Urban heat island prediction in the mediterranean context: an evaluation of the urban weather generator model

The lack of urban-specific climate data is, today, one of the major limits for an accurate estimation of the building energy performance in the urban context. The urban climate is substantially modified by the “heat island” effect that determines an increase of the air temperature compared to the surrounding rural areas. By contrast, the weather data used to run the energy simulations normally refer to rural or suburban weather stations, causing relevant errors in the energy assessment, especially in hot climates.This study aims at evaluating the accuracy of the Urban Weather Generator (UWG), a model developed for generating urban weather files from rural weather data in order to improve the accuracy of the building energy simulations in the urban context.To this purpose, the model predictions have been compared to actual observations in different urban sites in Rome and Barcelona. The comparisons have been conducted for one year of observations for each site, focusing the analysis on summer and winter months. The model accuracy has been assessed through statistical analysis of the average error.Results show that the UWG model is able to capture the general trend of the urban temperature. The accuracy of the prediction increases for urban sites located in a rather homogeneous urban fabric in terms of building density and vegetation coverage. In these situations, the model allows a good prediction of the urban air temperatures with low computational requirement and it can be a useful tool to improve the accuracy of urban energy analysis.