Onslow pumped storage

Onslow ongoing

* Environment Minister David Parker has stated that Onslow pumped storage is the only way to achieve net zero emissions by 2050.

* April 1 2021. No – it’s not an April Fool. This is from an item in April 1 Energy News.
Big electricity users seeking new supply contracts are being confronted with major price hikes as lake levels remain low, gas production continues to slide and the country heads into winter. Energy News understands that large commercial and industrial customers have been told their power bills will increase by 70 per cent or more. One major North Island customer is facing a 100 per cent increase, according to information provided to Energy News.
The availability of pumped storage at Onslow coupled with 100% renewable electricity would be a definite advantage to major electricity-using industries in New Zealand. The present situation where dry-year electricity prices are set by expensive fossil fuel power generation, is doing them no good at all. For example, in mid-April, steel prices were high in New Zealand due to demand from infrastructure builds. However, NZ Steel had to cut back on steel production because of high electricity prices impacting their production costs. With a view to the long term, the Major Electricity User’s Group might consider contributing to the cost of Onslow construction.

* The MBIE NZ Battery group now has a link to sign up to for news on the NZ Battery project, which will include any updates on Onslow pumped storage investigations.
* Transmission lines upgrade from Roxburgh, motivated in part by the possibility of Onslow, is already bringing some economic benefit to the Central Otago region.


1. Background
Government policy is for New Zealand to transition toward a low-emissions economy, making use of renewable electricity to replace fossil fuel use in transport and industrial heating. Hydro will remain a significant component of the required increased power generation. There is thus a need to buffer any transformed economy against future dry years, given limited national hydro storage capacity of less than 4.5 TWh. Dry years happen from time to time and public calls for reduced electricity consumption were made in 1992, 2001, 2003 and 2008. Pumped storage at Lake Onslow is a favoured option for providing backup power in dry years. Funding has been set aside for Onslow feasibility investigations, along with evaluating other energy storage options. Further information is in the NZ Battery web page. Indirect mention of Onslow is also seen in the supporting statements by Jacinda Ardern (9th paragraph) below the declaration of a national climate emergency.

2. Onslow energy storage potential
I was the first to note (in 2005) the possibility of pumped hydro storage in the Lake Onslow basin, published in a brief paper in the Journal of Hydrology (NZ). Energy storage would be achieved by raising the present 8.3 km2 Onslow reservoir (Fig.1), with water obtained by pumping up from either the Clutha River or from Lake Roxburgh. The main energy storage potential is in the Onslow Basin, requiring an earth dam at the Teviot River exit from the Onslow reservoir. In addition, there is possibility for a further storage increment to be added from the adjacent Manorburn basin to the north. The latter would require constructing a short tunnel through the low saddle separating the two basins, with a new concrete dam constructed at the present Upper Manorburn Dam.

Fig. 1. Energy storage potential of the Onslow Basin as a function of water level above the existing Onslow reservoir. From Majeed (2019). Note: 1,000 GWh = 1 TWh.

Various amounts of energy storage capacity can be achieved, depending on the specified upper and lower Onslow levels, and whether Manorburn is included or not. It would seem desirable to aim for a large storage capacity, given that the intended use of the scheme is as dry year buffer in the absence of fossil fuel backup. One configuration would be to construct a new 50-metre concrete dam at the present upper Manorburn Dam, build a 100-metre earth dam at the Teviot River exit of Onslow, and have a 4 km narrow-diameter tunnel linking the Onslow and Manorburn reservoirs at 740 metres elevation. Given an upper water level for both reservoirs at 780 metres, this would give about 8 TWh total energy storage capacity, given Onslow dry year drawdown permitted to 720 metres.

It would seem logical to make the extension through to the Manorburn basin, given the relatively small incremental cost of constructing a 50 metre concrete dam and a short tunnel. The above configuration would gain 1.0 TWh of energy storage capacity from Manorburn, which is more than the energy storage capacity of Lake Taupo. The Manorburn increment would be less if the upper operating limit of Onslow was lower. However, an upper limit for Onslow at 760 metres would still enable about 0.3 TWh as an increment from Manorburn, which is similar to the energy storage capacity of Lake Hawea.
The extent of an expanded Lake Onslow raised to 780 metres is shown in Fig. 2.

Fig. 2. Lake Onslow extent if raised to 780 metres elevation. From Majeed (2019). A connecting tunnel is shown here linking to Lake Roxburgh. An alternative tunnel would link Onslow with the Clutha River below Roxburgh. Map source: LINZ

The significant energy storage potential at Onslow derives from permitting a large water level drawdown in dry years. This could never be contemplated with existing scenic hydro lakes like Tekapo or Hawea because of the unacceptable visual impact of large areas of exposed silt and gravel around the lake margins. In contrast, the Onslow basin is bounded by schist rock hill country with thin soils that would be removed before lake filling to avoid dust issues during the rare drawdowns. A major dry year drawdown would see the lowered lake surrounded by an extensive zone of bare schist until the levels rose again.

3. Onslow pumped storage operation
The 2005 paper visualised Onslow pumped storage operating with some combination of seasonal water storage and holding water back against dry years. However, a recent PhD thesis (Majeed, 2019) recognised that the scheme could simultaneously operate in an enhanced commercial way, with income from a variety of on-going operations including frequency keeping, spinning reserve, security against outages, and buffering short-term wind power fluctuations. That is, the pumped storage system would always be in some form of continuous operation within the New Zealand electricity market, not used just two or three times a year as has been suggested.

