Water resource implications of decadal climate variability
south-western Australia
Sadler, Ruprecht and Leathersich
Summary
This paper examines the climatic uncertainties which have faced water managers in South Western Australia over the last two and a half decades in respect to the ongoing augmentation of sources of the Perth Water Supply. The decisions which have been made and the high stakes associated with those decisions are traced. The paper illustrates how a precautionary approach to these decisions has steered the water supply through extreme decadal scale drought and shows why water managers in the south western Australia are key partners in the of the Indian Ocean Climate Initiative.
South western Australia is a region of winter rainfall and dry summers. Rivers in the region are not perennial. They flow in winter and spring and cease in summer. Water supply systems therefore must be built with enough reservoir capacity to capture water not only to supply the following summer demand season, but also to carry the system through a possible succession of dry years. Perth's supply comes from a mix of surface and groundwater supplies. The main groundwater systems are shallow unconfined aquifers linked to important wetlands. Similar storage issues exist with groundwater sources as well as surface water sources.
A feature of Perth's water system is that sources available for supply augmentation, as demand grows, are not large in relation to demand. Water managers are thus under constant pressure to satisfy demand growth by planning, gaining public approval, designing and developing regular augmentation of sources. This process takes many years for each individual source increment and involves large amounts of capital. Planning which ensures augmentation is well matched to demand growth is therefore essential to maintain service and to control financial and environmental costs.
As a fundamental underpinning of such planning, water managers need good definitions of climate variability and associated streamflow and groundwater recharge under which the system may be expected to operate. The normal basis of climate definition is that past climate is the best available indicator of the future. In the case of Perth, designs in the 70s were based on periodic updates of the climatic data from 1911, a point in time related to data quality, to the latest year. However, in the 1980's water managers saw reason to question this basic assumption.
During the 60's and early 70's the Perth water system experienced a number of good replenishing years, but in the mid and late 70's a severe drought was experienced at a time of expected rapid growth in demand. Introduction of pay-for-use pricing in that period coupled with: user education; restricting garden watering to cooler hours; some reserve system capacity and a consumer trend into back yard bores, enabled the system to cope with the drought without undue hardship. However, the situation, which is illustrated in Figs. 1,2 and 3, was sufficiently severe to cause re-examination of design policy and concern for the possible implications of an extended drought.
Concurrent with these concerns, international and national concern on greenhouse gas and climate variation gained increasing prominence after the Villach meeting of 1985. Early attempts at modelling these effects, including CSIRO produced scenarios for the National Greenhouse87 Conference: Planning for Climate Change, suggested that an anomalous decline might occur in rainfall of south western Australia. The drought of the late 70s had not abated, and Western Australian water managers were invited to prepare a paper for Greenhouse87 (Sadler, Mauger, Stokes) exploring the possible impacts of climate change on Perth water system.
Figure1 shows the basic climate assumption made for that paper, namely a steady linear decline in mean annual rainfall from 1970 to 2040. Using this conference scenario as a starting point, assessments of decline in streamflow and groundwater recharge were made. These translated into scenarios in which system yield would decline by around 40% at 2040. Projections of demand growth to 2040 were made in order to establish notional water system designs, with and without the Greenhouse87 scenario. By this means the cost and system implications of such a future climate scenario were estimated.
For a 40% decline in streamflow and groundwater, and the demand growth nearest to current projections, the present value of the extra cost of awater system for the Greenhouse87 scenario was estimated to be in the order of $500M.
Subsequently, whilst knowing that the 1987 greenhouse scenario was technically speculative, it was accepted by water managers as yet another factor indicating that the climate base of the water system may be in need of revision. It was decided that from 1987 some progressive downward stepping of the mean annual flows would be introduced to future source planning on the basis that such actions and the state of climate knowledge would need to be reviewed periodically.
Figures 2 and 3 show the climatic outcomes from the 80's to 2000 which were actually experienced in the wake of the 1987 decisions. Whatever the cause, these outcomes represent, in hindsight, an undeniable reason for significantly writing down the water system yields on a long term basis. For the last 25 years, no annual inflow (Figure3) has exceeded the long run mean and the mean annual inflow for 1975-1999 is 172GL or as low as 59% of the 1911-1999 mean of 291GL.
The drought sequence of the last 25 years, is of such duration that, despite the scale of reduction, it may now be argued to constitute a realistic basis for water system design, even if due to natural variability. Significantly, this outcome involves an effective decline in system yield of the same scale as the extreme Greenhouse 87 scenario was projected to reach only by 2040. It is fortuitous that water managers had taken their precautionary approach since the 1980s. There have been several de-ratings of the system yield since that time, and consequent acceleration of development programs. These have enable the system to struggle gracefully through a period which otherwise could have experienced a catastrophic system failure.
It is not argued that this situation is a consequence of greenhouse gas emissions. Current research under the Indian Ocean Climate Initiative would not support this assumption or discount greenhouse effects as a possible component . IOCI research is searching for natural explanations as the primary cause, encouraging investigation to take a wide scope, including the influences of the Indian and Southern Oceans. A primary goal of water managers and of IOCI partners is to establish improved climate baselines which are appropriate for future planning.
Whilst not accrediting the past 25 years to greenhouse emissions it would be equally invalid to ignore this issue, particularly when recognising that the climatic definitions being introduced now, are setting a baseline for decades ahead. In most major international GCMs the south west of Australia, is the one part of the globe, which consistently displays declining rainfall under various greenhouse assumptions.
Water managers do not have the luxury of proven models. They must make decisions which weigh up risks as best they can. It is likely that their adopted climate base for the future will be some engineering mix of assumptions about decadal variability and a risk of some component of rainfall decline from greenhouse impacts.
In dollar terms and also in environmental and social terms the stakes in these issues are high. The adjustments to Perth's water system over the past 25 years have been very large. If not "anticipated", the drought would have been catastrophic. If done unnecessarily, the changes would have been very costly.
Current demand forecasts are slightly above, and the current system de-rating is similar to, those of the Greenhouse87 at 2040. It can be inferred for the new baselines, that the cost implications of the changes to date, plus those changes necessary in the future as demand grows, are of a similar order of magnitude to those estimated for Greenhouse 87. In other words a present value dollar cost of these adjustments as a result of re-defined climate variability, in order of magnitude terms, is $500M.
The significance of improved definitions of climate variability, involving appreciation of phenomena of decadal variability, are illustrated for south western Australia. Large costs and considerable uncertainty are associated with decisions based on climate definitions. National research and development on such phenomena is well justified.
B.S. Sadler, G.W. Mauger and R.A. Stokes: The water resource implications of a drying climate in south-west Western Australia. CSIRO Australia 1988, Greenhouse: planning for climate change, pp296-311 (ISBN 0 643 04863 4).
Ruprecht, J.K., Bates, B.C. and Stokes, R.A., eds: 1996, Climate Variability and Water Resources Workshop, Water and Rivers Commission, Water Resources Technical Report Series No WRT 5. (ISBN 0 7309 7253 4).
Indian Ocean Climate Initiative Panel, 1999: Towards Understanding Climate Variability in south western Australia - Research Reports on the first Phase of the Indian Ocean Climate Initiative. Water & Rivers Commission (ISBN: 0 643 06599 7)
Ruprecht, J.K., The Effect of Climate Variability on Streamflows in South Western Australia. Water and Rivers Commission, 1999. Unpublished Report SWH 25.