Revisiting large-scale interception patterns constrained by a synthesis of global experimental data

Abstract. Rainfall interception loss remains one of the most uncertain fluxes in global water balance, hindering management forested regions and precluding an accurate formulation climate models. Here, a synthesis data from past field experiments conducted worldwide is performed, resulting meta-analysis comprising 166 forest sites 17 agricultural plots. This used to constrain process-based model driven by satellite-observed vegetation dynamics, potential evaporation precipitation. The considers sub-grid heterogeneity dynamics formulates rainfall for tall short separately. A global, 40-year (1980–2019), 0.1∘ spatial resolution, daily temporal resolution dataset created, analysed validated against situ data. validation shows good consistency between modelled observations over vegetation, both terms correlations bias. While underestimation found degree which it responds representativeness errors difficulties inherent measurement vegetated ecosystems unclear. Global estimates are compared existing datasets, showing overall comparable patterns. According our findings, averages 73.81 mm yr−1 or 10.96 × 103 km3 yr−1, accounting 10.53 % continental approximately 14.06 terrestrial evaporation. seasonal variability follows annual cycle canopy cover, precipitation, atmospheric demand water. Tropical rainforests show low intra-annual variability, patterns dictated rainfall. Interception strong variance among types biomes, supported modelling new will serve as benchmark future investigations facilitate large-scale hydrological research.