Multi-Scale Correlation between Soil Loss and Natural Rainfall on Sloping Farmland Using the Hilbert–Huang Transform in Southwestern China

The Hilbert–Huang transform (HHT) has been used as a powerful tool for analyzing nonlinear and nonstationary time series. Soil loss is controlled by complicated physical processes thus fluctuates with nonlinearity nonstationarity over time. In order to further clarify the relationship between rainfall, surface runoff, sediment yield, this study adopted HHT analyze these characteristics through multiple scales investigated their time-dependent intrinsic correlation (TDIC) in A six-year (2015–2020) was conducted on sloping farmlands explore relationships soil rainfall southwest China. Time series of were identified relevant at different based method HHT. Local runoff carried out TDIC. original decomposed into eight mode functions (IMFs) residue ensemble empirical decomposition (EEMD). indicated that increased then decreased during maize-growing season from 2015 2020, whereas gradually decreased. IMF1 IMF2 accounted nearly 80% temporal variations loss, indicating variables varied most short scales. TDIC analysis showed strong positive correlations prevailed entire domain IMF2, rapid response Time-varying observed IMF3–IMF5 At IMF7 scale, an evident switchover nature years 2018 2019; could be related sudden rainstorm under low vegetation coverage conditions. EEMD-based effective means runoff. Our results provide better understanding Short-term heavy are important factors causing serious water scale mountainous region yellow soil, which great significance construction regional erosion prediction model.