Recha, Wambongo C. S.2011-08-162011-08-162011-08-16http://ir-library.ku.ac.ke/handle/123456789/812Department of Geography, 149p. 2007, S 600.7.C54R4Seasonal climate forecasts is an adaptation strategy to climate variability expected to be useful in farm management. The study was based in semi-arid southeast Kenya. The specific objectives of the study are:(i) To identify sea surface temperatures (SSTs) influencing seasonal rainfall and vegetation conditions in SE Kenya;(ii) To examine farmers' access and use of seasonal climate forecast information and;(iii) To identify socio-economic constraints in access and use of seasonal climate forecast information. Objectives were achieved by employed numerical and field survey scientific approaches. Objective 1 was achieved by assembling rainfall & SSTs (1961-2004), and NDVI data (1981-2004) for Katumani and Makindu synoptic stations. To analyse these set of data, F-Test, Stepwise multiple regression, correlation and distribution approaches were utilized. Objectives 2 and 3 were obtained by conducting field survey in 60 households. Questionnaires were prepared and issued to interviewees before and after the release of October- December (OND) rainfall forecast by Kenya Meteorological Department in 2004. Field survey data was analysed using Statistical Package for Social Scientists. Descriptive statistics were presented and hypothesis tested using chi-square test. The study reveals that rainfall variability is persistent in southeast Kenya, a phenomenon that continues to pose a challenge to agricultural production. OND rainfall in southeast Kenya has a low but significant relationship with SSTs and the chances of having an accurate forecast are highest during ENSO events. Vegetation conditions are influenced by inter and intra annual rainfall variability. NDVI data had a low correlation with rainfall predictors but peak-months OND NDVI showed a strong correlation with SOI, NAT and NfN04 of the months of June to September. The coefficient of determination, r2 was 0.65 and 0.39 for Katumani and Makindu respectively. Forecast verification scores in both sites showed rainfall forecast to have a better skill than peak-months NDVI forecast. The study has illustrated that direct application of climate forecasts to simulate vegetation conditions is not viable. However, the better skill for predicting OND rainfall and its significant correlation with growing season NDVI, suggest that in the event of an accurate rainfall forecast, users should anticipate greener vegetation conditions and better yields. Field survey results show that 77% of farmers in southeast Kenya access forecasts and their sources main are radio and neighbours. Although farmers access forecast, only 37% have confidence in it. Despite lack of confidence, 75% farmers rate seasonal climate forecasts to be useful, suggesting that downscaling of forecasts at a local level can enhance forecast credibility and therefore adoption. Chi-square test results show that there is an association between access and confidence in seasonal climate forecasts. Thus, farmers with confidence in forecasts are likely to seek information and adopt it in decision- making. There was no difference in the decision to change farm management strategies by agro-ecological zones. Farmers showed flexibility and potential to respond to climate forecasts when they altered planting dates, area planted and crop cultivars. The study further established that access and use of seasonal climate forecasts are not necessarily influenced by socio-economic constraints. However, interplay of various socio-economic constraints can impede or enhance use of forecast information. These factors are labour and timeliness of forecast (0.357, p<0.05), study site and inaccessibility to appropriate seeds (0.37, p<0.05) and, extension and income (0.48, p<0.05). Income, draft power, labour, forecast inaccuracy and access to appropriate seeds were the most mentioned constraints in the adoption of seasonal climate forecasts. To ensure access and use of seasonal climate, it is recommended that (1) seasonal climate forecast be downscaled to a local level, (2) Farmers be trained on the concepts and terminologies used in the dissemination of seasonal forecasts, (3) Dissemination of climate forecasts should be through extension officers and meteorological experts and, (4) Farmers should be assisted to access low interest credit facilities through community based organisations. It is suggested that future studies should include SSTs from the Indian Ocean (including Indian Ocean Dipole) in simulating seasonal forecast for southeast Kenya and attempts should be made to link seasonal forecast and number of rainy days.enCrops and climate--Kenya, South--East//Climate changes--Kenya, South--EastThesis