Abstract

In this study, discharge at the outlet of Xijiang River, the biggest sub-basin of the Zhujiang River, was simulated and projected from 1961 to 2099 using the hydrological model HBV-D. The model uses precipitation and temperature data from CISRO/MK3–5, MPI/ECHAM5, and NCAR/CCSM3 under three greenhouse gas emission scenarios (SRES A2, A1B, B1). The results in water resources and flood frequency suggest that annual precipitation and annual runoff would increase after 2050 relative to the reference period of 1961–1990. In addition, increasing trends have been projected in area averaged monthly precipitation and runoff from May to October, while decreasing trends in those from December to February. More often and larger floods would occur in future. Potential increase in runoff during the low-flow season could ease the pressure of water demand until 2030, but the increase in runoff in the high-flow season, with more often and larger floods, more pressure on flood control after 2050 is expected.

Keywords

Zhujiang River Basin ; hydrological model HBV-D ; flood ; low flow ; projection

1. Introduction

Under the background of global warming, air temperatures are rising and more heterogeneous precipitations occur in most basins in China, which cause floods and droughts. Water security, food security and ecoenvironment are threatened along with abrupt increasing in water demand. The Zhujiang River is located in southern China with annual runoff of 3,360 billion m3 . It is the second river in terms of water runoff next to the Yangtze River in China [ Yao, 2004  and Tong, 2007 ]. Recently, seasonal change of runoff becomes more significant due to climate change, which brings more often and large floods, and consecutive dry periods [ Tong , 2007 ]. Especially since 2003, consecutive precipitation and below-normal runoff occurred more often in winter and spring. Analysis of changes in precipitation, temperature and runoff [ Yu and Xie, 2007 , Dong, 2006 , Shi et al., 2007 , Wang et al., 2006 , Liu and Chen, 2007  and Dai et al., 2007 ], and impacting factors on runoff [ Peng et al., 2006 , You et al., 2005 , You et al., 2006  and Qin et al., 2008 ] have been focused. Few researches are concerned with change in future climate and runoff in this region. With the increasing in temperature and precipitation [ Liu et al. , 2009 ], hydrological processes will become more complicated. Research on trends in water resources and hydrological extremes in the future under climate change background is needed by flood control and drought relief sectors. In addition, it is the important scientific basis of designing strategies for climate change mitigation and adaptation.

Different climate models, climate change scenarios and hydrological models are used to study impacts of climate change on water resources and assessment of uncertainty impacts in different river basins [ Koutsoyiannis and Efstratiadis, 2007 , Minville et al., 2008  and Preston and Jones, 2008 ]. It is indicated that the largest uncertainty comes from the GCM structure, followed by the emission scenarios, and hydrological modeling [ Booij, 2005 , Gosling et al., 2011  and Philips and Gleckler, 2006 ]. So impacts studies based on results from only one GCM should be interpreted with caution [ Prudhommed et al. , 2003 ]. In this study, the semi-distributed hydrological model HBV-D driven by output from three GCMs under SRES A2, A1B and B1 scenarios is used to simulate daily discharge from the Xijiang River, the largest tributary stream of the Zhujiang River. Based on the simulations, future changes in water resources and floods are analyzed.

2. Basin introduction

The Zhujiang River has its source in Yunnan province, China. It drains the provinces of Yunnan, Guizhou, Guangxi, Guangdong, Hunan, and Jiangxi in China, and a part of northeastern Vietnam, and flows into the South China Sea. The basin covers approximately 453,700 km2 (excluding the Leizhou Peninsula), within which 442,100 km2 are in China [ Dai et al. , 2007 ]. It is characterized by mountainous area and Karst area in the northwest but mainly lower mountains and hills in the southeast [ Tong, 2007  and Xie et al., 2002 ]. Due to the mountains, there are 48,000 km2 cropland, and 126,000 km2 forest [ Tong , 2007 ]. The Zhujiang River consists of the Xijiang River, Beijiang River, Dongjiang River, and some smaller rivers within the Zhujiang River Delta. Xijiang River (Fig. 1 a) is the largest tributary among them, with a length of 2,214 km and a drainage area of 353,100 km2 [ Sun et al. , 2006 ], with about 2,300 billion m3 of water discharge every year [ Li and Zhang , 2006 ].


