Int. J. Water, Vol. 11, No. 1, 2017 1 Hydrological drought assessment in Aksu River basin Jun Hou Kunming University of Science and Technology, Kunming, 650000, China and State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, Institute of Water Resources and Hydropower Research, 1-A Fuxing Road, Haidian District, Beijing 100038, China Email: firstname.lastname@example.org Baisha Weng State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, Institute of Water Resources and Hydropower Research, 1-A Fuxing Road, Haidian District, Beijing 100038, China Email: email@example.com Xiaogang Liu Kunming University of Science and Technology, Kunming, 650000, China Email: firstname.lastname@example.org Jun Yin, Zhiyong Yang and Denghua Yan* State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, Institute of Water Resources and Hydropower Research, 1-A Fuxing Road, Haidian District, Beijing 100038, China Email: email@example.com Email: firstname.lastname@example.org Email: email@example.com *Corresponding author Copyright © 2017 Inderscience Enterprises Ltd. 2 J. Hou et al. Abstract: Based on the monthly stream flow date of three representative hydrological stations, in the Aksu River basin during the period of 1957–2010, i.e., ShaliGuilanke station, XieHela station and XiDaqiao station, this paper calculates river inflow anomaly index and stream flow drought index (SDI). The simulation results of these two indexes were compared and the results showed that the SDI could reflect the drought situation better. Runs theory was used to analyse the SDI, getting the drought duration and droughts intensity in Aksu River basin. The results indicate that drought in Aksu River basin occurs mainly in spring (March–May). During 1957–2010, drought occurred 50 times, approximately once per year, with an average interval time of eight to nine months. After the 1990s, drought duration became shorter and intensity decreased in ShaliGuilanke station and Xi Daqiao station, but drought risk frequency and drought intensity became larger. Drought situation is increasingly grim in XiDaqiao station of China. Keywords: river inflow anomaly index; stream flow drought index; SDI; runs theory; Aksu River basin. Reference to this paper should be made as follows: Hou, J., Weng, B., Liu, X., Yin, J., Yang, Z. and Yan, D. (2017) ‘Hydrological drought assessment in Aksu River basin’, Int. J. Water, Vol. 11, No. 1, pp.1–13. Biographical notes: Jun Hou is a student and his major is Water Resource. Baisha Weng is an Engineer in Water Resource. Xiaogang Liu is a Professor in Water Saving Irrigation. Jun Yin is a student and her major is Water Resource. Zhiyong Yang is a Professor in Water Resource. Denghua Yan is a Professor is Water Resource. 1 Introduction Stream runoff is an important basis for comprehensive utilisation and management of water resources, whose changes are driven by both climate change and human activities (Wang et al., 2011). With the deepening impacts of climate change, extent and frequency of drought events are becoming larger (Yan et al., 2014). One of the distinct results of climate change is regional water cycle change and the corresponding change in runoff coefficient (Xu, 2000). Evaluating the drought in the study region can strengthen its drought prevention and management, promoting the rational distribution of water resources. Hydrological drought assessment in Aksu River basin 3 A series of achievements of hydrological drought research has been made all around the world. Nalbantis and Tsakiris (2009) adopted the same method of standardised precipitation index (SPI) to calculate multiple time scale of stream flow drought index (SDI) in hydrological year. Zhou et al. (2011) used the runoff as hydrological drought index to identify hydrological drought variable feature via two methods of river inflow anomaly index and runoff cumulative frequency; the curve fitting method was