Water balance approach to maize yield forecasting for katumani area

Abstract

In this study, maize yields at Katumani have been related to two sets of weather indices. These indices, known as the water requirement satisfaction indices were computed according to a model developed by Frere and Popov (1979). One index was computed using crop coefficients given by Food and Agricultural Organization (FAO, 1986) and-the other by Mugah and Stewart (1982). The indices were set at 100 at the time of emergence and started reducing the moment the crop underwent water stress. The analysis showed that changes of available soil water during the flowering to wax ripeness phenological phases were the most sensitive as they affected the final maize yields the most. This was followed by the tasselling to flowering interphase. When the ratio of the interphase available water to the interphase water requirement was correlated to yields, the flowering to wax ripeness interphase showed the highest correlation. Correlating the interphase soil water deficits to the yields, results showed that the crops experienced water stress in the last three interphases i.e. tasselling to flowering, flowering to wax ripeness and wax ripeness to full ripeness. Flowering to wax ripeness interphases had the highest negative correlation, followed by tasselling to flowering and finally wax ripeness to full ripeness. A modification to this model was done with the aim of improving on the correlation between the indices and yields. This modification took into consideration the interphase( s) at which the crop experienced water stress during the computation of the indices. Interphase weights or coefficients were developed and used in the computation of the final indices. There was a considerable improvement both in the correlation coefficients obtained between the indices and yield, and the coefficient of determination. Much of the yield variation at Katumani seemed to be due to water stress experienced at the tasselling to flowering and flowering to wax ripeness interphases. The equations developed when maize yields (dependent variable) was regressed against the interphase modified indices (independent variable) were used to predict yields for 1989 to 1991. This study also tested the suitability of cultivating Katumani Composite B maize using the FAD criteria. The long rains were found to be suitable. The short rains were just below the threshold probability of success for a good crop. Also investigated in this study were the most appropriate sowing dates using the modified indices. The results showed that the earlier the planting dates in a given season, the higher the chances for a successful crop and hence, good yields to be attained.