Guest Post by Geoff Russell. Geoff is a mathematician and computer programmer and is a member of Animal Liberation SA. His recently published book is CSIRO Perfidy. This article follows on from his previous: What price of Indian independence? Greenpeace under the spotlight
In my previous BNC piece I examined the feasibility of two ways of producing a per-capita electricity supply in India which was roughly equivalent to that in Australia in 2011. The assumption was that such a supply would reduce vast amounts of suffering and transform India into a first world country over an implementation period of about 40 years.
I was also critical of Greenpeace India for its Euro-centric view of nuclear power. There is something bizarre about Greenpeace India identifying with the wave of street marches in Europe over a non-fatal nuclear reactor failure under extraordinary circumstances in Japan when a quarter of a million Indian children between 1 and 5 years old die every year due to cooking smoke because they don’t have electricity.
This post is a continuation. Part II if you like. It will have two goals:
- Motivate the scenario approach in more detail.
- To examine a third scenario. This isn’t some straw man of my own making, but comes from Greenpeace and the European Renewable Energy Council.Greenpeace International has a detailed conceptual energy plan for India which involves a phase out of nuclear power and building a sustainable energy infrastructure for a projected 1.6 billion people in 2050. Page numbers below are from that plan.The global sustainable CO2 emission level, as defined by Greenpeace, is about 1.3 tonnes of CO2 per person (ignoring non-CO2 forcings for now) (p.8). This is a little higher than the 1 tonne figure I used in my BNC piece, but near enough not to be an issue. In any event, Greenpeace’s plan for India would result in greenhouse gas emissions from energy production of 1 tonne per person per annum in 2050. The plan is cogniscent of India’s biomass cooking problems and has sufficient detail (52 pages, including a full energy spreadsheet) to expose all the assumptions behind its non-nuclear future. This is a professionally produced report done by EREC (European Renewable Energy Council, a body representing the European renewable energy industries) in conjunction with Greenpeace International. The Indian plan is part of a global vision with similar reports and recommendations for other regions.
I’ll call this plan the EREC/Greenpeace (ERGreen) scenario and discuss it shortly, but first there were comments in response to the previous post which indicate that I need to at least motivate, if not fully justify, this kind of high level scenario thinking when people are anxious to go straight for detailed costs of specific projects.
Scenarios, goals and mountain climbing
Getting to the top of a mountain from some point down below is a tricky problem when you don’t have a map. Even worse when nobody has a map; and worse still when nobody has ever done the climb. Early explorers failed in attempts to cross the Blue Mountains in Sydney by following rivers up-stream and getting stuck at water falls they couldn’t climb. If you are sitting in your tent in a densely forested base camp and you can’t see the summit, what do you do? Even more of a problem is that even if you can see a peak, it may not be the summit you are looking at but just some little pimple obscuring your view of the real top point.
This is a pretty good metaphor for energy planning. My strategy of looking at simple scenarios is like trying to get the lay of the land before heading off. It’s picking a few points seen dimly through the fog and estimating their relative height. Otherwise the tendency is to head off in the steepest direction in the hope that it will get you somewhere high.
But the 3-dimensional mountain climbing analogy isn’t perfect. What we really face is a problem with many more dimensions. We are simultaneously concerned with CO2 emissions, short lived emissions of things like methane and ozone, financial cost, job creation, biodiversity impacts, health impacts, disaster risks and construction speed. Mountain climbers have it easy.
Comparing two scenarios
The ERGreen document is very much a scenario document, but far more detailed than my blog piece. It’s a money document rather than a physical feasibility document. There is no specification of the area of Solar PV (Photo Voltaic) farms, or of wind farms. There is no estimate of the required amount of steel or concrete needed. Average factors are used to estimate emissions from each technology rather than physical simulations. It is assumed that a properly structured feed in tariff will drive adequate investment with the developed world guaranteeing the tariff for the next two decades. Living in a country where greed is the dominant cultural artifact, I suspect this is a questionable funding strategy … but nor is it clear that India can totally self-fund a transition to a first world standard of living without using the same cheap and dirty technologies which the current first world used and is still using.
