A Case Study of Crop Model Applications in an Increasing Diversity of Genetically Modified Traits.

GOSSYM is a mechanistic, process level cotton simulation model which can simulate crop growth, development, and final yield under a wide range of environmental, soil, management conditions for most cotton cultivars. The model requires a variety file to simulate each cultivar. The variety file contains 60 parameters which represents growth and developmental traits of a variety and enables GOSSYM to simulate variety response to the applied set of inputs and weather. Recently, the use of Genetically Modified (GM) cultivars has increased in US Cottonbelt and in 2010 GM cotton was cultivated on 93% of total cultivated area. At present, GOSSYM database has no variety files representing GM cotton cultivars.

A study was conducted to develop variety files for six GM cultivars. Several experiments were conducted between 2001 and 2006 to obtain data needed to generate cultivar-specific information and to validate the model under wide range of environmental and management scenarios in Mississippi. GM cultivars were selected considering their popularity among the growers. Weather and soil and management practices data required as input for the model was also collected at each of these locations. During 2001-2002 growing season, three experiments consisting of four nitrogen (N) rates, two irrigation levels and four plant growth regulator (PIX) application treatments were studied in two different soil types. During 2003-2004 growing season, interaction between four N and four PIX levels were investigated on two different soil types. During 2005-2006 growing season, four levels of N applied to four cultivars grown on two different soil types. In each experiment, plant height and numbers of main stem nodes, squares, bolls and biomass of various parts of plant were measured several times during the growing season. In addition, major phenological stages such as days to square, flower and open boll were recorded in each experiment along with lint yield. This data represents carbohydrate partitioning and plant development throughout the growing season.

The cotton model, GOSSYM, was used to simulate the growth, development and yield of six GM cultivars across years and management practices. We found that 25 cultivar parameters of variety file that included phenology, growth, and partitioning coefficients, are needed to be changed to simulate the GM cultivars. Graphical and statistical methods were used to test the predictive capacity of the model for GM cultivars. The simulated growth, phenology, and yield data are very close with observed data showing the robustness of the functions used in the model. The correlation coefficients between observed and simulated yields for all varieties ranged between 0.56 and 0.87. This study also proves that, the validated cotton simulation model with GM cultivar variety files will be immensely beneficial to several federal and state agencies and consultants who are currently using the GOSSYM cotton model for natural resource management as an on-farm decision aid, and in climate change impact analysis and assessment scenarios.