Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Adaptive Cropping Systems Laboratory » Research » Publications at this Location » Publication #253715
Title: Evaluation of modeling bias and uncertainty using different field capacity estimates in the APEX model

Author
item NEMES, ATTILA - University Of Maryland
item Pachepsky, Yakov
item Timlin, Dennis

Submitted to: American Society of Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 5/12/2010
Publication Date: 10/30/2010
Citation: Nemes, A., Pachepsky, Y.A., Timlin, D.J. 2010. Evaluation of modeling bias and uncertainty using different field capacity estimates in the APEX model. [abstract] American Society of Agronomy. 327-14..

Interpretive Summary:

Technical Abstract: The Conservation Effects Assessment Project (CEAP) uses the Agricultural Policy Environmental Extender (APEX) simulation model to assess the effects and benefits of Farm Bill conservation programs on U.S. croplands. The APEX model uses the 'field capacity' (FC) concept to represent soil water conditions. Recommendations differ world-wide on how to represent FC of the soil in a 'bucket' type water movement model like APEX is. In the U.S., FC is usually approximated by a laboratory measured or estimated water retention value at -33 kPa pressure head. However, there are continued reports that there is no single pressure head that can be reliably used to approximate FC for all soil textures. We tested using (1) APEX's built in estimation functions; (2) a k-nearest neighbor based estimation proposed by Nemes et al., 2009 and (3) a field measurement based correction factor to -33 kPa estimates to parameterize APEX. We subsequently ran the APEX model to characterize the impact of parameterization on the bias and uncertainty of modeled outputs using simulations of nutrient and water management and crop growth under different weather conditions. Regression tree methodology and commonly used statistical measures were used to delineate groups of soil and weather conditions for which a significant impact by different parameterization scenarios can be expected. While the general estimation uncertainty was less affected, substantial differences were found in the range of model estimates when the different FC estimates were used. A large proportion of these findings could be explained by systematic differences between FC estimates that were identified earlier. Care should be taken when parameterizing the hydrology component of models in large scale studies, in which the use of measured soil hydraulic properties is not a feasible option.