Simulating effect of anthropogenic activities and climate variation on Liulin Springs discharge depletion by using the ARIMAX model Yonghong Hao, 1 * Jing Wu, 2 Qingxia Sun, 3 Yuen Zhu, 4,5 Yan Liu, 6 Zhongtang Li 7 and Tian-Chyi J. Yeh 8 1 Tianjin Key Laboratory of Water Environment and Resources, Tianjin Normal University, No. 393 Binshuixi Road, Xiqing District, Tianjin 300387, China 2 Department of Statistics, Tianjin University of Finance and Economics, No. 25 Zhujiang Road, Hexi District, Tianjin 300222, China 3 Department of Accounting, Tianjin University of Finance and Economics, No. 25 Zhujiang Road, Hexi District, Tianjin 300222, China 4 College of Environment and Resources, Shanxi University, No. 92 Wucheng Road, Taiyuan 030006, Shanxi Province, China 5 School of Environment, Beijing Normal University, No. 19 Xinjiekouwai Street, Beijing 100875, China 6 College of Mathematical Science, Tianjin Normal University, No. 393 Binshuixi Road, Xiqing District, Tianjin 300387, China 7 Department of Environmental Engineering, College of Science and Engineering, Jinan University, No. 601 West Huangpuda Street, Guangzhou 510632, China 8 Department of Hydrology and Water Resources, The University of Arizona, John Harshbarger Building, 1133 E. North Campus Drive Tucson, AZ, 85721, USA Abstract: Based on the groundwater development process, and regional economic and social developing history, we divided the spring hydrological process of the Liulin Springs Basin into two periods: pre-1973 and post-1974. In the first period (i.e. 1957–1973), the spring discharge was affected by climate variation alone, and in the second period (i.e. 1974 –2009), the spring discharge charge was influenced by both climate variation and human activities. A piecewise analysis strategy was used to differentiate the contribution of anthropogenic activities from climate variation on karst spring discharge depletion in the second period. Then, the ARIMAX model was applied to spring flow time series of the first period to develop a model for the effects of climate variation only. Using this model, we estimated the spring discharge in the second period solely under the influence of climate variation. Based on the water budget, we subtracted observed spring discharge from the estimated spring discharge and acquired the contribution of human activities on spring discharge depletion for the second period. The results of the analysis indicated that the contribution of climate variation to the spring discharge depletion is0.20 m 3 /s from 1970s to 2000s. The contribution of anthropogenic activities to the spring flow depletion was 2.56 m 3 /s in 2000s, which was about 13 times more than that of climate variation. Our analysis further indicates that groundwater exploitation only accounts for 29% of the spring flow depletion due to the effects of human activities. The remaining 71% of the depletion is likely to be caused by other human activities, including dam building, dewatering during coal mining, and deforestation. Copyright © 2012 John Wiley & Sons, Ltd. KEY WORDS karst spring; climate variation; anthropogenic activities; ARIMAX model; piecewise analysis; Liulin Springs Received 20 September 2011; Accepted 11 April 2012 INTRODUCTION Since the 1950s, karst spring discharge in northern China has been declining (Guo et al., 2005). Similar problems can be found in many other countries, such as the USA (Beynen et al., 2007), France (Gams et al., 1993), Italy (Sauro, 1993), Germany (Heinz et al., 2008), and Serbia (Jemcov, 2007). Worldwide, 25% of the population is supplied largely or entirely by groundwater from karst aquifers (Ford and Williams, 1989; Jaquet et al., 2004). Water resources in this area are important, and the depletion of karst groundwater has become a severe environmental and social issue. Karst aquifers are generally considered to be particu- larly vulnerable to environmental change (White, 1988; Leibundgut, 1998; Butscher and Huggenberger, 2008). Groundwater depletion is a response of karst aquifer to climate and human influences (Guo et al., 2005; Hao et al., 2009). Based on the projected climatic variation scenarios determined by general circulation models, Loáiciga et al. (2000) simulated the potential impacts on groundwater in the Edwards Balcones Fault Zone (BFZ) karst aquifer in Texas, USA. Their simulation results indicated that a scenario with doubled atmospheric CO 2 could exacerbate negative impacts and water shortages in the Edwards BFZ aquifer, even if ground- water pumping rate does not increase from its present average rate. A cross-correlation analysis by Chen et al. (2004) that examined the importance of climate on groundwater level variation indicated that both precipita- tion and annual mean temperature were strongly correlated with annual groundwater levels in the upper carbonate aquifer in southern Manitoba, Canada. Ma et al. (2004) applied a seasonal decomposition method to study *Correspondence to: Yonghong Hao, Tianjin Key Laboratory of Water Environment and Resources, Tianjin Normal University, No. 393 Binshuixi Road, Xiqing District, Tianjin 300387, China. E-mail: haoyh@sxu.edu.cn; haoyhong@yahoo.com. HYDROLOGICAL PROCESSES Hydrol. Process. 27, 2605–2613 (2013) Published online 8 June 2012 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/hyp.9381 Copyright © 2012 John Wiley & Sons, Ltd.