were then determined. The results indicate that: (a) the effective drought propagation rate was mainly affected by the characteristics of meteorological drought events and the sensitivity of hydrological drought to meteorological drought; (b) the DRT in the Pearl River Basin (PRB) was mainly concentrated in 2–5 months, and the drought translation rate in the PRB was relatively large; and (c) the duration of hydrological drought events was longer in the sub‐regions with smaller meteorological drought trigger thresholds.

Key Points

• Directed information transfer index was introduced in drought propagation study to build a new drought response time evaluation system for the first time

• Trigger thresholds from meteorological to hydrological drought were determined by a drought propagation model

• Sub‐regions with smaller meteorological drought trigger thresholds had longer durations of hydrological drought events