Managing the Health Impacts of Climate Change through Sector Projects Figure 9 17 Estimating the Future Burden of Diseases STEP 1: Select the disease of interest (e.g., dengue, cholera, etc.) Estimates using historical data Future projections STEP 2: Select the period of time (e.g., 5 years, 20 years, etc.) STEP 5: Select the period of time (e.g., 20 years, 50 years, etc.) STEP 3: Specify the possible nature of the relationship between climate and burden of disease, controlling for the impact of all other variables. For example: STEP 6: Project in the future non-climate independent variables (e.g., education level, income per capita, etc.) D = F(Temperature, Rainfall, Education level, Income per capita) where D is the burden of diseases (no. of cases per 10,000). STEP 4: Quantify the relationship between climate and burden of disease. STEP 7: Estimate the burden of disease in the future without climate change, using the estimated relationship from STEP 4, and the value of the future non-climate independent variables from STEP 6. STEP 8: Select the emissions scenario and the global circulation models (GCMs) from which climate variables of interest will be projected. Selecting more than one emissions scenario and GCM may be of interest to provide a probable range of impacts, and not simply a point estimate. STEP 9: Select climate variable of interest. This may depend on the nature of the disease of interest and its expected correlation with climate variables. STEP 10: Project the values of the climate variables of interest over the selected period of time from the selected emissions scenario and GCMs. STEP 11: Estimate the burden of disease in the future with climate change, using the estimated relationship from STEP 4, the value of the future non-climate independent variables from STEP 6, and the value of the future climate variables from STEP 10. Source: Author. (iii) 9 many instances would be too large to project health impacts; and (b) there are 22 GCMs with their own sets of (uncertain) projections about future climate variables. At the global level, it is generally found that the model projections by 2050 do not diverge significantly insofar as temperature increases are concerned, but do vary significantly for precipitation changes. Selecting a method of downscaling GCM projections. The problem that pertains to the coarse resolution of GCMs can be overcome by downscaling.9 Downscaling increases both spatial resolution (e.g., from hundreds to tens of kilometers) and temporal resolution (e.g., from monthly to daily). There are two main approaches for downscaling: dynamical downscaling GCMs are run at coarse spatial resolution (typically of the order 50,000 square kilometers [19,000 square miles]) and are unable to resolve important subgrid scale features such as topography. As a result, GCMs cannot be used for local impact studies. To overcome this problem, downscaling methods are developed to obtain local surface weather from regional atmospheric variables that are provided by GCMs.