Evaluating the Time-Lag and Cumulative Effects of Climatic Factors on Vegetation Dynamics of Different Land-Use Types in the Circum-Tarim Basin

Abstract

Monitoring vegetation dynamics in drylands and disentangling their driving mechanisms are essential for deciphering vegetation–climate feedbacks and supporting targeted ecological restoration. Taking the Circum-Tarim Basin as the study region, this study adopted the modified soil-adjusted vegetation index (MSAVI) as a vegetation proxy and integrated multiple statistical approaches to quantify spatiotemporal vegetation variations across four land-use types from 2000 to 2021. The relative importance of time-lag and cumulative climatic effects on vegetation growth were evaluated for specific land-use types. The main results show that: (1) Widespread vegetation greening occurred across the study area, and croplands dominated this positive trend. (2) Vegetation generally exhibited more immediate responses to climate change in cropland and woodland, whereas grassland and unused land showed more heterogeneous, lagged responses. (3) Despite the inherently low absolute R2 values typical of hyper-arid pixel-scale modeling, climate models incorporating cumulative and lagged effects showed substantial relative improvements in explanatory power (ΔR2) compared to contemporaneous models, particularly in cropland and woodland. These improvements were consistently retained under out-of-sample validation (positiveΔCV-R2). These findings highlight the importance of considering land-use-specific lagged and cumulative climate effects when assessing dryland vegetation–climate interactions, which provides useful implications for ecological restoration and climate adaptation in arid basins.