On the Value of Effective Parameters Obtained Under Reduced Spatial Scale Mismatch to Represent Soil Moisture - Evapotranspiration Interactions in Land Surface Models.

Abstract:

Soil moisture - evapotranspiration interactions have been studied in several regions characterized by distinct soil properties, land cover types, and climate. Such interactions are usually assessed with measurements representing different spatial footprint. For example, soil moisture measurements obtained from point-scale sensors or with remote sensing products are typically compared with evapotranspiration measurements obtained with eddy covariance systems. With efforts to develop hydrometeorological models capable of simulating processes at hyper-resolution (i.e., 1 sq-km), novel approaches for intermediate-scale soil moisture measurements give us new opportunities to evaluate the representation of soil moisture and evapotranspiration processes at similar spatial scales. Here, we evaluate the performance of the Joint UK Land Environment Simulator (JULES) in which key parameters are determined effectively based on both soil moisture and evapotranspiration measurements obtained with similar horizontal footprint. We use soil moisture data from selected sites in the COsmic-ray Soil Moisture Observing System (COSMOS) network in combination with co-located Ameriflux eddy covariance towers to constrain key parameters in JULES assuming their similar horizontal footprint of hundreds of meters. The COSMOS-Ameriflux sites are characterized by distinct soils, land cover types, and climate. In addition, point-scale soil moisture at each site is also used in combination with eddy covariance measurements to constrain JULES parameters, while recognizing its much smaller support volume. Each individual site is ranked based on differences in soil moisture dynamics from both point-scale and intermediate-scale measurements. Model calibration is carried out by optimizing JULES performance against (1) point-scale soil moisture only, (2) cosmic-ray soil moisture only, (3) point-scale soil moisture and surface fluxes, and (4) cosmic-ray soil moisture and surface fluxes. We then evaluate how “informative” each measurement type or combination is when constraining model parameters, and consequently, its potential impact on simultaneously simulating soil moisture dynamics and energy partitioning at sub-kilometer scale.