Piping hot water

I picked this up from a post in Facebook … the source of truth and accuracy. I have heard quite a lot about the abundance of water that we are “letting run to the sea” and similar nonsense. So I need to write something about it.

Water flowing to the sea down the Murray?

Yes, a lot of water does run into the sea. Some even from the Murray-Darling does get there. However this does not mean that the water is economically transportable and able to be managed with a “simple pipeline”. You might see a very large flow down the Murrumbidgee or Murray rivers right now and think that this is water going to the Murray Mouth. Reality is that this water is flowing from large dams in the upper catchments (Hume Weir for the Murray and Burrinjuck for the Murrumbidgee) to be pumped out or diverted for irrigation. The only efficient way to get the required water to the irrigation areas is, perversely, to almost flood the river channels. I saw this in action recently while travelling from Canberra to Melbourne.

So here are a few facts:


  • About 7.1 Gl of water is flowing into South Australia per day.
  • 1.2 Gl is delivered via the Murrumbidgee as part of the Inter Valley Trade arrangements. This comes from Burrinjuck and Blowering Dams. losses are in the order of 0.3 Gl and that means 1.5 Gl is released to achieve the required output
  • About 3.4 Gl is released from MDBA storages that flow from the Hume Weir
  • There is an amount around 1 Gl released from the Snowy Scheme
  • The Goulburn is contributing 1.6 Gl
  • other Victorian tributaries contribute about 1.4 Gl
  • The Darling is contributing 0 – in fact taking some of the water out of the Murray
  • Overall that is 8.6 – the balance is what is evaporating

Outputs in the lower river

  • Of the 7.1 Gl a day going into SA:
  • 3.3 Gl a day goes to SA consumption – essentially a large chunk of the Adelaide water supply
  • 2.9 Gl a day is lost to seepage and evaporation
  • the remaining 0.9 Gl is environmental water. This is water to maintain sufficient level in Lake Alexandrina to prevent it becoming too salty to support its natural ecosystem
  • There is no flow out of the Murray Mouth. There is an occasional release of water through the Barrages at low tide to reduce salinity. That is all.

So 7.1 Gl seems like a lot of water. So lets see what the diversions for irrigation are. Firstly lets see what the security of water entitlements are.

NSW – Murray Valley   Victorian – Murray Valley
High security 97% General security 0%   High reliability 56% Low reliability 0%
NSW – Murrumbidgee Valley   Victorian – Goulburn Valley
High security 95% General security 6%   High reliability 64% Low reliability 0%
NSW  – Lower Darling   South Australia – Murray Valley
High security 30% General security 0%   High security 100%

Major Diversions from Murray and Lower Darling (GL) *

New South Wales This Week From 1 July 2019 Victoria This Week From 1 July 2019
Murray Irrig. Ltd (Net) 1.8 98 Yarrawonga Main Channel (net) 3.5 63
Wakool Sys Allowance 2.3 24 Torrumbarry System + Nyah (net) 0 141
Western Murray Irrigation 0.6 10 Sunraysia Pumped Districts 6.2 47
Licensed Pumps 4.7 56 Licensed pumps – GMW (Nyah+u/s) 1 10
Lower Darling 0.0 0 Licensed pumps – LMW 4.6 169
TOTAL 9.4 188 TOTAL 15.3 430

The tables above are from the December 18 MDBA River Murray Weekly Report. Victoria and NSW have different regulation for entitlements however it is clear that plenty of water from the MDBA Storage is consumed for irrigation. Then we need to look at the NSW data.

Consumption expected for irrigation (annual) Gl
Conveyance 265
Losses (transmission, evaporation, operational) 278
Total Losses553
Announced High Security (95%) 348
Announced General Security (6%) 113
Total Irrigation462

So there is 1015 Gl for the year and about double the daily average is used in Summer. That gives 5.6Gl a day down the Murrumbidge for irrigation purposes. 1.2 Gl for Adelaide, the SA consumption and maintaining the health of the Coorong and Lake Alexandrina. The comparison is a little less extreme for the Victorian situation. More of the water that goes down the stream reaches SA and there is less consumption (15.3 Gl a week compared to 39 Gl for the Murrumbidgee).

