Onslow pumped storage

Personal note. The idea of pumped storage at Lake Onslow has an unusual origin. New Zealand’s major hydro schemes started from government engineers and planners working away in large teams, sometimes in secret. Recognition of the pumped storage potential at Onslow came from a single university academic with qualifications in Geography from the University of Otago – that’s me. I much prefer to write about Onslow and not about myself. However, I am often asked by journalists about how it began, so I’ve added something here. Skip to the next section to read about the Onslow scheme itself.

The starting point of a long journey was a University of Otago field camp back in the early 1970’s. I was standing at the shoreline of Lake Hawea with a group of students and Dave Murray, who was then lecturer in hydrology in the Department of Geography. Looking at the lake, he said “they run this like a bathtub”.

The “they” he was referring to back then was the former NZED.

I found something very disquieting about a beautiful Otago lake being equated to a bathtub for the purpose of hydro storage. However, it wasn’t obvious that anything could be done about it and so the years went past.

I went on to play a small organisational role in the protest movement that stopped the proposed aluminium smelter at Aramoana. Sadly, the power component of that – the Clyde dam – was not stopped and construction went ahead when the Social Credit party supported Muldoon to pass the legislation. The flooding of the Cromwell Gorge and the loss of the confluence of the two rivers at Cromwell is an environmental tragedy on a par with the level of destruction that would have happened from raising Lake Manapouri. Even now, many of a certain age will break a little inside when driving the new highway between Clyde and Cromwell.

In 1980 I took up a position as lecturer in hydrology at the University of Waikato. I departed from Otago without ever having visited Lake Onslow and with only a vague idea where it was.

The next time of note was the 1992 electricity crisis. That was associated with a period of low inflows into the main hydro lakes. By then I had developed an interest in optimal water resource use and 1992 showed that our spectacular hydro dams were far from optimal. In dry times like 1992 there was not enough water to be released from the storage lakes to meet the electricity demand without power saving campaigns. While in wet years there was significant energy loss from hydro dam spill.

At this point I should be writing  “… and then I noticed that the Onslow Basin had massive energy storage capacity…”. You really would think that somebody with a PhD in Geography would be able to glance at a topographic map of Central Otago and come to the right conclusion.

But the opportunity was lost. The 1992 electricity crisis had such an impact on the nation that I have no doubt that, if recognised, the Onslow pumped storage scheme would have been constructed and operational by about 2002. On the plus side, if Onslow had been built back then it would have been sold to Contact Energy as part of the sale of the Clutha Power Scheme, thus giving a private company a massive degree of control over the New Zealand electricity market.

Sometime around 2003 I finally noticed the energy storage potential of the Onslow schist rock basin and made my first visit there. Here at last was the possibility to avoid dry year crises and at the same time shift some storage operations away from the South Island hydro lakes, perhaps even restoring some.

However, by then it was too late. The 1992 crisis had faded from memory and the ECNZ had been split into the big energy players we know today. They operate independently and competitively in a market environment and have no more obligation toward developing security of national electricity supply than Fonterra is obligated to provide cheap milk for all New Zealanders. Nor do they need to take note of any form of national renewable energy strategy. In fact, a cynic might suggest that their on-going opposition to Onslow is to ensure that a component of fossil fuels remains in power generation to enable the highest possible electricity prices and maximum profits to them. See, for example, recent comments by Trustpower.

I wrote up a short note in 2005 pointing out the energy storage potential at Onslow, which was later expanded into a detailed PhD thesis in 2019 by my research student Mohammed Majeed.

I expected that would be the end of the Onslow story, but climate change and emissions concerns now introduced a new element. I presented a paper to the 2019 New Zealand Hydrological Society Conference in Rotorua under the title 100% carbon-free power in non-normal hydrological years: the role of water in New Zealand and the world. That was the first public presentation of the possibility of 100% renewable electricity in New Zealand from Onslow pumped storage. It was then realised in government circles that Onslow could take over from fossil fuels as dry year buffer, with Keith Turner (former Meridian CEO) making a strong advocacy of the dry year insurance aspect. The 100% renewable electricity concept was announced as Labour election policy in July 2020 and an MBIE entity “The NZ Battery” was subsequently funded to carry out an initial investigation, to be completed by May 2022.

An indication of whether Onslow will proceed or not will be if the 100% renewable electricity policy, announced in the 2020 election, is maintained. In an October 2021 document for public comment, 100% renewable electricity is now referenced as “aspirational”. However, the importance of dry year buffering is still recognised and a final decision on the technology / policy combination will now not be made until 2025, following the completion of the first two phases of the NZ Battery project. If it is decided at that time to keep an element of fossil fuel power generation for dry year buffer then it is difficult to see Onslow going ahead. Equivalently, if the 100% renewable electricity policy is maintained then it not obvious how security of power supply could be provided without Onslow in the mix.

If Onslow pumped storage becomes a reality then its influence will be multi-faceted and massive for decades to come. That may sound like an ego statement but Onslow has never been about any single individual – it has always been about Otago.

Otago is different. Those who grew up there will understand. However, its past hydro developments have always involved a flurry of activity for a few years, followed by the silence of another flooded valley and another set of pylons going north.

Onslow is the opportunity for Otago get something back. In effect, it could amount to losing one element in the landscape in order to restore an even more important one. So we return to Lake Hawea. Of course there would be many details and compensations to be worked through, but restoring the lake and replanting its flooded forests would be a logical end to the Onslow story, which started on that cold lake shoreline all those years ago.

