Multifractal characteristics of multiscale drought in the Yellow River Basin, China

Abstract

As a climate-sensitive area in China, the Yellow River Basin (YRB) is increasingly affected by droughts resulting from climate change and frequent human activities. Droughts in the YRB are non-stationary, yet the temporal scaling properties of their complex fluctuations are still unclear. To address this deficiency, we investigated the spatiotemporal characteristics of multifractal nature and their sources through the multifractal detrended fluctuation analysis (MFDFA). Our research was conducted utilizing the Standardized Precipitation Evapotranspiration Index (SPEI) derived for 80 meteorological stations located within the YRB between 1961 and 2017 on various timescales. The main findings were as follows: (1) YRB droughts exhibited persistence since the Hurst exponent exceeded 0.5, indicating that future drought trends were expected to be consistent with the present. Drought persistence increased with increasing timescales, with the strongest persistence recorded in the mid-temperate semi-arid zone. (2) Multifractality was observed in all the SPEI series, as evidenced by the Generalized Hurst Exponent [h(q)] varied nonlinearly with q. Multifractal intensity strengthen with increasing drought duration, as measured by the width in the multifractal spectrum presented in detail at different timescales. (3) Small fluctuations dominated the multifractal behaviors of droughts as most of the multifractal spectrums were right-skewed. (4) In contrast to the original series, the shuffle and surrogate procedures significantly narrowed the width of the multifractal spectrum (while spectrum width remained greater than 0.05), revealing that the multifractality of the drought originated from the long-range correlations (LRC) and the broad probability density distribution (PDF). Moreover, the spectrum width of the shuffled sequences was more variable, demonstrating that LRC was a dominant source of multifractality. This study contributes to a better understanding of the complex fluctuations of drought, as well as a new perspective for improving the accuracy of drought prediction.