The Weather Research and Forecasting Model was used to study its accuracy and representation in modelling a study area within a complex wind system as well as the effects on the model when using different input data and physics schemes. The complex wind system consists of diurnal mesoscale effects from the nearby Pyrenees mountain range and diurnal effects from the planetary boundary layer. A total of six different simulations were performed. The model was able to represent the study area but the results could be improved as there were inaccuracies in wind speed and wind direction associated with the planetary boundary layer. The model was especially challenged at predicting the wind speed and wind direction in the layer from the top of the planetary boundary layer to few hundred meters above it. The comparisons based on planetary boundary layer height is however complicated by the fact that there are different definitions in effect. The choice of model physics schemes and input data led to some differences in the results and warrants consideration when conducting similar simulations. / Prognosmodellen WRF (Weather Research and Forecasting Model) användes för att undersöka hur väl den kunde representera ett område inom ett komplext vindsystem och även hur modellen påverkas av olika val vad gäller drivningsdata och fysikscheman. Det som utgör det komplexa vindsystemet är dygnsvarierande effekter från det atmosfäriska gränsskiktet och dygnsvarierande mesoskaliga effekter från den närliggande bergskedjan Pyrenéerna. Totalt genomfördes sex olika simuleringar. Prognosmodellen kunde representera området men med förbättringsbara resultat eftersom det fanns fel i vindhastighet och vindriktning relaterande till det atmosfäriska gränsskiktet. Modellen var speciellt utmanad i förutsägandet av vindhastighet och vindriktning i ett lager några hundra meter ovanför det atmosfäriska gränsskiktet. En tolkning baserad på atmosfärisk gränsskiktshöjd är dock svår eftersom det fanns flera definitioner var toppen på det atmosfäriska gränsskiktet låg. Val om prognosmodellens fysikscheman och drivningsdata orsakade en skillnad i resultat sinsemellan. Dessa val bör därför noggrannt uppmärksammas för simuleringar under liknande förutsättningar.

Deep artificial neural networks is a type of machine learning which can be used to find and utilize patterns in data. One of their many applications is as method for regression analysis. In this thesis deep artificial neural networks were implemented in the application of estimating the error of surface temperature forecasts as produced by a numerical weather prediction model. An ability to estimate the error of forecasts is synonymous with the ability to reduce forecast errors as the estimated error can be offset from the actual forecast. Six years of forecast data from the period 2010--2015 produced by the European Centre for Medium-Range Weather Forecasts' (ECMWF) numerical weather prediction model together with data from fourteen meteorological observational stations were used to train and evaluate error-predicting deep neural networks. The neural networks were able to reduce the forecast errors for all the locations that were tested to a varying extent. The largest reduction in error was by 83.0\% of the original error or a 16.7\degcs decrease in the mean-square error. The performance of the neural networks' error reduction ability was compared with that of a contemporary Kalman filter as implemented by the Swedish Meteorological and Hydrological Institute (SMHI). It was shown that the neural network implementation had superior performance for six out of seven of the evaluated stations where the Kalman filter had marginally better performance at one station.