Impact of climate change on freshwater resources: an inter-basin comparison using a consistent set of scenarios

Daniel Kingston1, Richard Taylor1, Martin Todd1, Julian Thompson1, Walter Collischonn2, Denis Hughes3, Marcio Nobrega2, Robin Thorne4, Hongmei Xu5

1Department of Geography, University College London, London, UK. Email d.kingston@ucl.ac.uk

1. Introduction • There is significant uncertainty regarding the potential impacts of climate change on global freshwater resources

• Such uncertainty arises, in part, from the diverse range of climate and developmental scenarios applied by different studies

• Under the NERC QUEST-GSI project, we address this research issue by applying a common suite of climate and developmental scenarios

• These are used to drive a series of hydrological models across a range of different river basins

• Subsequent research will also involve comparison of basin-scale hydrological models with a global-scale hydrological model (MacPDM, Arnell 2003)

4. QUEST-GSI Basins

2. Climate scenarios •  High resolution (0.5° lat/lon) climate scenarios are produced by

pattern-scaling GCM-generated patterns of global climate change (Climgen; Osborn et al. 2008)

•  Scenarios are designed to explore three aspects of uncertainty in projected climate change:

1.  Greenhouse gas emissions

•  Four IPCC SRES emissions trajectories (A1b, A2, B1 and B2)

2.  Climate sensitivity

•  Time-slice scenarios of prescribed changes in global mean temperature from the 1961-90 baseline

3.  GCMs

•  Emissions and prescribed warming scenarios will be created based on results from seven different GCMs (all from the IPCC AR4)

•  CCCMA CGCM31, CSIRO Mk30, HadCM3, HadGEM1, IPSL CM4, MPI ECHAM5, NCAR CCSM30

3. Methods • Hydrological models are calibrated for the 1961-90 period using CRU TS 3.0 gridded climate data (at 0.5° lat/lon resolution)

• All models are semi-distributed

• Liard and Mekong use SLURP model

• Mitano, Xiangxi and Huangfuchuan use SWAT model

• Okavango uses Pitman model

• Grande uses MGB-IPH model

• Results are presented here for SRES scenario A1b for 2040-69

5. Results: SRES A1b for 2040-69

Liard

6. Summary • Substantial variation occurs in projected changes in river discharge between the seven different GCMs

• No overall consistency in the direction of change for all river basins

• Variation between GCMs at the monthly level is replicated for average annual runoff

• Variation between GCMs is generally a function of differing precipitation totals (rather than temperature)

• The potential evapotranspiration method used by each model is also a substantial source of uncertainty

Affiliations 2 Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil 3 Rhodes University, Grahamstown, South Africa 4 McMaster University, Hamilton, Canada 5 National Climate Centre, Beijing, China

References

Arnell 2003. Hydrology and Earth Systems Science 7, 619-641.

Osborn T. et al. 2008, http://www.cru.uea.ac.uk/~timo/climgen/

Okavango

Huangfuchuan

Xiangxi

Grande

Mekong

Mitano

Liard (Mackenzie)

Grande (Parana)

Okavango

Mitano (Nile)

Huangfuchuan (Yellow R)

Xiangxi

(Yangtze)

Mekong