utilised to determine the distribution of individual drought variables, then a copula function was used to build a joint distribution of drought duration and drought intensity based on this. Yin et al. (2009) used truncation level to get hydrological drought events, and applied Markov chain and first-order autoregressive model to simulate the stochastic compound runoff, discussing drought strength, drought intensity, drought duration and risk probability statistics. Studies on drought index showed that if drought indicators were established objectively and reasonably, the drought process could be reflected accurately. Otherwise, some drought processes will be missed and some non-dry processes could unnecessarily be added (Wang et al., 2007). This paper took Aksu River basin as the case study, and evaluated the Aksu River basin hydrological drought by combining long sequences of hydrological data. 2 Case study Aksu River basin is located between 75° 35’E~80° 59’E, 40°17’N~42°27’N, covering an area of 52,000 square kilometres, of which 33,000 square kilometres are in China, rest of them in Kyrgyzstan. In general, the northwest area is higher than the southeast area with huge difference in landform. Located in the deep Eurasian continent, it has high evaporation and low precipitation, resulting a dry climate. Sunshine hours are long, thus having rich heat resources. Climate in this area changes sharply with cold winter and hot summer. Temperature range between day and night is large while average wind speed is small. The average precipitation is 64 mm per year. Average evaporation is 1,890 mm and average temperature range 9.2°C~12°C. Aksu River is one of the three major international rivers in Xinjiang, originated from Cork Hill in Kyrgyzstan, which is located between 75°35’E~82°00’E, 40°00’N~42°27’N. Kun Ma Li Ke River and Tuoshigan River are two main tributaries of the Aksu River. The main supply water of north branch Kun Ma Li Ke River is glacier melt, whose catchment area is 12,816 km2. The full-length from headwaters to the confluence of the two rivers is 293 km, with annual runoff of 460 billion m3. The main supply water in west branch Tuoshigan River is rain and snow, whose catchment area is 12,816 km2. The full length from source to confluence is 457 km, with annual runoff of 2.576 billion m3. 4 J. Hou et al. Figure 1 3 The location of Aksu River basin Methodology 3.1 River inflow anomaly index River inflow anomaly index (Wang et al., 2007) is an index to express the increment and decrement of runoff in a certain period compared with normal years. It can reflect the drought caused by runoff abnormalities directly. It can be expressed as follows: I = ( R − R ) / R × 100% where I represents the river inflow anomaly index R represents the current river flow R represents average flow of many years in the same period. River inflow anomaly indexes according to the classification of precipitation anomaly in classification meteorological drought (GB/T 20481-2006, 2006), for various drought levels are shown in Table 1. Hydrological drought assessment in Aksu River basin Table 1 5 Drought level of river inflow anomaly index Drought level Light drought Medium drought Serious drought Severe drought Anomaly percentage (%) –10~–30 –31~–50 –51~–80 < –80 3.2 Stream flow drought index SDI (Weng, 2013) used Γ distribution to describe the change of runoff, standardising runoff skewed distribution normality. Finally, normalised runoff cumulative frequency distribution was used to divide the drought level. Assuming a time-scale runoff is x, then probability density function f(x) which meets Γ distribution is as follows: f ( x) = 1 β α Γ(α ) xα −1e− x / β ( x > 0, α > 0, β > 0) Normalise Γ distribution get SDI: SDI = S t − ( c2 t + c1 ) t + c0 [( d3t + d 2 ) t + d1 ] + 1 where t = 2 ln( P ), c0 = 0.802853, c1 = 0.0100328, c2 = 0.010328, d1 = 1.432788, d2 = 0.189269, d3 = 0.001308. When