Onslow pumped storage would operate at perhaps 75% efficiency, taking system losses and lake evaporation into account. That is, the scheme must purchase more electricity from the grid than it will sell. However, this energy difference would be offset in financial terms by pumping at times of lower prices and generating at times of higher prices. This price-based pump-generating sequence would fit the local hydrological conditions because lower electricity prices generally coincide with higher Clutha River flows and vice versa. Downstream of Roxburgh, the effect of Onslow operation would be to make Clutha high flows lower and Clutha low flows higher, staying within the normal flow range of the river.

Market-based operation of Onslow pumped storage is likely to result in some overall net power gain to the nation. This is because Onslow would purchase power for pumping at low-price times when South Island hydro lake levels are high, resulting in spring/summer release of some of their water for power generation. This reduces the frequency of high hydro lake water levels, in turn reducing spill loss from any sudden floods into already-high lakes. Simulations in the Majeed thesis indicate that an additional 100 MW could be gained on average from the Waitaki power scheme if it was operated to minimise spill by keeping Tekapo and Pukaki lake levels near to their mid-range.

4. Onslow discussions
There was a government announcement in July 2020 to consider pumped storage buffering as a means of achieving 100% renewable electricity in New Zealand, with $30 million allocated for an initial evaluation over a range of sites. This was followed by an election announcement in September 2020 that pumped storage could be used to achieve 100% renewable electricity by 2030, with $70 million set aside for Onslow site studies if initial feasibility evaluations were favourable. There seems general local support for the project in the Teviot and general Otago region, particularly with respect to employment prospects. However, various topics of concern have also been raised. See the below listing (4.1- 4.8), together with my comments. It should also be mentioned in passing that some other issues have been raised more recently in the media, where the national party energy spokesperson Barbara Kuriger is reported as stating that battery options should be considered. This seems unrealistic because the cost of an Onslow-equivalent battery would be in excess of a trillion dollars. She also raised concerns about seismic risks due to proximity to the alpine fault. However, the Onslow dam would be an earth dam like Benmore – which can be readily designed against massive shakes. The newly-found tectonic concerns of the National party are admirable, but were remarkably absent when they authorised construction of the concrete Clyde Dam, nearer to the alpine fault. Her final concern was the flooding of two valuable wetlands. This is certainly unavoidable if Onslow proceeds, but a case can be made for a series of environmental mitigations, including restoration of an even greater area of wetland – discussed at the end of this post.

4.1 Focus should be on reducing emissions generally rather than seeking 100% renewable electricity from pumped storage
This argument notes that electricity generation accounts for only a small proportion of our non-agricultural greenhouse emissions. Therefore, billions spent constructing a large pumped storage scheme like Onslow to eliminate fossil fuel power generation does not seem an optimal way to reduce emissions. See, for example, the article by the BusinessNZ Energy Council, the comment by Genesis Energy, the Newsroom item by Marc Daalder, and the Climate Change Commission 2021 Draft Advice for Consultation.
This line of reasoning assumes that Onslow is motivated primarily toward an end-goal of achieving 100% renewable electricity generation after 2030 for its own sake. While “100% renewable power by 2030” is a compact election statement, it distracts from the main climate-related purpose of Onslow pumped storage. That is, to provide the dry year buffer for a future low-emissions economy based on increased renewable electricity generation that will replace fossil energy sources in transport and industrial heating. If Onslow is constructed then the coincidental elimination of relatively small amounts of emissions from power generation does give us a good branding opportunity as a nation with 100% renewable power. However, it is not a single goal to be sought in its own right at great expense. In other words, it is not an either/or situation. The 100% renewable power goal and significant emission reduction can be achieved at the same time. It should be noted though that 100% renewable power is more than just a branding opportunity. There is also the unquantifiable aspect of national pride. This is because there will many of us who will take some pleasure in the knowledge that when we turn on a computer or recharge an EV, not a single electron will have been derived from fossil fuels.

4.2 There are better dry year solutions than Onslow pumped storage
Given that we wish to transition to a low emissions economy based on renewable power, the issue is then is whether it is Onslow or something else that is the best way to achieve the required dry-year buffering that will enable the transformation to proceed without climatic disruption. The selection decision is currently the task of the NZ Battery project. It has been suggested that there are lots of low-hanging fruit when it comes to dry year solutions. In fact, given the magnitude of the problem, it is difficult to find the required fruit at any height on the tree. Some of the suggested alternatives are listed below with comments. The selected option is obviously important because it will determine the form of the New Zealand energy scene for decades to come.

(i) Buffering using fossil fuels
This is the position taken by the Climate Change Commission’s 2021 Draft Advice Report. It does seem strange that the CCC should take this view, which is in direct conflict with stated government policy to shift toward 100% renewable power (confirmed in February 2021). It will be interesting to see whether the final version of the report maintains the conflict, or (more sensibly) suggests instead awaiting the outcome of the NZ Battery feasibility review of non-fossil dry year backup. The negative aspects of fossil fuel buffering are: (i) fossil fuels produce emissions and are ultimately unsustainable into the future, particularly noting the present restrictions on gas exploration, (ii) infrastructure issues can create uncertainty with gas supply (the Pohokura marine pipeline is presently experiencing issues relating to scale accumulation), (iii) reduced post-Tiwai wholesale electricity prices may make fossil fuel power generation uneconomic, so there is no certainty of always having this backup against future dry years (iv) fossil fuels as generators of last resort ensures high electricity prices at times of reduced power availability.
With respect to (iv), the effect is well illustrated for early April 2021. Low lake levels, low inflows, and reduced gas supply contributed to particularly high electricity prices (Fig. 3), enabling Genesis to generate significant profit by burning tons of cheap Indonesian coal at Huntly.

