Using of CRU and GPCC data base in the analysis of long-term droughts in the Urmia Lake basin

Background and objectives: Drought is a natural event that faces many countries with water shortage every year. Arid and semi-arid climate and improper distribution of rainfall in terms of space and time have increased the negative effect of water resources in Iran. In the present study, the meteorological drought condition of the Urmia Lake catchment was investigated using network data of precipitation, temperature and evapotranspiration in 10 synoptic stations from 1955- 2019. For this purpose, the performance of network data of CRU (Climatic Research Unit) and GPCC (Global Precipitation Climatology Center) in estimating climatic parameters using ground data was evaluated. Then, network data was used to calculate RDI (Reconnaissance Drought Index) and drought monitoring was observed and analyzed during the statistical period study.
Materials and methods: The watershed is located in the northwest of Iran in the provinces of West and East Azerbaijan and Kurdistan, and its western border is the heights of Iran and Turkey. In the present research, at first, by referring to the Meteorological Organization of the Iran, the rainfall and temperature data of the studied stations were received and processed during a period of 65 years (1955-2019). Then, in order to introduce and use network data for places and times without statistics, from monthly precipitation data GPCC and (minimum, average and maximum) temperature components of CRU was used for 10 selected synoptic stations of Urmia Lake basin during the statistical period. In order to analyze the long-term drought and calculate the RDI index, the data of precipitation, temperature and evapotranspiration obtained from the network data were used.
Results: In this study, two general approaches were used to calibrate the data used. The first is that all ground data extracted from meteorological stations are drawn in against the network data and a regression relationship is fitted to them. In the second, the monthly changes of the network data are taken into consideration and the calibration is done for each month separately. Therefore, monthly recalibration was done in all stations for the available data and the results showed that the calculation error in the CRU data for temperature and evapotranspiration was much smaller and the temperature data values were able to estimate evapotranspiration with less error and better performance. For example, in Urmia station for ETo estimation, the value of the RMSE evaluation index between ground and CRU data is 0.918 mm/ day. While after calibration, this value decreased to 0.671 mm/day. This process of reducing the error between ground data and CRU has been repeated in all the stations studied. In Piranshahr and Saqez stations, the estimation of reference evapotranspiration was associated with more error than other stations. So that the MAE standard in the mentioned stations before calibration was 1.087 and 0.965 mm/day, respectively, and after calibration, this index decreased to 0.309 and 0.467 mm/day. In this section, in addition to statistical indicators, a violin chart was also used to show the data distribution. In general, it can be concluded that the climate data obtained from the CRU and GPCC databases show a good agreement with the time values, but the bias correction in them should always be considered.
Conclusion: The analysis of droughts in the catchment area indicates that from 1998-2019, RDI has more negative values, which indicates severe drought, and in other words, human activities and climatic conditions in the region with It faces a crisis. The results of the present study show the appropriate performance of CRU and GPCC network data in estimating hydrological parameters and it is recommended to use the above databases in areas where long-term ground data is not available.