Aggregating over land surface heterogeneity systematically biases evapotranspiration estimates in large-scale evaporation models

Land surface models are highly uncertain in estimating evapotranspiration (ET) fluxes, and differ substantially in their projections of how ET will evolve in the future. Biases in estimated ET fluxes will affect the partitioning between sensible and latent heat, and thus alter simulated temperatures and model predictions of droughts and heatwaves. One potential source of bias is the "aggregation bias" that arises whenever nonlinear processes, such as those that regulate ET fluxes, are modeled using averages of heterogeneous inputs. Here we demonstrate that this aggregation bias leads to substantial overestimates in ET fluxes in a typical large-scale land surface model. The proposed methodology can be used to correct for aggregation biases in ET estimates by quantifying the effects of finer-resolution spatiotemporal variability in ET drivers at each modeling time step, without explicitly representing sub-grid heterogeneities in large-scale land surface models.