Thesis Abstract of AGS Students

Calibration of genetic coefficients of paddy rice (Oryza sativa L.) for validation of the CERES-Rice model in northern Thailand

Panya Mankeb (1993)

Crop simulation models can be used as a tool for agricultural risk analysis. They allow researchers to explore potential cropping location and appropriate farm management strategies. The CERES-Rice model is one such crop model. This model is intended to have global application, therefore, genetic coefficients are an essential model component. Crop performance is terms of genetic coefficients used in the model can be used as a tool in choosing varieties and extend the utility of field experimentation. Once a crop model has been validated, it can be used to match variety and site rather than carry out extend field experiments.

The objectives of this study are to calibrate suitable genetic coefficients of rice varieties which are commonly grown in Northern Thailand for the CERES-Rice model and to validate the model.

Field experiment was carried out from April to December 1991 under irrigated conditions at the Multiple Cropping Centre Experiment Station, Chiang Mai University. The experimental design was split plot with three replications in which planting dates (May, June, July, August and September) were main plots and varieties (Niaw San Pa Tong, and Kaow Dawk Mali 105 and RD 7) were sub plots. The model was validated using data from yield trail experiments consist of three varieties which are Niaw San Pa Tong, Kaow Dawk Mali 105 and RD 7 from San Pa Tong Rice Experiment Station, Chiang Mai.

The simulation results indicate that the model is able to simulate phenological events i.e., heading and maturity dates satisfactory for all varieties regardless of planting date. However, the simulation of crop growth and yield do not conform with the observed data except the 1000-grain weight and LAI of RD 7. The model overestimates tiller numbers m-2 and panicle numbers m-2. However, the model underestimates above ground biomass m-2 which is probably due to the model greatly underestimates stem dry masses but slightly overestimates leaf and panicle dry masses. In addition, the model does not permit partitioning of assimilate to stem after heading while the observed results show continously increase of stem dry masses after heading. The overestimation of crop growth and yield indicate the potential value of each variety. This is because the model does not take into account of pest, diseases and lodging damage. The model also assume that spikelets formed will develop into grains that will eventually be harvested. However, field observations show that there is yield loss due to unfilled spikelets, pest (rat, bird and insect), diseases, lodging and shattering damage.

Results of model validation illustrate that the model is capable of simulating phenological events satisfactory but overestimates grain yield. These results confirm the consistency of model performance using the calibrated set of genetic coefficients. Generally speaking, the model need further validation of nitrogen and water treatments in order to increase confidence level prior to the application of the model.

Back to Thesis Topics