Ensuring household food security in a rainfed agricultural livelihood requires availability of climate information regarding onset of seasonal rains allowing for timely preparation for planting. Due to increased irregularity of the onset, amount and length of seasonal rains, the prediction of onset of seasonal rains at sufficient lead times is increasingly becoming a very critical issue for farmers. Currently climate scientists are able to use sea surface temperatures as scientific indicators, to forecast rainfall amounts of above normal, normal and below normal averaged over a period of three months. Whereas this type of information is important, the primary climate information need of the farmers is knowing in advance the expected onset of seasonal rains. As a coping mechanism, farmers attempt to use their traditional indicators, particularly local winds and temperatures, to forecast this important climate element. However, identification, validation and improvement of these indicators had not been done. As a synergy to the farmers practice, records of winds, temperature and rainfall from the existing synoptic weather stations can be used to study these relationships on scientific basis. Although analysis of pentad rainfall totals of records from some of the existing weather stations have been done indicating onset of seasonal rains on average basis, practically these seasonal rains set in at different periods of each year. Currently there is no availability of models to predict the different periods when the rains can set in.
Therefore this study identifies details of how farmers traditionally use local temperatures and winds to forecast onset of first rains; validate the indigenous rainfall indicators for onset of first rains; and develop statistical models for forecasting of first rains. Identification of usage was achieved through conducting individual and group surveys of farmers in eastern (Tororo), Lake Victoria basin (Jinja), central (Wakiso) and western (Masindi) Uganda. Validation of indigenous rainfall indicators is based on the climate data from synoptic weather stations in the four regions. Model development was achieved by statistical linear regression of validated temperature and wind indicators with rainfall onset dates formatted in pentads.
Was this article helpful?