Livestock and Climate Change … Status update

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.

The United Nations report Livestock’s Long Shadow (LLS) came out in 2006 with an estimate that 18 percent of anthropogenic greenhouse gas emissions are attributable to livestock. If you exclude deforestation emissions, then the number drops to 14 percent. Some 95 percent of these emissions are direct emissions of methane or nitrous oxide with just 5 percent being from associated energy use as shown in the table which is a contraction of a table from LLS.

The reason the energy emissions are so small is that almost no processing is included. The energy associated with the refrigerated meat chain from abattoir to consumer, cooking costs, energy to build the trucks that carry the animals and later the meat etc. None of this was included.

A couple of years after the LLS report, BNC published a piece by myself, Barry Brook and Peter Singer which showed that Australia’s most powerful climate forcing was livestock and not coal fired power stations. The demonstration relied on the difference between radiative forcing, a concept roughly equivalent to warming and used by climate scientists, and the less accurate concept of carbon dioxide equivalent used in the Kyoto protocol.

Two years later and it’s time for an update. The NOAA chart shows that methane levels are rising again after a flat spot during the early 2000s, and the biggest single source of anthropogenic methane is livestock.

This update will look at implications of livestock growth predictions, the Goodland/Anhang photosynthesis imbalance theory, industry attempts to show beef is carbon friendly, and ruminant methane reduction research. I’d like to also cover black carbon and ozone issues, but that will have to wait. I have written a small section to explain why black carbon and ozone are really, really important, but the detail will have to wait.

Pelletier and Tyedmers PNAS paper

August 2010 saw the publication in the US Proceedings of the National Academy of Science (PNAS) of a paper looking at the implications of various livestock population scenarios out to 2050. It looked at the impact of the scenarios on postulated safe operating limits for three things:

1. human production of greenhouse gases,
2. our use of planetary biomass
3. and our production of reactive nitrogen.

These are all topics which will be familiar to some degree from earlier BNC posts (here for biomass, and here for nitrogen).

Currently the greenhouse gas emissions (CO2eq) for every person on the planet average about 4.8 tonnes a year. Australians are among the worst with emissions more than 5 times this. Nuclear powered France comes in at a little less than double the average. Typically, developing countries are below average emitters. For example Bangladesh emits about 0.9 tonnes per person per year. Sub-Saharan Africa comes in closer to average at around 4.5 with half that coming from land-use change … burning the landscape.

By 2050, the global average needs to be about 1. That’s right, just 1 tonne of CO2eq for each of the 8.9 billion people that P&T expect on the planet. That’s no change for Bangladesh and a 25 fold reduction for us.

P&T consider several scenarios:

1. No change in livestock populations or emissions. Since 95 percent of livestock emissions are direct rather than from energy use, a rebuilding of our energy infrastructures can’t make a dent in livestock emissions. So livestock, assuming no growth at all, will occupy about half of that 8.9 billion tonnes of CO2eq, leaving half a tonne per annum for everything else. Because this scenario postulates no increase in livestock but a growing population, the average consumption of meat per person must fall.
2. Growth in livestock as per FAO predictions. In this case livestock emissions will be 71 percent of each person’s allocated tonne of CO2eq.
3. The vegan. This is one of two endpoint scenarios. I’ve called it a vegan scenario, P&T call it their soy scenario because they assumed that all protein comes from soy and none from animal products. Soy protein to go for 8.9 billion can be generated for a greenhouse gas cost of just 0.1 tonne of CO2eq per person. Of course, it’s not necessary to eat soy to get protein. Protein is actually tough to avoid and humans don’t need much compared with, for example, goats.
4. Extreme meat. In this scenario P&T look at what happens when animal products supply all of a person’s protein requirements. They view this as the opposite endpoint of the vegan case, but it’s already happening in many first world countries. Australians, for example, average 73 grams of protein a day from animal products, which means all the rest of the food they eat contributes superfluous protein. In such a scenario, 92 percent of your tonne of CO2eq will come from the animal products in your diet. There won’t be much room left in your allocated tonne for anything else. As a footnote, I should say that P&T are using very low emission factors for their beef, so I’d say many Australians are already well over their tonne just from their meat alone.
5. Chicken substitution. This interesting scenario postulates that all livestock increases above the year 2000 level required to meet the demands of the extra 2050 population will come from chicken which has the lowest greenhouse gas footprint of any meat. Under this scenario, livestock consumes 62 percent of a person’s 1 tonne allocation.