Finally, a belated note on the human side. It is a simple matter for an academic to draw lines on maps and do energy calculations. But if Onslow goes ahead then the maps and lines impact real people and their livelihoods. In this regard, it is good to see that the NZ Battery group has been engaging at the local level. My favourite news item is here, where some of those effected note that they will be treated kindly. I hope that this will apply to all concerned if the scheme proceeds.


1. Background
Government policy is for New Zealand to transition toward a low-emissions economy, including 100% 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. Just a reminder here, this limited storage capacity has nothing to do with our lakes being “shallow”. Gravitational potential energy stored in a lake is proportional to (lake elevation x lake surface area x permitted water level variation). The depth of the lake doesn’t come into it.
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 presently one favoured option for providing backup power in future dry years, and will serve as the reference for alternatives. Onslow and other possible dry year alternatives are under current review and the next step is for a decision to be made on the preferred option – see the NZ Battery web page set up by MBIE. An overview of some possibilities is given in Section 7. A short overview of Onslow pumped storage can be found here.

2. Onslow energy storage potential
Energy storage at Onslow would be achieved by building an earth dam at the Teviot River outlet and raising the present 8.3 km2 Onslow reservoir to a higher level than its present 684 metres. Water would be pumped up into the lake through a tunnel from either the Clutha River or from Lake Roxburgh. The capacity to store energy (gravitational potential energy) depends on how much the lake water level is raised above the existing Onslow level of 684 metres. Fig. 1 shows the energy storage for an expanded Lake Onslow at varying water level elevations. Fig. 2 shows the shoreline of Lake Onslow if it was raised to 780 metres. The NZ Battery web page shows a similar figure with the lake outline at 760 metres. The effective energy storage is also a function of the minimum operating level because it would be unlikely for environmental reasons to have drawdown right back to the original Onslow reservoir level.
A further energy increment could be gained by constructing a short tunnel linking the upper Manorburn basin to the expanded Lake Onslow, coupled with increasing the height of the existing upper Manorburn Dam to 780 metres (existing upper Manorburn dam replaced with a 60-metre concrete dam). Combining Onslow and Manorburn, there is possibility for up to 8 TWh of effective storage.

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

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 would be to have the tunnel linking Lake Onslow with the Clutha River downstream of Roxburgh. The latter tunnel option might require construction of another dam on the Clutha. Map source: LINZ

Incidentally, the energy storage at Onslow would not be lost over time to a combination of evaporation and leakage, as has been recently suggested. If there was a leak at the bottom of Lake Onslow or it was susceptible to disappearing by evaporation, then the present Lake Onslow would have long since vanished. Onslow is not in fact in a dry location and has a similar rainfall to Auckland. An Onslow water budget analysis can be downloaded here. Put simply, stand a half-full bucket of water on a hill anywhere near Lake Onslow. Come back a year later and the bucket will have more water in it – not less.

3. Onslow pumped storage operation
In addition to its primary role of dry year buffer, Onslow pumped storage could also play various short-term roles. There would be little point in going to the expense of constructing a large pumped storage scheme and then have the pump/generating plant doing nothing until the next dry year comes along. In particular, with 1000 MW of installed generation capacity, the scheme could buffer the wind power fluctuations from about 2000 MW of new wind generation capacity. That would encourage the development of new wind generation in Southland and Otago. There is certainly no question that wind power needs to be buffered at various time scales when wind makes a contribution to the energy mix. For example, reduced mean wind speeds had a significant impact in Europe in 2021. At small time scales, pumped storage can reduce price spikes that might otherwise be generated from large scale wind power development.

Onslow would also aid the transition to a low-emissions economy by avoiding dry-year high wholesale electricity prices (as in the first half of 2021), which may tend to deter some industries shifting to electrification and away from fossil fuel energy. Interestingly, this critical need to avoid high electricity prices in dry years seems not to have been taken into account in the Climate Change Commission’s final 2021 report.

Onslow pumped storage would operate at perhaps 75% efficiency, taking system losses and lake evaporation into account. That is, it would have to purchase more electricity from the grid than it would sell and it would be more an electricity customer than an electricity competitor.

Nonetheless, 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 seasonal pumping at low-price times when South Island hydro lake levels are high, resulting in spring/summer release of some hydro lake 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 as much as an additional 100 MW could be gained on average if the Waitaki power scheme was operated just to minimise spill by keeping Lake Tekapo and Lake Pukaki levels near to their respective mid-ranges.

However, Onslow is first and foremost an energy storage system. It is in no way a government-subsidised new power scheme that would deter investment in future power generation, as has been claimed in the past. Raising Lake Pukaki for extra hydro storage did not deter generation development. Neither would raising Lake Onslow.

In the New Zealand electricity market environment that rewards generation at times of high prices, Onslow would be just one player. Sometimes the Onslow generation bid will win against other hydro. But sometimes, say, Waitaki generation would win and some water for generation would be released from Lake Pukaki. The net effect of such market operation is that in a dry year all the hydro lakes would decline along with Onslow, but decline at a lower rate than if Onslow was not there. With a return later to higher inflows, the hydro lakes would refill to provide security but Onslow might not start pumping until there was sufficient water being released from the hydro lakes to lower the prices again. The waiting time for this situation is dependent on the weather-related recovery time of the hydro lakes. For this reason, it is not helpful to ask the question “how long would Onslow take to refill after a dry year?” The minimum possible time to refill an empty Onslow is found by dividing the Onslow water storage capacity by the maximum pumping rate. However, the market would ensure that recovery would be a longer time than this, while the hydro lakes refilled first and then prices reduced. It should also be noted that the Onslow energy storage capacity is so high that is is extremely unlikely that both it and all the current hydro lakes could all be emptied in a dry year.

