About Us | Help Videos | Contact Us | Subscriptions

Agronomy Journal Abstract - Biometry, Modeling & Statistics

Uncertainty Analysis and Parameter Estimation for the CSM-CROPGRO-Cotton Model


This article in AJ

  1. Vol. 104 No. 5, p. 1363-1373
    Received: Oct 26, 2011

    * Corresponding author(s): tpathak2@unl.edu
Request Permissions

  1. Tapan B. Pathak *a,
  2. James W. Jonesb,
  3. Clyde W. Fraisseb,
  4. David Wrightc and
  5. Gerit Hoogenboomd
  1. a School of Natural Resources, Univ. of Nebraska, Lincoln, NE
    b Dep. of Agricultural and Biological Engineering, Univ. of Florida, Gainesville, FL
    c North Florida Education and Research Center, Univ. of Florida, Quincy, FL
    d AgWeatherNet, Washington State Univ., Prosser, WA


The Cropping System Model (CSM)-CROPGRO-Cotton model simulates growth, development, and yield for cotton (Gossypium hirsutum L.) and requires a large number of parameters and inputs. It is practically impossible to estimate all the parameters with a high level of accuracy. The objectives of this study were to estimate values and uncertainties of the CSM-CROPGRO-Cotton model genotype parameters and resulting uncertainties in the cotton growth and development model and to evaluate the model performance by comparing model predictions based on the estimated parameters using generalized likelihood uncertainty estimation (GLUE) with those based on a prior distribution and the default model parameters. Observations for this study were collected from four experiments at Quincy, FL, Citra, FL, and Griffin, GA, for Delta Pine 555 cotton. The results show that the estimated parameters improved model performance compared with default parameters. There was a noticeable reduction in parameter uncertainties and resulting model output uncertainties. Resulting uncertainty in model predictions of leaf area index, leaf weight, stem weight, and boll weight were reduced to 4 to 13% with posterior and 29 to 56% prior parameters, partially due to improved estimates and a considerable reduction in the uncertainty of the important model parameters such as the light extinction coefficient and specific leaf area. The GLUE technique improved the model performance by improving estimated values of model parameters, reducing model parameter uncertainties, and reducing model output uncertainties.

  Please view the pdf by using the Full Text (PDF) link under 'View' to the left.

Copyright © 2012. Copyright © 2012 by the American Society of Agronomy, Inc.