Optimal design of a tall tower observation network for greenhouse gases (CO2, CH4 and N2O) over Africa. Incremental optimisation was used to select sites from a candidate list to achieve maximum uncertainty reduction in the total CO2, CH4 and N2O fluxes over Africa. The Lagrangian particle dispersion model FLEXPART was used to determine the sensitivity of each candidate site to the fluxes. The code is provided to calculate the uncertainty reduction achieved by adding a site. This includes the sensitivity matrices for each candidate site and the uncertainties in the prior CO2, CH4 and N2O fluxes. The biomass burning observation errors assigned to each candidate site are also provided.