It is sometimes suggested that it would be better if an Onslow equivalent was located somewhere near Auckland, because that is where the power demand is. In fact, if nature had provided identical topography and hydrology just outside Auckland, it would still be much better to have the South Island location. The reason is not because of how Onslow would operate in a dry year, but how it would operate in a wet year.

Suppose the hypothetical Auckland pumped storage scheme was constructed and New Zealand has progressed to 100% renewable electricity. A South Island wet and windy year comes along, the hydro lakes are full, and the Cook Strait link is at capacity sending power to the North Island. However, there still remains unused South Island hydro generation capacity. This is the time when Onslow would be pumping at maximum. However, the hypothetical pumped storage scheme outside of Auckland could not access power for pumping because it is already being used for North Island domestic and industrial use. There would then be wind and/or water spill in the South Island because the surplus energy could not be stored there.

In contrast, sending power from Onslow to the North island in a dry year would not have the Cook Strait bottleneck because a “dry year” in effect means a South island dry year since that is where most hydro generation capacity is located. Power from Onslow would then occupy the transmission line space made available by the reduced hydro power output from the South Island stations.

What is required therefore is a South Island dry year buffer / South island wet year energy absorber, both aspects being provided by Onslow. It would make no more sense to try to do this from the North Island than it would to build a new power station in Invercargill to provide electricity for Auckland.

It is also nice to see that Transpower is upgrading transmission lines sufficiently to allow Manapouri Power to go north. This ensures transmission capacity for an Onslow component to offset a South Island dry year for power supply to the North Island. The item below is from Energy News.

4. Onslow impacts
If constructed, Onslow would be a major civil engineering scheme and environmental impact is unavoidable. There would, however, be little impact on the Clutha River itself because Onslow would generate or use power as a consequence of high pressure and not of high discharge. Also, Onslow operation would tend to shift the Clutha discharge more toward its mid-range flow. This is because the times of highest prices (generating mode) will tend to coincide with low Clutha flows. Conversely, times of lowest prices (pumping mode) will tend to coincide with high Clutha flows.

The main environmental impact would be the creation of the new lake, which would be much bigger than the original Onslow reservoir and comprise mainly lower-quality Clutha water.

In normal years, it would be expected that the new lake will have some seasonal fluctuation with levels not greatly below the maximum level. In dry years, there could be considerable lowering of water level and consequential reduction in lake area. For example, it can be seen from Fig. 1 that lowering the lake from 760 to 720 metres would expose about 30 square kilometres of previously submerged land. For this reason, it would be necessary during scheme construction to water-blast and remove the soil cover so that exposed land was always Central Otago schist rock. Otherwise there would be dust generation created from significant drawdowns.

An unavoidable consequence of any degree of raising of Lake Onslow would be loss of the existing Onslow wetlands (Fig. 3), which are themselves a remnant of the original Dismal Swamp wetlands that were flooded to create the present Onslow reservoir. In a similar way, creation of the nearby Loganburn Reservoir was at the expense of largely flooding the former Great Moss Swamp (Fig. 4).

Fig. 3. The Lake Onslow wetlands were once part of Dismal Swamp, now submerged. Map source: LINZ. The 800 and 700 metre elevation contours are shown here as darker lines. Contour intervals are at 20-metre vertical intervals, so the raised lake shoreline at maximum level can be easily traced for the 760 and 780 metre cases.

Fig. 4. The Great Moss Swamp is not so great anymore, mostly having been flooded by the Loganburn Reservoir. Map source: LINZ.

A further environmental impact of Onslow relates to water quality because any invasive plant species present in Clutha water will also be introduced by pumped storage into Lake Onslow and then to the Teviot River. Subject to further investigation, the Teviot River is not likely to be significantly impacted in this regard because its steep channel gradient will be amenable to flushing flows from time to time.

The various possibilities of Onslow environmental impact are presently the subject of a study by the Department of Conservation and will presumably appear on the NZ Battery web page in due course, forming part of the final decision as to whether to proceed with Onslow.

In a sense, Onslow pumped storage can be seen as a continuation of the many environmental impacts on the Otago natural environment. For example, the flooding of Dismal Swamp and Great Moss Swamp, the loss of the Cromwell Gorge river environment from the Clyde Dam, and the flooding of the original Lake Hawea shoreline for hydro storage. In this more environmentally aware age, the question then arises as to whether there could be acceptable Otago mitigations, to be funded as part of the total construction budget. The need for Onslow mitigations has already been noted by Jacinda Ardern during the 2020 election campaign. However, mitigations cannot be imposed and can only arise from joint community agreement. The possibilities outlined in the next section are therefore just indicative of examples which may or may not be worthy of further consideration. Of course, they are only relevant in the context of Onslow pumped storage proceeding.

5. Environmental mitigations?

(1) Could floating wetlands be established on the Loganburn Reservoir?
Subject to available Onslow offset funding and the support of the Loganburn Reservoir owners, it might be possible to construct floating wetlands over a portion of the reservoir. This would be analogous to a partial restoration of the Great Moss Swamp. There is a carbon sequestration aspect involved here also, in that organic fragments from the floating wetland would accumulate on the reservoir floor. There is also likely to be an effect of some reduced reservoir evaporation loss from the new wetlands. If a floating wetland were to be constructed, it would need to proceed by increments to ensure there was no impairment of the important role of the reservoir in the supply of summer irrigation water to the Maniototo.

