Understanding moisture stress on light use efficiency across terrestrial ecosystems based on global flux and remote-sensing data

Light use efficiency (LUE) is a key biophysical parameter characterizing the ability of plants to convert absorbed light to carbohydrate. However, the environmental regulations on LUE, especially moisture stress, are poorly understood, leading to large uncertainties in primary productivity estimated by LUE models. The objective of this study is to investigate the effects of moisture stress on LUE for a wide range of ecosystems on daily, 8 day, and monthly scales. Using the FLUXNET and Moderate Resolution Imagine Spectroradiometer data, we evaluated moisture stress along the soil-plant-atmosphere continuum, including soil water content (SWC) and soil water saturation (SWS), land surface wetness index (LSWI) and plant evaporative fraction (EF), and precipitation and daytime atmospheric vapor pressure deficit (VPD). We found that LUE was most responsive to plant moisture indicators (EF and LSWI), least responsive to soil moisture (SWC and SWS) variations with the atmospheric indicator (VPD) falling in between. LUE showed higher sensitivity to SWC than VPD only for grassland ecosystems. For evergreen forest, LUE had better association with VPD than LSWI. All moisture indicators (except soil indicators) were generally less effective in affecting LUE on the daily and 8 day scales than on the monthly scale. Our study highlights the complexity of moisture stress on LUE and suggests that a singlemoisture indicator or function in LUEmodels is not sufficient to capture the diverse responses of vegetation to moisture stress. LUE models should consider the variability identified in this study to more realistically reflect the environmental controls on ecosystem functions.