Common questions about climate change and Sydney's water balance.
Climate change refers to changes in our weather and environment caused by increasing levels of carbon dioxide and other greenhouse gases in the atmosphere. Human activity, and particularly the burning of fossil fuels for cars, buildings, industry and electricity generation, emits increasingly large amounts of these gases.
Greenhouse gases trap heat in the earth's atmosphere. Over time, more and more heat is retained, leading to an increase in the earth's average temperature - global warming. There is mounting evidence that our climate is changing rapidly and it is getting warmer.
Climate variability refers to normal variations in climate over a long period of time. Variability may be due to natural processes within the climate system, such as the El Niño-Southern Oscillation. Human induced external processes, such as increasing carbon dioxide (CO2) in the earth's atmosphere due to human activity, may change the properties of variability, including frequency, intensity and duration.
The Sydney Water Balance Project focused on future impacts of climate change in relation to water supply and demand. However it is important to note that naturally occurring climate variability will continue to have major impact on Sydney's water supply and demand. Further research will need to be undertaken to increase our understanding of climate variability and how this may affect Sydney's climate in coming decades.
Global Climate Models (GCMs) are the most advanced tools for investigating the causes of observed climate change and projecting future climate change. A GCM is a complex mathematical representation of the earth's climate system that reflects changes in climatic variables such as wind, temperature, humidity and rainfall. GCMs typically provide outputs at a resolution of around 200 km x 200 km. Worldwide, there are 23 GCMs that attempt to predict the changes in climate under different carbon emission scenarios.
Climate models are able to reproduce the significant features of the observed climate very well and there is a high level of confidence in their ability to provide credible quantitative estimates of future climate change, particularly at broad continental scale and above. The highest confidence is attached to results analysed at the coarsest space and time scales, such as global or hemispheric levels. Confidence decreases with finer scales, such as sub-continental or regional levels.
Modelling climate change impacts at the local or regional level using GCM outputs is less certain. At finer scales the magnitude of natural climate variability increases and regional climate signals, such as the El Niño-Southern Oscillation and Southern Annular Mode are easily masked. Furthermore, local influences on climate (such as regional topography or processes) become more important at finer spatial scales.
Runs of the CSIRO Mark 3.0 (Mk3) GCM were used in this study because it was the only readily available model which provided continuous daily data of the climate predictors for both present day and future climates. CSIRO Mk3 has also been found to be one of the best models in simulating Australia's climate and associated large-scale climate drivers.
Ideally, additional GCMs would have been used to provide a range of possibilities and to reduce uncertainty in the projections, but these were not readily available in the timeframe of the study.
Economists and other experts have developed a number of scenarios or possibilities for how the world might develop over the next century based on a set of assumptions, such as how fast population might increase and how quickly renewable energy sources might replace fossil fuels.
In 2000, a set of greenhouse gas and aerosol emission scenarios for the 21st century were developed for use in climate model simulations. Research groups around the world use these scenarios to project climate changes. The results are featured in the report of the Intergovernmental Panel on Climate Change (www.ipcc.ch).
This study used the B1, A1B and A2 emission scenarios, representing low, mid and high emission futures. At the beginning of this study (June 2006) it was considered that the B1, A1B and A2 scenarios would provide an adequate spread of possible futures, ranging from optimistic (B1) to pessimistic (A2).
However, recent thinking on emission scenarios suggests that the more pessimistic scenario (A1F1) may be the more realistic scenario for future conditions. In consideration of this the study presents the outcomes for A1B and A2. Results for B1 are available in appendix four of the report, and in associated technical papers (Mehorotra and Sharma 2010, SCA 2009, Sydney Water 2009).
Climate models are becoming more and more realistic, but significant uncertainty exists. Firstly, it is impossible to know future atmospheric greenhouse gas concentrations with great certainty since this depends both on economic growth rates and the extent of mitigation strategies adopted internationally. Secondly, there remain significant limitations in the capability of global climate models to replicate important features such as the impact of clouds and aerosols, which heavily influence the presence of drought events into the future. Finally there are significant limits in the capacity of downscaling methods to estimate climate impacts at the local or regional level.
The broad scale resolution of Global Climate Models (GCMs) is too coarse for driving hydrological models to project changes in rainfall patterns, as local features and dynamics are not well represented at this scale. Therefore, to obtain the finer resolution required for hydrological models, the coarser data outputs from GCMs are downscaled.
There are three general classes of downscaling techniques: perturbation (daily scaling), statistical and dynamical (physically-based).
