Spatial and temporal aggregation effects in a regional water supply planning model

Abstract

In applying mathematical programing water supply planning models, water supplies and demands must be spatially and temporally aggregated. An assessment is made of the effects of aggregation on the results of a planning model for West Africa. The assessment uses a detailed hydroeconomic simulation model to study the solution set of the planning model when it has been applied to several representative subareas. Comparisons are made between the design reliability intended for the supply system and the actual reliabilities of sources as estimated by the detailed simulation model. Comparisons are also made between the system suggested by the planning model and the estimated least cost reliable system. (The least cost reliable system was found using a trial and error search method and the detailed simulation model.) It was found that the aggregation assumptions primarily affect the modeling of sources that depend directly on the highly variable precipitation (e.g., ephemeral aquifers). These sources should, in general, be excluded from the planning model and analyzed separately. Overall, unless a planner is particularly interested in studying the use of such unreliable sources, the spatial and temporal aggregation assumptions of the planning model do not hinder its use as an effective regional planning tool.