Multiscale Assessments of Droughts: A Case Study in Xinjiang, China

Understanding the multi-scale variation of drought is essentially important in drought assessment. Now, a comprehensive assessment is still lacking on the meteorological, ecological and hydrological drought perspectives. In order to better investigate multi-scale droughts, we carried out a comprehensive analysis of their long-term variation based on the two drought indices and observation data in Xinjiang, China, from 1961-2015. The two indices are the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI). The results show that the SPI and SPEI are highly consistent for most stations and time scales in Xinjiang. Based on multi-scale and considered evaporative demand, the SPEI from 1961-2015 showed a wetting trend followed by a drying trend (as of 1997), giving an overall slight drying trend (-0.0122± 0.0043 per year) for the 54-year period. We assessed the sensitivity of the two drought indices to precipitation (P) and potential evapotranspiration (PET) and found that the SPEI shows different sensitivity to P and PET. In arid regions characterized by high PET, drought severity is mostly determined by changes in PET. The intensified warming and diminished precipitation in Xinjiang that have been observed over the past two decades have resulted in SPEI-drought severity. These changes also amplify the risk of ecological drought. However, the hydrological drought was highly complex and not entirely comparable to the SPEI and SPI droughts. Hydrological records indicate that runoff in most rivers in the Tianshan Mountains has increased, whereas runoff in the Kunlun Mountains is either stable or has slightly decreased over the past 20 years. A moderately high and statistically significant correlation between the runoff anomaly and the SPEI and SPI was revealed for four major rivers in the region. This implies that the accelerated river runoff in Xinjiang is a function of both precipitation and increasing glacier melt.