Neil Smith

In my view there are plenty of good reasons to oppose Gunns’ proposed pulp mill. Is its thirst for water from Lake Trevallyn one of them? I’m not talking about whether a pulp mill is a good use for that much water – I accept that it isn’t, because of the low value of the product to Tasmania and all the other undesirable side-effects of such a monster. But would the act of taking the water seriously disadvantage other users and potential users, or the environment? In an effort to shed some light on it I did a small independent study of stream flow data over the past 7 years (outlined below).

And of course the matter of water availability in the South Esk Catchment has wider implication than just the Gunns issue.

I looked only at data freely available on the internet. I have no contacts in DPIW, Hydro Tas or elsewhere. Should this article flush out better or more up-to-date information I’d be delighted to know it. Gunns’ published intention is to use 26 gigalitres per year (Gl/a), presumably on a continuous basis – so that they’d want it day in, day out, independent of seasons and river levels. They intend to build a pipeline capable of 40 Gl/a, originally for “possible future expansion”, but now it seems that they might actually want more from day 1 – so that they can attract a few irrigation customers along the way who might otherwise be reluctant to allow the pipeline easement.

In their “response to submissions under the EPBC Act” [1] prior to gaining federal approval Gunns stated that, using annual averages over a number of years, the 26 Gl/a would represent “only 1.0% of the flow into the Tamar Estuary” and further that “in a dry month, such as March 2007, the pulp mill off-take for that period would represent only an estimated 3% to 4% of the flows into the Tamar estuary”. At first sight this looks ridiculous – can the average flow in a dry month really be relied on to be a quarter to a third of the annual average, floods and all? Tasmanian streams just aren’t like that.

But of course there are two distinct components of the flow into Lake Trevallyn – the “natural” flow (after all the irrigation entitlements are deducted), and that which flows at the discretion of Hydro Tasmania by the operation of the Poatina power station. Gunns may be approximately correct with their percentages provided Poatina continues to be used during dry periods to the same extent as it has in the past. This flow can often dwarf the natural flow in the South Esk catchment.

To date (and especially pre-Basslink) Poatina has been one of the two mainstays of electricity supply during dry times when “run-of-river” stations can produce only a small fraction of their rated power. Its large-capacity storage (Great Lake) is considered inter-annual; it can absorb the largest runoff events without spill (in fact it has never been full) and represents 48% of the energy storage capabilities of the Hydro system [2].

But Hydro Tasmania has long been engaged in an increasingly difficult balancing act in trying to protect storage levels whilst “keeping the lights on” and maintaining their balance sheet. According to their 2006-07 annual report [3], rainfall averaged over all catchments was only 69.5% of the annual average, and water storages dropped to less than 20% for the first time since 1967. Blackouts were avoided but at the cost of over $100 million in electricity imports via Basslink and buying gas to fuel the Bell Bay power station, rather than running down water storages to even more dangerous levels. At least Basslink provides the physical means of doing it. 2007-08 can only have been worse. At the end of July 2008 Great Lake was sitting at 19.1%, up from a low earlier this year of under 13%.

Now we see the Premier announcing a grand plan to “drought proof” the Midlands – with the Mercury (12/8) even suggesting that farms from Campbell Town to Oatlands could become “intensive horticultural enterprises”. Leaving aside the issue of whether the Midlands is the best place for such activities, is the water really available? Is enough likely to be available every year as climate change bites? It is perhaps significant that the plan now includes an intake from the South Esk river (presumably a winter intake into storage at local Midlands sites) – as well as the previously-mooted water from the Poatina tailrace. Is this possibly an admission of the likelihood that Great Lake levels might be so low at times that operation of Poatina will be impossibly risky? If so, Gunns’ percentages don’t look nearly as tiny. And the availability of water to other legitimate users is not necessarily a given.

The attitude of the present Tasmanian government towards helping Gunns by building the pipeline for them (with the – most useful – compulsory land acquisition thrown in) is still uncertain. But should they try to go this way they may need to cover their embarrassment by offering water to a suitably large number of other users on the East Tamar as well, and this would likely imply a drain from Lake Trevallyn way over 26 Gl/a, possibly even over 40 Gl/a.

The natural flow

Let’s look first at the “natural” (non-Poatina) flow into Lake Trevallyn. The major streams have had telemetered gauging stations in place for many years, and the data are available on the DPIW site [4]. Data from 4 stations (South Esk at Perth, Meander at Strathbridge, Macquarie downstream of Elizabeth River, and Liffey at Carrick) for the period January 2000 to July 2008 were downloaded and summed in a spreadsheet. The sum of these flows should be an approximation of the flow into Lake Trevallyn (but a bit high, when the effect of extraction for irrigation purposes below the gauging stations is considered).