(a) Geographic location of the Zhujiang and Xijiang River Basins, (b) elevation ...


Figure 1.

(a) Geographic location of the Zhujiang and Xijiang River Basins, (b) elevation and distribution of meteorological stations, (c) river systems and hydrological stations in the Xijiang River Basin

A tropical to sub-tropical climate prevails while the East Asian monsoon has strong influences on the basin [ Fischer et al. , 2011 ]. The annual mean temperature varies from 14°C in the west to 22°C in the east, the relative humidity varies from 71% to 80% [ Tong , 2007 ], and the annual average precipitation is 1,200 mm in the west and more than 2,000 mm around the coastline [ Gemmer et al. , 2011 ]. The seasonal precipitation varies significantly, mainly due to the impact of the East Asian monsoon. Precipitation during flood season accounts for 72% of the annual amount. With significant inter-annual variability, the annual precipitation in wet years is 6–7 times as the amount in dry years. As a result, flood or drought disasters are likely to occur [ Liu and Chen , 2007 ].

3. Available data

3.1. Observations

Daily temperature and precipitation records of 33 meteorological stations in the Xijiang River Basin during 1961–2006 are employed. The datasets are provided by the National Meteorological Information Center (NMIC) of the China Meteorological Administration. The datasets have been controlled on their quality [ Qian and Lin , 2005 ]. There are 109 hydrological stations in the basin, among which 23 measure the discharge through a reservoir, and 82 measure the discharge through a river. Because of limited data availability, the data of daily stream flow from 1961 to 2006 through the outlet (at Gaoyao hydrological station) is used to calibrate and validate the hydrological model. The locations of the meteorological and hydrological stations are shown in Figure 1 b and 1 c.

3.2. GCMs

Daily temperature and precipitation from three GCMs for three greenhouse gas (GHG) emission scenarios (SRES A2, A1B, and B1) in the 21st century are used to drive the hydrological model HBV-D. These data are provided by the IPCC Data Distribution Centre (Table 1 ). In order to keep a consistent length of the three datasets, the end of the data is set at 2099.

Table 1. Basic features of the three GCMs of the IPCC AR4 and GHG emission scenarios
Country Model name GCM OCM Sea ice Land model Scenario
Australia CSIRO-MK3–5 T63L18, 1.875°×1.875° MOM2.2 L31, 1.875°×0.925° A2, A1B, B1
Germany MPI-ECHAM5 ECHAM5, T63L32, 2°×2° OM L41, 1.0°×1.0° OM A1B
USA NCAR-CCSM3_0 CAM3 T85L26, 1.4°×1.4° POP1.4.3 L40, (0.3°–1.0° )×1.0° CSIM5.0, T85 CLM3.0 A2, A1B, B1

4. Methodology

4.1. Hydrological model

A distributed hydrological model is an important tool used to simulate stream flow under current and future climate conditions. In general, a semi-distributed hydrological model is widely applied because of simple parameters and low requirements for data. HBV and its various versions have been successfully applied in over one hundred basins worldwide [ Xu , 2009 ]. Being one of the newer versions, HBV-D is appropriate for large scale basins [ Krysanova et al. , 1999 ]. In this study, the daily stream flow through Gaoyao station from 1976 to 1990 is used to calibrate HBV-D, while the data for 1961–1975 and 1991–2006 are used to validate the model. The values of the Nash-Sutcliffe coefficient are greater than 0.80, the volume error is smaller than 5%, and the correlation between observation and simulation is significant at 99% confidence level (Table 2 ). Moreover, the flow duration curves (daily runoff depths) of the simulation match the observations well (Fig. 2 ). It is suggested that HBV-D can well simulate daily discharge of the Xijiang River.

Table 2. Performance assessment of HBV-D for daily stream flow
Period Nash-Sutcliffe Volume error (%) Correlation coefficient
Calibration (1976–1990) 0.807 –0.7 0.907
Validation (1961–1975) 0.868 –0.4 0.934
Validation (1991–2006) 0.861 –3.9 0.932


Flow duration curves of observed and simulated daily flow through Gaoyao station ...


