Hydrological variability and Zambezi hydropower development

Hydropower generation is fundamentally dependent on river flows. As described above, natural flows in the Zambezi Basin vary seasonally, among years, and over longer-term climatic cycles, which include periods of prolonged drought. This variability has had a significant impact on the operation of existing large dams in the Zambezi River Basin, with respect to meeting firm power requirements and total power generation goals during droughts, and also with respect to managing extreme flooding events.

Extreme flooding events, a natural feature of the Zambezi River system, have become more costly downstream since the construction of large dams. Reservoir outflow capacity is inadequate to discharge the maximum probable inflows, and each dam follows a design flood rule curve to prevent over-topping that frequently result in poorly timed or sudden water releases. Increased spillway capacity, proposed for Cahora Bassa Dam, would eliminate the need for a rule curve and allow outflows to ebb and flow more gradually.

Substantial evaporative water losses from large reservoirs, especially Kariba and Cahora Bassa, reduce water availability in the basin and will increase with climate change. Evaporation from these two reservoirs currently results in an 11% reduction in mean annual flows in the Zambezi River. These water losses serve to further increase the risk of shortfalls in power generation, in addition to their significant impact on downstream ecosystem functions and values.

The major existing hydropower dams on the Zambezi were designed based on an inadequate time series of inflows to adequately characterize the full range of natural variability experienced over the past century. As a result, firm power production is vulnerable to periods of prolonged droughts, and dam safety and downstream flood risk is vulnerable to extreme flooding events. Large reservoirs associated with hydropower dams can mitigate this risk somewhat by smoothing out seasonal variations and, in the case of enormous reservoirs with high storage to inflow ratios, such as Kariba, some annual fluctuations. The flow series observed on the Zambezi River and the Kafue River show extended periods of above and below normal flow, however. Hydropower models using this entire time series for the Zambezi River Basin (Shawinigan-Lavalin and Hidrotécnica Portuguesa 1990; Beilfuss 2010) indicate that at the end of periods of prolonged drought, including 1907-1924 and 1981-1995, reservoir levels fall to the minimum supply level and turbine discharges may be curtailed for prolonged periods.

Recently proposed hydropower projects have access to a more complete hydrological record of the past century and can presumably better account for long-term patterns of inflow variability. If past infrastructure and energy commitments had been based on our present knowledge of this variability, for example, the impact of drought on energy production might have been reduced. As we discuss in the next section, however, historic flows from the past century are not a reliable indicator of future mean annual flows or the seasonal and annual fluctuations in runoff variability expected over the next century with global climate change.