1Department of Statistics, University of Pune, Pune, Maharashtra
2Department of Statistics, Annamalai University, Anna Nagar 6080 02, Tamil Nadu
This paper seeks to answer two questions: first, how early can we predict jowar yield? and second, how much of variation in yield can be attributed to weather? Early prediction of crop yield is of considerable interest to the agriculturists, economists, policy planners, etc. Many attempts at early (preharvest) forecast have led to limited success. One common direction of such efforts is through regressing yield on biometrical characters of the crop. The other direction uses weather variables, mainly rainfall. This study has been based on a combination of biometrical and weather variables. The selected multiple regression equations have given the R-values ranging between 0.4 and 0.98. To answer the second question, the authors have carried out the analysis of variance (ANOVA) under a nested mode!. The contribution of weather to variability in yield has turned out to be about 39 per cent. The contribution due to variability between stations has been found to be about 43 per cent.
