Tuesday, 24 January 2017: 11:45 AM
604 (Washington State Convention Center )
Ian Baker, Colorado State Univ., Fort Collins, CO; and P. J. Sellers, S. Denning, I. medina, P. kraus, K. Haynes, N. Geyer, and
D. C. Smith
The lack of sub-grid variability in General Circulation Models (GCMs) that use a single value to represent heterogeneous quantities has been a persistent problem. Soil moisture is a strong control on evapotranspiration (ET), and can vary on horizontal scales from 1-10 meters, yet is usually resolved as a single, vertically resolved column in models. Convective precipitation may influence an extremely small fraction of a GCM gridcell with spatial resolution on the order of 10’s to 100’s of km. If convective precipitation is allowed to fall on the entire gridcell, it falls as mist and evaporates off leaves with little or no chance to reach the ground and infiltrate. Cumulus precipitation can be concentrated on a fraction of the gridcell, allowing saturation of leaf interception storage, runoff, and infiltration into the ground, but in this case the infiltration reaches the entire gridcell. Both of these solutions are unrealistic.
We have developed a method to represent sub-grid soil moisture variability in a spectrum of wetness ‘bins’, with variable relative area. Each bin is associated with a observationally-based function that describes soil moisture control on ET, and the bins method has been demonstrated competent in site-level evaluations. Bins also provides a means to couple cumulus precipitation to the surface in a realistic manner, allowing convection to impact a fraction of the total gridcell. Here, we perform a decoupled (land-only) simulation of surface behavior over the ARM-Southern Great Plains domain, driving the model with MERRA surface reanalysis products. The bins simulations are compared with traditional simulations at individual sites as well as satellite products that observe soil moisture and Solar-Induced Fluorescence. We show that subgrid heterogeneity in soil moisture, and its influence on surface status, Bowen ratio and carbon flux, is represented by the bins simulations in a way that traditional land models are unable to reproduce.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner