Evaluating and improving prediction of land surface forcing over the U.S. Southern Great Plains

 

Authors

Justin Bagley — Lawrence Berkeley National Laboratory
Ian N. Williams — Lawrence Berkeley National Laboratory
Yaqiong Lu — National Center for Atmospheric Research
Sebastien Christophe Biraud — Lawrence Berkeley National Laboratory
Lara Kueppers — Lawrence Berkeley National Laboratory
Margaret S. Torn — Lawrence Berkeley National Laboratory

Category

Boundary layer structure, including land-atmosphere interactions and turbulence

Description

Poor representation of the land surface can contribute to prediction biases in weather and climate models. Accurate prediction of land surface forcing is challenging due to spatial heterogeneity and temporal variability in land surface characteristics. We used 13 years of measurements from the Atmospheric Radiation Measurements (ARM) Climate Research Facility, to characterize land surface forcing over the U.S. Southern Great Plains (SGP) and to evaluate and improve a version of the Community Land Model (CLM) commonly used in cloud and precipitation studies over land. Additionally, we advanced the capability of characterizing the land surface at high spatiotemporal resolution over the SGP, by implementing and improving a winter wheat plant functional type (PFT) in CLM. Observed latent and sensible heat fluxes were strongly influenced by differences in leaf area index (LAI) seasonality between winter wheat and other vegetation types, at times by more than 100 W m-2. This impact was similar in magnitude to that of the strong east-west precipitation gradient across the region. Although spatially-averaged LAI in the default CLM was largely consistent with observations, the default CLM failed to capture the observed distinction in LAI seasonality between winter wheat and pasture grasses. These errors degraded the prediction of surface water and energy fluxes at the grid-scale, and point to needed improvements in subgrid-scale representation of vegetation. Toward this goal, we implemented and improved a winter wheat PFT in a version of CLM with prognostic LAI. The improved prognostic version substantially improved subgrid-scale representation of LAI compared to the default CLM, and shows promise as an approach to represent land surface forcing at high spatiotemporal resolution in weather and climate models.