The statistical downscaling framework developed by UNSW (Mehrotra and Sharma 2010) consistes of daily mathematical models for rainfall occurrences, rainfall amounts (volume), temperature and pan evaporation based on daily atmospheric variable outputs from the CSIRO Mk3 model.
Applying a direct percentage change to historical rainfall, temperature and any other variables is the simplest method for accounting climate change. This method ignores the changes in rainfall over time and the magnitude and frequency of events. It can underestimate the influence of larger rainfall events in the impact on total runoff. Daily statistical downscaling accounts for these changes by using the daily atmospheric variables.
Climate modelling is characterised by uncertainty at three levels:
Emission scenarios: It is not possible to estimate future atmospheric greenhouse gas concentrations with great certainty since this depends both on economic growth rates and the extent of mitigation strategies adopted internationally.
GCM performance: There remain significant limits in the capability of GCMs to model important features such as the impact of clouds and aerosols. Additionally, key forcing parameters such as solar radiation and volcanic activity cannot be predicted into the future.
Downscaling limitations: There remain significant limits in the capability of downscaling methods to estimate climate impacts at the local or regional level.
The climate period for 1960-2002 was used as the baseline in the study as it was representative (at the commencement of the study) of the recent average climate in the Sydney region. This baseline did not include the recent drought of 2001-2007 because the study commenced before the drought ended.
Averaged over a large number of global climate models, there is a tendency for a small increase in summer total rainfall but a small reduction in winter total rainfall, leading overall to a small change in annual total rainfall.
There may be little change in the number of wet days for both 2030 and 2070 but the amount of rain in these days may increase by about four percent by 2030 and two percent by 2070.
Sydney may see an increase in the number of extreme rainfall days (where more than 40 mm falls) across all seasons with a maximum increase of about 30 percent in spring in 2030 and about 20 percent in autumn in 2070.
Longer dry spells (of 15 days or more) could increase in 2030 and 2070. By 2070 Sydney is likely to see longer dry spells interrupted by heavier rainfall events.
The best estimate for an increase in Sydney's daily maximum temperature is 0.5o C by 2030 and 1.5o C by 2070.
Hot days, where the daily maximum temperature is above 35o C, are projected to increase to four days each summer in 2030 and seven days each summer in 2070 (up from the current average of three days each summer). The frequency of hot spells (periods when four to seven days each have a daily maximum temperature greater than 27o C) is also projected to increase.
The Sydney Water Balance study is the only report to date that provides a forecast on the impacts of climate change on the Sydney Water supply system. A modified system yield (system output) was used to determine yield changes.
On average it is forecast that there will be a reduction in system output for both 2030 and 2070, however there is considerable uncertainty surrounding the projections. The projections indicate a decrease in annual rainfall and runoff in the inland catchments and a minor increase in the more coastal catchments by 2030.
Sydney's water supply system uses records from between 1909 and 2008. The climate change study began in 2006 and at the time global climate model (GCM) simulations were only available to 2002. To ensure consistency a shorter 43-year period, based on the availability of downscaled GCM simulation between 1960 and 2002 was used to assess the possible impact of climate change on Sydney's future water supply. This period has a range of climate characteristics which provide sufficient representation of the current climatic period.
The modelling to determine the impacts on supply in the future used the base case scenario from the 2006 Metropolitan Water Plan.
Outdoor water use (mainly from garden watering) is strongly influenced by rainfall and evaporation, particular in summer. If it rains more, there is less need to water gardens, so less demand for water. On other hand, drier and hotter conditions with high evaporation will increase the need to water gardens, so higher demand for water. Similarly, cooling towers in commercial buildings use more water when it is hot and dry.
Long-term changes in the climate therefore have the potential to increase the underlying demand for water in Sydney Water’s area of operation. A drier, hotter future may mean a greater demand for water if all other factors hold constant.
The increase in water demand for Sydney will be influenced more by natural climate variability than human induced climate change impacts.
In general, the increase in average demand due to climate change is relatively small, about 1.1 percent by 2030 and 3.9 percent by 2070, under a medium to high (A2) emission scenario.
The highest increase in average annual demand due to climate change (from the current climate demand of 639GL/year) is about 25 GL/year 2070, under a medium to high (A2) emission scenario. This is much less than the estimated range for the variability in annual demand (52 billion litres/year in 2030 and 73 billion litres/year in 2070).
There are a number of parameters that are considered in climate change studies which aim to determine impacts at the regional or local level, including historical and future timeframes, global climate model/s used, emission scenarios and type of downscaling methods used.
Depending on research budgets and timeframes, different studies for the same region may use different scientific methods and may therefore result in different outcomes.