Evaluation of suitability using typical drought index in source region of the Yellow River

Abstract: As the most important catchment of the Yellow River, the drought of the source region of the Yellow River has a significant impact on water resources regulation and ecological environment protection. Although, as a foundation of drought estimates, drought index is capable of describing the intensity, range and starting and ending time of the drought, due to differences of the methodology and background in drought indexes, widely used drought index based on meteorological measurements cannot precisely depict the temporal and intensity characteristic of the drought, evaluating the performance of drought index is essential for drought monitoring and diagnosis. Therefore, soil moisture anomaly percentage index (SMAPI) from a well-instrumented regional-scale soil moisture and temperature monitoring network was identified as the reference of drought index to evaluate five droughts indices-the palmer drought severity index (PDSI), the self-calibrating palmer drought severity index (SC-PDSI), standardized precipitation index (SPI), standardized precipitation evapotranspiration index (SPEI), and monthly land water storage anomaly from gravity recovery and climate experiment (GRACE). The time series of SMAPI showed that during the whole study period there were five stages which were respectively from July 2008 to August 2009 (the first stage), September 2009 to March 2011 (the second stage), April 2011 to December 2011(the third stage), January 2012 to June 2016 (the fourth stage) and from July 2016 to June 2017 (the fifth stage). Results indicated against the first stage, two distinct features of SMAPI were presented with an initial increasing trend and increasing again after a decreasing trend at the end of the first stage. At this stage, a slight drought event occurred in the cold winter of 2008 (SMAPI=-17%). Against the second stage, except August of 2010, the SMAPI showed a slowly drying trend, especially at the end of 2010, that all SMAPI of different soil depths decreased to approximately -20% indicates during this period there was a stable drought event. Against the third stage, the study area presented a clear wetting process, and in the fall of 2011 occurring an extremely wetting event, which was the most humid month overall phases (SMAPI at 0.05 m was equal to 52%). The drought event with the longest duration was from January 2012 to June 2016 (the fourth stage), and in accordance with minimum SMAPI (SMAPI at 0.05 m, 0.10 m, and 0.20 m were equal to -47%, -43%, -41%, respectively) at three soil depths, the severest drought occurred in August 2015. The drought began to mitigate in the last stage, and concurrent SMAPI increased to larger than 5% in January 2017. The estimation of the five indexes indicated SC-PDSI had a similar trend with PDSI, but SC-PDSI showed a more stable characteristic in comparison of PDSI, and thus SC-PDSI performed the optimum effect in the Yellow River source region, but according to an existing classification of drought, it would generally overestimate the intensity of drought event, and the future work should thus need to define a new classification of drought for SC-PDSI. Against an identical time-scale parameter k, the time series of SPI and SPEI showed similar characteristics, but SPI was more suitable for assessing the intensity of drought in this study region. But when air temperature was less than 0, it would result in a large error in study area. Terrestrial water storage anomaly from gravity recovery and climate experiment, as a space-borne remote sensing observation product, showed a significant agreement with the SMAPI (correlation coefficient was 0.37, P<0.01). Relevant research should primarily focus on improving the existing drought assessment approach or developing a more suitable drought index for the source catchment of the Yellow River in future.