Observational Evaluation of Land-Atmosphere Coupling in Model Hindcasts on the U.S. Southern Great Plains

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 5 January 2015
Phoenix Convention Center - West and North Buildings
Thomas J. Phillips, LLNL, Livermore, CA; and S. A. Klein

In a recent study of observed features of land-atmosphere coupling (LAC) at the ARM Southern Great Plains (ARM SGP) site in northern Oklahoma (Phillips and Klein, 2014 Journal of Geophysical Research), we identified statistically significant interactions between 1997-2008 summertime daily averages of soil moisture (at 10 cm depth) and a number of surface atmospheric variables, such as surface evaporation, relative humidity, and temperature.

Here we will report on an evaluation of similar features of LAC simulated by version 5 of the global Community Atmosphere Model (CAM5), coupled to its native CLM4 land model, and downscaled to the vicinity of the ARM SGP site. In these case studies, the CAM5 was initialized from a 6-hourly atmospheric reanalysis for each day of the years 2008 and 2009 (where the CLM4 land state was equilibrated to the atmospheric model state), thus permitting a close comparison of the modeled and observed summer daily average features of the LAC in these years. Correlation coefficients R and “sensitivity indices” I (a measure of the comparative change of an atmospheric variable for a one-standard-deviation change in soil moisture) provided quantitative measures of the respective coupling strengths.

Such a comparison of observed versus modeled LAC is complicated by differences in atmospheric forcings of the land; for example, the CAM5's summertime precipitation is too scant, and thus the model's upper soil layer often is drier than observed. The modeled daily average covariations of soil moisture with lower atmospheric variables also display less coherence (lower R values), but sometimes greater “sensitivity” (higher I values) than are observed at the ARM SGP site. Since the observational estimate of LAC may itself be sensitive to soil moisture measurement biases, we also will report on a planned investigation of the dependence of LAC on several alternative choices of soil moisture data sets local to the ARM SGP site.

Acknowledgments This work was funded by the U.S. Department of Energy Office of Science and was performed at the Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.