Fig. 3. Snapshot of wholesale electricity prices on April 8, 2021. Prices exceeding $120 /MWh are regarded as “high”. High wholesale prices eventually contribute to increases in home power bills. Source: WITS Free To Air

Use of fossil fuel buffering would be justifiable as a temporary measure if there was some cost-effective giant battery or similar on the technological horizon, capable of storing TWh energy magnitudes. Left-field breakthroughs are always possible but there appears no indication as yet of any new large-scale energy storage technology. That being the case, the finite nature of the fossil fuels means that adopting fossil fuel dry year buffering now means that later generations would have to face the dry year issue all over again.

(ii) Renewables overbuild
This concept involves having redundant renewable generation capacity (probably mostly wind power) in normal years so that there is still power available in dry years. Not surprisingly, an investigation in the ICCC Accelerated Electrification Report (p. 97) found this would lead to significant retail electricity price increases: 14% for residential consumers and 39% for industrial consumers. Expressed another way, the emissions reduction cost was estimated at $1,280 per ton (p.60). These figures are the basis of the statement that “getting the last few percent to 100% renewable power will be very expensive”. The overbuild price figures were incorrectly cited by Judith Collins as applying to Onslow pumped storage, in the TVNZ leader’s election debate on 22 September. The $1,280 figure was cited by the National Party in January 2021 as the cost of 100% renewables, still making the error of assuming that 100% renewable power can only be achieved by emissions overbuild, ignoring pumped storage. After all this time, it is disappointing to see the renewable overbuild figures of 14 and 39% still being referenced in the media (March 10, 2021) as the “cost of 100% renewable power”. Most recently (March 30) a Newsroom article references the Climate Change Commission still resurrecting the $1,200 value as “the abatement cost required to achieve 100 percent renewable generation by 2030”. All that is required to refute the $1,200 value once and for all is a quick visit to the NZ Battery web page where the various alternatives are listed in a figure, together with error bounds. The overbuild option is $1,200 with a large error range. The Onslow pumped storage option is $250 with a small error range. Sadly, the $1,200 value seems to have acquired a life of its own. I’ll report it here when it next appears in the media, as it surely will.

(iii) Green hydrogen production and storage
This approach would use renewable-derived power to produce hydrogen from water electrolysis. The hydrogen is stored in some form and then later used when needed to generate electricity in a dry year. However, the energy inefficiencies in the conversion and storage processes means that green hydrogen storage will never be a viable option for the amount of energy needed to buffer a dry year.

(iv) Demand management
Improved electricity efficiencies, power sell-back from home and car batteries and various other methods have potential to reduce winter peak demands. However, the magnitudes of demand interruption required to offset a dry year, with the associated economic impact, would seem to make this dry year option untenable. In mid-April 2021, high electricity prices (caused in part from low lake levels) forced NZ Steel to reduce output of much-needed steel for infrastructure builds. This “demand management” via the electricity market has had the effect of some reduction in electricity use, but at what price to the economy?
Contact Energy and Meridian Energy have suggested that setting up post-Tiwai green hydrogen production in Southland could also serve as dry year buffer. That is, the plant could close for some duration in dry years and the power diverted for use elsewhere in the country. A variation on the theme advocates a similar demand interruption process using dry year reduced operation of the Tiwai smelter. However, we could visualise a future dry period when Te Anau and Manapouri lake levels have reached their lower permitted range and river inflows are greatly reduced. Switching off hydrogen / aluminium production in this situation will be of little help because there would be minimal Manapouri power available to send north. It is also unlikely that the shortfall could be made up by other South Island stations. This is because when dry conditions occur they extend over wide areas in the same climatic zone. That is, low power output from Manapouri is likely to coincide with low power output from the Waitaki and Clutha schemes as well, which will also lead to higher electricity prices.
Fig. 4 illustrates the problem of demand reduction as a dry year mechanism. The Manapouri – Te Anau system is shown as at 27/02/21. However, suppose it is actually some time in the future and that Manapouri power had been supplying a large hydrogen production plant in Southland. Manapouri / Te Anau storage levels are falling, as are river inflows. If that trend were to continue, the lake levels would soon be at the bottom of their operating range and the Manapouri power output would be constrained to the water supplied by the limited lake inflows from dwindling rivers. That is, Manapouri power available to the nation from hydrogen shut-down would be minimal against a dry year in this scenario. In contrast, Onslow would have 1000 MW available for an extended duration and without the economic impact of industry closure.

Fig.4. Manapouri / Te Anau levels and inflows, February 27 2021. Source: Energy Link

(v) Alternative pumped storage options
The possibility of multiple smaller pumped storage schemes instead of Onslow has been advocated by Dougal McQueen, and also mentioned by the NZ Battery. However, the problem with a cluster of smaller schemes as dry year buffer is evident from the energy storage capacities listed in Fig. 5. The magnitude of a dry year event was estimated by the ICCC Accelerated Electrification Report (p.64) as a deficit of 3 TWh. More recently, the NZ Battery has noted that they require a minimum of 5 TWh of energy storage capacity. There is potential for a maximum of 2.7 TWh of energy storage capacity in the upper Ngaruroro River near Taupo, with Lake Taupo as the lower reservoir. However, Ngaruroro pumped storage is likely to be contentious for environmental reasons. It would require a long 30 km tunnel linking to Lake Taupo and may need to have a dual purpose involving water transfer to Hawke’s Bay for economic viability. The required 5 TWh of energy storage could only be achieved by Onslow + Ngaruroro, or by Onslow alone. Given construction expenses, it seems unlikely that two pumped storage will be built in New Zealand for dry year buffer.
Outside of Ngaruroro, it does not seem possible to specify any set of smaller pumped storage sites that have combined energy storage capacity anywhere near 5 TWh. No such grouping of sites has been publicly identified to date. There is also an economic aspect because each site would require specific engineering construction and consenting processes. It could happen that one or more smaller pumped storage schemes could play a useful North Island role as alternatives to gas peakers. However, this is a separate use from provision of dry year buffer.