F > 0.5, S = 1; when F < 0.5, S = –1. SDI was divided according to the classification of precipitation drought index in Classification Meteorological Drought (GB/T 20481-2006, 2006), drought level was shown in Table 2: Table 2 Drought level of SDI Drought level SDI Light drought Medium drought Serious drought Severe drought –1~–0.5 –1.5––1 –2~–1.5 < –2 3.3 Runs theory Drought duration and drought intensity are two indicators describing drought phenomenon. Drought duration refers to the period of drought events from the beginning to the end. Drought intensity could be obtained by dividing drought severity by drought duration (Mishra and Singh, 2010, 2011). When the index is smaller or equal to R0, drought occurs. For drought events who has only one period, only when the index is smaller than R1, could it be defined as a drought events. Otherwise, it will be neglected since it is only a small drought event. In two drought events, when only one drought indicator periods is larger than R0 but smaller than R2, we assume that these two droughts are subordinate droughts, which can be combined into one (Fleig et al., 2006). After the combination, drought duration is D = d1 + d2 + 1, and drought intensity is S = s1 + s2. J. Hou et al. 6 Figure 2 4 Identification of drought duration and drought (see online version for colours) Results and discussion 4.1 The simulation effect of two drought index in Aksu River basin Based on the monthly runoff data of three representative hydrological stations between 1957–2010 in Aksu River basin, the three rivers’ inflow anomaly index and SDI were calculated. Statistical results show that drought in Aksu River basin occurs mainly in spring (March–May). The results are shown in Figure 3 and Figure 4. According to the historical drought conditions in Aksu River basin, the simulation results of river inflow anomaly index and SDI in Shali Guilanke station, Xie He La station and Xi Da Qiao station were compared with each other. It is found that SDI can reflect the drought situation in Aksu River basin better. Between 1957–2012, there were 25 droughts events in Aksu River basin, which are listed in Table 3. Most of them occurred in spring. From the inter-annual point of view, river inflow anomaly index corresponds with the actual drought for only ten years, and SDI corresponds with the actual drought 16 times. They are shown in Figure 5. From the monthly-scale data respect, the SDI can reflect Aksu River basin drought conditions better, which are in Figure 4 and Figure 5. SDI used Γ distribution to describe the change of runoff, and standardised runoff skewed distribution normality. Finally normalised runoff cumulative frequency distribution was used to divide the drought level. SDI indicators make droughts which occurred in the same area but different times comparable. At the same time, droughts occurred in different parts but the same period become comparable. Hydrological drought assessment in Aksu River basin Figure 3 Main river of river inflow anomaly index in Aksu River basin, (a) Shali Guilanke station (b) Xi Da Qiao station (c) Xie He La station (a) (b) (c) 7 8 Figure 4 J. Hou et al. Main river of SDI in Aksu River basin, (a) Shali Guilanke station (b) Xi Da Qiao station (c) Xie He La station (a) (b) (c) Hydrological drought assessment in Aksu River basin Table 3 9 The drought conditions in Aksu River basin Age Drought conditions Age 1957 1959 1961 1962 1965 1968 1974 1979 1980 1981 1985 1991 A Wa Ti A Wa Ti Ku Che Ku Che Ku Che A Wa Ti Wu Shi A Wa Ti, Aksu Ke Ping, Ku Cheetc Ke Ping Aksu Aksu 1993 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Drought conditions Ku Che, BaiCheng, WuQia Aksu Aksu Ku Che, ShaYa, Xin He Aksu Aksu Aksu Ku Che Aksu Aksu Ku Che, A Wa Ti Aksu Wen Su, ShaYa Source: Tan et al. (2013) Figure 5 Comparison of river inflow anomaly index and SDI, (a) Shali Guilanke station (b) Xi Da Qiao station (c) Xie He La station (a) (b) 10 Figure 5 J. Hou et al. Comparison of river inflow anomaly index and SDI, (a) Shali Guilanke station (b) Xi Da Qiao station (c) Xie He La station (continued) (c) 4.2 Temporal evolution of hydrological droughts in Aksu River basin Combined with monthly sequence data of SDI between 1957–2010 in Shali Guilanke station, Xie He La station and Xi Da Qiao station, threshold levels R0 = –0.5, R1 = –1, R2 = –2 were selected to extract drought time. There were 45 drought events during 1957–2010 in Shali Guilanke station, and average drought duration was 5.75 months. The average drought intensity was 4.93, and drought duration maximum was 19 months which occurred in 1960-October to 1962-April, corresponding to the historical drought intensity maximum value 18.56, and average drought interval is 8.4 months. Xi Da Qiao station showed 50 drought events, whose average drought duration was 3.78 months and average drought intensity was 3.90. Maximum drought duration was 12 months from in 2009-December to 2010-December, corresponding to the historical drought intensity maximum value 30.80, and average drought interval was 9.12 months. There were 50 drought events during 1957–2010 in Xie He La station, and average drought duration was 4.62 months. The average drought intensity was 3.78, and drought duration maximum was 14 months which occurred in 1963-July to 1964-August, corresponding to the historical drought intensity maximum value 8.26, and average drought interval is 8.26 months. Detailed statistics are shown in Table 4. Table 4 Drought events statistics of hydrological stations in Aksu River Drought times Drought duration mean Drought intensity mean Drought duration maximum Drought intensity maximum Drought duration mean Shali Guilanke 45 5.75 4.93 19 18.56 8.4 Xi Da Qiao 50 3.78 3.90 13 30.8 9.12 Xie He La 50 4.62 3.78 14 11.88 8.26 Station In the 1950s, 1960s, 1970s, 1980s, 1990s, and 2000s, drought events for each age changes in hydrological stations statistical results are shown in Table 5. For Shali Guilanke station, the frequency of drought was the highest in the 1980s, and the drought Hydrological drought assessment in Aksu River basin 11 intensity was largest in the 1950s. Drought times minimum and drought intensity minimum were in the 2000s. To Xi Da Qiao station, drought times maximum was in the 1970s and 1990s, drought intensity minimum was in the 1970s, and drought intensity maximum in the 2000s. To Xie Hela station, drought times maximum was in the 1970s, drought times minimum but drought intensity maximum were in the 1950s, drought intensity minimum was in the 1990s. By comprehensive analysis of drought characteristics of each hydrological station, we found that for Shali Guilanke station, drought occurred frequently before the 1980s. Besides, the duration was longer and intensity was greater. Drought intensity became smaller and drought events became fewer after the 1990s. For Xie Hela station, the duration of drought became longer and intensity became greater before the 1980s; entering the humid period after the 1990s, the characteristic of drought became shorter and intensity became lower. The reason is that in the past ten years, as temperatures rose, runoff of Kun Ma Li Ke River and Tuoshigan River, which was supplied mainly with glacier melt water, increased significantly (Wang et al., 2008). After the 1990s, drought happened frequently, drought intensity became larger and drought situation became increasingly grim in Xi Daqiao station. Table 5 Ages 1950s 1960s 1970s 1980s 1990s 2000s Drought events age change of Hydrological station in Aksu River basin Drought Drought duration intensity maximum maximum Drought duration mean Drought times Drought duration mean Drought intensity mean Shali Guilanke 4 12 –11.44 19 –18.56 4 Xi Da Qiao 5 4.4 –4.55 9 –9.35 9.4 Xie He La 4 6.25 –5.87 10 –11.88 6.75 Shali Guilanke 7 4.85 –3.78 9 –8.67 9.85 Xi Da Qiao 8 