The development of a Nonstationary Standardized Precipitation Index using climate covariates: A case study in the middle and lower reaches of Yangtze River Basin, China

JOURNAL OF HYDROLOGY

卷:588

Article

Song, Zhihong^1,2  Xia, Jun^1,2,3  She, Dunxian^1,2  Zhang, Liping^1,2  Hu, Chen^1,2  Zhao, Lin⁴ 

[1] Wuhan Univ, State Key Lab Water Resources & Hydropower Engn S, Wuhan 430072, Peoples R China

[2] Wuhan Univ, Hubei Key Lab Water Syst Sci Sponge City Construc, Wuhan 430072, Peoples R China

[3] Chinese Acad Sci, Key Lab Water Cycle & Related Land Surface Proc, Beijing 10010, Peoples R China

[4] Wuhan Univ, Sch Resource & Environm Sci, Wuhan 430072, Peoples R China

关键词: Meteorological drought;    Nonstationarity;    Nonlinear dependence;    Climate change;    Standardized Precipitation Index;   

DOI  :  10.1016/j.jhydrol.2020.125115

来源: Elsevier

【 摘 要 】

The widely used probability based drought indices for drought characterization usually underlie the assumption of the considered hydrometeorological variables to be stationary, which is largely challenged under the current global change. Incorporating the parameters of probability distribution with covariates, such as time or climate indices, has become a common way to develop drought index considering nonstationarity. However most previous drought studies emphasized more on the linear relationship between distribution parameters and related covariates, which may neglect the inherent nonlinear relationship and lost useful information for drought assessment. Our study aims to develop a Nonstationary Standardized Precipitation Index (NSPI) with distribution parameters nonlinearly varying with potential influencing climate indices. The middle and lower reaches of the Yangtze River Basin (MLRYRB) of China are selected as the case study area to examine the performance of the NSPI with the performance of the traditional standardized precipitation index (SPI). Results indicate that the NSPI (with climate indices as covariates) is more robust than both the NSPI (with time as the covariate) and the traditional SPI. Furthermore, the nonlinear dependence of distribution parameters on climate indices considered in the nonstationary model can be more suitable for drought assessment than the linear dependence of distribution parameters. Besides, there exists a higher frequency of extreme drought by the NSPI than the SPI in recent decades in most regions of the MLRYRB. Although the NSPI is a lithe computationally expensive with more required inputs of climate variables, it can depict the influence of changing environment on the drought occurrence, thus can be a feasible alternative for drought assessment considering nonstationarity in view of future change and provide valuable support for further studies.

【 授权许可】

Free