Multi-criteria reservoir water management

Abstract

The Plastiras dam was constructed in the late 1950s mainly for electric power production,
 but it has also partially covered irrigation needs and water supply of the plain of Thessaly.
 Later, the site has been designated as an environment conservation zone because of
 ecological and landscape values, while tourist activities have been developed around the
 reservoir. Irrigation of agricultural land, hydroelectric production, drinking water supply,
 tourism, ecosystem water quality and scenery conservation have evidently been conflicting
 targets for many years. Good management would require a multi-criteria decision making.
 Historical data show that the irregular water release has resulted in a great annual
 fluctuation of the reservoir water level. This situation could be improved by a rational
 management of abstractions. Apparently, higher release leads simultaneously to more
 power production and to irrigation of a larger agricultural land. Moreover, demands for
 electricity and for irrigation are partially competing to each other, due to different optimal
 time schedules of releases. On the other hand, higher water release leads to lower water
 level in the reservoir and, therefore, it decreases the beauty of the scenery and
 deteriorates the trophic state of the lake. Such degradation affects the tourist potential as
 well as the quality of drinking water supplied by the reservoir.
 A multi-criteria approach uses different scenarios for the minimum permissible water level
 of the reservoir, if a constant annual release is applied. The minimum level concept is a
 simple and functional tool, because it is understood by people, easily certified and
 incorporated into regulations. The quantity of water that would be yearly available is a
 function of the minimum level allowed. The water quality depends upon the trophic state of
 the lake, mainly the concentration of chlorophyll-a, which determines the state of
 eutrophication and is estimated by water quality simulation models, taking into account
 pollutant loads such as nitrogen and phosphorus. The value of the landscape is much
 depending on the water level of the lake, because for lower levels a dead-zone appears
 between the surface of the water and the surrounding vegetation. When this dead zone is
 large, it seems lifeless and the lake appears partially empty. Quantification of this visual
 effect is not easy, but it is possible to establish a correspondence between the aesthetic
 assessment of the scenery and the minimum allowed reservoir level.
 Using results from hydrological analysis, water quality models and landscape evaluation,
 it seems possible to construct a multi-criteria table with different criteria described against
 alternatives and with a plot of three relative indices against the minimum level allowed.
 However, decision making has to take into account the fact that comparison or merging of
 indices corresponding to different criteria analysis encompasses a degree of
 arbitrariness. More objective decisions would be possible if different benefits and costs
 were measured in a common unit. Moreover, management will be sensitive to different
 social pressures.