Fig. 5. Comparison of energy storage capacities. Onslow and Ngaruroro are potential pumped storage schemes. The others are existing hydro storage lakes. The 8 TWh for Onslow is with respect to the combined Onslow-Manorburn system described in Section 2.

One alternative pumped storage system that has been advocated is for between Lakes Tekapo and Pukaki as upper and lower reservoirs, respectively. However, it is not clear how this scheme could provide a dry year buffer because Lake Tekapo does not have sufficient energy storage capacity. Also, keeping Tekapo levels high as dry year reserve takes it out of its ordinary seasonal operation, runs the risk of generating spill from high inflows, and would lead to increased erosion along the lake shoreline. Further, there would be times when the pumped storage system could not operate. When Lake Tekapo is at the top of its range, the low power prices at the time could not be utlised for pumping because there would be no available storage in the lake to accommodate the pumped water from Lake Pukaki. Onslow operation would have the opposite effect, with a reduced frequency of times of high water levels in both Tekapo and Pukaki.

(vi) Batteries for dry year storage.
Surprisingly, this rather silly notion still appears in the media from time to time. Large batteries could play a very useful role in New Zealand to maintain short-term stability. However, confusion can arise over the specifications, because batteries are often just reported just as their power output along the lines of “200 MW battery”. Unfortunately, such descriptions fail to mention how long the MW output could be maintained for. The big South Australia Tesla battery has energy storage capacity of 0.00013 TWh. The NZ Battery page lists 5 TWh as the minimum required storage for a dry year. To get that amount of energy storage we would need 38,000 of those Tesla batteries. Suppose we purchase them for NZ$100 million each – you can confirm the price from the web. That comes to 3.8 trillion NZ dollars all up. And they would all need to be replaced after some period, for another 3.8 trillion dollars. Put another way, the Interim Climate Change Committee puts the per unit marginal emissions abatement cost from battery storage at $89,000 compared to $250 for Onslow pumped storage. The NZ Battery web page is perhaps unfortunate in its choice of name.

(vii) Enabling massive dry year drawdown of existing hydro lakes.
This would most likely be the cheapest of all the options in pure dollar terms. For example, the Waikato River outlet at Lake Taupo might be deepened to allow dry year drawdown down to about 12 metres below the present minimum level. That would give 5 TWh of new energy storage in the North Island but with a very bad visual impact of exposed lake bed and seems unlikely to be ever agreed to.

4.3 Pumped storage at Onslow is uneconomic
The construction cost of Onslow is likely to be in excess of 4 billion dollars. In terms of cost per unit of stored energy, Onslow (in whatever final configuration) is likely to be much more cost effective than the Snowy 2.0 scheme in Australia. This scheme is located in a national park, will probably cost in excess of $10b, and has only 0.35 TWh of energy storage capacity.
Quantifying the various economic benefits of Onslow pumped storage is multi-faceted and requires careful study. The most detailed economic evaluation to date is in the thesis by Majeed (2019). Taking into account various aspects including aiding new wind power generation, increased irrigation water, flood peak reduction, and reduced power prices, the thesis abstract concludes that Onslow pumped storage is economic. Final economic evaluation will need to await analysis from the NZ Battery project.

Two new economic factors have recently come into play: a national downturn associated with Covid-19 impact, and the announced future closure of the Tiwai Point aluminium smelter at the end of 2024. A large engineering project like Onslow construction will provide economic stimulus to Central Otago for some years. Also, there is the additional factor of long-term stimulus for the Southland economy. Tiwai closure will provide new power for Southland industry but power reliability is a factor also. Fiordland is wet on average but there can be dry periods as well. Low lake levels in March 2006 resulted in a partial shutdown of the Tiwai smelter and record low lake and river levels were recorded in April 2017. If a decision is announced that Onslow will be built, marketing could begin immediately for Southland as an optimal cool-temperature location for electricity-consumptive industry, with power reliability guaranteed by a mega- pumped storage scheme just to the north.

Negative comments concerning Onslow pumped storage economics are sometimes superficial. The former National Party energy and resources spokesman Jonathan Young was reported as stating that the Productivity Commission had looked into the Lake Onslow scheme in 2018 but found it “didn’t make sense economically”. In fact the “looked into” by the Productivity Commission Report was only a short footnote on the bottom of p.395 as a personal communication by Sapere consultants who suggested that Onslow pumped storage would impact existing hydro-generators. No supporting argument was given in the footnote and the nature of the impact was not specified, although it was presumably concern over reduced wholesale electricity prices.

