ملخص
Abstract This study investigated the use of Regional Climate Model )Regcm-4.4( to predict the temperature and Sunshine duration. Its objective is to determine the suitability of the model for the study area (Libya) by comparing the actual data of 15 meteorological stations from 1979 to 2009 (31 years), and the reference model data for the same period of time. The study showed that the efficiency of the model and its effectiveness to work with the data of the study area with high degree of accuracy which we can summarize in the following sig-nificant point: The accuracy of the model reaches 97.35% for predicting the average annual temperature, while the highest accuracy of the seasonal prediction of the average tempera-ture was during winter season 98.32%, but spring season has the lowest accuracy reaching up to 92.17%. On the other hand, the highest accuracy regarding months was in December 99.19% and the lowest accuracy was in April 90.84%. While the accuracy for prediction of the average annual maximum temperature is about 95.48%, the accuracy of the seasonal prediction for the highest temperature recorded in autumn season was about 99.86%, the lowest season was spring, about 92.56%. It is clear that the monthly prediction for the average maximum temperature is in October was the highest accurate month approximately 99.69%, While the lowest value in April is the accurate month reaching 91.60%. On the other hand, the accuracy of the prediction for average annual minimum tem-perature is about 90.54%, While the seasonal prediction of the minimum temperature is the highest accuracy during the autumn season with a precision of 98.29%, spring season has the lowest accuracy by about 81.05%. The highest accuracy of the model is expected with the average monthly minimum temperature during September at 99.24%, while January is the lowest month of accuracy of about 76.80%. In addition that the response of the aforementioned model to the time changes (annu-al, seasonal, monthly) and the matching of the pattern of general change of observed data, the model shows a significant variation in the spatial changes of the average temperature observed in the northern regions, It gives a lower expectation than the observed in South-West region, And the prediction is very close to the observed and does not exceed the dif-ference C ° 1 ± in the South-East region, It should be noted that the model is less accurate with the increase in the rise of stations from sea level, as the results were weak in stations that increased more than (600m), as in the Nalut (621m) and Shehat (621m) and Ghat sta-tions (692m). On the other hand, the accuracy of the prediction model with an average sunshine duration is less than the prediction with average temperature, so that the accuracy of the prediction model with average sunshine duration was about 74.62% for the same period of time against 97.35% for temperature, while the highest accuracy of the seasonal prediction of the average sunshine duration is in summer was about 81.05%, The model shows that the highest accuracy of the average monthly sunshine duration the in August was about 84.16%, and the least accurate was in March about 65.41%. The not response of the model to the spatial variations of the average sunshine duration is shown to be greater than that observed in all regions but it can be seen that the small difference between expected and observed in the southward direction towards dry regions, This is due to the lack of sensitivi-ty of the model to the moisture element, which is the main component of the cloud cover, which in turn obscures the sun and thus reduces the sunshine duration, which is an im-portant element in the factors feeding the model. The model's response to time and spatial changes (annual, seasonal, monthly) shows that the pattern of change of expectation is consistent with the observed data, indicating the suitability of the model for the study area. By comparing the model outputs, the scenario (RCP8.5) is the closest to the observed and reference data of the model on the study area.