遥感降水产品对黄河源区水文干旱特征的模拟性能分析

Simulation performance of remote sensing precipitation products on hydrological drought characteristics in the source region of the Yellow River

遥感降水产品为稀缺资料地区水文气象灾害机理与预警研究提供了重要的数据资料,但不同遥感降水产品的性能存在较大区域异质性。利用遥感降水产品开展水文气象相关的研究和应用前,需要对其性能进行综合评估。基于此,以稀缺资料的黄河源区为研究区,利用1983—2018年的观测降水数据(CMA)驱动并率定ABCD水文模型,并利用标准化径流指数(SRI),评估3套典型遥感降水产品(PERSIANN-CDR、CHIRPS v2.0、MSWEP v2.0)对水文干旱的模拟性能。利用游程理论识别水文干旱事件,阐明遥感降水对水文干旱特征的捕捉能力。结果表明:(1)3套遥感降水产品均能较好地捕获CMA多年均值的时空分布格局。CHIRPS产品的水文模拟性能(纳什效率系数NSE=0.72)高于其他2套产品。(2)CMA和降水产品模拟的4个尺度的SRI(SRI1、SRI3、SRI6和SRI12)均呈显著增加趋势(P<0.01),表明近36 a源区河川径流增加,水文干旱趋缓,但降水产品均高估了SRI,表明对黄河源区降水产品的偏差校正有待开展。基本统计指标方面,MSWEP产品计算的SRI与CMA的最为一致,性能最佳,但在年尺度(SRI12)上,PERSIANN表现最优。(3)3套产品均高估了SRI1和SRI3的干旱历时和烈度,MSWEP产品对SRI6的模拟性能最优,PERSIANN对SRI12的模拟性能最优。研究结果可为黄河源区水文干旱研究的降水产品数据的选择提供科学决策支持。

In regions with scarce data,remote sensing precipitation products provide crucial data for the development of the hydrometeorological disaster mechanism and early warning studies.However,the performance of various remote sensing precipitation products exhibits regional heterogeneity.Comprehensively evaluating the performance of remote sensing precipitation products is critical for their use in hydrometeorological-related research and application.Based on this assumption,the study investigated the source region of the Yellow River of China by using the observed precipitation data(CMA)from 1983 to 2018 to drive and calibrate the ABCD hydrological model.Furthermore,the standardized runoff index(SRI)was used to evaluate the simulation performance of three sets of typical remote sensing precipitation products(PERSIANN-CDR,CHIRPS v2.0,MSWEP v2.0)on hydrological drought.Furthermore,hydrological drought events were identified by using run theory,and the potency of remote sensing precipitation to capture hydrological drought characteristics was investigated.The results revealed that:(1)The three precipitation products can accurately capture the temporal and spatial distribution pattern of CMA’s multiyear mean value.Furthermore,the CHIRPS product(Nash-Sutcliffe efficiency coefficient,NSE=0.72)outperformed other two products in term of hydrological simulation.(2)The SRI values(SRI1,SRI3,SRI6,and SRI12)of the four scales simulated by CMA and three sets of remote sensing precipitation products revealed a significant increase trend(P<0.01),which indicated that the river runoff in the source region increased in the last 36 years,and the hydrological drought slowed down.However,the SRI values of the three sets of remote sensing precipitation products were overestimated.This result revealed that the deviation correction of precipitation products in the source area of the Yellow River is necessary.In terms of basic statistical indicators,the SRI calculated by the MSWEP product was the most consistent with CMA and was the best.However,on an annual scale(SRI12),the PERSIANN product achieved the best performance.(3)Three sets of remote sensing precipitation products overestimated the drought duration and intensity of SRI1 and SRI3;the MSWEP product achieved the best simulation performance for SRI6;and the PERSIANN product has the best simulation performance for SRI12.The results of this study can provide scientific decision support for the selection of precipitation product data for studying hydrological drought in the source region of the Yellow River.