Eddy covariance (EC) is one of the most effective ways to directly observe evaporation from a lake surface. However, the deployment of EC systems on lakes is costly and technically challenging which engenders a need for accurate modelling of evaporation from reservoirs for effective management. This study aims to 1) refactor the Canadian Small Lakes Model (CSLM) into modern high-level programming languages in open-source repositories and 2) evaluate evaporation estimates from the CSLM using nine years of EC observations of a pit-lake in Northern Alberta. The CSLM is a 1-D physical lake model simulating a mixing layer and arbitrary thick skin layer which interfaces with the atmosphere and includes a module for ice dynamics. It was developed to interface with the Canadian Global Coupled models as part of the surface classification scheme, and thus utilizes widely accessible forcing data. In this study the CSLM evaporation estimates are also compared to a commonly used bulk transfer method of estimating evaporation. In general, the CSLM had smaller open-water season error (RMSE of 0.70 mm day<sup>-1</sup>) than the bulk transfer method (RMSE of 0.83 mm day<sup>-1</sup>). However, if EC data is available, further improvement can be gained by using an Artificial Neural Network to adjust the modelled fluxes (RMSE of 0.51 mm day<sup>-1</sup>). This final step can be very useful for gap-filling missing data from lake observation networks as there has been recent attention on the limited coverage of direct open water evaporation observations in the literature.