Format

Send to

Choose Destination
Sci Total Environ. 2019 Jan 10;647:981-991. doi: 10.1016/j.scitotenv.2018.07.273. Epub 2018 Jul 31.

On the mechanisms of two composite methods for construction of multivariate drought indices.

Author information

1
State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China; School of Earth Sciences and Engineering, Hohai University, Nanjing 211100, China.
2
College of Hydrometeorology, Nanjing University of Information Science & Technology, Nanjing 210044, China.
3
State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China. Electronic address: rll@hhu.edu.cn.
4
School of Earth Sciences and Engineering, Hohai University, Nanjing 211100, China.
5
Department of Biological and Agricultural Engineering, Texas A&M University, College Station, TX 77843-2117, USA.
6
State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China.

Abstract

Droughts are comprehensive and complex issues that need to be characterized from a multivariate perspective. In recent years, a number of composite indices have been proposed for drought characterization. However, rare studies have systematically compared similarities and dissimilarities of these indices, and they have provided little insights into the combination mechanisms. To address this issue, two widely used combination approaches, namely the principal component analysis (PCA) and copula based joint probability distribution were employed, with the corresponding integrated product denoted as the Aggregate Drought Index (ADI) and Joint Drought Deficit Index (JDI). Five constituents for constructing ADI and JDI were derived from the variable infiltration capacity model (VIC) monthly simulations over the Yellow River basin (YRB), China, including precipitation (P), actual evapotranspiration (ET), soil moisture of top two layers, and runoff (during 1961-2012). Results showed that the behavioral patterns of ADI and JDI may not be easily influenced by the variation of one single element, and they represented comprehensive moisture status well. A further comparison between these two composite indices suggested that ADI and JDI behaved similarly in most areas of YRB, with some dissimilarities in the source region. The particular behavior of ET was responsible for the inconsistency. Comparing to other regions, an enhanced role of potential evapotranspiration (PET) was imposed on ET in the source region, leading to a poor relationship of ET with P and other hydrological variables. Accordingly, when constructing composite drought indices, the drought information indicated by ET was more easily abandoned by ADI but reserved in JDI. This study clearly demonstrates the mechanisms of two common integrated approaches in blending different drought information, which has significant implications for composite drought indices construction and application, and potentially provides some valuable references for the improvement of monitoring techniques in future drought related researches.

KEYWORDS:

Composite drought index; Copula; Dissimilarity; Principle component analysis; Similarity

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center