One of the challenges in the numerical weather models is the adequate representation of soil-vegetation-atmosphere interaction at different spatial scales, including scenarios with heterogeneous land cover and complex mountainous terrain. The interaction determines the energy, mass and momentum exchange at the surface and could affect different variables including precipitation, temperature, and wind. In order to quantify the long-term climate impact of changes in local land use and to assess the role of topography, two numerical experiments were examined. The first experiment allows assessing the continuous growth of urban areas within the Aburrá Valley, a complex terrain region located in Colombian Andes. The Weather Research Forecast model (WRF) is used as the basis of the experiment. The basic setup involves two nested domains, one representing the continental scale (18 km) and the other the regional scale (2 km). The second experiment allows drastic topography modification, including changing the valley configuration to a plateau. The control run for both experiments corresponds to a climatological scenario. In both experiments, the boundary conditions correspond to the climatological continental domain output. Surface temperature, surface winds, and precipitation are used as the main variables to compare both experiments relative to the control run. The results of the first experiment show a strong relationship between land cover and the variables, especially for surface temperature and wind speed, due to the strong forcing land cover imposes on the albedo, heat capacity and surface roughness, changing temperature and wind speed magnitudes. The second experiment removes the winds spatial variability related to hill slopes, the direction and magnitude is modulated only by the trade winds and roughness of land cover.