It is a truth universally acknowledged – amongst my friends and relations at least – that organic agriculture is better for the planet. Environmentally-conscious consumers typically are prepared to pay a hefty premium for organic meat and vegetables, whilst baby foods are nearly all organic these days – reflecting the equally widespread belief that organic is healthier due to the absence of synthetic pesticides and fertilisers. Everyone wants the best for their young children, and the best must surely be the most natural.
These beliefs are remarkably persistent, despite strong scientific evidence which refutes them. That natural necessarily equals more safe than artificial is a fallacy. In 2009 a major study for the UK Food Standards Agency found that there was no nutritional or health benefits to organic. Indeed there is strong counter-evidence, as relatives of those who died from eating organic bean sprouts in Germany last year can attest – as I understand it, the bean sprouts likely harboured toxic e-coli bacteria passed on via animal manure added to the parent plant. This use of manure rather than synthetic fertilisers is celebrated by organic proponents, but likely caused dozens of deaths and thousands of injuries in this instance. (Imagine if the sprouts had been GMO!)
I got into an argument on Twitter about this at the end of last week, because I retweeted a document discovered by David Tribe (a.k.a. GMO Pundit) revealing the funding sources behind the proposition for mandatory labelling of GMO foods in California – Big Organic and Big Quacka (‘natural health’ types) have poured $100,000s into the campaign, far outspending the biotech and grocers campaigns who oppose the proposition. The idea that consumers have a ‘right to choose’ and therefore GMOs should all be labelled irritates me – so I tweeted that organic should be labelled with an environmental warning due to its relative land-use inefficiency. This was picked up by Simon Singh, whose tweet was called “pathetic” by the Soil Association’s president Monty Don. (Simon has now blogged about this, posing two important questions for Monty Don to answer – no response as of yet.)
Lots of organic enthusiasts tweeted back at me that I had my facts wrong, or was not considering wider issues. I asked for references, and one proponent sent me the link to a piece in Nature by Verena Seufert, Navin Ramankutty and Jonathan Foley (Nature 485, 229-232, 10 May 2012, doi:10.1038/nature11069) – which I had just tweeted myself as support for my own proposal. The paper is entitled ‘Comparing the yields of conventional and organic agriculture’ and is a meta-study looking at the relevant published literature on yield comparisons world-wide (typically evidence is cited from a single farm). Here is the major conclusion:
The average organic-to-conventional yield ratio from our meta-analysis is 0.75 (with a 95% confidence interval of 0.71 to 0.79); that is, overall, organic yields are 25% lower than conventional
This is just the overall average, however: as the figure below shows, there is considerable variability amongst different crops.
As Seufert et al write,
The performance of organic systems varies substantially across crop types and species. For example, yields of organic fruits and oilseed crops show a small (−3% and −11% respectively), but not statistically significant, difference to conventional crops, whereas organic cereals and vegetables have significantly lower yields than conventional crops (−26% and −33% respectively)
This largely seems to be because organic perennial systems do better than annuals, perhaps because of nutrient shortages in systems with a higher turnover – indeed, the authors conclude the nitrogen limitation in organic systems is probably the key factor. The study also shows great variability amongst soil types and water management strategies – for the latter organic yield is -35% compared to conventional for irrigated systems, but only -17% under rainfed conditions.
when only the most comparable conventional and organic systems are considered the yield difference is as high as 34%
Even then, I was concerned that this might be understating the case. In conventional systems nitrogen is captured from the air via synthetic fertilisers, whilst in organic systems additional nitrogen must either be grown in situ with leguminous crops – thereby forgoing a fruit or cereals harvest on the land for part of the time in rotation – or imported from elsewhere via animal manures. Were these indirect land footprint issues considered?
The question did not seem to be answered by the Nature paper, so I emailed Jonathan Foley (who I know; Jon then shared with his co-authors) to ask directly. His reply, and that of lead author Verena Seufert, is worth quoting in detail:
Jon Foley (Institute on the Environment, University of Minnesota, USA):
- The original study in Nature did not consider the *land* requirement for growing organic and conventional crops. Only the yield differences on the two different kinds of farming systems.
- Farms also have a “shadow” amount of land use, which is associated with the inputs they use — whether it’s the land used to generate nutrient inputs, biocidies, energy, etc. Just think of the footprint of land needed to make the stuff you use on the farm.
- As you already figured out, the organic systems probably have a fairly large amount of “shadow” land, particularly if they use manure from other fields as an input on their own fields. That is extremely hard to quantify right now, as we often do not know (at larger scales, especially) where the organic nutrients are coming from. Further, the organic nutrients might be a mix of cover crops (legumes), compost (from on farm wastes — but this is a mostly closed material cycle), and manure from another field or farm.
- You can imagine that, for manure based systems, that the amount of shadow land is quite large indeed. Perhaps even larger than the organic farm field itself. But for other organic nitrogen sources, it could be far smaller.
