Unraveling the interplay between NDVI, soil moisture, and snowmelt: A comprehensive analysis of the Tibetan Plateau agroecosystem

Abstract

The rapid changing climate conditions within Tibetan Plateau determine the complex interaction between vegetation succession and agricultural water resources, including soil moisture and snowmelt. While previous studies have primarily focused on the coupling relationship between NDVI and soil moisture, snowmelt, as a critical water source in plateau ecosystems, plays an equally important role in regulating the water cycle. This study integrates MODIS remote sensing images and ERA5-Land meteorological reanalysis datasets to establish a ternary system encompassing NDVI, soil moisture, and snowmelt. Using geostatistical methods such as trend analysis, cross-correlation, random forest algorithm, and Granger causality, we explore the temporal dynamics and causal relationships among these ecological variables. Results indicate an overall increase in NDVI, a consistent decrease in snowmelt, and spatially heterogeneous changes in soil moisture across the Tibetan Plateau from 2001 to 2023. NDVI and soil moisture exhibit mostly instantaneous responses, with a brief one-month time-lag effect, while NDVI demonstrates a more pronounced lagged response to snowmelt. In grassland ecosystems, soil moisture lags behind snowmelt, whereas in woodlands, snowmelt lags behind soil moisture. Transitional vegetation zones reveal a regulatory feedback loop, where snowmelt predominantly influences soil moisture, which subsequently transitions to a bidirectional feedback mechanism between soil moisture and snowmelt as vegetation succession in woodland ecosystems. This study provides new insights into the feedback processes between vegetation growth and water resources in different ecological zones of Tibetan Plateau, guiding water management and sustainable development for agroecosystem.