Rajarathinam Arunachalam

Trends in Crop Response in Long-term Field Experiment: Statistical Modeling Rajarathinam Arunachalam

In long-term experiments on fertilizer nutrients and manures, the original data are of zig-zag in nature and there is a need to apply trend-fitting techniques to develop smooth trends to enable interpretation of the data in a better way. Very few attempts have been made in this direction. Present investigation on two long-term experiments was undertaken (i) to fit the trends in crop yields due to various treatments and (ii) to study the characteristics and to compare the trends for assessing the crop response to manure and fertilizer nutrients. The treatment-wise yield data for two crop sequences viz., groundnut-wheat-fodder sorghum at Junagadh and paddy-wheat-greengram at Navsari centers of the then Gujarat Agricultural University for 14 to 18 years were used. Two different approaches viz., parametric and non-parametric regression methods were employed to fit the trend on yield due to different treatments over the years. In parametric regression approach, polynomial and Auto Regressive Integrated Moving Average (ARIMA) methods were employed. Eight different polynomial models with varying degrees were tried using the moving average values. Among the eight polynomial m