Evidence has been provided that land-cover changes such as deforestation can have an impact on cloudiness and precipitation. However, conflicting results have been obtained at different scales and places, highlighting our poor understanding of the physical processes involved. Here we focus on mesoscale summer cloudiness in a temperate region, as influenced by a large forest massif (the Landes forest in France). Our study is based on an up-to-date atmosphere-surface mesoscale model (Meso-NH coupled with SURFEX). Based on observational data, we first optimise the model configuration for our purpose, and show that with a 500 m horizontal resolution we can successfully simulate the higher summer cloud cover observed over the forest, compared to its surroundings. Secondly, we investigate the physical processes leading to cloud formation in a representative case study. Based on a comparative analysis of diagnostics and budgets over forest and non-forest areas, we find that the larger sensible heat flux over the forest and its higher roughness are the main drivers of cloudiness, enhancing vertical velocity and boundary-layer mixing. Third, we simulate the impact of the 2009 Klaus storm that led to the loss of about one third of the trees. Considering fifteen representative convective summer days, we show that the model simulates well the resulting decrease in summer cloudiness that was reported in a previous study based on satellite observations. As a complementary tool, the mesoscale simulations allow to quantify the impacts of the Klaus storm windthrough on the diurnal cycle of the boundary layer.