A method for identifying soil properties influencing crop yield: the example of maize in the Savanna Belt of South-western Nigeria

Abstract

This paper describes a method of identifying the crucial soil properties influencing yields of crops, using the example of maize in south-western Nigeria. The method involves first, the examination of the relationship between soil properties and the maize parameters using the simple bivariate correlation analysis and the multiple regression model. It also involves, in part, the computation of an index of soil variable contribution to maize prediction in order to identify the crucial soil properties influencing the yield of the crop. In the current study, the multiple regression model reduces the 20 soil properties analysed to the 14 that have made significant contributions to the crop's prediction. The 14 variables were later reduced to two using the maize-prediction index. The index has the merits of being very simple to compute and the ability to reduce the numerous variables to few significant ones as far as the crop under investigation is concerned.