Multi-variate evaluation of land surface model performance in a semi-arid, complex terrain pine forest

Significant challenges inhibit progress in understanding and simulating land-atmosphere exchanges of energy and moisture in complex terrain regions due to strong spatial heterogeneity in atmospheric forcing conditions as well as in land surface characteristics. In this study we present a new comprehensive data set for land surface model evaluation collected from a semi-arid, sparse canopy, ponderosa pine forest located in the Front Range-foothill region of the Colorado Rocky Mountains. The observing network was designed to characterize footprint scale variations in precipitation, soil moisture and soil temperature. Area integrated measures of moisture variables such as near surface water content and snow water equivalent are also being collected using a cosmic-ray neutron detection system. Additional land-surface model prognostic variables such as infra-red skin temperature, plant transpiration via sapflow measurements and sensible and latent heat fluxes are being measured as well. Using this dataset we evaluate the performance of 2 recently-updated community land surface schemes (Noah and CLM) in their ability to represent the salient details of patch scale and area integrated energy and water balances and land surface fluxes. Emphasis in the evaluation is placed on understanding the necessary level of heterogeneity representation in forcing data and in simulating the appropriate partitioning of moisture fluxes between evaporation, transpiration and, in the case of wintertime conditions, sublimation from an ephemeral snow pack.