Abstract
Due to the differences in the green water (GW) budget patterns of different vegetation, improper vegetation restoration may not only fail to improve the ecological environment but also cause irreversible damage to ecologically vulnerable areas, especially when vegetation restoration continues to be implemented in the future, and the pressure on water scarcity increases further. However, there is a lack of standardized research on the differences in the patterns of recharge, consumption, and efficient use of GW in typical vegetation. This makes the research results vary and cannot provide direct support for water management decision-making. Therefore, in this study, 30-year-old woodlands (R. pseudoacacia and P. orientalis) and two typical grasslands (I. cylindrican and M. sativa) that are similar to each other except for species were selected in a headwater catchment in the rain-fed agricultural area. A new GW concept and assessment framework was constructed to study the GW of long-term revegetation using a combination of field experiments and model simulations during the 2019-2020 growing season. The study findings comprise the following: (1) High-efficiency green water (GW(H)), low-efficiency green water (GW(L)), ineffective green water (GW(I)), and available green water storage (GW(A)) in the four sample plots during the study period were defined, separated, and compared. (2) An analysis of GW(A) variations under different water scenarios. (3) The establishment of GW(H) and GW(L) thresholds. (4) Strategies to reduce GW(I) and optimize GW potential while maintaining soil erosion prevention measures. (5) Suggestions for vegetation restoration species based on diverse factors. This research enhances comprehension of the impact of vegetation restoration on green water dynamics in ecologically vulnerable areas such as the rain-fed agricultural zone of the Loess Plateau.