Introducing and Evaluation of LARS-WG Model for Modeling Meteorological Parameters of Khorasan Province (1961-2003)

In spite of considerable increase in the resolution of general circulation and regional models, none of these models are capable to predict and downscale the meteorological outputs in the scale of meteorological station. In this regard different dynamical and statistical models have been developed for downscaling outputs of GCMs. These models are capable to simulate missed data as well. In this paper a brief history of some common Weather Generatorsare firstly reviewed and then skill of LARS-WG statistical model in producing the historical daily data of four synoptic stations of Khorasan province including Mashhad, Birjand, Sabzevar and Torbat-Heydariyeh are investigated. By comparing observed data and the data produced by model, the greatest error was found in rainfall data. Results show that, LARS-WGcan produce minimum and maximum temperature accurately. Skill of the model for modeling the variance of the four mentioned data is acceptable in %95 confidence levels. The obvious errors of rainfall, minimum temperature, maximum temperature and radiation were found in Torbat-e-heydariyeh with 0.6 mm, Sabzevar with 0.9 ÚC, Birjand with 1.9 ÚC and radiation with 0.7 MJ/m2, respectively. So, it is highly recommended that LARS-WG could be used as a reliable tool for future climate assessment over Khorasan province. Before running the model, a climate change scenario obtained from GCM coarse grid is needed.