Wheat yield modeling in relation to agrometeorological data for western zone of Haryana
Loading...
Date
2014
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
CCSHAU
Abstract
An efficient crop forecasting infrastructure is pre-requisite for information system about food
supply, especially export-import policies, procurement and price fixation. Various statistical
approaches are in vogue for arriving at crop forecasts. The present study has been conducted to develop
trend-agromet models on agro-climatic zone basis for pre-harvest wheat yield forecasting at district
level in Haryana. The fortnightly weather data have been used for the development of zonal yield
models following two different statistical procedures viz., multiple linear regression and principal
component analysis technique. The developed zonal models are based on the time-series data (1978-79
to 2009-2010) on weather parameters and trend predicted yield and the data from 2010-11 to 2012-13
were used for validation of the models. Year/time was included to take care of variation between
districts within the zone as the weather data is not available for all the districts, however, the zonal
model utilized the same weather data in the adjoining districts under the zone thus a longer series could
be obtained in a relatively shorter period and also provided the basis to use multivariate statistical
analyses.
Forecast models via step-wise regression have been fitted, taking yield as the dependent
variable and weather indices/parameters or principal component scores and trend yield as the
regressors. The significance of the Bartlett’s test of sphericity confirmed the presence of
multicollinearity among the weather variables used in regression analysis. The performance of the
developed models has been compared on the basis of different statistics i.e. adj-R2, percent deviation of
the forecast from the observed yield and root mean square error (RMSE). Adequacy of the fitted model
was examined through histogram, normal- probability plot and residual plot against fitted values. A
perusal of the results indicates the preference of using prediction equations based on principal
component scores over the model development period and the model testing period. From the fitted
models, wheat yield forecasts for the years 2010-11, 2011-12 and 2012-13 have been obtained for all
the districts of the western zone viz, Sirsa, Fatehabad, Hisar and Bhiwani.
The results indicate that the district-level yield prediction gives good agreement with DOA
yield estimates for almost all the four districts in western zone of Haryana. Moreover, the developed
models provide reliable forecasts of wheat yield about 4-5 weeks in advance before the harvest of the
crop while real time yield data are obtained quite late after the actual harvest of the crop.
Description
Keywords
Yields, Forecasting, Crops, Wheats, Statistical methods, Environmental degradation, Crop yield, Yield forecasting, Land management, Productivity