The possibility of reduced wholesale electricity prices may have prompted Contact Energy to state that Onslow pumped hydro could temporarily paralyse investment in renewables, a theme also taken up by Mike Hosking on Newstalk ZB. However, it is interesting to note the warning by Meridian Energy that public investment in pumped storage might instead push up electricity prices, thus slowing down the electrification of the economy. ACT party leader David Seymour expressed similar sentiments with respect to electricity price increase. The Climate Change Commission in its 2021 Draft Advice for Consultation also raises the specter of dry year buffer costing taxpayers billions of dollars (p.112). It seems strange however, that such statements could be made with apparent certainty in the absence of knowledge of how Onslow would be owned, operated, and the time period over which it would be paid for. Contact Energy at least have now backed away from their original assertion of Onslow paralysing renewable investment and propose instead that all dry year and firming options be first considered “dispassionately and scientifically” – see Contact canvasses firming, dry-year options (Energy News, 18 February, 2021).
The Ministry of Business, Innovation and Employment has commissioned a review on the likely effect of Onslow in the electricity market. It seems probable that some downward influence on wholesale prices will be confirmed because (i) electricity prices tend to be lower when the hydro lakes are full and Onslow will be near maximum capacity most of the time, more than all the existing hydro lakes put together, (ii) Onslow will presumably generate when the prices are high, effectively giving a hard upper price bound which will be below that required to initiate thermal generation from coal and gas, which will be removed from power generation.

In terms of paying for construction costs, Keith Turner has argued that a long payback time would make Onslow invisible to retail electricity prices. A similar conclusion concerning the merit of long payback times on electricity prices was reached in a preliminary study by the University of Auckland.
However, any downward effect by Onslow on electricity prices is likely to be relatively minor in the long term, compared the effect of northward transfer of Manapouri power. This is because the Onslow scheme would be a net consumer of power, not a net generator. In particular, it would be a purchaser when electricity prices trend down, thus creating a price floor. Contrary to the suggestion of paralysing investment in renewables, this pricing situation is likely to be strongly encouraging of renewable investments and not in any way crowding it out. For example, wind power must be sold when the wind blows, regardless of how low the electricity price may be at the time. Having both an upper and lower ceiling to power prices will enable a degree of planning of wind power developments. For example, Onslow is likely to be beneficial to the economics of proposed wind farms at Mahinerangi (Otago) and Kaiwera Downs (Southland).

In addition, having 1000 MW of power (available for generating or taken up by pumping) will buffer the destabilising effect of having an increased amount of intermittent renewables in future. This would avoid instability situations like that presently developing in Australia. It is easy to say that New Zealand needs thirteen large wind farms as part of the move toward a low emissions economy, but the grid instability potential has to be considered as well. For example, 500 new MW of wind generation capacity needs to be supported by 250 MW of hydro power generation capacity to offset wind speed fluctuations – which could conveniently come from pumped storage.

Reduced variability of wholesale electricity prices is likely to have various positive effects on the national economy. For example, major electricity users like the pulp and paper industry will appreciate not having to plan against the possibility of dry year price spikes, as happened in 2008. Similarly, knowing there will be a hard price ceiling on wholesale electricity prices will encourage transition away from existing fossil fuel use for industrial heating requirements.

As a generator, Contact Energy in particular stands to gain from Onslow construction. Apart from a small amount of storage in Lake Hawea, Contact’s Clutha hydro power scheme (Clyde and Roxburgh stations) is run of the river and there is spill loss from time to time. Pumping from an Onslow intake at Lake Roxburgh will reduce flow at the Roxburgh dam (including possible spill) at times of low electricity prices. That same water will generate more income from Roxburgh power generation when later returned to Lake Roxburgh from Lake Onslow when prices are higher.

Finally, part of the economics of setting up Onslow will involve purchasing a significant one-off energy increment over the initial duration of time when the lake height is increased by increments to its final mean operating level. Taking into account the dead storage volume needed to reach minimum operating level, this could involve the purchase of more than 8 TWh from the grid. This would seem a good initial use of some of the post-Tiwai Manapouri power. If a decision was made for pumped storage at Onslow (construction starting in 2023), there may be just a few years between first pumping of Clutha water and Tiwai closure at the end of 2024.

An extended period of payback of Onslow construction is something like paying an insurance policy (against future dry years). There is an added benefit, though, in that the insurance still remains after the last payment.

4.4 Onslow would “stand the electricity market on its head
This provocative headline is a reflection of genuine concern over the magnitude of Onslow pumped hydro and how it might impact the electricity market. However, regardless of storage capacity, Onslow can only interact with the market through power purchase or power selling to a limit of around 1000 MW – not much more than the generating capacity of the Huntly Power Station. But to help offset concerns, it is desirable that Onslow should be notified as operating in a commercial way in the electricity market. That is, Onslow would purchase power for pumping when prices are low and generate for selling when prices are high. This will reduce the variability of wholesale electricity prices but Onslow otherwise would be just another player in the market and not a major disruption to it. Put another way, if Onslow pumped storage had been constructed before the electricity market was established, the presence of Onslow would not have prevented the electricity market from being set up.

4.5 Availability of Manapouri power post-Tiwai means Onslow pumped storage may be unnecessary
This aspect was mentioned by Janet Stephenson (Otago University) as part of an article advocating that demand-side management should also be considered. However, to compare Onslow and Manapouri is to confuse power and energy. Manapouri is a power scheme – it’s purpose is to generate electricity. If constructed, Onslow would be a hydro lake for storing energy, no different from any other hydro lake except that water is pumped up to it. A power station cannot substitute for a hydro lake or vice versa – they are different entities.
In fact, it happens that Manapouri power going north (after transmission upgrades) will make Onlow more urgent. This is because Onslow pumped storage would primarily buffer South Island hydro generation against dry years. The significant Manapouri increment of hydro power output from the same climatic region would make a South Island dry year impact greater because there would then be more hydro generating capacity that is not generating power.

