Recent increase in soil moisture levels concerning climate variability in the karst region of southwest China using wavelet coherence and multi-linear regression

Abstract

Classifying climate-induced variations in soil moisture (SM) is vital in the face of climate change to enhance vegetation restoration and decrease rocky desertification in karst areas. In this paper, the Monotonic Mann-Kendall and Sen’s Slope test are used to assess the changes in SM and regional climatic variables from 1981 to 2019 across the karst region of southwest China (KRSC) using the ERA5 SM at 0–7, 7–28 and 28–100 cm. By utilizing wavelet coherence and multi-linear regression (MLR), we assess the association of SM with local climatic factors and teleconnections. Our findings indicate increased SM, NDVI, temperature (TMP), and precipitation (PCP) from 2002 to 2019 in the KRSC. The combined variance of the first and second EOF for 0–7, 7–28, and 28–100 cm showed 55 %, 52 %, and 44 %, respectively. The findings revealed a dominant influence of PCP and sensible heat fluxes (SHF) in SM for 8–120 months (∼1–10 years) periodic cycles, whereas the teleconnections showed consistent patterns in karst and non-karst regions. For integrated factors, PCP + actual evapotranspiration (AET), NDVI + SHF and long-wave radiation (LWR) + SHF showed mean significant coherences between 0.70 and 0.80, whereas ENSO + PDO + NAO (0.71–0.76) explained high coherence in comparison to other combinations. The karst SM showed a significant correlation with PCP (0.63–0.89), LWR (0.59–0.77), and SHF (0.76–0.95), while non-karst exhibited a significant correlation with ENSO (0.07–0.49), PDO (0.16–0.41) and PNA (0.10–0.38) in all seasons. The MLR identified PCP, SHF and LWR as the leading influencing variables in the karst region, whereas PCP and LWR showed more influence in the non-karst region for 0–7 cm. Our findings give valuable insights regarding sustainable SM management to promote vegetation and reduce rocky desertification.