Maize and sorghum are the leading staple foods in Tanzania. For example, maize is planted on more than 45% of all cultivated land in Tanzania, which accounts for more than 70% of the cereal grain produced in the country. Fortunately, the production of maize has been growing steadily from 715,000 tonnes in 1971 to 6,300,000 tonnes in 2020, where 85% of the production is done by small-holder farmers. Despite the necessity of yield forecasts, according to Tanzania’s Ministry of Agriculture’s Division of National Food Security (personal communications), no operational yield forecasting system exists at the subnational level. Crops such as maize and sorghum yields in the region’s semi-arid agroecosystems are constrained by highly variable rainfall, which may be worsened by climate change. During the rainy season, farmers in Tanzania are seeing rising temperatures and irregular rainfall. Under both the business-as-usual climate change scenario and the more hopeful scenario in terms of emissions reduction interventions, this is anticipated to deteriorate in the next decades. Regardless of which climate change scenario is used, the impact on sorghum and maize output is projected to be low.
There have been efforts from different organizations and forums in Tanzania to improve resilience to extreme climate variability that are ongoing through the issuance of pre-seasonal climate forecasts based on both dynamic and statistical techniques. East African Climate Center of Excellence (EACCE) and Greater Horn of Africa Climate Outlook Forum (GHACOFs) accredited by the World Meteorological Organisation (WMO) and the Intergovernmental Authority for Development – Climate Prediction and Application Centre (ICPAC) provide weekly, monthly and seasonal forecasts to create resilience in regions deeply affected by climate change and extreme weather.
However, these forecasts are general and only focused on the country’s level; they do not cover the farm or district level. Moreover, the seasonal climate impact outlook is only based on subjective judgement rather than using explicit quantitative methods. Developing AI/ML-driven prediction models, quantitative weekly, pre-seasonal and monthly forecasts of crop yield predictions and communication of uncertainties can be incorporated with the EACCE or GHACOFs to enhance the use of seasonal climate prediction by providing direct impacts on maize and sorghum production in Tanzania.