(2) Shifting Waitaki River flows back to the original seasonal discharge regime
This is not a planned and funded mitigation as such, but would arise in any case from the market operation of pumped storage at Onslow. 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, the Waitaki River flows now have winter flows higher than before and summer flows lower than before. 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 would re-create the former Waitaki River natural seasonal flow regime.

(3) Could Lake Hawea be restored?
Pumped storage at Onslow would be spectacular engineering with considerable impact. Unlike previous engineering projects in Otago, a case could be made for an equally spectacular environmental mitigation: the restoration of Lake Hawea. Using Otago’s 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 storage cycles on the new shorelines are greatly in excess of the small natural variations in the original lakes.

Onslow has multiple national advantages including emission reduction and avoiding periods of high power prices. From a South Island viewpoint, Onslow pumped storage could be seen as a means to a specific environmental end: to start the process of rehabilitation of the South Island scenic lakes that were flooded for hydro storage. As noted at the start of this post, the hope behind the original 2005 paper was that the very large Onslow-Manorburn reservoir would enable the shoreline impact of seasonal hydro water level variation to be transferred away from the scenic lakes to the much more resistant hard base rock shorelines around an expanded Onslow reservoir. Onslow pumped storage, if constructed, will not be on the grandiose scale of the original paper. However, some of the original motivating philosophy could still remain: the complete restoration of Lake Hawea.

Lake Hawea is the only natural lake in the Clutha catchment that has been flooded for hydro power storage operations, enabling some limited control of water flows to Contact Energy’s Clyde and Roxburgh power stations. The other large lakes in the catchment, Wanaka and Wakatipu, are uncontrolled natural lakes. 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 much shorter duration. The present Hawea 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. The drowned forests at the head of Lake Hawea is still visible protruding above the water (Fig. 5), but have received little publicity due to difficulty of access. Fig. 6 shows how the map of Lake Hawea would appear if it was lowered back to its original shoreline. The new areas of land (bright green on the map) would have potential for planting of significant areas of silver beech forest, perhaps even more extensive than the original forest cover. Fig. 7 illustrates wave erosion effects along the lake shore, and the modified flow of the Hawea River.

Fig. 5. Only the upper branches are visible of this submerged native forest, drowned by raising Lake Hawea for hydro storage. Picture source here. The forest is obviously not taking in carbon dioxide, but it will have been producing methane.

Fig. 6. Lake Hawea (blue) as it would appear if restored to its original level. The green zones, presently underwater, are the main areas that were submerged when the lake was raised in 1958. These localities would be available (where terrain permits) for parkland and forest re-establishment. The Hunter River would discharge at the original shoreline just south of the Little Hopwood Stream. A former small lagoon near The Neck would be present again. Hunter Valley river channel patterns are indicative only. Map source: LINZ.

Fig. 7. Hydro operation impacts at Hawea: wave-eroded shoreline near Hawea township (above), river flow fluctuations in the Hawea River (below). River flow data from the Otago Regional Council web page.

Hydro storage in Hawea and the other South Island lakes has been an essential part of maintaining resilient 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 any significant amount of Onslow energy storage and 1000 MW of Onslow installed capacity, there would no longer be a need for the 0.35 TWh seasonal storage component at Lake Hawea. Fig. 8 shows some relative storage magnitudes. The energy storage capacity of Lake Hawea could be achieved by extending lake Onslow from 760 metres to about 763 metres.
From a commercial viewpoint, the 1000 MW of Onslow installed capacity will massively flatten out the present seasonal electricity price differential. That is, winter hydro water will not be very different in value from summer hydro water. The Clutha power scheme could then operate quite differently, with water only held back briefly in a restored Lake Hawea during flood events. The spring-summer surplus power from the Roxburgh and Clyde stations is then likely to be purchased by the Onslow operators for pumping. This might generate more income for Contact Energy from the resulting higher summer prices. That is, even at present Lake Hawea storage gives only limited seasonal control of discharge at the Roxburgh and Clyde stations because the main flow contributions are from the uncontrolled discharges out of Lakes Wanaka and Hawea (Fig. 9).

Fig. 8. Lake Hawea storage and other hydro storage magnitudes. Ngaruroro is a potential pumped storage reservoir near Taupo, but unlikely to progress because of both engineering and environmental considerations. Onslow energy storage capacity (if constructed) is still uncertain. It is shown here as 8 TWh and is unlikely to be less than 5 TWh.

Fig. 9. Mean Clutha flow as recorded at Clyde (blue), and natural Clutha flow at Clyde as estimated in the absence of Lake Hawea storage. (a) Flow means for June, July, and August, (b) Flow means for October, November, and December. Data kindly provided by Contact Energy, with no implication of support of any of the content of this post.

If restored, Lake Hawea would lowered by increments (to avoid dust issues for Hawea Township) back to its original 327-metre level and its 90 km of flooded shoreline restored. The Hunter River would then extend to the south by about 9 km from its present delta. Forest replanting would be required to restore the regions of the drowned forests and re-create the original biodiversity. The new forest would also represent a useful carbon biomass gain and generate carbon credits.

Lake Hawea outlet water releases from a restored lake would be maintained at about equal to lake inflows, so the Hawea River would largely revert to its pre-hydro flow pattern.

With Onslow pumped storage operational, there would be no obvious national, regional, or commercial good in keeping Lake Hawea in its present flooded state. However, artificial groundwater recharge or other mitigations may be required to offset any groundwater level reductions impacting domestic or irrigation use.

(4) Could a new wetland/lake complex and upland reserve be created near Onslow?
The gentle topography of the upper portion of Bonds Creek (north east of Onslow) could enable a low dam to create of an extensive shallow lake (Fig. 10). Subject to willingness of the landowners concerned, an Onslow mitigation fund might purchase the land and create the new lake which would be a natural lake for practical purposes.

