Some statistical models for crop yield forecasting
Loading...
Date
2009-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
Crop yield forecasting is an important aspect for a developing economy so that adequate planning exercise is undertaken for sustainable growth and overall development of the country. Weather fluctuations affect crop yield significantly during different stages of crop growing season, therefore several studies have been carried out to forecast crop yield using weather parameters. However, such forecast studies based on statistical models need to be done on continuing basis and for different agro-climatic zones, due to visible effects of changing environment conditions and weather shifts at different locations and areas. Therefore, present study was undertaken for forecasting yield of two major crops viz. rice and wheat based on time series data for 27 years (w.e.f. 1981-82 to 2007-08) of yield and weather parameters obtained from G. B. Pant
University of Agriculture and Technology, Pantnagar, District Udham Singh Nagar, Uttarakhand, India. In this investigation an attempt has been made to develop statistical models for crop yield forecasting using various statistical techniques incorporating technical and statistical indicators. Two computer programs were developed in FORTRAN 77 language for generating the predictors used in proposed modified models. On the basis of developed statistical models following conclusions were drawn. I. Rice yield was found to have significant positive linear correlation with Average Weekly Maximum Temperature (4th week), Average Weekly Relative Humidity at 14.00 hrs (13th week) and Average Weekly Total Rainfall (13th week). However, rice yield was found to have significant non linear relationship (of the type Y =ab X ) with Average Weekly Max. Temp (4th week and 13th weeks), Average Weekly Total Rainfall (10th week and 13th weeks), Average Weekly Relative Humidity at 14.00 hrs (13th weeks) and Average Weekly Sunshine hrs (13th weeks). II. Wheat yield was found to have significant and positive linear correlation with Average Weekly Minimum Temperature (10th week), Average Weekly Sunshine hrs (20th week), and Average Weekly Pan Evaporation (20th week). However, wheat yield was found to have significant non linear relationship (of the type Y =ab X ) with Average Weekly Max. Temp (16th, 17th and 18th weeks), Average Weekly Min. Temp (10th week and 18th weeks), Average Weekly Relative Humidity at 7.00 hrs (20th weeks), Average Weekly Sunshine hrs (20th weeks) and Average Weekly Pan evaporation (20th weeks).
III. For forecasting the yield of rice and wheat various statistical models viz. Model I
and Model II (using Linear and Non Linear Regression Analysis), Model III (MLR) and Model IV developed by Hendricks and Scholl (1943) and Model V developed by Agrawal et al (2001) were applied. In addition to these models, two other modified models (Modified Model IV and Model VI) were suggested.
IV. It was found that proposed Modified Model VI (A3) based on technical and statistical indicators for forecasting the rice yield was better than Model V (A5) suggested by Agrawal et al. (2001).
V. It was found that proposed Modified Model VI (B5) based on technical and statistical indicators for forecasting the wheat yield was also better than Model V (B6) suggested by Agrawal et al. (2001).