全球变化知识资源中心

Future optimal rule curves are required for operating reservoir under uncertainty situation. This study applied the genetic algorithm (GA) connected a reservoir simulation model with smoothing function and adjusting of expert participation to search for future optimal reservoir rule curves for the Ubolrat Reservoir, which is located in the Northeast of Thailand, in period 2015-2064. The future optimal rule curves considered the impact of climate change with the PRECIS model under two emission scenarios, A2 and B2, and future land use maps using the CA Markov model. The future streamflow into the reservoir was determined using the Soil and Water Assessment Tool (SWAT) model. The results showed that the average future streamflow for the A2 and B2 were increased in comparison to the baseline year (1997-2014) due to the increase of average rainfall and temperatures, including widest land use change from rice and forest area to the sugarcane. The smoothing function applied to be the constraint can reduce the fluctuation of the upper and lower obtained rule curves from the GA. Further, the future rule curves can mitigate the frequency of water shortage situations and the releases of excess water during the increasing streamflow impacted by both climate scenarios and land use changes situation more than the existing and historic rule curves. Moreover, adjusted future rule curves with the expert participation were acceptable to operate the reservoir.