Assessing the Impact of Climate Change on the Water Resources of the River Basin,

Yoichi FUJIHARA1, Kenji TANAKA2, Tsugihiro WATANABE3, and Toshiharu KOJIRI4 1,3Research Institute for Humanity and Nature, 457-4 Kamigamo-motoyama, Kita-Ku, Kyoto 603-8047 2,4Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji 611-0011

1. Introduction mm in the northern area. The annual inflow at the Seyhan Dam ranges between 3.7 and 7.3 Gm3 and The Intergovernmental Panel on Climate averages 5.5 Gm3. The Seyhan and Catalan Dams Change (IPCC) Third Assessment Report con- have storage capacities of 0.8 and 1.6 Gm3, re- cluded that there was evidence that most of the spectively. The stored water is used mainly for warming observed over the last 50 years is at- irrigation. According to the 1990 statistics, the tributable to human activities. With the expected amount of irrigation water used annually is about build-up of greenhouse gases in the atmosphere, 1.4 Gm3, and it is increasing annually (Figure 1). it is anticipated that the climate will continue to The amount of domestic water used annually is 0.1 change throughout the 21st century. Moreover, it Gm3 according to the 2003 statistics. is thought that global warming will have a signifi-

) 10.0 cant impact on the hydrology and water resources 3 Annual Inflow m

9 9.0 of river basins. Irrigation Water 8.0 Basins that have a large fraction of runoff driven Domestic Water 7.0 by snowmelt, such as the Basin 6.0 in Turkey, will be especially sensitive to global 5.0 warming, because the temperature determines the 4.0 fraction of precipitation that falls as snow and the 3.0 timing of snowmelt. In this paper, the climate 2.0 projected using two general circulation models 1.0 Annual Inflow at the Seyhan Dam (10 0.0 (GCMs) under the Special Report on Emissions 1990 1992 1994 1996 1998 2000 2002 2004 Scenarios (SRES) A2 emissions scenario was used Fig.1 Annual inflow, irrigation water, and to drive hydrologic models to assess the impact of domestic water use at the Seyhan Dam. climate change on the water resources of the Sey- han River Basin. 3. Approach

2. Study Basin 3.1 Downscaling Method The raw outputs of GCMs are inadequate for as- The Seyhan River Basin (21,700 km2) is lo- sessing the impact of climate change on the hy- cated in southern Turkey between 34.25-37.0◦E drology and water resources of river basins, be- and 36.5-39.25◦N. The lower basin is dominated cause the temporal and spatial resolution of GCMs by the Mediterranean climate, while the middle is too coarse compared to those of hydrologic and upper basins are influenced by the Continental models that are applied to river basins. This study climate. applied a dynamic downscaling method called The annual precipitation is about 700 mm in the pseudo warming (Sato et al., 2006) to connect the coastal area, increases to approximately 1,000 mm output of the raw GCMs and river basin hydro- at higher elevations, and decreases to about 400 logic models. 600 600 Observed Observed

/s) Simulated (Present) Simulated (Present) /s) 3 3

400 400

200 200 Inflow at the Station 1818 (m Inflow at the Station 1818 (m

0 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1 2 3 4 5 6 7 8 9 10 11 12 (a) Monthly inflow (b) 10-year average monthly inflow Fig.2 Simulated hydrograph at Station 1818.

The pseudo warming downscaling method is as discharge. Nevertheless, since the input data were follows. For the current climate simulation, the downscaled data, the hydrologic models repro- pseudo-warming method uses reanalysis data as duce the river discharge at station 1818. The an- a boundary forcing of the regional climate model nual inflow at the Seyhan Dam is shown in Fig- (RCM). A specially created boundary condition, ure 3. This figure shows that the simulated results in which changes in meteorological variables pro- agree with the observed data. jected in a GCM simulation are added to reanaly- sis data, is used to simulate global warming.

) 10.0 3 Observed The GCMs used in this study were m 9 9.0 Simulated (Present) MRI-CGCM2 (Yukimoto et al., 2001) and 8.0 CCSR/NIES/FRCGC-MIROC (K-1 Model De- 7.0 velopers, 2004) under SRES A2. The downscaled 6.0 data covered two subset periods (the 10 years 5.0 present and the 10 years future; Kimura et al., 4.0 2007), and were used to drive hydrologic mod- 3.0 2.0 els to assess the impact of climate change on the 1.0 water resources of the Seyhan River Basin. Annual Inflow at the Seyhan Dam (10 0.0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

