Page 176 - Climate Change and Food Systems

climate change and food systems: global assessments and implications for food security and trade
figure 4
Mean daily maximum temperature (oC) for the warmest month, 1950-2000
Source: WorldClim version 1.4 (Hijmans et al., 2005)
Some of the most extreme temperature increases are found in the ECHAM GCM. When averaged by region, this always proves to be the hottest model but the increases are particularly strong in this GCM for Southern Africa. When combined with the drying trend that was shown previously, this suggests that if the ECHAM model proves to be accurate, Southern Africa may be the hardest hit in terms of rainfed crop production.
MIROC also shows some extreme temperature increases, but these tend to be focused in North Africa, and in the northern portions of West Africa. While the calculations for West Africa according
to this GCM appear to present challenges for cropping, much of the extreme temperature increase appears in parts of the region which
are already too dry for rainfed crops. This is a mitigating factor, as well, for Southern Africa in the ECHAM model, because the highest temperature increases are also in very dry areas.
The CSIRO GCM generally predicts relatively modest temperature increases. This is especially the case for East Africa and large portions of West Africa.
Finally, the CNRM GCM resembles the ECHAM GCM, except in Southern Africa, where it is
more moderate in its projections for temperature increases.
Considering all the models together, Southern Africa is projected generally to be the hardest hit in terms of temperature increases.
3. DSSAT Crop Model results
DSSAT is a crop modelling software package that was used in the three monographs on climate change impacts on agriculture in Africa on which this chapter is based. The crops analysed in those monographs using DSSAT are maize, wheat, rice, soybeans, groundnuts and sorghum. DSSAT takes into account soil characteristics and weather, as well as crop variety and farming practices.
DSSAT has its own daily weather data generator, which was applied in the study. For each month, climate data were provided, consisting of mean precipitation, number of rainy days, solar radiation, mean daily high temperature, and mean daily
low temperature. From these data, the software programme stochastically generates daily weather data that are based on the monthly statistics.
Using the daily data for the climate of 1950 to 2000, thirty years of weather were simulated, and yields were computed for each of those years, taking the average of weather outcomes. The same procedure was applied for each GCM, to generate climate assumptions for the year 2050. The mean yield results were compared, gridcell- by-gridcell, to determine how yields would change between 2000 and 2050 as a result of climate change.
In this particular analysis, it was assumed that there would be no adaptation. This implies that the model did not allow for changing cultivars
or fertilizer regimens or, in the case of rainfed crops, for switching to some kind of water supplementation, such as irrigation.
The analysis focused on 10 kilometre gridcells for East and Southern Africa, and 30 kilometre gridcells for West Africa (except for sorghum, which was done at 10 kilometre resolution).
In the case of East Africa, a grid was overlaid
156