- Compared to industrial nitrogen, organic nitrogen sources are probably much more land intensive — although of course they are less energy intensive, probably better for the environment in other ways, etc.
So it is fair to say that organic systems use more land than their conventional counterparts: some of this has been quantified (the *direct* land use difference, which is roughly 20-30%, but varies by crop) and some of this has not (the *indirect* land use effects, counting where the nutrients ultimately came from).
We have been thinking about this, and hope to address this in a future paper.
Verena Seufert, Department of Geography and Global Environmental and Climate Change Center, McGill University, Montreal, Canada:
We did indeed so far only look at the direct land use, i.e. the productivity per unit area, not accounting for indirect land use to produce nutrient inputs (e.g. animal manure) for organic or conventional systems. What we did account for (at least to some degree) is the influence of non-food crop rotations in organic systems (e.g. a rotation of alfalfa in between cash crops to provide nutrients and to incorporate as green manure). Our study showed that organic systems that had a longer period of non-food crops in their rotation compared to their conventional counterparts (e.g. when an organic corn-soybean-alfalfa rotation was compared to a conventional corn-soy rotation) had a similar yield difference to conventional systems (i.e. yield ratio) as those organic systems that had a similar length of non-food crops as the conventional system (e.g. when an organic wheat-sunflower rotation was compared to a conventional wheat-sunflower rotation).
We conclude from this result that the yield of organic systems does not depend on whether they use longer non-food crop rotations than conventional systems or not. It thus appears possible to design productive organic systems without needing longer periods of non-food crops.Of course if organic systems do not implement a non-food crop rotation they need to get their nutrients from other sources, e.g. cover crops (which have no additional land costs), animal manure (which has additional land cost for growing animal feed) or compost (which could come from on-farm recycling or from municipal waste and does not necessarily require additional land).
The question how much land organic or conventional agriculture would require is an interesting question and a natural next step to the yield analysis. So far a couple of studies have discussed the issue, e.g. Badgley et al. tried to quantify the nutrient availability from leguminous cover crops, arguing that these could provide sufficient nutrient inputs; or David Connor, who criticized the Badgley analysis in a 2008 paper in Field Crops Research 106, 187-190, arguing that the main cost of organic agriculture is the additional land it requires to grow organic N inputs. But both of these papers have I think not yet given a satisfying answer to the question of organic land requirements.
But as Jon pointed out, any comparison of total organic and conventional land use also needs to take into account the land needed to produce conventional inputs. If we discuss indirect land use we need to be fair and assess this for both systems we are comparing.
So I think my case is made. Organic agriculture is significantly less efficient in land-use terms than conventional. (And the picture could be even worse than -34% in comparative terms, given that indirect land use was not taken into account in the Nature study.) On the other hand, there is no compelling scientific evidence that GM crops are in any way more dangerous than their conventional alternatives, and therefore they do not require labelling.
Having said that, starting a fight between organic enthusiasts and those who care about land use is not the point – we need to avoid zero-sum, black-and-white thinking, and take the best from both systems. Moreover, we need to bear in mind trade-offs in all the ‘planetary boundary’ areas – including water, greenhouse gases (emitted by artificial nitrogen production), eutrophication of water ecosystems due to chemical fertiliser overuse and so on. Verena Seufert puts is very well at the end of her email reply:
In any case, I think that the question of total land required to feed the world conventionally or organically also somehow risks leading into a polarized either-or debate. While in fact I don’t think there will be a one-size-fits-all solution. We try to argue in our paper that instead of looking for a ‘winner’ in the organic vs. conventional debate, we should learn from the successes and failures of both systems. Thats why we try to emphasize the results of our categorical analysis rather than the overall yield difference in our paper. It’s much more interesting to learn that organic systems have a relative yield advantage in rainfed systems or that they are often nitrogen limited than to know that overall the yield difference between organic and conventional is around 25%.
This allows us to learn that organic management and increased soil organic matter can be beneficial under rainfed conditions or that we need to improve organic nutrient management to increase organic productivity. By learning from these successes and failures of the different farming systems we can improve organic and conventional management or we can create hybrid systems that can potentially balance the benefits of organic & conventional management.
Update: 17 July 2012
The third co-author of the Nature paper discussed above, Navin Ramankutty – Canada Research Chair in Land Use and Global Environmental Change at McGill University in Canada – has also sent me an email response, reproduced below. I am very grateful that all three co-authors have contributed to this post, and also to those who have posted comments below the line.
Navin Ramankutty writes:
Here’s my 2 cents on your question, to add to what Jon & Verena have already said. As they said (and as Monty Don did in his tweet), our paper only looked at a small piece of the puzzle on the role of organic. To perform a rigorous study, using 100s of data points, we could only focus on a narrow question for which data had been painstakingly collected by various studies. But it is certainly not enough, we need to compile more data/information on a variety of other factors.
