Surface Water Hydrology

For my final project in Surface Water Hydrology, I tested whether China's Grain for Green Program — the largest deliberate land-cover transformation in human history — has produced a hydrologically detectable water debt across the Loess Plateau, and whether that debt scales with local aridity. Using the full 2002–2024 GRACE and GRACE-FO satellite gravity record, I documented a domain-mean total water storage decline of −0.38 cm equivalent water height per year (Mann-Kendall p < 0.0001, −8.4 cm cumulative). Precipitation showed no significant trend, and a Clausius-Clapeyron analysis demonstrated that warming could explain at most 10.8% of the observed evapotranspiration increase — leaving roughly 89% requiring a vegetation-driven mechanism.

The central empirical contribution is a vegetation × aridity interaction regression across 96 grid cells (β₃ = −62.7, p < 0.0001, adj. R² = 0.53). The interaction term shows that the same degree of NDVI greening produces disproportionately greater water depletion in drier cells — roughly 13-fold amplification of the per-unit greening cost between sub-humid and arid grid cells. To my knowledge it is the first satellite-scale quantification of the aridity-conditioned water debt of afforestation.

The paper has been submitted to Earth's Future for first-author publication. The implications matter beyond the Loess Plateau: the Bonn Challenge commits nations to 350 million hectares of restoration by 2030, much of it in semi-arid drylands. The aridity index offers a satellite-available criterion for distinguishing where afforestation is hydrologically self-sustaining from where it accumulates a long-term debt against deep legacy soil moisture.