4.6 Onslow is in the wrong island
This issue has been raised as part of a general comment on Onslow. The confusion here is that Onslow is often mistaken to be a hydro power station. After all, it will have tunnels and turbines and water flows, just like the Manapouri power station. You would have to be pretty silly though to locate a big new power station in the southern South Island when the main power demand is in the northern North Island. However, Onslow would not be a power station, it would be an energy storage station. In fact, the simple physics of pushing water up a hill and running it back down again later means that Onslow would be a net consumer of power. Onslow as proposed is simply an insurance against South Island dry years. That is, another hydro lake is created in the South Island to go along with Lake Pukaki, Lake Tekapo, and the others. The only difference is that the existing hydro lakes get their water from inflowing rivers while Onslow would get its water from being pumped up to it from the Clutha.
That power from pumping will almost always come from the nearby Waitaki, Clutha, and Manapouri stations, so there is no issue of transmission losses to the North Island that was no already there before. There is also no net energy issue as long as there is not too much energy lost in the to-and-fro process. The cost aspect of having to buy more power than is generated is offset by buying power when prices are low and selling when prices are high. And it turns out that Onslow is likely to give a net energy gain to the nation because its operation in the electricity market will reduce spill in wet years from existing hydro lakes – as noted at the end of Section 3.

From the South Island environmental aspect, it could be argued that the last thing the South Island needs is yet another hydro lake. It has to be keep in mind, however, that the South Island hydro lakes were once natural lakes and they have been flooded by raising their levels to store water for hydro power. In contrast, the existing Lake Onslow is an artificial reservoir. Also, the large energy storage capacity of Onslow means that an argument can be made that Lake Hawea is no longer needed for hydro storage and could be restored back to its original shoreline, presently submerged under 18 metres of water [see 4.8 (3)]. That is, there is the possibility of regaining a natural South Island lake for the cost of enlarging an existing man-made reservoir.
Still on the South Island environmental theme, it is desirable for a scheme like Onslow to be located in the same climatic zone as the hydro power stations that are being buffered against low river flows. When the power is needed there will be low Clutha River flows, so the discharge released from Onslow generation will not cause the river to go above its normal discharge range. Likewise, pumping will correspond to times of high Clutha flow because electricity prices will be low at such times. The water volume impact on the Clutha River will be minor in fact because Onslow would be a small power in terms of water throughput – its power output capability comes from high turbine pressures rather than high water discharges. Whether generating or pumping, its impact on Clutha River flows will be difficult to detect.

There is one issue that will arise from a dominance of South Island power supply in the absence in future of fossil fuel generation in the North Island. That is, there will be concerns over the risk of a power outage at a time when Auckland city is mainly reliant on South Island hydro power. One possibility here is to contract Mercury to maintain Lake Taupo with sufficient water as security. That is, always having potential to ramp up output from the Waikato stations. Another possibility is to have some small North Island pumped storage schemes capable of significant power output over the short time period of the outage.
There will never be a situation, however, when Onslow in an extreme dry year will “keep the country going”, with an image of all the hydro lakes at zero storage and a full Onslow generating at the maximum rate. As a dry year builds develops, the market will ensure Onslow generates preferentially and more water will be held back in Taupo and the other hydro lakes. That is, the hydro lakes will all tend to decline proportionately. So an outage at Onslow does not mean the lights would go out in Auckland.

4.7 Onslow is a return to “Think Big
The Onslow pumped storage scheme would be significant, and any large project has potential to be seen in hindsight as a large mistake. However, it works both ways. For example, the Benmore Dam is a large government-constructed edifice that has served the country well in supplying renewable electricity. On the other hand, the Clyde Dam should never have been built and only passed through parliament with the aid of two votes from the Social Credit party. In contrast, Onslow will be subject to thorough review. Setting up a $30 million preliminary study of Onslow and other dry year alternatives gives good opportunity to overview all aspects before billions of construction dollars are committed.

4.8 The environmental impact at Onslow is large and would be controversial
It can be seen from Fig.1 that Lake Onslow at 780 metres elevation would represent 75 km2 of flooded land, including the present reservoir. That is, it would be an artificial reservoir the same size as Lake Benmore. In addition there would be a further area of flooded land if the Manorburn-Greenland reservoir is raised to 780 metres.

In terms of environmental negatives, the area of inundated land is significant. This would include in particular the loss of existing wetlands at the southern end of the present Onslow reservoir. Also, during rare extreme dry years there may be up to 60 metres of visible water level drawdown and rock exposure, depending on the final permitted water level range. There is also concern over the viability of the endangered Teviot flathead fish population, which is restricted to the local area.

On the other hand, there are potential for environmental gains to be made as well and it becomes a decision as to whether the positives outweigh the negatives. At the larger scale, Onslow would play a significant role as an enabler of the transition to a low emissions economy through providing dry year security. On the other hand, there are undoubted environmental impacts with respect to the Onslow wetlands loss. David Parker has noted the environmental gains of net zero emissions. However, there is also the question of whether local Otago environmental (and economic) gains could offset the impacts. That is, achieving an overall consensus that a net positive environmental outcome will be arrived at. Such environmental offsets have the potential to be as impressive in an environmental sense as Onslow pumped storage is from an engineering viewpoint.

Of course, the offsets cannot be imposed from above or simply defined to be environmental offsets unless there is strong local support. This is well summarised by Jacinda Ardern (September 10, 2020).

Here you will have the potential with one project to make sure that we are overcoming a significant issue that has an impact on our climate and our emissions profile. But how do you mitigate the environmental impacts? Those are the debates and conversations we need to have.

There has been much published in the media about Onslow and its possible raising for pumped storage. However, there has been minimal discussion about possible environmental offset options, if the scheme were to go ahead. Ideally, the cost of environmental offsets should be considered as part of the overall Onslow construction budget. The list below is only indicative and the various possibilities may or may not be considered for further investigation, depending on feasibility and degree of local support.

