While researching desalination, I found this interesting presentation (big PPP file) from the Water Corporation, including (slide 22) an estimated cost of $6.10 for a Kimberley pipeline with a capital cost of $11 billion[1]. It’s from a 2004 GHD report, so there’s no question of its being a political stunt. The pumping requires two 600 MW power stations, which suggests (if my mental arithmetic is right) around 50kw-hours for each Kl of water delivered. I’ve extracted the relevant slide here.

As an aside, I saw in the Fin that the proponents are suggesting diverting as much as 80Gl of the flow to irrigation along the way, leaving only 120/Gl for Perth. It’s unrealistic to expect any contribution to the capital costs from irrigation users, so the entire capital cost would have to be spread over the 120/Gl supplied to Perth, and the unit pumping cost would probably also rise. The delivered cost could easily exceed $10/Kl.

Update As Rob Corr points out, towing icebergs from Antarctica would be a lot cheaper.

Further update Rob’s source puts the iceberg option at $3.20/kl, but I’ve found another study that puts it at $17.00. So a canal at $10/kl would split the difference between these two.

fn1. Either the implied rate of return here is well below the one I used or the capacity is higher. I’d guess they are assuming bond financing and low depreciation. Still this sounds like a more realistic figure than the $2 billion Tenix estimate being used by Barnett. The power stations alone would chew up most of that.

Reader Nic White asks for some comments on WA Premier Geoff Gallop’s desalination plan, and I’m happy to oblige, as this is a topic I’ve been meaning to do some work on. We’re talking here about about desalinating seawater or groundwater for human use, rather than schemes for reducing salt inflows to river systems like the Murray-Darling, another big topic in itself.

There are two basic ways of going about this. One is distillation. The most common approach to distillation is to evaporate the water, leaving salt and other pollutants behind, and capture the steam, but freezing and vapour compression. The other is to separate pure water using reverse osmosis or electrolysis. The first approach is the traditional one, but it’s inherently energy-intensive, so unless you have a cheap source of waste heat, it’s becoming outdated. The top candidate at present is reverse osmosis, which involves passing the water through a membrane, and using pressure to reverse the normal osmotic flow from high salt concentration to low. The current energy cost is about 4.5KwH for each Kl of seawater desalinated (the cost increases with the salinity of the source). Electrical energy is required, so this would come in around 25c/kl. The main operating cost comes from the need to replace the membranes.

I found this report (1.2 Mb PDF) which focuses on small scale plants for remote areas, up to 50/kl a day or around 15Ml/year. I assume there are significant scale economies beyond this, but it’s worth noting that, unlike large-scale engineering works, desalination is an incremental option, so you shouldn’t have problems of excess capacity. Operating costs are estimated at 65c/kl for a source with 2000mg/l up to $1.89 for 35 000 mg/l (seawater), for an output salinity less than 500mg/l. I’d guess the optimal way to go would be to accept more saline output and dilute it with fresh water. At a rough guess, I think a larger scale plant could produce water with operating costs of $1/kl

Capital costs are about $1600/kl/day or about $5/kl/year for small scale plants. That’s $5 million for each GL of annual capacity, compared to $10/GL in the unlikely event that the canal alternative could be delivered for $2 billion[1]. Assuming BOOT financing as I did for the canal (a high-cost option, but I want to be fair), the annual capital charge would be around 70c/kl, for a total of $1.70/kl, before reticulation and any additional treatment.

At that price, desalination is a pretty expensive option, and I’d expect to see some fairly dramatic reductions in optional water uses, like watering lawns. Before going to seawater desalination on a large scale, it would be sensible to work through the cheaper options, such as conservation, repurchase of irrigation water and use of groundwater in appropriate locations. This large PowerPoint file has some interesting data.

The availability of desalination as a backstop also suggests we need to take a sceptical look some of the more overblown rhetoric implying that urban Australians are going to run out of water. If we conservatively put the cost of large scale desalination up to $2.50/kl and assume water use of 200kl/person/year (you could manage a suburban lifestyle, including a water-efficient garden but no lawn on half that) it’s still only $500/person/year or $1500 for a three-person household. Not trivial, but cheaper than broadband or cable TV.

fn1. In my post, I suggested staged construction costs of $3 billion (still v. conservative) which gives a capital cost of $15/kL capacity and an annual capital charge of about $2/kl.