Overcoming the Problem of Scale: Evaluation of an Agroecosystem Model Using Cosmic-Ray Neutron Soil Moisture.

The properties of the land surface affect the interaction of the surface and the atmosphere. The partitioning of absorbed shortwave radiation into emitted radiation, sensible heat flux, latent heat flux, and soil heat flux is determined by the presence of soil moisture. When the land surface is dry, there will be higher sensible heat flux and emitted radiation. However, when water is present in the soil, energy will be used to change the phase of water from solid to liquid and liquid to gas. The resulting latent heat flux moves water into the atmosphere and changes the way in which the atmosphere is heated, subsequently affecting weather and climate. In the Midwest U.S., the management of fields changes the latent heat flux through different crop choices, planting and harvest date, fertilizer application, and tile drainage. Therefore, land surface models which simulate land-atmosphere interactions need to be evaluated at the field-scale. Agro-IBIS is a land surface model that is able to incorporate changes in vegetation growth as well as management practices, which in turn impact soil moisture available for latent heat flux. In measuring soil moisture, a consistent challenge is the representative scale of the instrument, which is often a point. A newer method of obtaining soil moisture at the field-scale is to use a cosmic-ray neutron detector, which is sensitive to an area of ~700 m in diameter and a depth of ~20 cm. We performed a soil texture sensitivity analysis using Agro-IBIS to determine what soil texture best represented, in terms of hydraulic properties, the variety of soils found within the field-scale. For a few soil textures, the results show an RMSE of less than 0.04 cm³ cm^-3 and accuracy within 10 bu/ac of maize yield, demonstrating that the Agro-IBIS is in good agreement with the field-scale observations of soil moisture and yield for the growing season in 2011. At point measurement sites, however, the model performance is mixed because of the variability in soil moisture and yield with soil texture and landscape position. We argue that land surface models and weather and climate models should be evaluated at the field-scale instead of at individual points.