The Analysis of the Hydrological Droughts in the Northern Part of Lake Urmia

Hydrological droughts occur after meteorological and agricultural droughts. In other words, this kind of drought is the final stage of the drought cycle and is affected by global climate change. Nowadays, many studies on hydrological droughts are based on the Threshold Level method, which is based on the Run Theory. According to this view, a drought is described as the length of a period in which hydrological variables, or discharges, are below the specified threshold level. To the researchers' knowledge, there have never been a major research on hydrological droughts in Iran and studies in this field are very limited in comparison with other types of droughts. Most of the researches in hydrological droughts have been done in the monthly and annual time intervals. However, the present study was conducted to investigate the hydrological droughts on a daily basis with the objectives of identifying and monitoring them and determining their occurrence and severity in the studied area.

In this research, the Threshold Level Method was used to identify the hydrological droughts in the DaryanChai sub-basin in the northern part of Lake Urmia. Using the computer program NIZOWKA2003 based on the partial time series (PDS), the droughts in the hydrometric station of the Daryan during a thirty one year period (1982-2014) were calculated. Hydrological drought characteristics including time of occurrence, duration, severity, and minimum observed flow during dry period were calculated based on the NIZOWKA2003 software. In addition, frequency analysis, probability of the number of drought occurrences, duration, and volume deficiency in different return periods were calculated using the mentioned computer program. Here, the threshold level was chosen based on the flow duration curve (FDC), which was based on the daily flow rates of the selected hydrometric station and prepared using the NIZOWKA2003 program. The IC method was used to integrate the minor and mutual dependency periods of hydrological drought. Based on the characteristics of dry periods, the probabilistic distribution and frequency analysis of dry periods, the probability of the occurrence of drought and characteristics of dry periods (duration and volume deficit), fitting different types of statistical distributions and selecting the most suitable ones based on the fitting values (chi Score) were calculated using the NIZOWKA2003 computer program.

The results of the calculations showed that a total of 38 periods of a hydrological drought occurred in this river. Approximately about 20 events from all detected drought occurrences lasted more than 200 days. The longest period of drought with 577 days lasted between the end of June 1988 and March 1990. Less than two months later, the second largest hydrological period with 365 days lasted between the beginning of May 1990 and the end of June 1991. Regardless of these two months, the largest hydrological drought period with 950 days was between 1988 and 1991. The severest hydrological periods were respectively 8987 and 6133 cubic meters, coincided with the largest hydrological drought periods.  Moreover, the calculations showed that this river's water volume has fallen by 117 million cubic meters over the period of 31 years.   Based on the results of Chi-square test and Akaya's criterion, Poisson distribution was the most appropriate distribution of the probability of the occurrence of drought occurrence. Among the various probabilistic distributions, Weibull's distribution was the most appropriate distribution of the duration of the hydrological drought events. Accordingly, the probability of the occurrence of a dry period of at least 13 day duration is likely to occur in this river. Additionally, the maximum observation dryness event of 578 days was estimated with a probability of non-exceeding 0.99 % at the Daryan hydro-station. The probable distribution of Weibull was selected as the most suitable distribution of the hydrological drought severity. Based on the Weibull method, the maximum water deficit volume was estimated at 13730 cubic meters with a low probability (0.01%). In contrast, the highest probability of the severity of drought (about 70%) in the Daryan Chai was estimated at 253,000 cubic meters. The results of the calculation of the return periods showed that in half of the cases, the probability of the occurrence of a dry period was with a duration of 170 days and an intensity of 2276 cubic meters with a 2-year return period.

Until the last decade, little research has been conducted on the hydrological droughts in Iran.In very few studies, the analysis of the low flows from the point of view of frequency analysis has been addressed.The lack or shortage of hydrometric data in many parts of the world has been the main limitation in hydrological drought studies. In this research, the Threshold Level Method was used to extract thehydrological dry periods and their characteristics. Despite the fact that this method has been extensively used in drought studies, its selection is still one of the controversial issues in such studies. Using the daily time series can provide complete and accurate information from the start and end dates of drought events. However, the use of this time basis in drought studies, especially hydrological droughts, is not common in the world and in Iran.The only studies in Iran, based on the daily time series, were conducted by Bayzidi and Saghafian (2011), Mesbahzadeh et al. (2017), and Mostafa Zadeh et al. (2018).  Almost all other studies have been based on monthly or annual basis. However, the use of a daily time series is often associated with the problem of having minor andmutual dependency periods. There are several ways to overcome this problem. The Interevent Critria. (IC) is one of these methods which was used in this research.The efficiency of this method has been previously reported by Zelenhasićand Salvai(1987), Hisdal et al. (2003), Tallaksen et al. (2004), and Baiziidi and Bayzidi and Saghafian (2011).