3.75 –3.9 11 –11.54 8.75 Xie He La 10 5.1 –4.275 14 –10.86 6.1 Shali Guilanke 8 7.4 –5.71 18 –13.5 7.375 Xi Da Qiao 10 3.3 –2.58 6 –4.33 8.6 Xie He La 11 5.27 –4.02 11 –7.86 6.18 Shali Guilanke 12 5.75 –5.62 12 –16.78 3.83 Xi Da Qiao 8 2.875 –2.84 6 –6.64 14 Xie He La 10 4.4 –3.237 10 –7.26 7 Shali Guilanke 9 2.45 –2.02 11 –4.82 15 Xi Da Qiao 10 3.7 –3.285 7 –5.73 7.6 Xie He La 7 3 –2.66 7 –8.23 12.25 Shali Guilanke 3 2.67 –2.17 10 –3.42 7.33 Xi Da Qiao 9 4.89 –6.6 13 –30.8 7.22 Xie He La 5 2.28 –2.37 4 –5.2 12 Station 12 J. Hou et al. 5 Conclusions 1 The calculated results of river inflow anomaly index and SDI indicated that drought in Aksu River basin occurs mainly in spring (March–May). 2 According to the recording of the historical drought conditions in Aksu River basin, the simulation results of two indexes were compared and we found that SDI can reflect the drought situation better. 3 Runs theory was used to analyse SDI, getting the drought duration and droughts intensity in Aksu River basin. The result indicates that between 1957–2010, there were 50 drought events in Aksu River Basin, once a year on average and the average drought duration was eight to nine months. 4 After entering the 1990s, drought duration became shorter and intensity decreased in Shali Guilanke station and Xi Da Qiao station, but drought risk became frequent, drought intensity became larger and drought situation became increasingly grim in Xi Da Qiao station. Acknowledgments This research was partially sponsored by the National Science and Technology Support Program Project (2013BAC10B01), Twelve-Five Science and Technology Support Program (2012BAC19B03), NSFC (51309129). References Fleig, A.K., Tallaksen, L.M., Hisdal, H. et al. (2006) ‘A global evaluation of streamflow drought characteristics’, Hydrology and Earth System Sciences, Vol. 10, No. 4, pp.535–552. GB/T 20481-2006 (2006) Classification Meteorological Drought, China. Mishra, A.K. and Singh, V.P. (2010) ‘A review of drought concepts’, Journal of Hydrology, Vol. 391, pp.203–210. Mishra, A.K. and Singh, V.P. (2011) ‘Drought modeling – a review’, Journal of Hydrology, Vol. 403, pp.157–175. Nalbantis, I. and Tsakiris, G. (2009) ‘Assessment of hydrological drought revisited’, Water Resources Management, Vol. 23, No. 5, pp.881–897. Tan, X., Xue, L., Wang, X. et al. (2013) Risk Assessment of Drought Warning and Disaster Effects in Tahe Basin, Southeast University Press, Nan Jing. Wang, G., Shen, Y., Su, H. et al. (2008) ‘Runoff changes in Aksu River basin during 1956–2006 and their impacts on water availability for Tarim River’, Journal of Glaciology and Geocrylogy, Vol. 30, No. 4, pp.652–568. Wang, G., Yan, D-h., Huang, Z-f. et al. (2011) ‘Analysis on the long-term evolution of runoff volume and its affecting factors in the Luanhe River basin’, Arid Zone Research, Vol. 28, No. 6, pp.998–1003. Wang, J., Yong, G., Zhou, Y. et al. (2007) ‘Research progress and prospect of drought index’, Arid Land Geography, Vol. 30, No. 1, pp.60–65. Weng, B. (2013) Generalized Drought Risk Assessment and Its Coping Strategies: Taking Dongliao River Basin for An Example, China Institute of Water Resources and Hydropower Research, Bei Jing. Hydrological drought assessment in Aksu River basin 13 Xu, C.Y. (2000) ‘Modeling the effects of climate change on water resources in central Sweden’, Water Resource Mgmt., Vol. 14, No. 3, pp.177–189. Yan, D.H., Weng, B.S., Wang, G. et al. (2014) ‘Theoretical framework of generalized watershed drought risk evaluation and adaptive strategy based on water resources system’, Natural Hazards, doi:10.1007/s11069-014-1108-5. Yin, Z., Huang, W. and Chen, J. (2009) ‘Analysis of reservoir runoff regulation effects on hydrologic drought’, Hydrology, Vol. 29, No. 2, pp.1–44. Zhou, Y., Yuan, X., Jin, J. et al. (2011) ‘Regional hydrological drought frequency based on copular’, Scientia Geographica Sinica, Vol. 31, No. 11, pp.1383–1388.