Full text available at http://search.proquest.com/openview/c37e2519021a867311275c213dc68008/1?pq-origsite=gscholar
This paper describes how farmers in the South-western Highland of Tanzania predict rainfall using local
environmental indicators and astronomical factors. The perceptions of the local communities on
conventional weather and climate forecasts were also assessed. The study was conducted in Rungwe
and Kilolo districts in Mbeya and Iringa regions respectively. Participatory rural appraisal methods, key
informant interviews and focus group discussions were used in data collection and the collected data
was analyzed using Statistical package for social science. It has been found that plant phenology is
widely used by local communities in both districts in seasonal rainfall forecasting. Early and significant
flowering of Mihemi (Erythrina abyssinica) and Mikwe (Brachystegia speciformis) trees from July to
November has been identified to be one of the signals of good rainfall season. The behaviour of
Dudumizi bird has been singled out as one of the best indicator for rainfall. Both Indigenous Knowledge
specialists and TMA experts have predicted 2009/2010 rainfall season to feature normal to above
normal rainfall. Systematic documentation and subsequent integration of indigenous knowledge into
conventional weather forecasting system is recommended as one of the strategy that could help to
improve the accuracy of seasonal rainfall forecasts under a changing climate.