Simulation of surface energy fluxes and stratification of a small boreal lake by a set of one-dimensional models

Five one-dimensional (1D) lake models were run for the open water season in 2006 for Lake Valkea-Kotinen (Finland) using on-lake measured meteorological forcing. The model results were validated using measurements of water temperature and of eddy covariance (EC) fluxes. The surface temperature is satisfactorily simulated by all models showing slight overestimation (by 0.1–1.1°C). Both sensible and latent heat fluxes are positively biased in respect to EC data, consistent with earlier studies. However, correlation coefficients between EC-fluxes and those simulated are relatively high ranging from 0.55 to 0.74. The skill to simulate vertical temperature profiles by different models is assessed as well. It is found that the lake models underestimate the EC-derived surface drag coefficient, however providing realistic temperature profiles. It is argued that the real momentum flux from the atmosphere is larger than simulated, however it is split up between the wave development and the acceleration of lake currents. Adopting the simple parameterisation for momentum flux partitioning in one of the models showed that this mechanism can be significant. Finally, the effect of including the lake bathymetry data in k-ɛ models was the drastic overheating of water below the thermocline. This is likely to be caused by omitting the heat flux at the lake margins. Thus, the parameterisation of heat flux at the lake’s margins should be included in the models; otherwise it is recommended to neglect bathymetry effects for such small water bodies as the Lake Valkea-Kotinen. Keywords: lakes, numerical modelling, stratification, surface fluxes (Published: 16 January 2014) Citation: Tellus A 2014, 66 , 21389, http://dx.doi.org/10.3402/tellusa.v66.21389 This publication is part of a Thematic Cluster entitled "Parameterization of lakes in numerical weather prediction and climate models". Read the other papers from this thematic cluster here