An LLS reply …

The P&T paper prompted a response from FAO’s LLS authors Steinfeld and Gerber who argued that changes are already in place to make meat more environmentally friendly … by making it more animal abusive with a shift to pigs and chickens. It was a qualitative reply to a set of quantitative predictions.

Steinfeld’s use of the word “shift” is vague and a little misleading. The global cattle population has increased in each of the past four decades. True, it hasn’t grown as fast as pigs and chickens, but this doesn’t mean that there has been any global substitution of pigs and chickens for cattle. Pigs have long been the dominant meat globally with pig and chicken meat production combined being triple the size of beef. Without a reduction in cattle, and there is no evidence of this globally, it’s difficult to envisage any significant decrease in overall livestock impacts in the 3 areas P&T consider. Steinfeld’s response is really no more than hand-waving.

Is “abusive” a defensible description of the pig and chicken industries? Absolutely and it has solid scientific support. A major study looking at chickens in the UK found that only a few percent can walk normally. The researchers looked at 51,000 birds in 206 flocks. Apart from the few percent walking normally, the rest suffered varying degrees of crippling because their muscle growth exceeds that of their skeletal system. Scientists have also shown that the birds are in pain. They will quickly learn to select drinking water laced with analgesic when given a choice. This is a heavily globalised industry so there are good reasons to believe this study is representative. The situation with pigs is little better … typically 60 percent of breeding sows are replaced each year with lameness being a common reason.

P&T on Biomass

P&T use an estimate of the maximum sustainable limit on human biomass use by 2050 of about 9.7 billion tonnes annually which is consistent with a a land use stabilisation scenario of the IPCC. Livestock already use about 7 billion tonnes of biomass. But far more critical is which 7 billion. Almost 3 billion of those 7 are crop residues. This increases erosion and damages soil structure, ultimately reducing the productive capacity of plants which provides the bulk (83 percent) of human food. In wealthy countries we compensate with more fertiliser, but in poorer countries low productivity and hunger are the result. It’s tough to decide which of the livestock industries’ devastating impacts on animals, people and the planet is worse … animal suffering, species extinction, bowel cancer or methane’s contribution to climate change. But exacerbation of hunger and poverty in poor countries by the apparently innocuous process of eating crop residues possibly trumps all four.

Worldwatch and mega fauna

In 2009 Worldwatch published a paper by Robert Goodland and Jeff Anhang which put livestock emissions at 51 percent of anthropogenic emissions. I’ve mentioned this before. A large part of the difference between the 18 percent and 51 percent figures comes down to the treatment of livestock respiration … the carbon dioxide breathed out by animals (and us). If livestock are gradually reducing planetary plant capital, then at least some of their respiratory CO2 should counted. Current IPCC accounting rules are predicated on an assumption that respired carbon is matched (or exceeded) by plant growth.

Measuring such matters is difficult at best, but an interesting study came out in 2010 which throws light on the issue from a different angle. Put simply, the study estimated that humans and their domesticated biomass black holes … livestock … are consuming about 6 times more plant matter than the long extinct mega-herbivores in their hey-day back in the Pleistocene. This makes it clear just how unprecedented our current impact on the planet is, regardless of how the detail pans out.

Carbon neutral cattle stations

The last few years have seen a number of studies trying to show that various grazing systems are carbon friendly. This one from Mark Liebig at the US Department of Agriculture is an example and contains references to quite a few more. Here’s an Australian one from the Queensland Department of Primary Industries and Fisheries.

I’ve dealt with these issues in some detail before. But here’s an executive summary.

The first fundamental problem is in how the system boundary is defined. Consider methane. We used to have a planet with a rough balance between the methane sources and sinks. The terrestrial sources were things like wetlands and there were also terrestrial methane sinks like forests and some grasslands some of the time. The terrestrial sinks were always significant but were about a third the size of the atmospheric sinks. Regardless of the relative size of the sinks, it seems pretty obvious that you just can’t run around finding methane sinks and sticking cattle on them to make the area “methane neutral”. Ditto carbon in general. This is as silly as drawing a system boundary around a rechargeable plug-in electric vehicle and calling it pollution free.