Figure 2.

Flow duration curves of observed and simulated daily flow through Gaoyao station for the (a) calibration period (1976–1990), and (b), (c) validation periods (1961–1975, 1991–2006, respectively)

4.2. Trends and flood frequency analysis

Daily stream flow is simulated by HBV-D, which is driven by output of three GCMs under SRES A2, A1B and B1 scenarios. High flow concentrates from June to August, while low flow is apparent from December to February. The Mann-Kendall test [ Kendall , 1975 ] is used to identify trends in such indicators as annual and monthly precipitation and runoff, runoff during high-flow (JJA) and low-flow (DJF) seasons, and annual maximum 1-day and 7-day runoff. The general extreme value (GEV) distribution is used to estimate flood frequency [ Ding and Jiang , 2009 ]. Changes in flood frequency are analyzed for the near-term (2011–2050) and long-term (2051–2099) periods relative to the baseline period (1961–1990).

5. Changes in water resources

5.1. Annual and monthly precipitation and runoff

The regional mean precipitation is computed by averaging the precipitation at all grid points in the Xijiang River Basin for each GCM. Upward trends have been observed in monthly precipitation from May to October and annual precipitation during the 21st century, while downward trends in monthly precipitation from November to next February are observed only under certain scenarios. The direction of the trend varies under different scenarios in March and April (Fig. 3 a). Tendencies in runoff are similar to those in precipitation (Fig. 3 b), i.e., upward tendency in runoff during the high-flow season but downward tendency during the low-flow season.


Statistics of Mann-Kendall test for basin-averaged annual and monthly (a) ...


Figure 3.

Statistics of Mann-Kendall test for basin-averaged annual and monthly (a) precipitation and (b) runoff in the Xijiang River Basin in the 21st century (2011–2099)

Relative to the baseline period, the direction of changes in annual precipitation/runoff is uncertain before 2050 in regard to the increasing trends in annual precipitation/runoff over the whole 21st century. They could increase by 2.5%/7.5%, respectively, as projected by NCAR under SRES B1, or decrease by 7.0%/13.1%, respectively, as projected by CSIRO under SRES A2 (Table 3 ). However, the direction of trend is consistent with almost all projections after 2050. Six of seven projections show positive trends. The ensemble mean of annual precipitation/runoff shows an increase by 6.9%/12.5%, respectively.

Table 3. Anomaly percentage of projected annual precipitation and runoff relative to the 1961–1990 mean, before and after 2050 in the 21st century (%)
Period Variable NCAR MPI CSRIO Ensemble mean
A2 A1B B1 A1B A2 A1B B1
2011–2050 Precipitation 1.0 2.3 2.5 –0.8 –7.0 –6.3 –4.7 –1.9
Runoff –2.2 3.0 7.5 –3.9 –13.1 –9.7 –12.1 –4.3
2051–2099 Precipitation 14.6 16.4 8.6 8.1 0.6 –0.4 0.3 6.9
Runoff 26.4 29.2 15.5 13.9 0.2 3.2 –0.5 12.5

The annual precipitation projected by NCAR under SRES A2 will slightly increase (1.0%) before 2050, while annual runoff will slightly decrease (2.2%). Though some studies suggested that the effects of CO2 enrichment may lead to reduced evaporation, and hence either greater increases or smaller decreases in the volume of runoff [ Kundzewicz et al. , 2007 ], still further explanations are needed.

5.2. Runoff during high-flow season and low-flow season

5.2.1. High-flow season

Floods often happen in the high-flow season when the runoff accounts for 50%–60% of its annual amount [ Dong, 2006  and Shi et al., 2007 ]. An upward tendency in runoff has been observed significant at 99% confidence level during the high-flow season over the 21st century (Table 4 ). In comparison to the baseline period, the ensemble mean will decrease before 2050, but increase afterwards (Fig. 4 a).

