Evaluation and comparison of the accuracy of the CORDEX database's summer precipitation network data with station data (Case study: summer precipitation of South East of Iran)

South East of Iran will yearly receive part of its water requirements from the summer precipitation, which is mostly due to the monsoon expansion. These precipitations are related to the temperature changes in the Indian Ocean level (Aramash et al., 2016: 2). Producing the accurate climate data is one of the main goals of the forecasting and modeling centers. Among the most important advantages of this data can mention the following: better estimation of the climatic variables in the regions which have no station, possibility to study the climate more appropriately and evaluate the fluctuations and changes of climate elements (Forsythe et al., 2015).

In this research, we used daily precipitation data of 6 synoptic stations and the outputs of various CORDEX models in the South Asia during the statistical period of 30 years (2005-1976) in the studied region.

Evaluation of the accuracy of the global models' output on the basis of the downscaling of the studied regional model on a daily basis during the monsoon period (June, July, August and September) in (1976-2005) indicated that according to the different indicators the CanESM2, CSIRO and NorESM1 models have respectively more accurate estimation of precipitation values at most studied stations than other models. The correlation coefficient of none of the models has a good relationship. The observed and estimated precipitation values of the studied models in the form of the average activity months of the monsoon system showed the accuracy of the investigated models at each station is different but, in general, the three models CanESM2, CSIRO and NorESM1 are more appropriate in the estimation at the most stations than other models. A number of macro criteria must be considered to select the proper model; the 500 HPA Geo-potential height and the sea level pressure have been used in this research for this purpose. In fact, investigating the model's ability to simulate the 500 height as well as the earth is preferable over precipitation and temperature.

The results have suggested that the accuracy of the studied models at a variety of stations is different but, totally and at most of the stations, the three models CanESM2, CSIRO and NorESM1 have respectively better estimation of monsoon precipitation values than the other models. We cannot definitely prefer one model over the others according to the error estimation criteria as well as the comparison of different models with each other. We have shown in this research that the large scale variables such as Geo-potential height and pressure are more predictable than the precipitation and CanESM2, CSIRO-Mk, IPSL-CM2A-MR, MIROC2 and MPI ESM-LR models can be introduced as suitable models to simulate and predict climatic parameters of sea level pressure and 500 HPA Geo-potential in the monsoon system activity governing on the South East of Iran.