The other standout issue is that of alternatives. A long term study in South East Spain illustrates the kinds of gains in carbon sequestration that can be made in suitable areas when pastures are converted to woodland.

These are not subtle points and when groups of well trained and intelligent scientists do substantial studies involving plenty of time, money and equipment and forget to mention these issues in the subsequent paper, then you have to wonder if the work is politically rather than scientifically motivated. In the case of the Queensland study mentioned above, Principal Scientist Gerard Bisshop resigned from the Queensland Government in protest.

Is Queenland a developing country hunting CDM credits?

It’s interesting to compare the Queenland study with appropriate United Nations Framework Convention on Climate Change Clean Development Mechanism (CDM) rules. These are the rules which developing countries need to follow in order to get tradeable carbon offset credits for reforestation work on agricultural or pastoral land. The rules are clear and require that the land is degraded and degrading with crops or livestock numbers having declined by at least 30 percent or more during the five years preceding the project. i.e., an ongoing cattle (or wheat) property can’t get credits for woody regrowth because it happened to decide not to bulldoze one or more paddocks. But the above linked Queensland DPI paper claims credit for precisely this kind of non-activity. It wouldn’t qualify for actual tradeable credits for a developing country under CDM and it shouldn’t appear in dodgy pro-industry advocacy papers by Queensland Government scientists.

And the CFI? … Carbon Farming Initiative

The Australian Government has recently released its Carbon Farming Initiative (CFI) with a consultation paper and draft legislation for public comment. Despite the draft legislation being 331 pages of very precise legal language, the real substance is still being developed in Methodology documents being developed “with industry”. I’ll try and describe its key features as the substance emerges in a future post.

Less methane per moo

I feel compelled to give an update on the scientific progress to feed or engineer ruminants to produce less methane. Stories about breakthroughs in this area appear about as regularly as brain dead suggestions that eating more kangaroos will save the planet. Here’s a research result that might prompt just such a breakthrough story:

Many members of a series of … [compounds] were potent inhibitors of methanogenesis by rumen contents in vitro. The most potent compound inhibited methane production by 70% or more at a concentration of 1 microgram/ml (approximately 2.5 mumol/l).

Sound good? A massive 70 percent reduction in methane …

Yes it’s in vitro, but that’s a minor issue, think of the 70 percent!

This quote comes from a piece of research almost 30 years ago. Livestock farmers have long considered methane production from an animal as wasted energy and scientists have been working on a solution to this “problem” for a very long time. I traced papers back to the 1980s before losing patience. Apart from a huge wad of basic science about rumen function, I found papers on all kinds of things people have added to sheep or cattle feed to reduce methane production. The list included: mustard oil, horseradish oil, glucosinolate, cashew nut oil, linseed oil, coconut oil, krabok oil, palm fronds, soy oil, whole soybeans, antibiotics (various), fumaric acid, canola oil, copra meal, monolaurin, bromochloromethane, monensin, Yucca schidigera, Quillaja saponaria … and these are just the things with common names rather than chemical formulas. Monensin deserves a little footnote … it’s from a group of chemicals called ionophores, some of which have antibiotic properties. Meat companies are using ionophores as growth promoters while still using “raised without antibiotics” labels. One news story about huge US chicken producer Tysons doing this is here, but, curiously, the original Associated Press source seems to have been pulled from the website.

Many of the compounds reduce methane emissions but none seems to have made it into active use for all kinds of practical reasons. Seriously, who would waste cashews on cattle? A further persistant problem is that the organism populations in rumens adapt and can make any impacts short lived. Antibiotics work to some degree but create an evolutionary arms race in the rumen which will eventually see resistant organisms emerge. We don’t need any more antibiotic resistant bugs on the planet.

Apart from adding things to ruminant feed, the other approach under research is artificial selection. Some animals eat less per kilogram of growth and thereby generate less methane. Breed from these animals and hopefully the offspring will be similar. It’s rather trickier than it sounds and the gains will be small. How small?

Genetic approaches

Back in 2006 an Australian study did modelling suggesting that breeding for low methane production might yield a cumulative reduction in methane emissions in the Australian herd by 3.1 percent by 2025. This is hardly earth shattering, but seems to be enough to secure funding for this approach.

Since then various genes have been isolated that are associated with feed use efficiency (which is linked to methane production) with one study finding that these genes account for about 6.9 percent of the variability and a 2010 study found 150 gene factors (SNPs for those who know about such things) that account for 36 percent of the variability. How do you breed herds with these factors? That’s a good question.

