Yield forecast model for wheat (Triticum aestivum) crop

Abstract

The experimental data collected from experimental plots of the university, Junagarh as a part of a research project
on cropping system were utilised to develop wheat (Triticum aestivum L.emend. Fiori & Paol). forecast model based
on the biometrical characters such as number of tillers, height of plants and the length of earhead. Two types of
functions, ie linear and quadratic were fitted. It was found that the explanatory variables could explain the variation in
yield up to 62% if they were considered to be linearly related hut 68% provided the function was of quadratic in nature.
The linear type of model was more valid than a quadratic one. This forecast model could successfully be used for
obtaining advance estimates of wheat yield tor a period of 4 years. Our results also show that forecast models with
longer time series data are preferred because of the yield forecast under them is very close to the observed yield.