Fig. 10. Maximum extent of a new lake (dark blue) that could be created from a low dam (red) in the headwaters of Bonds Creek, with lake level at 760 metres asl. Map source: LINZ

If the lake was in fact constructed, it would be a first in New Zealand – a new lake created purely for environmental gain. Something in the same philosophy is seen in the recent purchase of a dairy farm for conversion to a wetland to improve the water quality of Lake Horowhenua.
If there was willingness from all concerned, ecological studies would need to be initiated to establish the likely extent of new Bonds Creek wetlands and also evaluate its possible trout fishery development. Ideally, the entire catchment area of the new lake could be set aside as an Otago upland reserve, permanently ensuring no future water quality impact from agricultural development.

6. 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 still 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 has any right to expect 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 serious contender to handle the issue of dry year buffering. A range of some possible dry year options, including pumped storage, is listed below.

7. Selected dry year options

Gas backup. This is the approach favoured by the Climate Change Commission, despite being in opposition to the government policy of 100% renewable electricity. The argument is based on the idea that eliminating the final few percent of fossil fuel power generation will be very expensive. However, gas has its own issues as dry year backup. A report by Concept Consulting notes that gas suppliers may be reluctant to enter into contracts for the inevitably unpredictable dry years, as opposed to them providing steady gas supply to industry. Also, gas infrastructure is not necessarily reliable at times of dry year urgency. For example the high wholesale electricity prices that have impacted New Zealand industry this year have arisen from a combination of low hydro inflows and gas supply issues from the Pohokura field. There is also the uncertainty factor factor in gambling that a large new gas field could be found to enable future security of electricity supply within some reasonable time (see this link to a short video of a parliamentary debate in August, 2021).

Battery storage. Large batteries could be used in principal to store some power output from normal years, with their energy later released in dry years to make up for decreased hydro output. The attraction is that the batteries could be located anywhere with minimal environmental impact. Unfortunately, the economics are not good. The world’s largest battery is the Tesla battery that was set up in South Australia for a purchase cost of about 100 million NZ$.  It has energy storage capacity of 0.0002 TWh. The needed 5 TWh of new storage capacity would therefore require purchase of 25,000 of those batteries for a cost of 2.5 trillion dollars, excluding setting up costs. Also, batteries don’t last forever and a further 2.5 trillion dollars would have to be found later for replacements. Even allowing for improved technology, batteries will not be used to buffer New Zealand dry years.

Hydrogen storage. In this model, renewable power in normal years is used to create green hydrogen by water electrolysis. The hydrogen is stored and in dry years used to generate electricity – by using fuel cells, for example. There are two negative aspects to this approach. Firstly, there is the issue of finding some means of safely storing the large amounts of hydrogen, or hydrogen derivatives, which would equate to 5 TWh of chemical potential energy. Secondly, the process of converting from electricity to hydrogen and back to electricity is very inefficient and would lead to high electricity prices. The Interim Climate Change Committee accelerated electrification report estimates 14% conversion efficiency. This would require considerable additional renewable generation capacity just to provide the extra power to make up for what is lost in the inefficiencies.

Renewables overbuild. This approach to 100% renewable electricity involves constructing more renewable generating capacity, mostly wind, than what would be required in normal years. There would then be sufficient generation capacity in place to offset the decline in hydro lake levels in a dry year. However, this is an expensive option because it involves construction of generating capacity that is not used for most of the time. The Interim Climate Change Committee accelerated electrification report estimates that renewables overbuild would increase the electricity price for household and industrial users by 14% and 39%, respectively. That report estimates the cost per tonne (beyond natural gas) of carbon dioxide removal by renewables overbuild is about $1,200. Renewables overbuild is one mechanism of achieving 100% renewable electricity. Pumped storage is another. However, the $1,200 figure is often misrepresented as “the high cost shown by expert analysis of getting the last few percent to achieve 100% renewable electricity” without reference to the fact the $1,200 is specific to renewables overbuild only and not to pumped storage. See Section 8 for an overview of how the misconception developed.

Demand interruption. The concept here is that agreements are set up for large electricity users to shut down all or part of their plant for a time during a dry year, presumably with financial compensation to the industries concerned. An agreement of this type is currently in place between New Zealand Aluminium Smelters Ltd and Meridian Energy, with respect to the Tiwai Point aluminium smelter. Meridian and Contact Energy are presently scoping the possibility of setting up hydrogen production in Southland with dry year demand interruption. However, it is difficult to see how this type of approach would make a substantial contribution to dry year buffering. Setting up a new electricity-intensive industry would have the effect of drawing down the hydro lakes at a greater rate than if the industry was not there. Shutting down production when the lakes became low would provide additional power to the grid, but only for that brief period of time before the depleted lakes went down to zero storage.

Extracting additional storage from existing lakes. This is the cheapest option by far, involving little more than redefining permitted hydro lake operating ranges to extend some way below their present consented minimum levels. As an extreme example, up to 4 TWh could be achieved in principle by permitting the operating range of Lake Pukaki to extend down to the original natural lake level. The constraining factor in this approach is the environmental impact of the lakes having shorelines of extensive areas of exposed lake bed in a dry year. However, the impacts would be location-dependent. For example, increasing energy storage capacity at Lake Taupo by 1 TWh could be achieved by allowing 2.4 metres of lake level fall below its present minimum consented operating level. In contrast, one additional TWh for Lake Pukaki would require 8.6 metres of water level fall, giving a much greater visual environmental impact. This relates to water volumes. Even though Lake Pukaki is higher, Lake Taupo more than makes up for that by its larger surface area. Mercury Energy has been reported as not ruling out the possibility of seeking dry year lowered operating levels for Lake Taupo.