As I was reading these papers it wasn’t always clear what was happening over time in the same animal. A group of Swiss researchers seemed to have the same interest. They examined animals from 3 popular dairy breeds over 41 weeks and found plenty of variability at different times but very consistent long term averages. They concluded:

The apparent lack of persistence of individual animal differences in methane yields suggests that genetic determination of this trait is of minor importance in dairy cows.

Measuring methane

A recent news article raised the issue of the accuracy of measurements of methane from cattle. The researcher involved was Canadian PhD student Jennifer Ellis. She has published a couple of papers, one of which I have, but she hasn’t (yet) responded to email requests, so I don’t have the other. The news article states:

Statistics have been giving us a bum steer when they state how much cattle methane emissions contribute to global warming, a new study shows. That’s because mathematical equations used to predict cows’ methane emissions are inaccurate and don’t take into account factors such as dietary changes, said Jennifer Ellis, lead author of the study and a PhD student at the University of Guelph.

The paper of Ellis’s that I do have investigates the accuracy of various equations used to predict methane from cattle. The first step is to look at the accuracy of the various methods of measuring methane from actual animals … there are three. Ellis and her 12 co-authors are happy that all three methods give the same results. So there isn’t a problem measuring methane emissions from cattle. But of course, you can’t measure emissions from every animal. You need to use the measurements on some animals to come up with equations which predict national herd emissions on the basis of livestock genetics and lifestyle. Is it true, as the article implies, that after 30 years of work, nobody has fitted any half reasonable models to the data?

In a 2008 paper, Ellis’s Ph.D supervisor, Associate Professor Ermias Kebreab calculated that current IPCC equations overestimate the emissions from US dairy cattle by about 12.5 percent and underestimate the emissions from US feedlot beef animals by about 9 percent. Since the ratio of beef to dairy cattle is probably about 10 to 1 then, the equations will give a net underestimate of total cattle emissions in any place with US feedlot conditions … and cattle emissions dominate livestock emissions. So the equations may be off by a few percent, but hardly enough to warrant the bum steer claim, especially since they look to be underestimating cattle methane.

But getting back to Ellis and her paper that I do have. She tested a swag of different equations proposed by many different researchers against a database of actual measurements from animals. Here’s a couple of her pictures to give you an indication of the findings. Each little diamond is a single animal. The graph below is from one equation and you can see from the predominance of diamonds below the observed=predicted line, that this equation underestimates emissions.

The equation used in the next figure on the right looks reasonably balanced.

It’s superficially clear that the range of emissions from various animals is large and not well captured by the available equations which are all pretty simple linear combinations of dietary components that researchers suspect are important. The equations were probably popped out of some linear regression software. But all you really need for large scale use is reasonable symmetry about the observed=predicted line. On the other hand for individual animals, the equations are frequently off by large amounts … factors of 2 or more. This indicates that it’s likely that nobody really has a clue about what causes the extreme variability between animals and that the dietary components featuring in the equations may not be among the causal factors driving that variability. The list of causal factors included in the equations includes starch, cellulose, forage, nonfiber carbohydrate, fat, dry matter intake and lignin. These are all common agricultural science suspects. There isn’t a single vitamin in the list or any of the other myriad suspects that feature in human nutritional work. My non-expert judgement is that this research, despite being decades old, is still in the very early stages. Make no mistake, I’m not maligning the scientists. This is very tough science, a ruminant is a wondrously complex creature and the research effort is tiny compared to, for example research into the diseases caused by red meat … like bowel cancer … or heart disease.

Industry responses

The livestock industry has countered LLS with all manner of rubbish. For example the NSW Farmer’s Association ran a line which confused US figures with global figures and left out nitrous oxide emissions altogether. Much of this confusion comes from one Frank Mitloehner who has been trumpetting a 3 percent figure at any journalist who will listen for some time now. While I’m sure Associate Professor Mitloehner knows what he’s talking about, journalists and NSW Farmer’s Association people tend to be easily confused.

Livestock methane emissions in the US are indeed about 3 percent of US emissions and are well below the 14 percent global average figure of LLS … about half of which was nitrous oxide as we saw in the first table.

