Vegetation structural change and carbon–water functional responses across arid Central Asia: spatiotemporal patterns and environmental associations

Abstract

Anthropogenic climate change is intensifying ecological stress in arid Central Asia, where vegetation structure and carbon–water-related functional responses are strongly constrained by water availability and atmospheric demand. Yet structural greening does not necessarily imply parallel changes in carbon–water-related functional dimensions. Here, we examined long-term relationships between leaf area index (LAI) and three complementary indicators of ecosystem carbon–water coupling—precipitation use efficiency (PUE), water use efficiency (WUE), and the transpiration-to-evapotranspiration ratio (TET)—and assessed their associations with climatic and soil-moisture conditions during 1982–2018. Using satellite, reanalysis, land-use, soil-moisture, evapotranspiration and climate records, we combined temporal trends, pixel-wise correlations, land-use transitions and ridge-regression-based relative statistical contributions across aridity gradients. Regional LAI increased only weakly, whereas WUE declined significantly (−0.023 decade −1 , p = 0.047). Spatially, significant greening was more widespread than browning (22.99% versus 11.31%), but declines in PUE, WUE and TET extended across more than half of the study region. Positive LAI–functional associations were concentrated in irrigated oases and cropland-dominated landscapes, whereas weaker or negative relationships occurred across parts of natural grassland and desert vegetation. Environmental associations were indicator-specific: surface soil moisture generally exhibited larger relative statistical contributions to PUE and WUE, whereas root-zone soil moisture showed relatively stronger associations with TET; sensitivity analyses broadly retained these rankings but indicated weaker depth-specific differentiation for WUE. Contribution direction also depended on the temporal trajectory of each environmental factor, with the largest-contribution factor showing trend-consistent signed contributions across 72.53–80.56% of the analyzed area. Along the aridity gradient, associations shifted from stronger soil-moisture dependence in drier environments toward greater temperature importance in wetter environments. These results indicate that, within the integrated dataset framework adopted here, vegetation greenness alone is insufficient to characterize carbon–water-related ecosystem functional responses across the dryland landscapes examined in arid Central Asia and should be interpreted together with hydroclimatic conditions, land use, and aridity context.