Assessment of WOFOST Model for Yield Prediction of Rice in Paddy Lands of Shaft City Using Spatial Analysis of Geographic Information System (GIS)

The present study was conducted to quantify rice yield in Shaft city, located in Gilan province, crop production zoning map was prepared through the Geographic Information System (GIS) approach.

Therefore, farm information of 100 points from the paddy fields of the region was received in two crop years (2012 and 2013) for simulation by WOFOST model. The available data were measured for two categories of local and high-yielding varieties in the region.

The results of the model did not show a significant difference between the observed values of rice grain yield with the simulated values (Sig <0.05). The value of coefficient of determination (R2) was obtained for both groups of varieties was about 0.80, and the root mean square of normalized error (NRMSE) was 13% and 7% for local and high-yielding varieties, respectively. Evaluation of statistical results indicates the favorable similarity of real and estimated data. Also, statistical indicators show that the model has less error in estimating yields of paddy fields with local varieties than high-yield varieties. In large-scale studies, the estimation error is mostly due to differences in crop management. The spatial distribution of yield has intensities and fluctuations in the study area, which can be in response to different amounts of inputs, delay in planting date, off-season rainfall and soil variability and soil physiographic status.

Therefore, the use of simulation models can be effective in achieving sustainable rice production and economic savings of agricultural inputs in the region and provide a landscape of production conditions for farmers and managers.