One way to think about it is that we have an agricultural production system which receives a lot of inputs such as land, water, nutrients, energy, labor, etc. It produces stuff we consume (yields per unit area), but also modifies the environment (biodiversity, water quality, carbon emissions, etc.), contributes to the economy, provides livelihoods, etc. In comparing different types of farming systems (organic, conventional, GM,..), we eventually need to look at all of these various factors, and their efficiencies. You have picked up on the indirect land use effect. But we also need to consider water-use efficiency, labor efficiency, the environmental impacts, the influence on price, livelihood impacts, health, etc.
We have really only scratched the surface so far, there’s a lot more work to be done. Our meta-analysis benefited from the various scientists around the world who spent years designing and conducting field experiments. We don’t have many such experiments yet for most of the other questions.
Update, 14 August 2012: I have received a response from Peter Melchett, policy director of the Soil Association – the primary organic certifier/lobby group in the UK. Here it is, in full with no editing:
Dear Mark Lynas,
You say your family and friends believe that ‘organic agriculture is better for the planet’ and that, like them, everyone assumes ‘that organic is healthier due to the absence of synthetic pesticides and fertilisers’. However, you know better because you know the scientific evidence, and you cite a 2009 study that found ‘there was no nutritional or health benefits to organic’. In fact your friends and family’s view is supported by more recent science, so maybe you could learn a thing or two from them.
In a major review paper published in 2011, Brandt et al (Brandt, K. C. Leifert, R. Sanderson and C. J. Seal. 2011. Agroecosystem Management and Nutritional Quality of Plant Foods: The Case of Organic Fruits and Vegetables, Critical Reviews in Plant Sciences, 30:1–2, 177–197) concluded that ‘A meta-analysis of the published comparisons of the content of secondary metabolites and vitamins in organically and conventionally produced fruits and vegetables showed that in organic produce the content of secondary metabolites is 12% higher than in corresponding conventional samples’ (a statistically significant difference). The argument had been over whether differences between organic and non-organic (conventional) food are significant – the fact that there are differences, and the scientific explanation for these differences, are well known. As Brandt et al state: ‘Organic and conventional crop management systems differ in terms of the fertilisers and plant protection methods used. Ecological and agronomic research on the effect of fertilization on plant composition shows that increasing availability of plant available nitrogen reduces the accumulation of defense-related secondary metabolites and vitamin C, while the contents of secondary metabolites such as carotenes that are not involved in defense against diseases and pests may increase. In relation to human health, increased intake of fruits and vegetables is linked to reduced risk of cancer and cardiovascular disease. This benefit may be primarily due to their content of defense-related secondary metabolites, since most other constituents of fruits and vegetables either are not unique to these foods or have been shown to not provide health benefits when the intake is increased’.
Your discussion of the land needed by organic farming is confused by two false assumptions. First, you seem to assume that the land on organic farms growing legumes to fix Nitrogen for subsequent crops (using the sun’s energy rather than fossil fuels to manufacture Nitrogen) is only fixing Nitrogen, Wrong. Fields of alfalfa (Lucerne) or red clover fix Nitrogen and feed livestock, fields of peas and beans fix Nitrogen and provide human or animal food. Second, you assume that organic systems must produce exactly the same quantity and type of food (and crucially animal feed) as non-organic systems. Wrong again. Intensive and industrial chicken, pig, beef and dairy systems are not permitted under organic standards, nor possible under organic systems, as they generally depend on widespread, prophylactic use of antibiotics, and on systems that have lower animal welfare and higher incidence of disease than is acceptable in organic farming. These animal systems consume huge quantities of grains and protein, otherwise consumable by people, and convert it very inefficiently into meat or dairy products. Half or more of our wheat crop in the UK goes for animal, not human consumption, so even with a 30% lower yield, more not less grains could be available for people to eat. True, our diets would change, to a far healthier balance of more fruit and vegetables, and less but better quality meat and dairy products. Your family and friends may be ahead of you here too, if they think that meat and milk from all or mainly grass fed beef, sheep and dairy cows (as is the case with organic systems) is healthier. A report just published by Compassion in World Farming says just that (http://www.ciwf.org.uk/yourfood/nutrition.aspx). They conclude ‘The industrial farming model is unsustainable and relentless in its exploitation of animals, land, energy and water. An urgent move from intensive to higher welfare farming is required to improve animal welfare, as well as to reduce environmental pollution and waste. Higher welfare farming produces healthier meat, eggs and milk than similar products from factory farms. Switching from factory farmed animal products to higher welfare meats, eggs and milk could contribute to improved consumer health, especially if consumed in moderation within a balanced diet, rich in fruit and vegetables.’
The ‘truth universally acknowledged’ that you dismissed is, I am happy to say, generally true after all. This is for publication, and I hope you will add my comments to the others you have been sent