Disentangling hydroclimatic controls on drought-induced vegetation productivity loss at the basin scale across China

Journal of Hydrology. 2026. Vol. 677. No. A. Article 135843.

Yin X., Wu Y., Kivalov S., Alexandrov G., Chen J., Liu S., Qiu L.

As a frequent extreme event under global climate change, drought significantly threatens the net primary productivity (NPP) of terrestrial ecosystems. Although numerous studies have reported drought-induced declines in NPP, the hydroclimatic mechanisms behind remain insufficiently understood. The strong interdependence among temperature, precipitation, and water availability has made it challenging to quantify their independent effects on vegetation productivity under drought stress. Here, we assessed basin-scale NPP responses to drought across mainland China from 1982 to 2018 and applied a ridge regression method to disentangle the individual impacts of multiple hydroclimatic drivers on NPP variability, which overcomes the limitations of conventional correlation-based analyses by accounting for multicollinearity, thereby allowing a more robust identification of the dominant controls on drought-induced NPP variations. Our results revealed pronounced spatial heterogeneity in NPP responses among major basins, with the strongest productivity losses in the Songliao River basin and the Yellow River basin (12.41 g·C·m⁻2·yr⁻1 and 11.71 g·C·m⁻2·yr⁻1, respectively). Ridge-derived coefficients indicated that water availability was the primary driver of NPP sensitivity to drought, and stronger water-availability control was typically associated with greater NPP losses during drought in water-limited basins. By integrating basin-scale analysis with a multivariate attribution framework, this study isolated the hydroclimatic controls of vegetation productivity under drought. The findings can improve understanding of terrestrial carbon dynamics and be informative for enhancing drought resilience and optimizing water–carbon management strategies in terrestrial ecosystems.

Keywords: Drought ; Net primary productivity (NPP) ; Water availability ; Ridge regression analysis