Abstract. Two key sources of uncertainty in projections of future runoff for climate change impact assessments are uncertainty between global climate models (GCMs) and within a GCM. Within-GCM uncertainty is the variability in GCM output that occurs when running a scenario multiple times but each run has slightly different, but equally plausible, initial conditions. The limited number of runs available for each GCM and scenario combination within the Coupled Model Intercomparison Project phase 3 (CMIP3) and phase 5 (CMIP5) data sets, limits the assessment of within-GCM uncertainty. In this second of two companion papers, the primary aim is to present a proof-of-concept approximation of within-GCM uncertainty for monthly precipitation and temperature projections and to assess the impact of within-GCM uncertainty on modelled runoff for climate change impact assessments. A secondary aim is to assess the impact of between-GCM uncertainty on modelled runoff. Here we approximate within-GCM uncertainty by developing non-stationary stochastic replicates of GCM monthly precipitation and temperature data. These replicates are input to an off-line hydrologic model to assess the impact of within-GCM uncertainty on projected annual runoff and reservoir yield. We adopt stochastic replicates of available GCM runs to approximate within-GCM uncertainty because large ensembles, hundreds of runs, for a given GCM and scenario are unavailable, other than the Climateprediction.net data set for the Hadley Centre GCM. To date within-GCM uncertainty has received little attention in the hydrologic climate change impact literature and this analysis provides an approximation of the uncertainty in projected runoff, and reservoir yield, due to within- and between-GCM uncertainty of precipitation and temperature projections. In the companion paper, McMahon et al. (2015) sought to reduce between-GCM uncertainty by removing poorly performing GCMs, resulting in a selection of five better performing GCMs from CMIP3 for use in this paper. Here we present within- and between-GCM uncertainty results in mean annual precipitation (MAP), mean annual temperature (MAT), mean annual runoff (MAR), the standard deviation of annual precipitation (SDP), standard deviation of runoff (SDR) and reservoir yield for five CMIP3 GCMs at 17 worldwide catchments. Based on 100 stochastic replicates of each GCM run at each catchment, within-GCM uncertainty was assessed in relative form as the standard deviation expressed as a percentage of the mean of the 100 replicate values of each variable. The average relative within-GCM uncertainties from the 17 catchments and 5 GCMs for 2015–2044 (A1B) were MAP 4.2%, SDP 14.2%, MAT 0.7%, MAR 10.1% and SDR 17.6%. The Gould–Dincer Gamma (G-DG) procedure was applied to each annual runoff time series for hypothetical reservoir capacities of 1 × MAR and 3 × MAR and the average uncertainties in reservoir yield due to within-GCM uncertainty from the 17 catchments and 5 GCMs were 25.1% (1 × MAR) and 11.9% (3 × MAR). Our approximation of within-GCM uncertainty is expected to be an underestimate due to not replicating the GCM trend. However, our results indicate that within-GCM uncertainty is important when interpreting climate change impact assessments. Approximately 95% of values of MAP, SDP, MAT, MAR, SDR and reservoir yield from 1 × MAR or 3 × MAR capacity reservoirs are expected to fall within twice their respective relative uncertainty (standard deviation/mean). Within-GCM uncertainty has significant implications for interpreting climate change impact assessments that report future changes within our range of uncertainty for a given variable – these projected changes may be due solely to within-GCM uncertainty. Since within-GCM variability is amplified from precipitation to runoff and then to reservoir yield, climate change impact assessments that do not take into account within-GCM uncertainty risk providing water resources management decision makers with a sense of certainty that is unjustified.
Re-evaluation of multi-purpose reservoir yield