Increasing storage from new or expanded lakes. There is little scope for significant water level increases in existing lakes because of economic, cultural, and environmental considerations. However, new lakes at sufficiently high elevations are a possibility. These high valley sites are not part of major river systems so it would take a long time for them to fill by water inflows from local streams. Equally, there would be long refill times following a dry year water release.
A convenient way around this is issue is to pump water into the upper valley from a lower elevation, and then run the water back down for power generation when needed in a dry year. This is pumped storage and the local streams and rainfall now only serve to offset evaporation losses from the new lake. Lake Onslow has received most discussion to date as it could accommodate the required 5 TWh of energy storage capacity. Genesis Energy may be considering pumping water from Lake Taupo up to a new lake to the east of the Desert Road.
Pumped storage would have minimal impact on average electricity prices and would avoid the high prices from fossil fuel dry year backup as at present. The Interim Climate Change Committee accelerated electrification report estimates Onslow pumped storage costing $250 per ton of emissions abatement (beyond the cost of natural gas), as compared to about $1,200 for renewables overbuild. The various option costs are listed below

Fig. 11. Marginal emissions abatement cost for the dry year solutions above the cost of natural gas ($ per abatement cost of CO2 equivalent). Modified from NZ Battery web page.

Slow-fill lakes. Slow-fill lakes are a special category of the previous class. Their long times to fill and refill means they could not be considered as being of use for normal dry year backup – say, once every 10 years. However, they could provide additional backup against extreme situations like multiple dry years that might occur only once every 30 years or so. Slow fill lakes do have one advantage in that no energy is expended in their water accumulation. As mentioned earlier, there are two basins that could be involved for Lake Onslow pumped storage – the Onslow basin itself and the smaller and higher Manorburn Basin. The latter could be used as an expanded slow-fill lake which could eventually store water from local streams with energy storage equivalent to Lake Tekapo. Similarly, a new lake near the Desert Road could be used as a slow-fill lake to avoid tunnel costs to Lake Taupo. In this case, dry year power power could be generated by releasing the water to pass first through the Tongariro Power Scheme to Lake Taupo, and then through the Waikato River power stations.

Combination solutions. The NZ Battery group may opt for some mix of the above. For example, one scenario might be constructing Onslow pumped storage with 5 TWh at Onslow itself, with slow-fill backup in the Manorburn Basin. In addition, a slow-fill lake might be constructed near the desert road. One useful potential addition could be having 1 TWh of additional storage at Lake Taupo as last-resort water. Hopefully the Taupo water drawdown would never be needed but would be available in the event of a major outage preventing power transfer from the South Island.

8. The “expensive” final percent to total renewable electricity: tracing the evolution of a fallacy

A $30m evaluation of 100% renewable electricity generation for New Zealand was announced by the government in July 2020. A further $70m was announced in September 2020, earmarked to enable more detailed study of specific dry year options.

Following those announcements, which were mostly concerned with pumped storage possibilities, there has been a degree of misinformed discussion over the cost of removing the final small contribution of fossil fuels in electricity generation.

The issues date back to the 2019 report by the Interim Climate Change Committee (ICCC).

Part of the report was concerned with the cost of introducing 100% renewable electricity. This cost was measured by marginal emissions abatement per tonne of carbon dioxide equivalent, and by estimated percentage increase in retail electricity prices as referenced against 2018 prices.

The report gave particular attention to renewables overbuild as a means to enable 100% renewable electricity through both normal and dry years. The overbuild model requires having excess renewable generating capacity in normal hydrological years. That is, constructing and maintaining generating capacity which is not operational for most of the time.

Unsurprisingly, the overbuild approach was found to be very expensive. Retail electricity prices were estimated by the ICCC report as being likely to increase by 14%, 29%, and 39% for residential, commercial, and industrial users, respectively, by 2035. The marginal emissions abatement cost was estimated as $1,280 per tonne of carbon dioxide equivalent.

Regrettably, much of the text in the ICCC report refers to the overbuild cost simply as “the cost of 100% renewable electricity”. The report then went on to recommend that the apparently expensive 100% renewable electricity goal should be dropped. Focus instead was to be on emissions reduction generally, because electricity generation produces only about 5% of total national emissions.

The ICCC report also recognised, however, that pumped storage might provide an alternative path to 100% renewable electricity. A recommendation was made for further investigation. In particular, Lake Onslow in Central Otago appeared a better economic prospect with a marginal emissions abatement cost of $250, much lower than the $1,280 overbuild value. The ICCC report did not estimate electricity prices arising from pumped storage but an independent review indicates that prices would be likely to either move downward or remain largely unchanged.

The government followed the ICCC recommendation and established the NZ Battery project to evaluate Onslow pumped storage and other approaches to renewable energy systems buffering against dry years.

However, given the lack of clarity in the ICCC report as to what their 100% renewable electricity cost actually referred to, it was inevitable that there would be some confusion in subsequent comments. For example, the New Zealand Initiative quickly congratulated the ICCC for supposedly identifying that establishing 100% renewable electricity would increase electricity prices.

A further example was the TVNZ leaders debate in September 2020, where Judith Collins incorrectly equated the ICCC overbuild electricity prices as applying to Onslow pumped storage.