I have worked throughout in units of cubic metres per second (cumecs) as appropriate to short-term flows and for compatibility with Hydro figures. Gunns’ 26 Gl/a equates to a continuous flow of 0.824 cumecs. In terms of the other commonly used unit, 1 cumec equals 86.4 megalitres per day. To keep the dataset manageable monthly averages were used, although daily averages and 15-minute raw data are both available.

Of the 103 months examined since Jan 2000, 16 were discarded because of significant gaps in the record, but enough information is available to show that this makes very little difference to the overall picture.

Of the 87 months analysed 25% of them (that is, 22 months) had a total average flow less than 5 cumecs, 50% less than 14 cumecs, and 75% less than 44 cumecs. There were 5 months with an average flow of greater than 100 cumecs (June to September 2003 and September and October 2005). The record was 167 cumecs (August 2003). There were nine months when the flow would have been insufficient to provide just the pulp mill requirement, the input to drinking water treatment stations, and the guaranteed environmental flow in Cataract Gorge. Gunns’ 0.824 cumecs, expressed as a percentage of the total monthly average natural inflow, ranges from 0.5% all the way up to 59%. There were 15 months when it would have been over 20% and 32 months when it would have been over 10%.

(The continuous environmental flow for Cataract Gorge guaranteed by Hydro Tas is 1.5 cumecs. With regard to drinking water offtakes from Lake Trevallyn, Esk Water has a total entitlement of 6000 Ml/a (0.19 cumecs) and this is the value assumed above. The latest (2006/07) Esk Water annual report reveals somewhat less than this (4064 Ml) actually sold to West Tamar and Meander Valley Councils in that year. But on the other hand, both the Reatta Road (WT) and Mount Leslie (MV) treatment stations are both sized for 20Ml/day, and if they are ever used to capacity in the future the total offtake would be 14600 Ml/a or 0.46 cumecs).

There is an obvious deficiency in using monthly average flow figures. High flows due to major rainfall events commonly last only 2 to 5 days, but the few high figures greatly inflate the average. Even in a fairly wet month, for example April 2003, with a monthly average inflow of 44 cumecs (which is higher than 75% of months during the study period) there were only 7 daily averages higher than the monthly average, and ten days with less than 8% of it – 3.5 cumecs – right down there with the dryest months.

The point I try to make is that because Lake Trevallyn (unlike Great Lake) is a small storage unable to accumulate many days of high flow, there may be many more periods where total inflows would be problematic for potential users than would be suggested by the monthly average figures alone.

And now we have ever-increasing areas of thirsty tree plantations being established in the upper catchments (of the Meander especially) – and to top it all off, statistics accumulated by Hydro Tas and others [3, p39] are now pointing to the reduction in rainfall being a feature of permanent climate change.

How much comes out of Poatina?

A 1999 Hydro Tasmania publication [5] provides a wealth of information relating to the history and operations of the Great Lake, Poatina and Trevallyn schemes, including water management and environmental issues. It states that between 620 and 730 million cubic metres of Great Lake water are diverted annually through the Poatina power station. If this were concentrated in, let’s say, the 8 dryest months it would equate to a continuous flow of between 29 and 34 cumecs – more than enough to satisfy all users (indeed most of it would flow through Trevallyn power station into the Tamar, utilising the final head of 112 metres for further electric generation). But in the present drought conditions – and probably for most years in the future – this sort of discharge (amounting to 20% to 24% of Great Lake’s capacity at its mythical full supply level annually) is vastly unrealistic.

So how much comes out when they do run it? With its huge 830 metre head, Poatina requires far fewer cumecs per megawatt than any other station in the State. Hydro [5] quote its “full gate” (i.e. flat-out) discharge as 52 cumecs. Comparing the rated output power (300 MW) with the potential energy loss of this water, the station would then be running at around 71% efficiency. But in common with all hydro generators it will have much improved efficiency at a considerably lower discharge (and a much lower efficiency if run at just a small fraction of that). Unless the load could not be serviced from elsewhere it would make no sense to operate this station at full gate. Running at the best “efficient load” generates less power but recovers more total energy in the long term from a given depletion of the valuable water store.