Table 4. Statistics of Mann-Kendall test of projected runoff in the high-flow season and low-flow season, and annual maximum 1-day and 7-day runoff in the 21st century
Variable NCAR MPI CSIRO
A2 A1B B1 A1B A2 A1B B1
High-flow 12.57 11.50 9.75 8.21 7.29 8.80 9.05
Low-flow –8.31 –3.03 0.56 –6.42 1.96 –11.09 6.27
Annual maximum 1-day runoff 12.96 5.41 5.66 11.60 8.34 9.19 8.17
Annual maximum 7-day runoff 13.10 5.83 6.16 11.83 8.35 9.34 8.20


Anomaly percentages of projected 30-year running average runoff relative to the ...


Figure 4.

Anomaly percentages of projected 30-year running average runoff relative to the 1961–1990 mean for the (a) high-flow season and (b) low-flow season in the 21st century

5.2.2. Low-flow season

The tendency in runoff during the low-flow season is uncertain (Table 4 ) for all projections. Upward trends have been observed for three of seven projections, while downward trends prevail for the other four (Table 4 ). Relative to the baseline period, a significant increase has been tested for the projections by CSIRO under SRES A2 and B1 scenarios, while only insignificant changes are obtained for the others (Fig. 4 b). Before 2030, the direction of change is the same for almost all projections, while it differs much more afterwards. The direction of change is different between the high-flow season and low-flow season during similar periods, which is probably because the sensitivity of seasonal precipitation and runoff to scenarios is to some degree different between summer and winter [ Kundzewicz et al. , 2007 ].

5.3. Floods

5.3.1. Flood intensity

A significant upward tendency at 99% confidence level has been found in the intensity of annual maximum 1-day and 7-day runoff for the 21st century (Table 4 ). The trends are illustrated in anomaly percentages in Figure 5 . Relative to the baseline period, the intensity of annual maximum 1-day runoff will gradually increase after 2015, while the 7-day runoff will increase after 2020. The magnitude of increase is significant in both indices after 2050. As a result, more focus should be put on flood control mechanisms, as flood control facilities will probably have to deal with more severe conditions after 2050.


Anomaly percentages of projected 30-year running averaged annual maximum (a) ...


Figure 5.

Anomaly percentages of projected 30-year running averaged annual maximum (a) 1-day and (b) 7-day runoff relative to the 1961–1990 mean

5.3.2. Frequency

The magnitudes of the annual maximum 1-day and 7-day runoff series are fitted using the GEV distribution. The frequency is then derived by using the GEV model. In comparison with the baseline period, the frequency for 30-year overlapping period will increase in most projections except those corresponding to MPI and CSIRO under SRES A1B (Fig. 6 ). Especially after 2050, floods will occur in less than 15 years, or even in 3–7 years as projected by NCAR for the three emission scenarios. The changes in flood become more consistent among models as lead time extension. And this is the case for annual runoff and high-flow season. Some studies suggest that this change is the result of CO2 enrichment, which may lead to reduced evaporation, increased runoff and concentrated heavy precipitation in some regions [ Bates et al. , 2008 ]. In addition, climate projections using multi-model ensembles also show increases in precipitation over East Asia over the 21st century [ Kundzewicz et al. , 2007 ].


Changes in projected annual maximum (a) 1-day and (b) 7-day runoff for the ...


Figure 6.

Changes in projected annual maximum (a) 1-day and (b) 7-day runoff for the 30-year overlapping period in the 21st century

6. Conclusions and discussion

(1) The trends in annual and monthly precipitation are consistent with those in runoff in the 21st century. Upward trends have been found in annual precipitation/runoff as well as in monthly time series from May to October, and downward trends appear in the time series from November to February. An increasing trend has been found in discharge during the high-flow season, as well as in the annual maximum 1-day and 7-day runoff series. However, there is high uncertainty in the trend of discharge during the low-flow season.

(2) In comparison with the baseline period, the potential increase in discharge during the low-flow season will reduce the water stress. But the potential increase in flood frequency and intensity as well as higher discharge during the high-flow season will threat the flood control and flood security after 2050. So, the probability of increases in discharge during the low-flow season in the near-term, and in flood frequency and intensity in the long-term must be accounted for when designing flood prevention projects and planning strategies for long-term water resources regulations.

The limitations in understanding the climate system and predicting climate change using GCMs lead to high uncertainty in these climate projections. Consequently, high uncertainty is apparent in impact assessment. In addition, the evapotranspiration hasn’t been analyzed due to limitation in data availability. This will lead to additional uncertainty in this study.

