The AgroEcoSystem (AgES) Response‐Function Model Simulates Layered Soil‐Water Dynamics in Semiarid Colorado: Sensitivity and Calibration. Issue 8 (21st August 2015)

• Record Type: Journal Article
• Authors: Green, Timothy R.; Erskine, Robert H.; Coleman, Michael L.; David, Olaf; Ascough, James C.; Kipka, Holm
• Is Part Of: Vadose zone journal. Volume 14:Issue 8(2015)
• Page Start: 1
• Page End: 16
• Publication Date: 2015-08-21
• Subjects: Soil science -- Periodicals; Zone of aeration -- Periodicals; Groundwater flow -- Periodicals; Groundwater flow; Zone of aeration; Periodicals; Electronic journals; 631.4
• DOI: 10.2136/vzj2014.09.0119
• Languages: English
• ISSNs: 1539-1663

Abstract

Abstract : Simulation of vertical soil hydrology is a critical component of predicting more complex multidimensional soil water dynamics in space and time. The AgroEcoSystem (AgES) model is identified here as a single land‐unit application of the three‐dimensional AgES‐W (Watershed) model. AgES simulates vertical soil water dynamics using global and layered soil response functions with conceptual storages as state variables. A detailed description of the response functions that control infiltration, evaporation, and soil‐water processes facilitates sensitivity analysis, model calibration, and evaluation against volumetric soil‐water content (SWC) at measured layers. The Object Modeling System links AgES to a Shuffled Complex Evolution calibration tool called Luca. We used Luca and fractional factorial experimental designs to analyze parameter sensitivities, then applied different strategies of implementing Luca to layered SWC data. The profile dynamics of the simulated SWC resulted in depth‐averaged Nash–Sutcliffe Efficiency (NSE) values of 0.60 to 0.95 for calibration in 2003 and 2005, and up to 0.80 for 4 yr used for (cross‐)evaluation. Using the 2005 calibration parameters, NSE became negative in 2009 and 2011 due to large negative values at some depths with low variance in SWC. Optimal parameter sets for each calibration year were not unique, and model results did not fully capture the measured dynamics. Even so, AgES simulations compared favorably with previous simulations of SWC at this site using a Richards' equation model. These results provide new understanding of the model responses and interactions between functions controlling the vertical flow and storage of water to aid watershed modeling.