An integrated approach for shaping drought characteristics at the watershed scale

Abstract

[EN] Different indicators such as precipitation, surface and groundwater availability, vegetation, and soil types are indispensable in developing land management plans for detecting, monitoring, and evaluating drought impacts at the watershed scale. Because of the complex interactions among these indicators, it is well known that one single hydrometeorological variable would be unable to capture all the aspects of drought characteristics. In this study, a methodology based on Principal Component Analysis (PCA)-Copula was proposed to assess the frequency of different drought events and improve mitigation in watersheds. A comprehensive approach for an Integrated Drought Indicator (IDI) was proposed using a combination of three hydrometeorological variables (precipitation, runoff, and soil moisture) based on PCA and tested for the upstream Nanpan river in China. IDI confirmed the cumulative contribution rate of the first and second principal components (85%), and its design included simulation of monthly average soil moisture content (from 0 to 0.4-meter soil layer) using Soil and Water Assessment Tool (SWAT) model. The parameters were calibrated using the SUFI-2 optimization algorithm in SWAT-CUP. Then, the IDI relative anomaly characterization was used to identify drought processes based on the Run theory. Two major drought characteristics, duration, and severity were abstracted from the observed drought events. Finally, two-dimensional copulas were applied for analyzing comprehensive drought characteristics in the region. Our results showed that significant differences in drought duration periods could be identified. Drought duration, predicted by runoff and soil moisture, were found longer than those assessed using the precipitation index. IDI suggest our proposed model was suitable to identify comprehensive drought events and describe the overall drought characteristics of the region. In the same way, PCA-Copula methodology was found with high potential for drought analysis in areas where no previous studies have been performed.