Accumulated soil moisture deficit better indicates the effect of soil water stress on light use efficiency of grasslands during drought years

Light use efficiency (LUE) models have been widely used in the estimation of gross primary productivity (GPP). However, many studies indicated that current LUE models generally overestimate GPP during drought years. Recent studies also showed that soil moisture (SM) regulated LUE during drought years, indicating the necessity of adding SM-related stress scalars in LUE models. Meanwhile, considering the persistence characteristic of drought, the effect of SM stress on LUE is expected to be not only instantaneous but also accumulated. However, it is unclear whether the accumulated SM stress can better capture LUE. Thus, based on observations from grassland flux sites in the United States and Australia, we compared the effectiveness of concurrent soil moisture deficit (SMD), lagged soil moisture deficit (LSMD), and accumulated soil moisture deficit (ASMD) in capturing actual LUE during drought years. ASMD had more stable and stronger linear correlations with LUE than SMD and LSMD at almost all the sites. Furthermore, ASMD outperformed water stress scalars commonly used in LUE models such as the Land Surface Water Index (LSWI) and vapor pressure deficit (VPD) for tracking actual LUE during drought years. The introduction of ASMD improved the accuracy of the LUE model for simulating grassland GPP during drought years (with a R2 and RMSE of 0.88 and 1.06 g C m−2 d−1). Our study highlights the accumulated effect of SM in regulating LUE, and proves that ASMD has the potential in improving the performance of the existing LUE models during drought years.