(1) An expanded Lake Onslow for recreation and ecological reserve
A new lake may be regarded as an asset, depending on nature of the land displaced. The present Onslow Reservoir flooded part of Dismal Swamp, but there would probably be opposition to any suggestion of draining the lake to re-establish the wetland. For the much larger new Onslow reservoir, it may be possible to establish extensive areas of artificial floating wetlands as part of an ecological reserve. Some reserve land could be set aside for establishment of one or more Teviot flathead sanctuaries above 780 metres, protected from predatory trout by constructed artificial waterfalls. Such sanctuaries could be established as part of the $30 million initial feasibility investigation. That would be a useful ecological legacy independent of whether pumped storage went ahead at Onslow.

(2) Shifting Waitaki River flows back to the original seasonal discharge regime
Like all the major rivers with headwaters in the Southern Alps, the natural flow regime of the Waitaki River is for high flows in spring and summer and low flows in winter when much of the mountain precipitation accumulates as snow. Lakes Tekapo and Pukaki are presently managed for hydro power by holding back high spring and summer inflows in anticipation of the lower winter inflows and increased winter power demand. That is, Waitaki River flows have been increased in winter and decreased in summer. As noted in Section 3, commercial Onslow operation would purchase power for pumping when wholesale electricity prices are low, leading to summer water releases from Lakes Tekapo and Pukaki. This shift back to higher summer flows to some degree re-creates the former Waitaki River natural seasonal flow environment.

(3) Potential for restoration of Lake Hawea
Using Lake Hawea and other South Island scenic lakes for seasonal hydro storage has had significant environmental impacts. The original lake shoreline environments had evolved and stabilised over thousands of years following the retreat of the glaciers. Raising the lakes for hydro storage creates enhanced local erosion from wave action on the soft glacial tills along the line of the new water levels that had never been part of a lake shoreline system. In addition, the imposed seasonal cycles on the new shorelines are greatly in excess of the small natural variations in the original lakes.

Lake Hawea is the only lake in the Clutha catchment that is used for hydro power storage operations, enabling some degree of control of water flows to Contact Energy’s Clyde and Roxburgh power stations. The other natural lakes in the catchment, Wanaka and Wakatipu are uncontrolled. Lake Hawea was raised in 1959 from its original 328 metres above sea level, with the highest water levels now extending to 346 metres. Seasonal water level fluctuations of 8 metres are presently permitted, compared to the original natural variations that were about 1 metre and of short duration. The present seasonal water level fluctuations have reduced the lake’s aquatic plant diversity compared to nearby Lake Wanaka. Also, periods of high lake levels concurrent with wind waves have created eroding banks near parts of the Hawea township and elsewhere. Lake Hawea remains beautiful but there are some who can still recall how it used to be.

Hydro storage in Hawea and the other South Island lakes is an essential part of maintaining national power supply. However, significant new hydro storage capacity at Lake Onslow offers an alternative as far as Lake Hawea is concerned. If there was, say, 8 TWh of storage capacity established at Onslow then the total Clutha catchment energy storage capacity would be 8.35 TWh. The decimal places represent Hawea storage. With 8 TWh of Onslow storage and at least 1000 MW of Onslow installed capacity, we can get by without the 0.35 TWh storage component at Lake Hawea. That is, Lake Hawea could lowered and restored back to its original natural shoreline at 328 metres. Outlet water releases would be maintained at about equal to lake inflows, so the Hawea River would largely revert to its pre-hydro state. In this way, the present Lake Hawea storage and seasonal variation would be transferred to Onslow. With Onslow operational, there would be no obvious national good in maintaining Lake Hawea in a flooded state.

Any Lake Hawea restoration process would probably need to be carried out by incremental lowering, concurrent with planting grass and other vegetation to avoid blowing dust. There would also be a need for replanting native trees at some locations around the restored lake margins so that the drowned forests of Lake Hawea can become real forests again. A map including the 328 metre contour will show the extent of new land area that would be gained around Hawea township.

A restored Lake Hawea could still be used for hydro power operation, but in a more benign way. The outlet dam could enable occasional major floods to be held back for short periods, giving brief water level rises. This water management regime will reduce spill at the Clyde and Roxburgh power stations when there are large outflows from Lakes Wanaka and Wakatipu. Our simulations indicate that a restored Lake Hawea managed in this way gives the possibility of a small increase in mean power output at Clyde and Roxburgh through spill reductions.

The role of Onslow pumped storage in Lake Hawea restoration would be with respect to specific arrangements with Contact Energy so that income from commercial operation of Roxburgh and Clyde is not compromised. The new water management would mean that power generated at the Roxburgh and Clyde stations from released Lake Hawea water will seldom coincide with times that are best for power sales income. This could necessitate some form of compensatory arrangement by which Contact Energy receives above-market rates from Onslow power purchases, for the relatively limited amount of power involved. The other commercial issue is that Contact Energy would not have a future option of setting up a small power station at the Hawea Dam. In the end, however, commercial considerations are not everything. If Lake Hawea is restored, shareholders may take some pride in the role played by Contact Energy in writing the final history of the rise and fall of Lake Hawea.

(4) Potential for restoration of Great Moss Swamp and buffering Taieri River drought flows for irrigation

The Great Moss Swamp was an iconic Otago upland wetland on the southern portion of the Rock and Pillar Range. However, only a remnant remains because because almost all of it was flooded by the Loganburn Reservoir (Fig. 6).