ERGreen express electrical energy in terawatt hours and overall energy in peta joules. I’ll use watt hours for everything … terawatt hours (TWh) for national figures and megawatt hours (MWh) for per capita figures. One TWh is a million MWh.
Here’s a summary table comparing my thumbnail nuclear (TNNUKE) plan with ERGreen. Firstly, note that FD=final demand. Energy statistics are complicated by all kinds of losses between energy as produced (usually called primary (P)) and the energy actually consumed (often called final demand(FD)). There are a few (P) figures below just to stress how much fossil fuel remains in ERGreen, but its generally only final demand that concerns me here.
ERGreen aims for 13,000 TWh in final energy for India by 2050 with 4,600 TWh coming from electricity. My nuclear scenario delivered 14,000 TWh of electricity and I didn’t specify any details about non-electrical energy.
I was hopeful that India’s population could be pegged to 1.4 billion. ERGreen is using 1.6 billion as their estimate by 2050. I’ll use their estimate in what follows. My 14,000 TWh is about 8.7 MWh of electricity per year for each of 1.6 billion people. This is rather less than the current Australian level of 11 MWh, but rather more than the current levels of 6.3 MWh in Spain and Denmark and 7.7 MWh in France.
TNNUKE’s goal was a modest but adequate first world electricity supply, plus some extra to deal with declining levels (or high cost) of oil availability. ERGreen is 2.8 MWh per person of electricity with the remaining energy from biomass and fossil fuel.
As you can see, ERGreen expects Indians in 2050 to manage with less than half the electricity of Spain and Denmark. This isn’t just an issue of household consumption but of the electricity to run factories, produce aluminium, mold plastics and all the other things required to develope the country’s housing and plumbing infrastructure. I’m all for frugality and efficiency, but the ERGreen electricity figure looks very low. I’ll discuss the energy efficiency aspects of ERGreen later.
Solar Photo Voltaic (PV) matters
Under ERGreen, Fossil fuels will provide 1,730 TWh of electricity coupled with a mix of renewables, the biggest being Solar PV which will provide 1,530 TWh per year by 2050. Based on current German technology, with 2 hectares producing 3.1 GWh per year, this is about a million hectares of Solar PV. The nuclear scenario was for 166 sites of 1000 hectares per site with an output of 14,000 TWh.
In a country as crowded as India, space is at a premium and even desert areas are seldom vacant. In any event ERGreen places a premium on having local energy supplies, so they wouldn’t just stick a million hectares of Solar PV panels in the Thar desert with HVDC connections to the rest of India. Nor is such a plan likely to be feasible. The Thar may be India’s biggest desert, and there’s 20 million hectares of it, but it is heavily used. Ber fruit trees grow in the Thar and can yield 10 tonne of fruit per hectare. Guar trees provide gum for the world’s ice cream makers and the Thar is also the home of India’s largest wildlife sanctuaries.
Finding spaces for a million hectares of PV panels will be tough, however you partition it. The entire area of India’s National Parks system is just under 4 million hectares. India’s average human density is currently 3.6 people per hectare, so a million hectares of panels will displace roughly 3.6 million people or a sizeable chunk of wildlife or domestic animals or a combination of the three. Displacing wildlife is usually the cheaper of the three options since their demands for compensation are zero.
It might be argued that using the average human density like I’ve just done is unrealistic, because India has many big cities which reduce the average density in the rest of the country. True enough, but ERGreen’s policy of keeping energy sources close to population centres guarantees they will be in denser rather than more sparsely occupied areas. This probably makes 3.6 million an under rather than an over estimate.
There are plenty of references to energy efficiency in ERGreen, but rather less that is concrete. We have had energy efficiency labels on white goods in Australia for decades and they have done nothing. Our per capita energy consumption has been rising for decades. Elsewhere it is the same. Among International Energy Agency (IEA) countries per capita energy consumption doubled between 1974 and 2006.