Taking the charitable view, the factual error by Collins might be put down to the National Party only having a quick skim though the ICCC report in the heat of an election campaign. However, the supposed high cost to the nation for getting the last few percent of renewable electricity soon became established as an apparent fact.

Thus the February 2021 draft version of the Climate Change Committee (CCC) report was approvingly supported by Contact Energy in its submission:

The Commission correctly identifies the very high abatement cost of $1,200 per tCO2e required to achieve 100% renewable generation by 2030.

So again, the ICCC figure of $1,280 is incorrectly referenced as the unavoidable price needed to achieve 100% renewable electricity. See also the Trustpower reference to the $1,200 figure in its submission to the CCC draft report.

The final version of the CCC report, presented in June, persisted with the $1,280 error and kept its original recommendation of maintaining a gas contribution to electricity generation. A footnote on p. 279 added that the 100% renewable electricity target should be treated only as “aspirational” because of its supposed upward pressure on electricity prices:

Work undertaken by the Interim Climate Change Committee (ICCC) demonstrated that moving from 98% renewable electricity to 100% renewable electricity would cost about $1,280 for every tonne of carbon dioxide abated, and would result in higher electricity prices. Higher electricity prices could make switching to electricity as a low-emissions fuel relatively less attractive.

This stays with the implication that the expensive renewable overbuild model was the only path that was identified in the ICCC report as being capable of achieving 100% renewable electricity generation. Indeed, the supposed  $1,200 expense is so great, and the emissions reduction so small, that the 100% renewable electricity goal was seen by one Newsroom article as being little more than a distracting red herring. Another Newsroom item restated the now familiar 14, 29, and 39% price increases as the “cost of 100% renewable electricity”.

Far from being a red herring, the presence or absence of a small component of fossil fuel for power generation is a critical factor as far as future energy-related emission reductions are concerned. This is because the New Zealand electricity market functions such that the wholesale electricity price at a given time reflects the highest marginal cost of generation. For example, burning natural gas does not provide a large component of total power generation or emissions contribution. However, its relatively high cost determined the high wholesale prices for all electricity in the first part of 2021 when lake levels were low.

Whether we wanted it or not, nature contrived in 2021 to demonstrate the high-price impacts of going into a dry year in the absence of renewable backup. This has implications for the transition to a low-carbon economy because, as the CCC report noted, higher electricity prices reduce the attractiveness of switching to electricity as a low-emissions fuel.

The major electricity generators should not be seen as unbiased commentators when they collectively oppose the 100% renewable electricity policy, with its implications of eliminating high spot prices in the electricity market.

If we can finally remove the distraction of the irrelevant and expensive overbuild model, the question still remains as to the best means of achieving the desired transition to 100% renewable electricity generation. Will the final few percent still turn out to be expensive or not? There is uncertainty here but the question is presently under active consideration by the MBI NZ Battery project, due to report in April or May 2022.

In the event that Onslow pumped storage plays a significant role, Keith Turner has raised the interesting possibility of painless payment of the billions involved in scheme construction while at the same time still having lower electricity prices. If achievable, that would indeed be a “just transition” toward reduced inequality in the spirit outlined by James Shaw.

In the meantime, it seems unlikely that the government will act on the CCC advice to abandon its 100% renewable electricity policy and keep a gas contribution. Its response to the CCC report is required before the end of year. If the Onslow scheme proceeds, the cost of converting the last few percent to 100% renewable electricity may well be measured in environmental terms as much as in dollars. As noted by Jacinda Ardern, the issue then is whether sufficient environmental mitigations can be identified and considered along with the project itself. Work along these lines is presently being undertaken at the University of Waikato.

9. Onslow in the long term – a speculation

If Onslow pumped storage is constructed, it is well situated to be in operation for a very long time. That is because Lake Onslow has only small inflowing streams which carry minimal sediment loads. In contrast, the Kawarau arm of Lake Dunstan is already accumulating the anticipated input of silt derived from the Shotover River.

The concept of pumped storage at Onslow has been around since 2005. However, motivation for its possible construction only dates from 2020. That is when its value was seen by the New Zealand government in terms of a possible means of dry year buffer for a future electrified low-emission economy with 100% renewable power generation.

However, it would be wrong to see Onslow as a means to transition to an end point of a sustainable nation based on renewable electricity. This is because renewable power is itself not sustainable and is only desirable in the sense of being a better option than fossil fuels. That is, in the long term the renewables scenario is something we transition “through” rather than “to”.

Of course, the sun will always shine and the winds blow. However, our electricity demand will always creep upwards and there must come a time when we reach peak renewables. It is neither desirable nor even possible to keep adding windfarms onto the next ridge, or dam the last stream. Variations on the renewable theme – offshore wind farms, better light bulbs, solar panels, tidal power in Cook Strait etc – are just kicking the can down the road and only delaying the inevitable crunch point of peak renewables.

In the long term, a nation with a large hydro and wind component is vulnerable to climatic shifts. In New Zealand’s case, we are dependent on the absence of any climatic shift resulting in a major weakening of the westerly wind system – the driving factor contributing to both our wind power and hydro rain. No amount of new hydro storage, Onslow or anywhere else, could offset such a climatic shift. The encouraging aspect is that the consensus appears to be that the westerlies will strengthen over the South Island. But what if it went in the other direction?

Our alternatives against a long-term negative climatic shift in the far future are not great because we are a remote Pacific island. An undersea power cable from Australia or Asia is a remote possibility. However, even if was feasible it would still be undesirable because we would be at the mercy of whoever was at the other end of the cable. Importing hydrogen, green or otherwise, for hydrogen power stations would be equally undesirable. This is because hydrogen is not amenable to large scale safe storage to provide buffer against the uncertainty of supply from a global hydrogen market. The “safe” aspect here is important because hydrogen and its derivatives are highly explosive and/or toxic. History tells us that if something can blow up then sooner or later, it will.

