Following several years of severe drought in California, water supplies have recovered to the extent that state officials recently suspended statewide restrictions on water use. During the drought, there had been repeated suggestions that hard-hit localities such as San Diego might consider investing in desalination plants. A common response was that California should take a hard look at Australia’s experience with large-scale desalination during its “millennium drought” from roughly 2004 to 2010.
As it happens, I became very familiar with Australia’s desalination adventure during my tenure at the University of Sydney. We can draw several lessons in risk management from this particular real-world example.
The millennium drought was unusually severe and long-lived––Sydney Water customers watched uneasily as the water level in Warragamba Dam continued to fall. Warragamba is a huge reservoir with a large catchment and storage capacity equivalent to about six years’ supply for Sydney. Managers had imposed restrictions on water use, and these restrictions became more stringent as time passed without a break in the drought. Planners developed a contingency plan for a $2 billion desalination plant to augment supply, and recommended that the plant be built in the event the water level fell below 30 per cent of capacity.
By 2007, the drought was showing no signs of letting up. Residents were tolerating the water restrictions fairly well, but becoming fearful that things might get worse. Engineers routinely calculate the 100-year drought and the 500-year drought (i.e. the worst drought with probability 0.01 and 0.002, respectively), and those numbers suggested that running out of water in Warragamba was very unlikely.
Nevertheless, as the length of drought began exceeding expectations, people began to question the strategy of waiting-out the drought. Was Sydney just experiencing an unlikely run of bad luck, or was the weather system really changing – as the climate scientists had been warning? Some commentators began speaking of drought as “the new normal”.
There is a powerful reason to delay a big-bang solution, such as a $2 billion desalination plant, as long as possible – if it rains in time, the problem goes away. State government officials started to feel the heat as the water level in storage dropped below 40 per cent of capacity, and officials do not like to be accused of fiddling while Rome burns. In due course, the state government succumbed to the “we have to do something” impulse and authorized construction of the Sydney Desalination plant (SDP) in 2007, when the dam level was still a few percentage points above the planners’ trigger.
The plant began operation in 2010 for a two-year proving period, after which it would be governed by rules calling for full-scale operation when the dam level falls to 70% of capacity, and return to stand-by mode when storage returns to 80%. Ironically, the drought had broken while SDP was under construction, and Warragamba was near full capacity for much of the proving period. Water restrictions were eased and, as it turned out, demand did not recover to its pre-drought level – people had learned ways to live just as well while using less water.
In retrospect the evidence suggests that the unprecedented low inflows in the 2000s really were just a run of bad luck, but we can sympathize at least a little with the embattled decision makers in 2007.
In May 2012, the newly-elected state government, which had opposed construction in 2007 when it was in opposition, sold the plant to a consortium of institutional and corporate investors led by the Ontario, Canada, teachers’ pension plan. The purchase price was $2.3 billion (just enough to cover the initial cost plus the transactions costs of the sale), and Sydney Water entered into a 50-year contract with the buyers requiring guaranteed payments of $190 million annually when the plant is in stand-by mode, plus substantial shut-down and start-up payments, all inflation-adjusted. Sydney Water committed to purchasing any water produced by the plant, at rates set by the independent regulatory authority to allow the new owners a 6.7% return on investment. SDP was duly put on stand-by in July 2012 and, given the current water level in Warragamba, it is unlikely to be needed again before 2020. Optimists point out the SDP provides some insurance against a future drought, but realists argue that desalination technology is progressing so fast that SDP will be too inefficient and expensive to use in the event it ever is needed.
So, a major investment was made in order to forestall a potentially devastating but quite unlikely water shortage. In the end, this investment turned out to be unnecessary, and it has little salvage value. While the water authority got its money back, its customers are saddled with substantial payments to the SDP’s new owners for fifty years beginning in 2012.
In a future post, I will consider the challenge faced by decision makers watching a bad situation unfold. It matters whether we are experiencing an unusual run of bad luck or the system itself has changed, and it is not easy for decision makers to figure out which it is.