Acknowledgements

Data from three GCMs were obtained from the IPCC Data Distribution Centre (http://ipcc-ddc.cru.uea.ac.uk ). This study was supported by the National Basic Research Program of China (No. 2010CB428401). The authors are thankful for the professional suggestions and comments on the original manuscript by the anonymous reviewers and for the English polishing by the expert and the editors.

References

  1. Bates et al., 2008 B. Bates, Z. Kundzewicz, S. Wu; Climate change and water; Technical paper of the Intergovernmental Panel on Climate Change, IPCC Secretariat, Geneva (2008), p. 210
  2. Booij, 2005 M.J. Booij; Impact of climate change on flooding assessed with diriment spatial model resolutions; Journal of Hydrology, 303 (2005), pp. 176–198
  3. Dai et al., 2007 S.-B. Dai, S.-L. Yang, A.-M. Cai; Variation of sediment discharge of the Pearl River Basin from 1955 to 2005; Acta Geographica Sinica (in Chinese), 62 (5) (2007), pp. 545–554
  4. Ding and Jiang, 2009 Y.-G. Ding, Z.-H. Jiang; Methods Introduction to Extreme Climate (in Chinese), China Meteorological Press (2009), p. 232
  5. Dong, 2006 D.-H. Dong; Discussion on flood control and drought resistance in the terms of annual extreme precipitation over plenty of and lack of precipitation regions in the Pearl River Basin; Pearl River (in Chinese), 5 (2006), pp. 33–34
  6. Fischer et al., 2011 T. Fischer, M. Gemmer, L. Liu, et al.; Temperature and precipitation trends and dryness/wetness pattern in the Zhujiang River Basin, South China, 1961–2007; Quaternary International, 244 (2011), pp. 138–148
  7. Gemmer et al., 2011 M. Gemmer, T. Fischer, T. Jiang, et al.; Trends of precipitation extremes in the Zhujiang River Basin, South China; J. Climate, 24 (3) (2011), pp. 750–761
  8. Gosling et al., 2011 S. Gosling, R. Taylor, N. Arnell, et al.; A comparative analysis of projected impacts of climate change on river runoff from global and catchment-scale hydrological models; Hydrology and Earth System Sciences, 15 (2011), pp. 279–294
  9. Kendall, 1975 M.G. Kendall; Rank Correlation Methods (4), Charles Griffin (1975), p. 202
  10. Koutsoyiannis and Efstratiadis, 2007 D. Koutsoyiannis, A. Efstratiadis; Uncertainty assessment of future hydroclimatic predictions: A comparison of probabilistic and scenario-based approaches; Journal of Hydrometeorology, 8 (2007), pp. 261–281
  11. Krysanova et al., 1999 V. Krysanova, A. Bronstert, D. Müller-Wohlfeil; Modelling river discharge for large drainage basins: From lumped to distributed approach; Hydrological Sciences Journal, 44 (2) (1999), pp. 313–331
  12. Kundzewicz et al., 2007 Z.W. Kundzewicz, L.J. Mata, N.W. Arnell, et al.; Freshwater resources and their management; M.L. Parry (Ed.), et al. , Climate Change 2007: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press (2007), pp. 173–210
  13. Li and Zhang, 2006 X.-Y. Li, P. Zhang; Analysis on the “05·06” heaviest rain-storm flood of Xijiang River; Acta Scientiarum Naturalium Universitatis SunYatseni (in Chinese), 44 (2) (2006), pp. 266–268
  14. Liu et al., 2009 L.-L. Liu, T. Jiang, F. Yuan; Observed (1961–2007) and projected (2011–2060) climate change in the Pearl River Basin; Advances in Climate Change Research (in Chinese), 5 (4) (2009), pp. 209–214
  15. Liu and Chen, 2007 Y.-Q. Liu, C.-M. Chen; Temporal variation of precipitation during flood season in the Pearl River Basin; Pearl River (in Chinese), 4 (2007), pp. 47–51
