QUEST-GSIImpacts of climate change on

global water resources

Nigel Arnell, Simon Gosling, Matt Charlton

QUEST-GSI water resources workshopUCL March 2009

Overview

• Impacts on global water resources

• Characterising adaptation?– Global reservoirs

Estimating impacts on water resources

1. Simulate river runoff across the global domain using a macro-scale hydrological model

2. Calculate indices of water resources stress

3. Determine change in indices

Macro-scale hydrological model

MacPDM- daily water balance accounting model- parameters estimated from spatial data sets

(soil, vegetation)- PE estimated using Penman-Monteith- soil moisture characteristics vary across a grid cell- no routing from cell to cell- not calibrated at the catchment scale

MacPDM performance

Tends to overestimate runoff in dry regions –evaporation of runoff and transmission loss

Intercomparison

long term mean annual runoff (NA - A1B)

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

WaterGAP MATSIRO MacPDM VIC

runo

ff [k

m3 /a

]

baseline

HadCM3Echam5

WATCH intercomparison, North America

Parameter uncertainty

S England

-5-4-3-2-1012345

% c

hang

e fr

om c

ore

So ut hern Eng land

0123456789

10

0 10 20 30 40 50

Number of r epeti tions

Annual r unof f Dr ought r unof f

10-year max mon 10-year max day

stochastic parametric

Change in runoff: priority A1b emissions

Change in runoff with increasing temp

Indicators of water resources stress

Indicator = water resources per capita by watershed

Indicator of impact of climate change= populations living in water-stressed watersheds, where runoff decreases significantly

Water-stressed: < 1000m3/capita/year

“significant decrease”: decrease in runoff greater than standard deviation of 30-year mean runoff

Distribution of stresses

A1b 2050 – core models

A1b 2050: increase in exposure

Impacts on global resources stress

Increase in stress

Regional impacts

Incorporating change in withdrawals

Data from Oki et al. (2008)- domestic/municipal- industrial- agricultural

Use different characterisations of SRES storylines

Threshold = withdrawals > 20% of available resources

A1 A2 B1 B22000 3824 3824 3824 38242025 5283 6202 5717 51672055 6634 8891 5938 6321

Total global withdrawals (km3/year)

Withdrawals/resources

HadCM3 2050 2oC

Characterising adaptation

How can we characterise adaptation across a large domain?

- Adaptation options and feasibility vary with local context

Use a hypothetical reservoir with defined yield and reliability

Simple reservoir model

Standard behaviour analysis

Yield=75% of average flow; reliability = 90%

Overview of performance indicators

Consistent with McMahon et al. (2007) figures

Indicator of impact – change in volume

Change in volume to maintain target yield and reliability

HadCM3 A1b 2050

Conclusions

• Broad-scale approach to estimating climate change impacts on water resources stresses

• Generalised approaches to characterising adaptation?