Assessment of serious water shortage in the Icelandic water resource system

Water resources are economically important and environmentally extremely vulnerable. The electrical power system in Iceland is hydropower based and due to the country's isolation, power import is not an option as elsewhere in Europe. In the hydropower system, a water shortage is met by flow augmentation from reservoirs. The management of these reservoirs are a human intervention in a natural flow and therefore necessarily limited by environmental regulations. During a heavy drought, the available water storage in the reservoir may not be sufficient to cater for the demand and consequently there will be a shortage of electrical power. This is politically acceptable as long as it only touches heavy industries but not power deliveries to the common market. Empty or near empty reservoirs cause power shortage that will be felt by homeowners and businesses, until spring thaw sets in and inflow to the reservoirs begins. If such a power shortage event occurs, it will cause heavy social problems and a political decision making will follow. It is commonly agreed, that management methods leading to such a disastrous event as a general power shortage in the whole country, are not acceptable. It is therefore very important to have mathematical tools to estimate the risk of water shortage in the system when searching for the best management method. In view of the fact that the subject is to estimate the risk of events that have to be very rare, i.e. with large recurrence time, stochastic simulation is used to produce synthetically run-off records with adequate length, in order to estimate very rare droughts. The method chosen is to make the run-off series stationary in the mean and the variance and simulating the resulting stationary process. When this method is chosen, future trends in the run-off from climate change and glacier reduction can easily be incorporated in the model. The probabilities of extreme droughts are calculated and their frequencies are compared to theoretical distributions. (c) 2005 Elsevier Ltd. All rights reserved.