Human needs

342 Gl is reserved/allocated to critical human needs across the Basin. The majority of this is used in the large cities, Adelaide, Canberra, Albury/Wodonga, Toowoomba, Shepparton, Bendigo, Wagga Wagga etc. It is not clear to me whether that number includes extractions of 70Gl to Melbourne from the Goulburn River. The fact that it is so hard to find information on community use of water and so easy to find information on what irrigator entitlement is very instructive.

Critical Human Needs are prioritised first and then other uses which are theoretically decided on economic grounds. I have written other articles on the relative value of water for different uses see: http://petaguy.info/blog/sustainability/water-use-in-the-murray-darling. In fact high security irrigation water gets prioritised over other uses that might deliver higher economic value because of the special status assigned to it in legislation. it is hard to argue that the 70Gl diverted to Melbourne or the amount allocated to Adelaide is bad economics; quite the reverse.


So, as with all things water, it is complicated. But with a lot of digging into the data, it is clear that water running in the major rivers is actually travelling to be used for high security allocation (mostly irrigation but also industrial) and critical human needs. Not to run into the sea.

A tiny fraction of the water going down the Murray and Murrumbidgee rivers is reaching the ocean. Around 40% is evaporating and nearly all the rest is being consumed. About 5 Gl a year is used to reduce salinity in the Coorong and Lake Alexandrina – less than is used for irrigation in a day on the Murrumbidgee alone. Life sustaining uses by basin communities consume around 120 Gl a year. One third of a Gl a day.

Around 60% of all available water is used for irrigation in the Murray Darling Basin. less than 10% is used for people who live in the Murray Darling Basin. Most of the unavailable cannot be stored or harnessed. The remainder can only be used for irrigation by degrading the natural environment – therefore affecting people’s lives and livelihood.

So what about pipes?

Pipes are a different thing again. Let’s see if we can find an answer to the question “Why don’t we pump water from the big rivers to the dry ones?”. The following are areas I will analyse and discuss:

  • Small municipal pipes
  • Industrial pipelines
  • Existing inter-valley pipes
  • Long pipes
  • Pipes over mountains
  • Economics

The Problem with Pipes

Pipes are a very good way of transporting water downhill and over relatively short distances. However, pipes are not that efficient for carrying a liquid over long distances. Some basics.

A typical irrigation channel in the Murray valley can move at least 100 cubic metres of water a minute or 0.1 Ml/m. gravity can move this much with a slope of a few mm per km. A pipe, is much smaller and you need to add energy to overcome friction. Rather than energy consumption, the measure used in pipes is pressure. 1 Bar is the same as atmospheric pressure. Two Bar of pressure is considered “good” pressure from a tap.

Taking a 50mm pipe that is 18 km long, and carrying 100l per minute, you need about 28 bar of pressure at the water tower. That water travels at 0.85 m/sec. You would not want to use a higher pressure than that at the tower.

Of course, you could go to a 100mm pipe and the pressure drop would only be 1 Bar. The problem is that a 100mm pipe costs about 7 times the cost of the 50 mm one when installation costs are taken into account. The pipe materials alone cost 5 times as much. In practice water networks use large main pipes (400 mm or more) to supply a suburb and a smaller one (100mm) to supply a street then 19mm or smaller to supply a house.

Small Municipal Pipes

Most pipes we come across are small municipal pipes that carry water to homes and factories. Their source water is usually quite close and high above the city or town. We rarely consider that the water needs to be pumped to local water towers and that there is a cost to that. After all, we need the water and are happy to pay for it. Even if that is $3-4,000 per mega litre. Most of that cost is for storage, pumping and treatment. Then there is a supply fee that covers some of the cost of maintaining the pipes and water towers (usually located on convenient hills nearby a suburb)

Within that substantial cost it is easy to forget that there is that inbuilt cost of transporting the water by pipe.

We know that copper pipes cost a lot. Just ask any plumber what it might cost for a 50m pipe – you are unlikely to get a quote for less than $1000 just to supply it. Steel pipe is cheaper to buy but will not last as long in the ground, especially.

Industrial Pipelines

Industrial pipelines are larger than the domestic ones because they need to carry more water. This usually means a higher fixed supply cost because the infrastructure costs typically 5-10 times that of a domestic connection to install and maintain. Cost of the pipe is roughly linear with length of the pipe and at least 7 times the cost to install each time you double it between 50mm and 400mm. Sizes above that cost about 10 times when you double the size and are of a different construction.