More careful climate corrected measures in these countries of per-capita final energy use (not just electricity) show a per-capita rise of 2.9 percent over the period 1990 to 2006, when climate change and energy efficiency were well and truly front and centre. Such a small rise looks promising for people who think efficiency can be more than just a slogan, but a recent study by Steven Davis of Standford University in the US shows that the real per-capita energy use was actually significantly higher because of the import of energy rich goods by the developed world. For example, Europeans imported goods generating 4 tonnes of CO2 per person per annum in 2004. This means that we need to add about 4 MWh of extra energy onto the usage of each European which turns a 2.9 percent rise into something more like a 30 percent rise. Another study found a 1.2 billion tonne transfer of greenhouse emissions from the developed world to the developing world between 1990 and 2008.
Nobody can reasonably be opposed to energy efficiency but betting the planet on it is naive. For many people, profligate consumption is a matter of pride. In my hometown of Adelaide, there are many small groups of energy efficiency enthusiasts but their voice is drowned out by hoons burning rubber in a thousand back streets every night of the year. An energy efficiency display draws small groups of the earnest while the V8 supercars pull crowds in the hundreds of thousands with celebrities and politicians from both major parties milking the photo opportunities.
Will India be different? I desperately hope so. But it would be a foolish to assume so.
ERGreen estimates transport energy requirements in India to rise by a factor of 6 by 2050. This is about half the increase predicted by the IEA in its reference scenario. ERGreen predicts the increase will be met by an almost doubling of crude oil use to 2,632 TWh.
Is a factor of 6 reasonable? India has 12 cars per 1000 people. Clearly most Indians walk, cycle of use public transport now. A factor of 6 increase by 2050 would see them at 72 cars per 1000, still well below the 2008 Chinese rate of 128. Of course, they could put most of their transport increases into public transport. What kind of policies will be required to prevent a rise in motor car use? None are suggested.
If ERGreen’s transport estimate is reasonable, then under the nuclear plan, there is capacity for electric vehicles recharging off peak from nuclear plants running at or close to full power 24×7. If the IEA demand forecast is closer to the mark, then we would need some 5000 TWh flowing from nukes to electric vehicles. This would constrain other energy uses but still looks possible. It would be totally impossible to meet such a demand from the ERGreen energy infrastructure.
Cooking and biomass
ERGreen mentions the biomass cooking health impacts but their scenario postulates an increase in biomass use. Of course, it is possible to cook and heat with wood in a way that doesn’t kill and sicken people. In the US, the Government is subsidising the cost of clean replacement wood heaters to the tune of $1000 each. Electric stoves and microwave ovens can be much cheaper than this. They simply don’t need the amount of material required by a well sealed wood stove with a proper flue.
So it isn’t at all clear that the ERGreen plan actually solves the wood smoke problem, nor is it clear where the expansion in biomass is going to come from. ERGreen makes the common mistake (p.37) of thinking of crop residues as fuel. But crop residues have essential functions of protecting soil structure and reducing erosion. Removing them exacerbates the negative nutrient balance of cropping.
Climate constraints on biomass use
The latest climate science has shown that in order to avoid dangerous climate change, the atmospheric concentration of CO2 must be reduced to about 350 ppm by the year 2150. This is explained in a 2008 paper by NASA’s James Hansen and 9 other authors.
The Hansen paper explains that 3 simultaneous constraints we must satisfy to enable us to achieve this 350 ppm goal by 2150:
- Phase out unsequestered coal use by 2030.
- Halve non-CO2 forcings like methane, ozone and black carbon.
- Rollback 200 years of deforestation to draw down additional CO2.
The picture tells the story, but for the details, you really need to read the actual paper. It’s worth the effort.
The first constraint one is looking globally doubtful and ERGreen fails to satisfy this constraint for India.
ERGreen contains nothing relating to constraints two … the short lived climate forcing constraint.
The reforestation constraint, number 3, places severe restrictions on any use of biomass. Some land deforested in the last 200 years is now cropped or settled. So land most available to be reforested is grazing land. Some grasslands will never reforest if destocked and choices will need to be made between grazing and biomass (for biofuel, not wood stove cooking) use.
All three actions are necessary. No single action is sufficient.
