Hydrology of Natural and Manmade Lakes (Proceedings of the Vienna Symposium, August 1991). IAHS Publ. no. 206,1991. Estimating the water exchange between a bay system and the main water body of a lake J. Virta, A.-R. Elo & K. Pulkkinen University of Helsinki, Department of Geophysics, Fabianinkatu 24 A, SF-00100 HELSINKI, FINLAND ABSTRACT There are some problems in the modelling of water exchange between a bay and the main body of a large lake during periods of temperature stratification. At this time the water exchange takes place as a two layered flow which is influenced by wind action and controlled by bottom topography, geometry of connecting sounds, temperature stratifi- cation and the inflow system. For testing and improving flow models a measuring program was established in a bay of Lake Pâijânne situated in South Finland. The surface area of this lake is 1100 km 2 and that of the bay about 40 km 2 . The measuring program included current measurements in sounds, temperature recordings and meteorological measurements. This paper reports on tests of a model for computing the hypolimnetic flow in sounds from thermocline height differences between ends of the sound and the density difference over the hypolimnion. Thermocline height difference is computed from an oscillatory model with wind velocity data. INTRODUCTION Numeric water current models which are used in Finnish lakes are usually based on the model of Simons with fixed layers (Virtanen el aL, 1986). This type of model seems to be most effective for short period of computations. Computational difficulties increase during the period of high stratification. The so-called water balance model has been developed for longer periods, for example for yearly averages. This model has also been applied for some Finnish lakes (Frisk, 1989). In this paper a model is proposed for hypolimnetic flow computation for periods lasting from some hours to several days suitable for use for the whole stratification period. It is specially intended for systems of lakes with consecutive basins connected by sounds. The model is not closed in the sense that it needs seasonal thermocline depth and seasonal density differences between hypolimnion and epilimnion which are obtained from measurements. The model may be combined with some temperature model in order to get a closed current model. 247