  16. Minville et al., 2008 M. Minville, F. Brissette, R. Leconte; Uncertainty of the impact of climate change on the hydrology of a Nordic watershed; Journal of Hydrology, 358 (2008), pp. 70–83
  17. Peng et al., 2006 D. Peng, J.-P. Wen, J. Tang; Climate background and strong portent of summer severe drought and flood in the Xijiang River Basin; Guangdong Meteorology (in Chinese), 28 (3) (2006), pp. 27–30
  18. Philips and Gleckler, 2006 T.J. Philips, P. Gleckler; Evaluation of continental precipitation in 20th century climate simulations: The utility of multimodel statistics; Water Resources Research, 42 (2006) http://dx.doi.org/10.1029/2005WR004313
  19. Preston and Jones, 2008 B.L. Preston, R.N. Jones; Evaluating sources of uncertainty in Australian runoff projections; Advances in Water Resources, 31 (2008), pp. 758–775
  20. Prudhommea et al., 2003 C. Prudhommea, D. Jakob, C. Svensson; Uncertainty and climate change impact on the flood regime of small UK catchments; Journal of Hydrology, 277 (2003), pp. 1–23
  21. Qian and Lin, 2005 W. Qian, X. Lin; Regional trends in recent precipitation indices in China; Meteorology Atmospheric Physics, 90 (2005), pp. 193–207
  22. Qin et al., 2008 X.-H. Qin, M.-Q. Jian, Y.-T. Qiao, et al.; Relationship between the temporal variations of the river discharge in the Xijiang River and Beijiang River Basins and the moisture sinks; Acta Scientiarum Naturalium Universitatis SunYatseni (in Chinese), 47 (2) (2008), pp. 136–139
  23. Shi et al., 2007 J.-Z. Shi, Z.-L. Wang, X.-H. Chen; Trends and spatial distribution of areal average temperature in the Zhujiang River; Pearl River (in Chinese), 5 (2007), pp. 4–12
  24. Sun et al., 2006 H.-G. Sun, J.-T. Han, S.-R. Zhang, et al.; The impact of extraordinary flood in “2005·06” on carbon flux from river; Chinese Science Bulletin (in Chinese), 51 (23) (2006), pp. 2773–2779
  25. Tong, 2007 J. Tong; The generalizations and hydrological character analysis of Pearl River Basin; Water Conservancy Science and Technology and Economy (in Chinese), 13 (1) (2007), pp. 31–33
  26. Wang et al., 2006 Z.-L. Wang, X.-H. Chen, L. Zhang; Spatio-temporal change characteristics of precipitation in the Pearl River Basin in recent 40 years; Journal of China Hydrology (in Chinese), 26 (6) (2006), pp. 71–75
  27. Xie et al., 2002 Z.-Q. Xie, Z.-M. Yao, J.-P. Li; Analysis on characteristics of floods in “1994·6” and “1998·6” and comparison; Hydrology (in Chinese), 22 (3) (2002), pp. 56–58
  28. Xu, 2009 Z. Xu; Hydrological Models (in Chinese), Science Press (2009), p. 525
  29. Yao, 2004 Z.-M. Yao; Analysis of recent variation of water resources and water consumption in the Pearl River Basin; Journal of China Hydrology (in Chinese), 24 (5) (2004), pp. 20–23
  30. You et al., 2005 D.-W. You, C.-L. Tang, S. Deng; Preliminary analysis on relationship between the climate warming and the discharge of Xijiang River in the recent 50 years; Guangdong Meteorology (in Chinese), 27 (4) (2005), pp. 4–6
  31. You et al., 2006 D.-W. You, C.-L. Tang, S. Deng; The early atmosphere-ocean characteristics for discharge abnormity of Xijiang and Beijiang River in winter; Guangdong Meteorology (in Chinese), 1 (2006), pp. 18–21
  32. Yu and Xie, 2007 Y.-G. Yu, H.-X. Xie; New hydrological characters of Pearl River since 1990s; Pearl River (in Chinese), 4 (2007), pp. 23–25
Back to Top

Document information

Published on 15/05/17
Submitted on 15/05/17

Licence: Other

Document Score

0

Views 9
Recommendations 0

Share this document

claim authorship

Are you one of the authors of this document?