Ecosystem root-zone water capacity is driven by water deficit and precipitation seasonality

In models that include the role of terrestrial vegetation in the water and carbon cycle, such as hydrological, vegetation, and Earth System models, it is common to find a variety of approaches to simulate the stress signal to plants due to reduced available water in the soil. These schemes depend on data availability, ecosystem type, and modelling assumptions. Our aim is to deliver an improved method to estimate root-zone water capacity (RZ _WC ) — a key parameter to estimate soil–water stress — that can be easily applied from local to global scales and used as a benchmark when assessing the outputs of related environmental models. We test our method using a function derived from the mass-balance approach, which adjusts RZ _WC following precipitation seasonality to account for the effects of soil–water stress on gross primary production of the vegetation. This method does not require calibration or soil information and quantifies RZ _WC by calculating the cumulated water deficit based on the balance between water influxes and effluxes. We used global datasets, including ERA5-Land, WFDE5, and PML-V2, to extract precipitation, runoff, and evapotranspiration time series to compute RZWC. We assessed our results using the outputs of a gross primary productivity (GPP) model based on least-cost optimality theory, i.e. P-model, which estimates GPP based on the least-cost optimality theory. We compared our approach with the P-model’s soil–water stress function at site and global scales and found consistently higher efficiency. Having verified the method, we then derived and analysed RZ _WC estimates across global terrestrial biomes.