The discharge from the full power station at efficient load [5] is 32.3 cumecs. Since there are actually six separate 50MW machine sets, we might expect the actual discharge at a given time to be around any of 5, 11, 16, 22, 27 or 32 cumecs, depending on how many machines were running (and that depends on the total Tasmanian load and any trading via Basslink). Or it might be none at all in the case that “run-of-river” stations had sufficient flow (and/or Basslink and Bell Bay were employed) to satisfy the load without having to deplete long-term storages.

Of the water which emerges from the tailrace, Hydro is committed to providing 12 gigalitres per annum to the Cressy-Longford irrigation scheme. Historically most of this has been concentrated in December/January with a flow rate of around 1.2 cumecs (implying 9 of the 12 gigalitres in these two months alone) [5]. But with increasingly dry seasons the need may have spread over a longer period. Scheme irrigators may have a good case for negotiating to purchase even more water from Hydro if the viability of the scheme appears threatened.

The recently-announced “drought-proofing” scheme for the Midlands will also require water from the tailrace – and presumably will want it during the dry part of the year, which is when the power station would be run the most and when the water is unavailable from other sources such as the South Esk.

So it may be that by no means all of the Great Lake water will actually end up in Lake Trevallyn for Gunns or any other user – and as mentioned above, if low rainfall conditions continue, the Hydro will be increasingly reluctant to operate Poatina extensively, especially as Basslink is now available to help meet the load.

For Gunns’ “4% of the flows into the Tamar Estuary” to be true in any of the ten dryest month in this study, Poatina would have to contribute around 18 cumecs, meaning that four of its six machines would have to run at efficient load almost continuously.

What about storage in Lake Trevallyn?

Lake Trevallyn is a steep-sided storage of area only 1.48 sq km [5]. Its full supply level is 126.49m. By agreement with Launceston City Council it is normally kept above 124.97m (although it can be drawn down lower by power station operation in anticipation of a flood, thus maximising generation potential). With this normal level range of only 1.52m, its working volume is 2.2 gigalitres (assuming vertical sides). Operation of all four 20MW machines in Trevallyn power station at efficient load (discharge of 61.4 cumecs) would deplete this in only ten hours.

For this reason Lake Trevallyn is not considered a substantial storage for the Hydro’s purposes. Trevallyn is generally run at a discharge rate equal to the inflow level (including any Poatina water)[5]. The availability of four separate machines means that this can usually be done with good efficiency. With sufficient flow to ensure a spill at Trevallyn Dam, the station would be run at full gate (after all, why not?).

However, with only the 2.5 cumecs drain required to satisfy environmental flows, Esk Water and Gunns, the 2.2 gigalitres would last ten days. In theory, Poatina could be run just as often as necessary to keep the lake level up for these users, and the Trevallyn power station run only when natural inflows made it worthwhile. So it would be possible for the Government to guarantee water for Gunns (if they wanted to) without really running Great Lake down significantly. There would be not much generation from Poatina (although its timing could be optimised) and even less from Trevallyn in dry periods, but Basslink might come to the rescue. There might be an element of “special deals for special mates” in a plan like this though, and that sort of thing doesn’t happen in Tasmania, does it?

Conclusions?

Nothing concrete.

The amount of water Gunns need can probably be found one way or another.

As reported in The Mercury, Hydro Tas appears to vacillate. On March 14 this year [6] it was “we don’t guarantee a water supply to anyone due to our lack of 100 per cent control of inflows and operational requirements”. On July 25th, after Great Lake levels had recovered somewhat, they said [7] “the level of surety of water delivery to Gunns is very high, much higher than to most irrigators”.

But if climate change is real and drought conditions like the present continue to be commonplace, any water use policies in central Tasmania need to be very carefully considered. A guaranteed supply for 30-50 years to an industry offering few benefits and a lot of detriments to society and the environment is, well, reckless.

References:
[1]. Mailed out to submitters but appears not to be available online.
[2]. http://www.hydro.com.au/home/Corporate/Generating_Power/Hydro_power/Catchment+Areas/Great+Lake+and+South+Esk+Catchment/Poatina.htm
[3]. http://www.hydro.com.au/documents/Corporate/Annual%20Reports/2007/hydro_ar.html
[4]. Stream flow data: http://water.dpiw.tas.gov.au/wist/
[5]. www.hydro.com.au/documents/Our%20Environment/SEGL%20Environmental%20Review.pdf
[6]. www.news.com.au/mercury/story/0,22884,23373359-3462,00.html
[7]. www.news.com.au/mercury/story/0,22884,24075270-921,00.html

Neil Smith is an electrical engineering consultant and renewable energy enthusiast, and has in his time sipped latte and hugged trees.