The statistical assessment of rainfall probability and its application in Uganda agriculture

Abstract

Rainfall and resulting soil moisture are probably the most important of the climatic factors affecting crop production. This is true even in temperate climates, subject as they are to additional hazards of winter temperatures and photo-periodic effects. Agriculturists in countries with a long-established agricultural tradition have the heritage of an accumulated weather knowledge on which to base their agricultural systems. Such information is not, of course, available for new or under-developed countries. In consequence means of rainfall or totals of means have had to be used. As might be expected, with tropical convection storms, such means, without associated statements of variability, may often be misleading. The concept of rainfall expectation at selected levels of probability has been shown to provide a completely objective and reliable estimate of rainfall to be expected, and so to assess long-term crop risk which is not apparent when means alone are used. Minimum and maximum rainfall expectation may be used in the preparation of rainfall maps, either annual, seasonal or even for shorter periods. Brief reference is also made to the application of these confidence limits to problems in agricultural meteorology. Given estimates of water required for optimum crop growth, it is a simple matter to calculate the chances of obtaining such amounts of rainfall. The value of the resulting probabilities in assessing crop risk was demonstrated by their relationship with some yield data. Much of the seasonal variation in cotton yields, itself of the order of some 200%, could be ascribed to the chances of obtaining an amount of rainfall equivalent to the average estimated crop water requirements. Critical examination of these probabilities, either as single or combined seasons, enables the agriculturist to formulate the agricultural system best suited to rainfall expectation of the region. In addition, it is now possible to delineate those areas best suited to perennial or double cropping systems in contrast to those where a single crop and succeeding fallow would make a more efficient use of limited water resources.