If you want a flow of 1 kl/sec with a 5km pipe length (from a 400mm main) and the source pressure is limited to 50 Bar you will need around 200mm pipe diameter. This means a cost equivalent to supplying a a water main for a street of 100 -200 houses. Roughly $100,000.

In reality industrial suburbs share the infrastructure costs so the individual costs are lower. However, I am focusing on costs of pipes.

Long Pipes

To get water from, say the Atherton Tablelands to the upper reaches of the Darling system you would have to pipe water about 1,600 km.

What happens when you want to put the same 1kl of water per minute through a 1600 km pipe? You will need to have a pressure of 230 Bar to push that through. However high pressure pipe costs more. Or you could go to 300mm pipe to get a pressure drop of 32 Bar. Either way the cost increases per km. So if we take the baseline of a 5km pipe of 200mm at $100,000 and factor the length and size increase in we get $100,000 x 4 x 320 = $128 million (rounding up the 3.94 factor for diameter increase – 1600/5=320)

Now 1 kl/sec is a trickle compared to the needs of irrigators. Lets see what happens when we scale up to one irrigation channel of flow. We need a 1500 mm pipe to handle that flow. The diameter is 5 times as big and just that factor would increase the cost by 25 times. However the construction of the pipes changes and the complexity of the pipe laying too for many reasons. Overall the cost is going to be something like 50 times and possibly 100 times if the terrain is difficult.

That means the cost is going to be at least $6.4 billion for the equivalent of a single irrigation channel flow of water. If we wanted to do the equivalent of 10 irrigation channels, which would be significantly increasing the capacity to irrigate in the Darling, the cost would be around $13 billion with a 3000mm pipe and much more risk involved.

Some further issues with long pipes include the fact that they will have to cross private property and impact the environment. Both these factors have potentially high costs for compensation.

Pipes over Mountains

Pipes going over mountains have additional problems. Pumping water uphill is expensive. Every kl of water pumped up a 100m rise takes about 100 million joules. If we want 100 kl per minute pumped then we have to provide 167 kW of power to do that. Every hour that happens, there is a consumption of 167 kW hours of electricity – to deliver 6 mega litres of water in that hour. The cost of that per hour is about $11,000. If it operated for a year at full capacity, 52 Gl of water would be pumped at a cost of at least $96 million dollars.

You do get some energy back again when the water flows down hill. It will be lost eventually because of friction and over a long pipe the friction losses are high. However, there are additional issues with water pressure and fluid dynamics that create significant risks in a long pipe with a large flow.

The take away is that mountains cost money when you put lots of water over them.

Case Study – North-South Pipeline

The Sugarloaf or North-South Pipeline travels 70 km and crosses the Great Dividing Range. It cost $750 million to construct. It uses about 260 MW hours of electricity to pump up to 75 GL per annum and less in wetter years. See the link below.



Lets work out how much energy it would cost to send water from just beyond Townsville to the Darling. Here are the key factors. 100 kl per minute sent 1600 km with the water pumped 100 m up and then dropping 300 m to an open outlet. It needs to run at a discharge rate of 1.7 kl per second and with a 1.5m diameter smooth concrete pipe most of the way and a high pressure metal pipe over the mountains, the flow rate is just under 1 m/sec. in this scenario the head loss is just over 1,000 m due to friction – resulting in a nett head loss of 800. I will optimistically assume that the overall drop is 200m and that there are no pumping losses to lift the water (which are actually large).

167 kWh x 8 = 1.3 MWh per hour to pump 6 Ml of water from source to outlet. This is using the calculation for a 100m drop/rise and multiplying it to reflect a 800m head loss. If operated full time the cost would be 11.4 GWh and a cost of $1.7 million to pump 52 Gl of water if electricity costs 15 cents a kWh.

The Bottom Line

The analysis preceding shows that there is a high cost to pumping significant amounts of water over long distances. There is a business case for doing this for augmenting city water supplies, however there is hardly a case for providing irrigation water at a cost of $32,850 a Gl. Irrigators buy water for much less than that.

My calculations are conservative. Water delivered through a long pipeline will be much more expensive. A 70 km pipeline cost over $10 million a km to build. A 1,600 km pipeline scales up to $16 billion at that rate – without complications. Operating costs for pumping water on top of the high capital cost demonstrate why this is not feasible.

I do not think this is a realistic solution to the water problem. As unpalatable as it may seem, the better option is likely to be to re-allocate water from low productivity irrigation and learn to farm sustainably – like the original owners did for thousands of years.