Runoff projections and hydrological impacts
Runoff projections
The Global Climate Models (GCM) used as part of the IPCC's work are the best available tools for modelling future climate scenarios. However, temperature and rainfall results produced by GCMs are at too coarse a spatial resolution (typically approximately 300 km by 300 km) to allow for an appropriate analysis of hydrologic impacts.
In order to manage water resources into the future, Department of Water and Energy undertook a study to investigate how future projections for temperature and rainfall were likely to impact on runoff and water availability for all of NSW (Vaze et al (2008) Future climate and runoff projections (~2030) for New South Wales and Australian Capital Territory).
The method used was to generate a reference (no climate change) time series of runoff estimates for 5 km by 5 km areas of land at daily time steps for the period 1895 to 2006. The runoff was generated using the historical daily rainfall record and estimated potential evapotranspiration (PET) applied to rainfall runoff models calibrated to over a hundred gauged catchments in NSW.
The study then generated comparable time series of climate change runoff estimates at a reference date of 2030 for the fifteen GCMs that had daily data available for the A1B emissions scenario (see www.ipcc.ch/ipccreports/ar4-syr.htm ) for the current and future time periods. A daily scaling method was used to adjust the historical daily rainfall record. The daily scaling method applies different scaling factors based on ratios of seasonal distributions of daily rainfall totals (for full details of this method refer to the original report). The method adjusts daily rainfall totals on a seasonal basis to be higher or lower than the historical, and maintains the inter-annual and interdecadal patterns.
This approach produced a range of changes to rainfall and resulting runoffs, from significantly wetter to significantly drier futures, reflecting the current level of uncertainty of rainfall projections (note however, there is greater consensus for temperature projections). However, by using only the results of the four best GCMs used by UNSW in the regional assessment project, the range of change was significantly narrowed.
While there are some differences between the results of this study and the regional assessment project (due to the different emissions scenario and time period used) these are considered to be minor.
Analysis of runoff estimates
The study extracted modelled runoff estimates across the State Plan regions for the 112 year modelling period used in the regional assessment project and the four selected GCMs (see Assessing the best models for predicting climate). Results were analysed for:
- percent change in average seasonal runoff depths, and
- percent change in total runoff depths for driest period of record during the 112 year record for period durations of 0.5, 1, 2, 3, 4, 5, 6 and 7 years.
Modelled historical reference and GCM runoff estimates were also extracted for gauged catchments for which the rainfall-runoff models had been calibrated. The catchments with the best reproduction of historical flow distributions were analysed for changes in:
- high flow magnitude, defined as percent change in the flow rate currently exceeded 1% of the time, and
- the frequency of occurrence of low flows, which is the percent change in the amount of time for the flow rate currently exceeded 90% of the time.
Presentation of analysis
Each GCM resulted in a different estimate of runoff changes. The reporting of these results is based on the following characteristics:
- the magnitude of average of the changes from the four GCMs
- the full range of changes (from lowest to highest % change)
- the degree of agreement between the four GCMs in terms of the direction of change.
For the magnitude of changes, the following descriptions are used for the average increase or decrease:
Description | Range for % change in: | ||
Runoff depths | High flow magnitude | Low flow frequency | |
Slight | 0-3 | 0-5 | 0-10 |
Minor | 3-6 | n/a | n/a |
Moderate | 6-9 | 5-10 | 10-20 |
Major | >9 | >10 | >20 |
Terminology for the degree of agreement is as follows:
Description | Number of GCMs |
Very likely | 4 |
Likely | 3 |
More likely than not | 2 |