Response of crop photosynthesis to carbon dioxide, temperature, and light : experimentation and modeling

V. Reddy,L. Pachepsky,B. Acock
DOI: https://doi.org/10.21273/HORTSCI.29.12.1415
1994-12-01
Abstract:ment in the prediction of overall fruit mass accretion were obtained by GAGossym. Running GAGossym with the full parameter set resulted in an overall fitness improvement of ≈25% over all output types normally used to assess fit. Our discovery of the potential of GAs to derive improved parameter sets from high-level data will have profound implications on the speed at which we can develop locality-specific systems and on the accuracy of these systems to perform at that specific site. The GA was restricted to investigate parameter ranges that were within the realm of the field-observed variability. That is, parameters continued to maintain the original biological interpretations. We have demonstrated that we can develop a fine-tuned system that responds to local environments optimally given logical constraints on parameter ranges and functional forms. GOSSYM is currently used throughout U.S. cotton-growing regions. A self-correcting feature was developed that allowed the model to rederive its parameters when field data indicated that forecasts were beyond an arbitrarily set threshold of acceptability. The process of finding the optimal configuration of parameter sets has been completely automated. The process of model correction thus depends only on the model’s perceived goodness of fit. If it does not fit data well, it will invoke its associated GA and evolve a better set of parameters. This research implies that models may be applicable and have improved forecast efficacy over a wider range of conditions than is presently the case. This improvement in performance and applicability will affect natural resource management.
What problem does this paper attempt to address?