The influence of a small lake on the mesoscale flow in a boreal forest boundary-layer

Many studies have reported on non-classical mesoscale circulations, forming as a consequence of surface heterogeneity. Obviously typical scales and types of the heterogeneity and the strength of the background forcing must govern such formations. An important example is that of a lake in an otherwise homogeneous landscape. In climate models, the majority of the Earth's lakes are sub-grid scale but may act as effective moisture sources. Resent experience from regional climate modeling in Sweden indicates that the abundance of lakes here must be considered for a proper representation of the regional climate.

This study focuses on the patchy mid-Swedish landscape in the boreal forest zone, with forest and agricultural land heterogeneously mixed. The background is from CFE-1 of the NOPEX experi-ment in Sweden. The core of this study are quasi-idealized model simulations of the flow over Lake Tämnaren, based one well documented day, June 13, 1994. To avoid ambiguity introduced by aspects of the land-surface scheme, temperature and humidity at the lower model boundary are prescribed. A simple time-dependent function for temperature and fraction of potential evaporation is found for each major land-cover type, and fitted to observed data at sites large enough for the model grid to be homogeneous. At the remaining grid points, with a mixture of surface types, specification uses a weighted area-average. After a validation using tower data and radio soundings, the background flow is altered in a series of sensitivity runs.

Lake Tämnaren is a 5x10 km large lake in northern Uppland, Sweden. It is relatively shallow, but aircraft measurements show that the lake has a significantly different surface temperature than the surroundings. In most flights, effects of the lake were visible in 100 m flight-level data. The simu-lations indicate that major effects on the flow by the lake, e.g. formation of a lake-breeze, are lim-ited to days with low winds and significant solar heating. During most conditions the effects of the lake were weak during daytime and stronger during nocturnal conditions. The main effect is in-duced by roughness differences while the effect of thermal differences were indirect.

At night, the lake surface is warm and the surrounding boundary layer is stably stratified. This causes a convective internal boundary layer. The air accelerates when flowing out over the smoother surface; this acceleration is intensified as the vertical transfer of momentum becomes more efficient. This produces divergence on the upstream half of the lake and convergence at the downstream shoreline where the air is decelerated. The corresponding vertical motions distort the flow aloft so that the air at ~100 m senses a "thermal slope" over the lake, reducing the wind speed. In the simulations, there is a dipole pattern below 100 m while above 100 m there is convergence.

During daytime conditions, the situation is opposite. Stable stratification over the lake inhibits vertical momentum, which counteracts the direct effect of the roughness and the dipole pattern is weakened, while conditions over land are convective. Only rarely, with low wind speeds, can a real lake-breeze circulation develop.