ERGreen shows no understanding of these climate science constraints.
India’s methane problem
India has a very serious methane problem due to the place of cattle in the Hindu religion. It’s 100 million buffaloes and 185 million cattle are much smaller than the feedlot monsters of Europe and the US. The methane output of India’s 185 million cattle is about the same as that of the US’s 95 million. I have no idea how to reduce these populations, but it’s a tough problem that needs addressing.
About 30 percent of Indians are vegetarian. But they aren’t vegan, so if they get wealthier, they may tend to consume more of those foods that are currently luxury foods … dairy products … more methane.
The other 70 percent of Indians are not vegetarians but eat very little meat because India doesn’t produce much. Average per capita meat consumption in India has declined by over 20 percent during the past two decades with the population increasing faster than the meat supply. Of the 57 grams of protein per person per day in the Indian food supply, meat provides just 1.2 grams (FAOstat). Nevertheless, animal fat consumption has increased by a similar percentage. The per-capita averages conceal differential intake changes in the rich and poor so India has an obesity epidemic among the rich while the poor are still stunted. The climate however, doesn’t care much about per-capita figures and will respond to forcings from increased animal numbers, particularly ruminants.
The bottom line here is that the ERGreen just squeezes in at the 1 tonne sustainable CO2 emissions figure and allows no room for any possible expansion in animal related climate forcings. Such an expansion is highly likely if the 70 percent of 1.6 billion Indians who eat meat want more of it and the 30 percent of vegetarians haven’t switched to being vegan.
The nuclear scenario has a reasonable buffer to allow for such an expansion in animal numbers, although any expansion would violate the strict climate constraint to reduce methane and tropospheric ozone. If the Indians must keep their bovines, then someone else will have to lose theirs! All of them.
Costs and footprints
My goal in considering rough scenarios was to avoid cost arguments. First see what is feasible and then worry about the money. If a plan won’t deliver a reasonable energy output for less than the 1 tonne of CO2 per-capita constraint, then it doesn’t matter how cheap it is. Likewise if we deal with CO2 and fail to reforest and cut non-CO2 forcings, then we are similarly in trouble.
Ultimately, however, cost will be critical. Money spent on energy infrastructure is money not spent on clean water or other essentials. But how can we estimate costs of technologies 20 and 30 years into the future? I would argue that physical constraints are the best guide to such guesstimation.
Andasol 1 is worth revisiting. This has a peak power output of 50 MW. To build it required 65,150 tonnes of concrete, 20,300 tonnes of steel and 6650 tonnes of glass sitting on close to 200 hectares with its power plant and appropriate spacing between the 50 hectares of mirrors. Efficiency gains are limited by the diffuse nature of the energy it is harvesting.
In TNNUKE, I postulated 166 sites of 1000 hectares for huge nuclear power sites. Most of that 1000 hectares could be wildlife preserve. The actual reactor and turbine building footprints are just a small part of the site. Water cooling pond areas could be significant depending on design choices.
If you really want localized power sources with a small footprint, then a small scale nuclear scenario is also possible.
The SSTAR is a design for a nuclear reactor which produces double the peak power of Andasol 1 and will produce it 24×7. The reactor plus steam generator weighs 500 tonnes. Couple it to an electricity turbine and it could occupy a large suburban block, with the reactor in a 30 metre hole underground. It would run for 30 years before being recycled. It will produce the same amount of energy annually as 4.8 Andasols built with 313,000 tonnes of concrete, 97,440 tonnes of steel and 32,000 tonnes of glass occupying a thousand hectares. Build enough SSTARs and each one will be much cheaper than almost anything weighing close on half a million tonnes. It’s that simple. The Toshiba 4S is another in a string of small nuclear reactor designs.
The top image is not of Andasol, but a large PV farm in Germany. Note the double decker bus in mid frame! The bottom image is of an SSTAR reactor.
An alternative way of estimating costs is to run some sort of regression curve through historical costs in an industry. This gives you a number instead of an estimate of relativities. This is what ERGreen does.
