Using ARM observations to test soil moisture dynamics in climate models

Abstract:

Potential changes in soil moisture may have significant societal impacts, as soil moisture directly influences agriculture. Soil moisture is also a critical factor in climate simulations as it is the moisture source for evapotranspiration over land. Climate model projections generally show reduced soil moisture in future warmer climate conditions, and the scale of potential adverse impacts means that validation of those projections is a science priority. Our understanding of soil moisture dynamics is hampered by limited suitable observational data, but the Southern Great Plains (SGP) Atmospheric Radiation Measurement (ARM) site offers a unique resource for this purpose, with over a decade of simultaneous measurements of soil moisture profiles and measurements of moisture fluxes and aboveground variables. In this work we use stations across SGP to identify statistical relationships in drivers of soil moisture dynamics. We also compare observed soil moisture dynamics to those in the Community Land Model version 4 (CLM4), running CLM4 in an offline mode with observationally derived atmospheric forcing, to identify similarities and discrepancies in resulting soil moisture evolution in observations and model. Preliminary comparison of metrics such as soil moisture characteristic time, soil moisture infiltration rate, etc. suggests that the governing hydrological and/or biophysical processes in models need improvements. The spatial heterogeneity of the SGP measurement stations also provides insight into the role of sub-grid scale features and the role of spatial resolution in producing accurate representations of soil moisture in climate models.