Fig. 6. The Great Moss Swamp is not so great anymore. Map source: LINZ.

The Loganburn Reservoir was the last government-constructed irrigation reservoir and presently maintains summer water supply to an irrigation scheme in the Upper Taieri Plain. At the time, the Great Moss Swamp was the only option for water storage. However, the loss of an extensive wetland seems a high environmental price to pay for a local irrigation scheme. Also, the reservoir is not an ideal water store because in major droughts there is not enough water able to be released for the irrigation need.

Constructing Onslow pumped storage is yet another major hydro system to be hosted by the Otago province for the national good. It seems reasonable therefore that a claim could be made for some degree of regional compensation. A possibility in this regard is shown in Fig. 7. It happens that the local topography would enable construction of a short 3 km tunnel that could discharge some Onslow water into the Taieri River via Riddles Creek. The tunnel would have to be constructed prior to raising the lake because the tunnel entrance will be well below the surface of the expanded lake.

Fig. 7. Onslow water could be supplied to the Taieri River through a short tunnel (arrowed). Map source: LINZ

Onslow water could be released free of charge into the Taieri to play the same water supply role as the Loganburn Reservoir, with the important improvement that water supply could be maintained through droughts. Additional Onslow water could be supplied from time to time to the Taieri River to maintain drought irrigation and environmental flows during significant regional droughts, which may become more frequent in Otago with climate change. An environmental advantage of this water augmentation is that the Great Moss Swamp could start to be restored, with the Loganburn Reservoir now greatly reduced in size to provide only a flood protection role. The amount of wetland restored in this way would be greater than the wetland area lost by raising Lake Onslow. There is also the possibility that the restored Great Moss Swamp, as a high fenland, could provide enhanced carbon accumulation.

Augmenting Taieri River drought flows would also help maintain the wetlands of the upper Taieri scroll plains during low flows, while also improving river water quality. Onslow water used in this way, free of charge, could only be sufficient to maintain existing irrigation schemes through rare drought events, as opposed to ongoing supply to new irrigation schemes.

The frequency and volumes of Onslow-augmented Taieri flows would have to be defined clearly from the outset, because there would be a small decrease in the pumped storage efficiency. However, the water volumes involved would be minor compared to pumped storage throughputs. A further factor would be the need for iwi considerations because in Taieri River low flows would be augmented with Clutha River water.

Fig. 8 gives an indication of of how water release from Lake Onslow could avoid irrigation restrictions during Taieri River low flows arising from droughts. There have been two significant Taieri low flow events in the last few decades, one at the start of 1999 and the other at the start of 2015. The graphs of Fig 8 show the effect of adding Onslow water to the upper Taieri River – as recorded at Canadian Flat. If Onslow pumped storage had been in operation, Onslow water could have been added to the low flows to bring the discharge up to 8 cubic metres per second during the the low flow periods. This would be sufficient to maintain irrigation use over the length of the Taieri River for the dry periods. This would only lower Lake Onslow by 1 metre and by 35 cm, for the 1999 and 2015 events respectively.
The water releases would therefore have minimal influence on pumped storage operations. Water releases from Onslow would have been only needed at those two times in the last 20 years. This suggests only minor influence on the overall economics of the scheme if it is constructed, even allowing some possible increase in drought frequency and water releases with climate change.

Fig 8. Illustration of hypothetical Onslow water release on the upper Taieri hydrograph (Canadian Flat) during the droughts of 1999 (upper) and 2015 (lower). Onslow water (orange) is released to top-up the low flows to 8 cubic metres per second for the duration of the river low flows (Jan 1 to April 14 for 1999, and Jan 1 to Feb 2, for 2015). Canadian Flat discharge data from NIWA.

(5) Possibility of creating new wetlands near Onslow
The gentle topography of the upper portion of Bonds Creek (north east of Onslow) would enable a low dam to create of an extensive shallow lake (Fig. 9).

Fig. 9. Maximum extent of a shallow lake (dark blue) that would be created from a low dam (red) in the headwaters of Bonds Creek, with lake level at 760 metres asl. Map source: LINZ

Any actual lake is likely to be smaller than the 760-metre lake shown in Fig. 9. This is because the purpose of the lake is to enable the establishment of new wetlands as offsets to those lost at Onslow, should Onslow be raised for pumped storage. That is, the object is not to achieve the largest possible lake surface area, but the largest possible area of new wetlands. The lake elevation should therefore be selected such that the lake maintains a high water table over extensive areas of near-flat relief suited to wetlands establishment, just as the present Lake Onslow does for the wetlands around its southern shoreline. The selected lake elevation is therefore of importance. There would be a need to take into account details of topography, soils, drainage and various other ecological factors. An ecological study in the Bonds Stream headwaters would establish the total area of new wetlands that could be created. The whole development would also require land purchase agreement from the landowners concerned. Subject to these provisos, development of Bonds Creek wetlands might provide a relatively low-cost ecological offset as compensation for wetlands lost around the present Lake Onslow shores.

5. Summary
The Onslow pumped storage scheme is a large and many-faceted project. It is perhaps natural that some should react with initial trepidation and caution. Solution of the New Zealand dry year problem is now recognised as essential, but it could happen that the NZ Battery review decides on some other course of action. However, if Onslow proceeds then there need not be economic losers in the process, with many different gains to be made in different aspects at both national and regional levels.
All this is not a personal advocacy of Onslow pumped storage as such. No individual can advocate large expenditure on a major project, as opposed to being spent in some other part of the economy. But within the stated goal of 100% renewable power, Onslow pumped storage is a prime contender to handle the associated issue of dry year buffering.