Observation-based assessments of surface flux partitioning regimes in 4 commonly-used land surface models

Flux partitions between surface water and energy terms are essentially important to the climate system. They can potentially affect assessments of climate risk projections in the future. However, the characterization of surface flux partitioning in numerical models is rarely evaluated due to the absence of large-scale observational evidence. Here, we use long-term satellite datasets and observational meteorological records to evaluate the flux partitioning regime presented in four widely-used Land surface models (LSMs) over two study regions (i.e., China and Continental U.S.). We show that the regime in LSMs differs significantly from satellite-based estimations, which can be due to unrealistic representations of land surface characteristics. The biases in models’ flux partitioning regime may lead to the underestimated potential for climate risks, especially over regions with typical land surface characteristics. The results highlight that particular attention should be paid to the calibration of surface flux partitioning regimes in LSMs. Large model spreads in surface flux partitioning strength and climate risk maps are also reported.