The bias correction is done for daily accumulated precipitation, mean, minimum and maximum daily temperature. All these variables are calculated from the raw 6-horly model output. The bias correction is done by relating each variable to the E-OBS observations during the referent period. The comparison is done for the daily values within the each month and for each grid point separately. This method assumes that daily mean, minimum and maximum temperature during one month all follow the Gaussian, while daily precipitation follows Gamma distribution. The first step was to calculate a cumulative density function (cdf) for the observed and simulated values of a variable during the same period, for each grid point over the domain of interest. The second step is comparing observed and modeled cdfs and constructing a correction function for each grid point. This correction function transfers modeled values to the correct (observed) ones with the same cdf value. Once the correction function is determined it is applied to the appropriate modeled datasets for the referent and future periods. The final products of the procedure are the fields of daily temperature and precipitation values over Serbia for the entire integration period. The correction is done separately for each variable and each month. Bias correction of precipitation is more complicated because it requires a special attention to be payed on the number of dry days. It is very important to equalise the number of dry days in the modeled and observed sample before building the cdfs. If the model simulates more wet days than observed, days with smallest amount of precipitation should be artificially dried. Otherwise, if the number of dry days is larger in the model than observed, randomly chosen days should be wetted. After the equalising the number of dry days, following steps are the same as described the above: calculating and comparing modeled and observed cdfs, constructing the corrective function and applying it to the entire model simulation.