It is common for US authors to confuse the US with the entire planet, but US livestock methane emissions are below average for a number of unsurprising reasons:

1. US cattle are predominantly grain fed (producing less methane) than cattle on pasture.
2. The US ratio of cattle to people is about 1 to 3 compared to 1 to 1 in Brazil and well above 1 to 1 in Australia.
3. US methane emissions from garbage are huge. In Australia, by comparison, and we are not noted for frugality, our garbage methane emissions are 1/6th of our livestock emissions. In the US, methane from garbage exceeds livestock methane.
4. The US imports about 10 percent (net) of its beef and almost all of its sheep meat … not that they eat much. So the emissions from that beef don’t appear in US figures.
5. Lastly, US advertisers and fast food chains may portray their home country as a hamburger culture, but Americans actually eat twice as much chicken as beef and almost no sheep meat at all. Australian ruminant meat intake is double that of the US.

It’s not so much that US livestock emissions are small, but that they are swamped by other profligate consumption emissions.

Tropospheric ozone and black carbon

Some readers will have seen the Supreme Master TV ads (on SBS in Australia) which proclaim that “If we cut methane emissions now, the worst climate change effects could go in a decade.”. The Supreme Master organisation has some of the look and feel of a religious sect to an athiest like me, but its backbone seems to be a large group of dedicated, caring and sharing volunteers who are refreshingly easy to deal with. I haven’t seen data to back their claim but here’s a claim from a paper co-authored by one of the world’s top climate scientists. Veerabhadran Ramanathan is Distinguished Professor of Climate and Atmospheric Sciences at the Scripps Institute of Oceanography at the University of California. Here’s the quote:

Fully applying existing emissions-control technologies could cut black carbon emissions by about 50 percent. And that would be enough to offset the warming effects of one to two decades worth of carbon dioxide emissions. Reducing the human-caused ozone in the lower atmosphere by about 50 percent, which could be possible through existing technologies, would offset about another decade’s worth. Within weeks, the heating effect of black carbon would lessen; within months, so, too, would the greenhouse effect of ozone. Within ten years, the earth’s overall warming trend would slow down, as would the retreat of sea ice and glaciers.

The detail will have to wait, but ending animal agriculture would be a big contribution to the reductions of both black carbon and ozone because the main cause of rising tropospheric ozone is rising methane and livestock is also a potent producer of black carbon via deforestation. The bottom line is that the Supreme Master claim is definitely plausible if Ramanathan’s modelling is accurate.

Concluding remarks

Livestock’s Long Shadow marks a watershed with livestock industry advocates giving solid numbers for many of livestock’s worst environmental impacts. The numbers were conservative as the subsequent Goodland/Anhang figures indicated. The report was also solid in its demonstration that more meat and dairy products requires more intensification if environmental impacts are to be minimised. Intensification is the industry’s euphemism for confinement, chemicals, crippling and suffering. The only way for meat and dairy to reduce its costs on the environment is to increase its cost on animals. For meat to become greener while continuing to be produced in vast quantities requires that it become crueller.

A population’s various impacts on the planet are largely dominated by what it eats. As people get richer, they can swamp this impact with other things. Private jets and a couple of Hummers will blow any environmental footprint budget. But for most of the planet’s 9 billion people in 2050, it will be food choices that dominate their impacts on the planet and it’s pretty clear that change is required. More of the same is both undesirable for most of the creatures involved and probably impossible. We need a substantive dietary transformation. The changes will be small in Bangladesh but large in Australia and other extreme-meat countries. As omnivores, unspecialised eaters, we have choices. We can trash the planet with our food choices, we can allow pigs, chickens and cattle to outbid the poor for food, or not.

Appendix … P&T figure with food supply characterists

P&T contained some fancy graphics which confused me a little … so I’ve tried to redraw, in the figure above, some of the data in what I think is a clearer form and added a heap detail about the global food supply … sticking to the law of conservation of confusion which allows the additon of extra data while keeping confusion constant :) I haven’t (yet) included nitrogen data. The data in the figure are generally the latest I could find, but the percentages of meat from different systems are ratios from LLS (circa 2000) and applied to the 268 million total meat production in 2007. The biomass data come from Kraussman. If you look at the red lines from “Meat” to “2796 Calories”, you will see it is about 8 percent of global calories. The other 8 percent, not highlighted with any lines, is dairy and eggs.