3.2 Hydrologic Model Fig.3 Annual inflow at the Seyhan Dam. We used a land surface model (Simple Bio- sphere including Urban Canopy (SiBUC); Tanaka 3.3 Reservoir Models and Ikebuchi, 1994) to estimate the surface en- We developed reservoir models to simulate the ergy and water balance components. In addition, reservoir operations of the Seyhan and Catalan we used the stream flow rooting model of Hydro- Dams. We examined the historical record, includ- BEAM (Kojiri et al., 1998) to simulate river dis- ing the inflow, water level, and dam discharge, and charge and incorporated a reservoir model in this interviewed the dam operators about the actual op- flow rooting model. erations. From these analyses, we used the follow- The region simulated was a 2.75 × 2.75◦ ing operational rule as a basic rule: water is stored area (34.25-37.0◦E and 36.5-39.25◦N) with a 5- to maintain a target operational water level and the minute latitude-longitude spatial resolution (33 × demand water is released regardless of the level. 33 grids). The simulated hydrograph at station The simulated river discharge using the flow 1818 is shown in Figure 2. There were some rooting model described in section 3.1 is input into discrepancies between the simulated and observed the reservoir models. The target operational wa- 1800 1000 ) Observed 3 Observed m Simulated 6 800 Simulated 1500 600 ) 3 m 6 1200 400

200 900 1000 1000 Reservoir Volume (10 0 Observed /s) Observed /s) Simulated 800 3 Simulated 800 3 600 600 600 Reservoir Volume (10 400 400 300 200 200 Dam Discharge (m Dam Discharge (m 0 0 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 (a) Catalan Dam (b) Seyhan Dam Fig.4 Simulated reservoir volume and discharge. ter level is the average of historical operational 3.4 Land and Water Use Scenarios records, and the demand water is the actual water The land and water use at the present period withdrawal for irrigation and domestic use. were the actual conditions in the Seyhan River The simulated reservoir volume and dam dis- Basin. For the future period, the following three charge at the Seyhan and Catalan Dams are shown scenarios were used: in Figure 4. The simulated volume and discharge (a)Future: The land and water use are the same agreed with the observed values. Although no re- as at present. sults are shown in here, we found that the estab- (b)Adaptation 1: The land and water use are un- lished reservoir models also reproduced the hydro- der low investment conditions. The cropping pat- electric generation quite well. Figure 5 shows the tern in the Lower Seyhan Irrigation Project (LSIP) simulated inflows with and without the reservoir simulated by Umetsu et al. (2007) is used to es- models. This figure clearly indicates that the reser- timate the water demand. In addition, the effects voir models can reproduce the actual reservoir op- of global warming on the irrigation water require- erations. ments are considered using the SiBUC simulation. (c)Adaptation 2: The land and water use are un-

600 Observed /s)

3 Simulated (with reservoir model) Simulated (without reservoir model) 400

200 Inflow at the Seyhan Dam (m

0 1 2 3 4 5 6 7 8 9 10 11 12

Fig.5 Simulated hydrograph at the Seyhan Dam. der high investment conditions, in which 25% of 120 the rain-fed winter wheat is converted to irrigated 100 Present crop, and citrus is cultivated in this area. The crop- Future (MRI) Future (CCSR) ping pattern in the LSIP simulated by Umetsu et 80 al. (2007) is used as a future scenario to calculate the water demand. The effects of global warming 60 on the irrigation water requirements are also con- 40

sidered using the SiBUC simulation. Monthly Precipitation (mm) 20

4. Results 0 Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.

4.1 Temperature, Precipitation and Stream Fig.7 Precipitation changes predicted under different models. Flow Changes The monthly mean temperatures are compared 400 in Figure 6. The average annual temperature change for the Seyhan River Basin was +2.0◦C Present 300 Future (MRI) /s) in the Meteorological Research Institute GCM 3 Future (CCSR) (MRI) and +2.7◦C in the Center for Climate Sys- tem Research GCM (CCSR). The monthly pre- 200 cipitation is compared in Figure 7. The aver-

age annual precipitation change for the Seyhan Monthly Inflow (m 100 River Basin was -159 mm in MRI and -161 mm in CCSR. The decreases in precipitation in Jan- 0 uary, April, October, November, and December 1 2 3 4 5 6 7 8 9 10 11 12 were greater than in the other months. Fig.8 Stream flow changes predicted under different models. The monthly mean inflow at station 1818 is shown in Figure 8, which shows that the future inflow will decrease remarkably compared to the 4.2 Water Resources System Effects present. In addition, the decreases in the April, The ratio of water withdrawal to discharge is May, and June inflow are greater than in the other shown in Figure 9. Many studies (e.g., Alcamo months, and the peak monthly inflow occurs ear- et al., 2003; Oki and Kanae, 2006) have reported lier than at present. that a region is considered highly water stressed if this index exceeds 0.4. The ratio is less than 0.4 at

30 present, while it ranges from 0.4 to 0.7 in the fu-

Present ture period, from 0.4 to 0.6 for Adaptation 1, and

) Future (MRI) from 0.5 to 1.0 for Adaptation 2. !n Future (CCSR) The reservoir volume at the Seyhan Dam is 20 shown in Figure 10. The reservoir volume for the future and Adaptation 1 is less than at present, and in a few cases, the reservoir is empty. By contrast, 10 in Adaptation 2, the reservoir is frequently empty.

