Abstract：
Although many drought indices have been developed to monitor drought conditions, their applicability differs across climatic regions, which results in different characteristics of drought index assessments at different timescales and over different record lengths. Therefore, the applicability of precipitation-based and precipitation-evapotranspiration/temperature-based drought indices was examined in this study using the Generalized Extreme Value Index (GEVI), Homogeneity Index (HI) of precipitation and temperature, K index (K), precipitation anomaly percentage (Pa), Standardized Precipitation Evapotranspiration Index (SPEI), Standardized Precipitation Index (SPI), and China-Z Index (CZI) over different record lengths and at a timescale range of 1-24 months. In addition, the results that were obtained using these indices were compared with the Self-Calibrating Palmer Drought Severity Index (scPDSI). The stability, accuracy, and consistency of the different drought indices were evaluated using precipitation and evapotranspiration/temperature data (1-24 months) collected from different climatic regions in the arid area of northwest China in the 1961-2017 period. The results indicated that the Pa, CZI, K, and SPEI were more stable; the SPI, SPEI, and HI were more accurate; and the GEVI, SPI, HI, and SPEI were more consistent with the scPDSI. In addition, the results indicated that it is more appropriate to select a long record length (> 35 years) to monitor drought when sufficient data are available. However, defining constant drought classes may not be appropriate for all drought timescales. In fact, precipitation and evapotranspiration data from different timescales had different optimal distribution functions. The drought indices also demonstrated that they were applicable to a temperate arid area in the study region. In addition, the HI and SPEI better captured the precipitation and evapotranspiration/temperature characteristics, while the CZI, K, and Pa overestimated or underestimated the frequency of different drought classes at different timescales to some extent in the study region. The results of this study suggest that greater priority should be given to the precipitation-evapotranspiration-based indices. In addition, it is suggested to change the drought index class thresholds in future related studies on drought event recognition at different timescales to ensure more accurate drought monitoring.
摘要：
尽管已经开发了许多干旱指数来监测干旱状况，它们的适用性在不同的气候区域不同，这导致在不同时间尺度和不同记录长度上的干旱指数评估的不同特征。因此,在这项研究中，使用广义极值指数（GEVI）检查了基于降水和基于降水-蒸散/温度的干旱指数的适用性，降水和温度的均匀性指数（HI），K指数（K），降水异常百分比(Pa)，标准化降水蒸散指数（SPEI），标准化降水指数(SPI)，和China-Z指数（CZI）在不同的记录长度和1-24个月的时间尺度范围内。此外,将使用这些指数获得的结果与自校准帕尔默干旱严重程度指数（scPDSI）进行比较。的稳定性,准确度,使用1961-2017年期间从中国西北干旱区不同气候区收集的降水和蒸散量/温度数据（1-24个月）评估了不同干旱指数的一致性。结果表明,CZI,K,SPEI更稳定；SPI，SPEI,和HI更准确；和GEVI，SPI,HI,SPEI与scPDSI更为一致。此外,结果表明，在有足够数据的情况下，选择较长的记录长度（>35年）来监测干旱更为合适。然而,定义恒定的干旱类别可能不适用于所有干旱时间尺度。事实上,不同时间尺度的降水和蒸散数据具有不同的最优分布函数。干旱指数还表明，它们适用于研究区域的温带干旱地区。此外,HI和SPEI更好地捕获了降水和蒸散/温度特征，而CZI,K,和Pa在一定程度上高估或低估了研究区域不同时间尺度下不同干旱等级的发生频率。这项研究的结果表明，应优先考虑基于降水-蒸散的指数。此外,建议在未来不同时间尺度的干旱事件识别相关研究中改变干旱指数类别阈值，以确保更准确的干旱监测。