ERGreen (p.17) assumes that the cost of electricity from Solar PV will be similar to electricity from coal by 2030. This is based on a statistic called the learning rate. A learning rate of 0.8 indicates a cost reduction of 20 percent with a doubling of installed capacity. Thus suppose you have 10GW installed and the cost is $D per GW and the learning rate of 0.8. Then after 3 doublings, with 80GW installed, the cost will halve (D*0.8*0.8*0.8=0.512D). After 3 more doublings, with 640 GW installed, it will halve again. However, according to a study by Manfred Lenzen, more recent data puts the learning rate at 0.9. This means it takes over 6 doublings to halve the price and Solar PV will never be as cheap as coal. To expect the learning rate to stay at 0.8 indefinitely is unrealistic.
ERGreen is predicated upon an unsubstantiated assumption that decentralised energy systems are more efficent and have security advantages. “Security” is a word with many meanings so I’m not quite sure how they are using it. There are certainly transmission loss reductions in decentralised systems but it isn’t obvious that they are not counterbalanced by the extra replication costs. E.g., is it always more energy efficient to build 5 power plants for 5 population centres than 1 big power plant and an appropriate distribution grid? I’d be very surprised if that question had the same answer under all configurations.
I won’t discuss the matter in detail, just provide a few motivating examples to show that this decentralisation dogma needs proper justification.
For example. How often have you received an email to the effect:
“Sorry, my hard drive died and I lost all my emails so …”.
Among people managing their own computers with email stored on their own hard drive, such events are common. Once people use a corporate email account with servers managed by people who know what they are doing, it becomes less common. As the size of corporations running servers gets bigger, you tend to get better computer system engineers designing failsafe data storage systems.
If you want robust systems, then big centralised ones work extremely well, precisely because brilliant engineers recognise the problems of having a single source of catastrophic failure and fret over the problems until they are solved.
Multiple points of failure mean multiple failures and nobody solves the problems because few people know enough at the local level to do things right. Is it any different with energy systems?
Here’s an energy based example of what I’m talking about. It concerns a solar power system in northern South Australia. In February 2011, Sarah Martin of the Adelaide Advertiser reported:
A SOLAR power station in the state’s Far North has been idle for more than a year. The station has been out of action for four of the past seven years.The Government-owned $3.7 million power plant at Umuwa, in the Anangu Pitjantjatjara Yankunytjatjara Lands, was upgraded in 2008, but was switched off just over a year later because of safety concerns.
Originally built in 2003, it also was shut down in 2005, for three years, before receiving a $1.2 million taxpayer-funded upgrade.
The report went on to quote various people finishing with the appropriate Minister, Grace Portolesi:
Ms Portolesi said the plant had been out of action because of “complex issues relating to meteorological conditions,” including electrical storms and wind-blown dust.
Wind-blown dust in outback Australia? Who could have anticipated that! Note carefully. In 2005 the system was shutdown awaiting an upgrade. Why do you shutdown a system in need of an upgrade? You might shut it down while actually doing the upgrade, but just while waiting for the money? That’s beyond me. No matter, in that same year the system won an Engineering excellence award.
A zillion little power stations is a recipe for a mass of problems like this. Each may be trivial to solve if you had the right people on the ground, but in outback Australia or villages in India, this isn’t always possible. Does anybody really think that profit oriented businesses want to pay very expensive engineers to drive or fly huge distances to fix modest sized installations serving a small community?
Distribution design decisions should be made using proper data. They should not be made on the basis of some knee jerk ideological slogan about local being best.
ERGreen only tackles one of the three areas in which we need action to avert dangerous climate change. Their plan involves plenty of fossil fuels, relies on both natural gas and oil being affordable and available and condemns India to staying well below first world standards by 2050. I’d much rather aim far higher.
Both TNNUKE or a small scale nuclear power plant scenario involving SSTAR or Toshiba 4S style technologies gives India a fighting chance of achieving a first world standard of living while allowing expansion of forest areas in line with the reforestation climate constraint. Removing biomass should substantially reduce black carbon emissions and go some way to meeting the constraint to reduce short lived climate forcings, although I have no clues on how to tackle India’s methane problem.