High flow, low flow and drought characteristics of Küçük Menderes river basin

Abstract

Water plays an important role in all living life. The importance of water in many areas such as the ecosystem, agriculture, livestock or industrial production has not changed from the past to the present, despite technological advances. Unfortunately, climate change, which is a natural occurrence in the world, and its consequences have accelerated due to anthropogenic activities, which has led to stress on water resources. Today, climate change and the resulting water deficit pose a great risk to societies, especially to those living in arid and semi-arid regions. In addition, due to the rapid growth of the world's population, the available water for human consumption is decreasing both in terms of resources and quality. According to the United Nations World Bank High Level Panel on Water (HLPW) final report, it is predicted that half of the world's population will be affected by the stress caused by water scarcity as of 2050, and 700 million people will be displaced by 2030 due to water scarcity. In light of all these considerations, climate change, drought, changes in river regimes and the consequences of drought on the ecosystem have been investigated under various disciplines. The increasing risks of water scarcity have also led to the development of new research methods and an increasing number of research papers. This study aims to examine critical dry periods and determine the high- and low-flow characteristics of the Küçük Menderes River Basin in western Turkey by using a variety of tools such as the Standardized Precipitation Index (SPI), flow duration curves, percentiles obtained from the flow duration curves, and the D-day low flows. In addition, the study uses frequency analysis to derive frequency curves for the percentiles and the D-day low flows. In the study, precipitation and streamflow data obtained from 13 meteorological stations and 5 gauging stations, respectively, were analyzed. Also, the Mann-Kendall trend test was applied on the SPI, percentiles, and D-day low flow time series to check the existence of a trend. Drought analysis was carried out using the SPI time series obtained from 13 meteorological stations in the river basin and its surroundings. SPI was calculated for k = 1, 3, 6, 9, 12, and 24-month time scales, and the calculated SPIk time series were visualized in graphs. To make the graphs easier to interpret, the thresholds at the levels of the moderate, severe, and extreme drought were plotted on the graphs. In addition, the longest dry periods were determined for each station and time scale. It was found that the determined dry periods were compatible between stations. Since the SPI values at the 12- and 24-month time scales represent hydrological drought, the informations obtained by SPI12 and SPI24 were also used to interpret the results of the streamflow analysis. The Mann-Kendall trend test was used on each time scale of SPI series to determine their trends. However, due to the differences in the record periods of the meteorological stations, both positive and negative trends were seen as a result of the Mann-Kendall trend test in the river basin. It is thought that the reason for this situation is the noncommon recording period of the gauges. The indices calculated using the streamflow or the precipitation data with common record periods demonstrate the same direction of a trend. This helped us to interpret the flow indices despite the limited length of the data. For the flow analysis, annual flow duration curves (FDC) were obtained from the daily streamflow data of 5 gauging stations. FDCs are practical tool to visualize the characteristics of rivers such as the relationship between the magnitude and frequency of streamflow discharges, the high- and low-flow values, or to understand morphometric characteristics of the region. Q1, Q5, Q10, Q50 Q90, Q95, and Q99 percentiles corresponding to the 1%, 5%, 10%, 50%, 90%, 95% and 99% exceedance percentages of time, respectively, were determined from FDCs, and the time series for each percentiles were obtained. The best-fit probabilty distribution functions among the Generalized Extreme Value (GEV), Log-Pearson Type III (LP3), 2-and 3-parameter Gamma (G2, G3), 2-and 3-parameter log-normal (LN2, LN3), 2-parameter Weibull (W2), and Gumbel (G) were choosen by using the Anderson-Darling test for the frequency analysis to calculate 2, 5, 10, 25, 50, and 100-year return period of each percentile. Some low-flow percentiles series such as Q90, Q95, and Q99 contain zero values. Frequency analysis was carried out for these series according to the total probability theorem by discarding zero values but taking their presence into account. It was seen that the LP3 was the most common probability distribution function that best-fit the data. After the calculation of the return periods, the percentiles-frequency curves were derived. By the consideration of the percentiles-frequency curves, the risks that may occur due to high- or low-flow can be predicted in the river basin and measures to be taken can be decided. Low-flow analysis has a substantial importance in water resources management, especially in a region where agricultural activities are important and water reserves are limited. Due to the complexity of drought effects on low flows, this study examined low flow analysis using the D-day low flow method in addition to the low-flow percentiles. Low-flow time series were derived for D = 1, 7, 14, 30, 90, and 273-day minimum discharges, and frequency analysis was performed for the minimum of the D-day average discharges. Since the low-flow time series contain many zero flows, the total probabilty theorem was applied for the frequency analysis. Using the Anderson-Darling method, LP3 was found the most common best-fit distribution function to the D-day low flow disharges. The Mann-Kendall trend test was applied to the percentiles and minimum of the D-day low flow discharges to examine whether the data tend to increase or decrease over time. According to the results obtained, in 2 stations, downward trends were observed in both the percentiles and the minimum of the D-day average discharges. In addition, a positive trend was observed in one station at high-flow percentiles. Since the record periods of this station and some meteorological stations are similar, the positive trends of the high-flow indices, and SPI12 and SPI24 were evaluated together, and it is thought that the reason for the positive trend was related to the length of the record period. It was concluded that the Küçük Menderes River Basin is moving towards a drier climatological character for which precautionary measures should be taken.