PREDICTION OF RICE YIELD WITH DSSAT CROP SIMULATION MODEL AND MULTIPLE LINEAR REGRESSION ANALYSIS

Rice is one of the most important food crop of India in term of area, production and consumer preference. West Bengal is one of the leading states of India which is known as bowl of rice, as it cultivate and produce a major portion of rice of the whole produce from 90% of its agricultural field. Weather variability has a significant impact on crop growth and development. Timely and accurate crop yield forecasts are essential for crop production, marketing, storage and transportation, decision making and managing the risk associated with these activities. Weather variable includes maximum and minimum air temperature, total solar radiation, relative humidity (morning and afternoon) and total rainfall. In this study two different approach CERES-Rice model and Multiple Linear Regressions analysis were used for predicting rice yield for West Medinipur district of West Benagal, India. DSSAT CERES-Rice model was used to simulate the seasonal yield with long term weather data and to establish the relationship between weather variables and crop yield and forecast rice yield (1983-2010). Statistical technique includes dependent variable namely yield data of a crop and independent variable namely daily weather data, arranged weekly from sowing to maturity and time (year) as technology trend. The relation between weather parameters and crop yield was determined by using statistical tool like correlation and linear regression analysis. The results of the study shows that the average relative root mean square error computed for the year 2011 and 2012 by regression analysis were 131.5 and 108.4 respectively. The Mean Absolute Error 11.7 & 9.47 for same duration.