If New Zealand continues down its renewables path then we will be sleepwalking our way to eventual nuclear power stations. That will be because by the time the renewable peak hits, it will most likely be too late to go to anything for the next power increment other than nuclear, unless we go back to fossil fuels.

It sounds radical but there is one alternative possibility – a globally-subsidised and emission-free solid fuel as a cheaper alternative to coal for power stations. Call it newfuel for now. Given plentiful supplies of newfuel and power stations to burn it in, there would be much less interest in maintaining or building coal-fired power stations. An essential requirement would be that the fuel is produced using emission-free energy.

Why would the new fuel be subsidised and who would subsidise it? It would need to be subsidised because the its production would be energy-inefficient. That is, it would take more energy to produce newfuel than the amount of energy it contains. In other words, like green hydrogen, the new fuel would be an energy vector and not an energy source. The subsidy for newfuel production would be as part of the fight against global warming. It should therefore be paid by those nations who have contributed the most to the human-generated greenhouse gases that have accumulated to date in the atmosphere.

New Zealand would stand to gain from a global newfuel scenario because we could avoid nuclear power. A retrofitted Huntly power station is a possibility, assuming thermal power stations. Another power station might be constructed near Balclutha for the South Island, taking advantage of the Clutha River for cooling. There would no longer be a need for dry year or seasonal hydro storage, so the raised scenic southern hydro lakes could be restored to their original natural shorelines. Hydro power would still be of importance for doing what it is good at – meeting rapid demand fluctuations such as the daily demand cycle and the variabilities of wind power. A newfuel power station might even be set up to offset some of the worst impacts of hydro power. For example, the Manapouri station might be permanently closed with the re-creation of the Waiau River seen as the greater national good.

What of Onslow pumped storage in a future newfuel scenario? The large pumping/generating capacity of Onslow would still be of value because it is well suited to smoothing power fluctuations and demand variability. However, its large energy storage capacity would no longer be required and the lake would now fluctuate only a few centimetres about its mean level. Future generations would determine what that level should be. Perhaps it would be the original reservoir level or maybe some higher level for recreational and ecological purposes.

A particular attribute of newfuel that would be helpful for purchasers would be ease of stockpiling. The idea would be to import more than needed and build up a multi-year stockpile to ensure security against any supply interruptions.

It would be helpful therefore if New Zealand could be a leader in both fuel development and its power stations. Our climate is our averaged weather and nothing we do here will change our weather. On the other hand, being part of fuel development means playing a direct role against global warming. We do have some precedent in that the first paper proposing a global new fuel (metallic silicon) originated in New Zealand.

An interesting feature of silicon as a fuel is that, lump for lump, it has the same energy on oxidation as burning high-grade coal. Its advantage is that it is abundantly available in the oxide form as desert sand and generates no emissions as a power station fuel. Its “ash” is essentially just fine sand and the silica fuel could be stockpiled indefinitely without degradation. A research project at the University of Waikato defined a method for industrial production of fuel silicon using a magnesium intermediary rather that carbon, which is presently used for high-grade silicon production.

There are many other new fuel possibilities also that have been subsequently proposed. A summary can be found here.

Like any commodity, the new fuel would require both a supplier and a market. New Zealand would not be the only country that would prefer to avoid nuclear power – Japan would be one of many others. A major research and development program would be needed, involving serious funding that could make Onslow expenditure seem insignificant. The development group would involve industrial chemists and combustion engineers, to first define the optimal fuel and then design and build a demonstration power station. There would also need to be movers and shakers who can make things happen to set up the global market. Taranaki would seem a logical home for such developments, given the present local skills base. The critical requirement for solid fuel development – silicon, aluminium or whatever it may be – will be milling it down to a sufficiently fine powder so the oxidation process can continue to completion quickly. Burning coal does not does not have this problem because the burning process exposes new coal surface to maintain the oxidation, at the expense of carbon dioxide production.

We shouldn’t be wary of breaking new ground on this. Just as Onslow may turn out to be the world’s largest pumped storage scheme by energy storage, there is no reason why the world’s first prototype new power station should not be built in Taranaki. Developing ability to create off-the-shelf power stations of the new type could be of particular value and it might be speculated that Taranaki could develop an export industry in the same way that Denmark now is a major supplier of wind turbines.

The technological developments involved for a demonstration power station are significant but not so daunting that they could not be achieved in New Zealand. It would be orders of magnitude simpler than the complexities of nuclear fusion, which might never happen for practical power generation and certainly not in time for climate change urgency.

Making all this happen requires significant funding partners who also stand to gain. There are major global funds looking for novel green investments. Also, the Middle East oil nations could be approached. They have available capital, they are looking for sustainable alternative exports as their oil and gas reserves decline, the waste heat from newfuel production process could be coupled with desalination for freshwater supply. There is also the practical issue that an increase in global warming to any extent could make those nations almost unlivable at times.

The source of energy for creating the new fuel at source left open. In the first instance it might come from natural gas to get the market established, to be followed soon after by a mix of renewable and nuclear power.

Of course, all this is only speculation about a future energy scenario that may be unrealistic and never go past a preliminary scoping evaluation. However, in an uncertain age of climate change we should follow up all possible avenues of investigation. Regardless of how our electricity production evolves, if Onslow is constructed then it will play a future role in some form and will become part of the landscape well into the future, though perhaps reduced in size in the end.