Search

Climate change is expected to impact east Africa in the coming decades, with an overall warming trend, but much is still uncertain. Climate researchers working in this field aim to quantify, understand and reduce this uncertainty. On a five to 40 year timescale, climate projections are uncertain because the models used to predict the expected response from greenhouse gas emissions cannot capture all the relevant processes sufficiently well. In addition, many of the standard climate models do not fully consider all regional factors causing local climate change. There are therefore several ‘burning questions’ that scientists hope to address.

Crop models are useful tools for simulating impacts of climate and agricultural practices on crops. Models have to demonstrate the ability to simulate actual crop growth response in particular environments before application. Data limitations in southern Africa frequently hinder adequate assessment of crop models before application. The DSSAT model was used to test the usefulness of crop models under data-limited dryland conditions of southern Africa by validation using data from experimental trial reports and district-wide crop yield estimates. Two crops each were selected in three locations to represent varying cropping and physical conditions in southern Africa, i.e. maize and sorghum (Mohale’s Hoek, Lesotho and Big Bend, Swaziland) and maize and groundnut (Lilongwe, Malawi). DSSAT performs well in simulating crop yields obtained from experimental trials. District-wide simulated mean crop yields were acceptable (relative difference ranged from −12.2% to +2.36%). However, the model’s capture of seasonal yield variation for some locations and crops was uncertain due to climate extremes. It was concluded that satisfactory crop model testing before application is possible and that DSSAT crop models are useful even under data-limited conditions.Keywords: climate, crop modelling, dryland farming, DSSAT, southern Africa