Evaluation of Tabriz Doppler Radar data Using Synoptic Concurrent Data by Analyzing Intensive Rainfall over Urmia Lake Basin in October 2014

Introduction
Considering the problem of continuous reduction of the water amount of urmian Lake, Identifying the distribution of rainfall in the basin area of Lake has a particular importance from the point of view of climate and hydrology. Doppler weather radar has an ability of the estimating of intensity and the accumulation of daily rainfall with suitable spatial and diurnal resolutions. In current study, radar rainfall data, observed at the Sahand station, were evaluated with 10 synoptic weather stations data inside the Urmia Lake Basin exampling some of intensive rainfall events. The compared models show that among synoptic stations Tabriz, Shabestar, Sahand, Urmia, and Bostanabad have a best fit with radar daily rainfall productions, having high-quality conformity in northwest of the study area. In contrast, low level of agreements between two sets of radar has been observed in mountainous area.
Due to the problem of continuous decreasing volume of Urmian lake water, accurate identification of the temporal distribution of rainfall can be very important from climatic and hydrological points of view. There are various ways to measure or estimate rainfall. Synoptic stations have a relatively low efficiency compared to radar and satellite due to their point and number limitation, relative to the area of the study area and other influential factors such as weather and human error. Tabriz Doppler Radar is one of the 12 radars of the National Radar Network of the Meteorological Organization of Iran, which works in the frequency band of Doppler C-type radars. The aim of this study was to investigate the efficacy and accuracy of radar-distance measurement tools in the study of heavy precipitation, which due to the infancy and lack of similar studies, the results can be used in future research.
Methodology
 The accumulative precipitation data of synoptic stations in the studied area and the product of the daily accumulative precipitation of Tabriz Doppler radar, which is produced by the radar equation, by converting the echo-return intensity of precipitation, have been used. In this study, the data of accumulative precipitation of synoptic stations of the study area and the product of daily accumulative precipitation of Tabriz Doppler radar have been used. With the help of radar software, the product of surface precipitation intensity is produced in a 24-hour period and its temporal resolution is 15 minutes. Other product specifications such as start time, spatial resolution, and maximum distance, frequency of repetition of sent waves, name of the saved file, color scale of the data and the name of the radar site next to the product are listed.
Radar accumulative rainfall on the most severe rainy day in Urmia Lake basin , the distance from the site of the radar site (concentric circles with a distance of 50 km from each other) and the location of the stations studied. Also, to compare the difference in estimation between radar and stations, error estimation indicators such as: mean error, absolute error mean, mean square error and Pearson correlation coefficient were used.
Results and discussion
 The October 14 to 21, 2014 heavy rainfall in Urmia Lake basin has been studied by various radar products and among them 24-hour collective rain product, due to compliance with the cumulative rainfall data of stations, for 10 synoptic stations around Lake Urmia. Due to the collision of the waves with mountains, the topography of the area has a significant impact on the accuracy of radar estimation. They are considered invisible spots; these points causes a lot of errors (in some cases even up to 100%). Therefore, to compare radar data with the station, the accuracy of the separate precipitation estimate at different stations was examined.
Conclusion
The 24-hour accumulative precipitation comparison of the stations northwest of Urmia (for the cities of Tabriz, Sahand and Shabestar)with radar estimates on the days of heavy rains in October 2014, was highly consistent and the only difference in radar estimates on 20 and 21 days, was about 5 mm that less than Measured by synoptic stations. The correlation coefficient between the data is 0.996, which confirms the closeness of the measurement values of the two methods. The remarkable point in the chart is the significant difference and jump in rainfall on October 19 compared to other days. An examination of the graphs of the cities of Salmas and Urmia in the west and Bostanabad in the east of Urmia Lake shows less accurate but acceptable estimates of rainfall and differs.
Conclusion: The comparative graph of rainfall in the Ajabshir city, despite its proximity to the radar site (50 km from the radar), shows a relatively large difference between the radar estimates and the stations. The most important cause of the error is the orientation of the southern Sahand Mountain. In moving to the more southern areas, the radar accuracy is lower, but the comparative rain chart of Ajabshir city, despite its proximity to the radar site, shows a significant difference. Overall, the results shows that: the southern regions, both due to the large distance from the radar and blocking effect of radar waves, almost all of the return waves are weakened from the targets and the radar estimates the amount of precipitation zero.