Monthly Mean Temperature ( The reliability of the dams is shown in Figure 11. The reliability (R) is defined using the follow- 0 Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. ing equation: R = Vs/Vd (1) Fig.6 Temperature changes predicted under different models. where Vs is the volume of water supplied, and Vd is the volume of water demanded. The index at 1.6 1.6

1.4 Present 1.4 Present Future Future 1.2 Adaptation 1 1.2 Adaptation 1 Adaptation 2 Adaptation 2 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2 Ratio of Water Withdrawals to Discharge Ratio of Water Withdrawals to Discharge 0.0 0.0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 (a) MRI (b) CCSR Fig.9 Ratio of water withdrawal to discharge. present is usually 1. This indicates that the dams tion 1; and can supply the entire demand. The value in the 3. The effects of global warming and the in- future and for Adaptation 1 is always 1 in MRI creased demand for water in the upper basin will and ranges from 1 to 0.95 in CCSR. By contrast, lead to water scarcity at the LSIP in the case of for Adaptation 2, the reliability is from 1 to 0.7 in Adaptation 2. MRI and CCSR. These results lead to the following conclusions. Acknowledgments Although the ratio of water withdrawal to dis- This research was supported financially by the charge will increase due to the effects of global Project Impact of Climate Changes on Agricul- warming (decreased discharge), it is possible to tural Production System in Arid Areas (ICCAP), supply the demand for water from the water re- administered by the Research Institute for Human- sources system in the future case and Adaptation ity and Nature (RIHN) and the Scientific and Tech- 1. By contrast, the effects of global warming and nical Research Council of Turkey (TUBITAK). In the increased demand for water in the upper basin addition, this research was also supported finan- will lead to water scarcity at the LSIP in Adapta- cially in part by Japan Society for the Promotion tion 2. of Science (JSPS) Grant-in-Aid No. 16380164.

5. Conclusions References In this study, the climate projected using two Alcamo, J., P. Doll, T. Henrichs, F. Kaspar, B. GCMs under SRES A2 was used to drive hy- Lehner, T. Rosch, and S. Siebert (2003) Global drologic models to assess the impact of climate estimates of water withdrawals and availability change on the water resources of the Seyhan River under current and future business-as-usual con- Basin. The results showed that: ditions, Hydrological Sciences Journal, 48(3), 1. Compared to the present, decreased precipita- 339-348. tion will result in a considerably decreased inflow, Kimura, F., A. Kitoh, A. Sumi, J. Asanuma, and in which the peak monthly inflow occurs earlier A. Yatagai (2007) Downscaling of the global than at present; warming projections to Turkey (in this volume). 2. The ratio of water withdrawal to discharge Kojiri, T., A. Tokai, and Y. Kinai (1998) Assess- will increase due to the effects of global warm- ment of river basin environment though simu- ing (decreased discharge), although it is possible lation with water quality and quantity, Annuals to supply the demand for water based on the water of Disaster Prevention Research Institute, Kyoto resources system in the future and using Adapta- University, 41(B-2), 119-134. 1000 Present Adaptation 1 1000 Present Adaptation 1 Future Adaptation 2 Future Adaptation 2 ) ) 3 3 m m 6 6

500 500 Reservoir Volume (10 Reservoir Volume (10

0 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 (a) MRI (b) CCSR Fig.10 Reservoir changes.

Present Future Adapt 1 Adapt 2 Present Future Adapt 1 Adapt 2

1 1

0.5 0.5 Volumetric reliability for LSIP Volumetric reliability for LSIP 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Year Year (a) MRI (b) CCSR Fig.11 Reliability changes.

K-1 Model Developers (2004) K-1 coupled sphere Model Including Urban Canopy model (MIROC) description, K-1 Technical Re- (SiBUC) for regional or basin-scale land sur- port 1, Center for Climate System Research, face processes, Proceedings of the International University of Tokyo. Symposium on GEWEX Asian Monsoon Ex- Oki, T., and S. Kanae (2006) Global hydrological periment, 59-62. cycles and world water resources, Science, 313, Umetsu et al. (2007) (in this volume). 1068-1072. Yukimoto, S., A. Noda, A. Kitoh, M. Sugi, Y. Sato, T., F. Kimura, and A. Kitoh (2007) Projec- Kitamura, M. Hosaka, K. Shibata, S. Maeda, tion of global warming onto regional precipi- and T. Uchiyama (2001) The New Meteoro- tation over Mongolia using a regional climate logical Research Institute Coupled GCM (MRI- model, Journal of Hydrology, 333, 144-154, CGCM2): Model climate and variability, Papers doi:10.1016/j.jhydrol.2006.07.023. in Meteorology and Geophysics, 51, 47-88. Tanaka, K., and S. Ikebuchi (1994) Simple Bio-