Building the SDGs on dodgy premises

by Luigi Guarino on May 29, 2017

A couple of things on the SDGs today for you to wade through.

First, from FAO, there’s “FAO and the SDGs — Indicators: Measuring up to the 2030 Agenda for Sustainable Development.” There’s a lot of sensible stuff in there on how to measure progress towards the SDG targets, goal by goal. I’m afraid, however, it lost me with the little sidebar I reproduce here. No, I didn’t know that. Mainly because that first bit is not true.

And then there’s an IIED Briefing on SDG2 in particular — that’s the hunger one. Surely they’ll stay away from dodgy numbers. Nope.

Genetic diversity reduces risk in agricultural systems and allows farmers to adapt to a changing environment, yet an estimated 75 per cent of crop diversity was lost between 1900 and 2000 with local varieties replaced by modern ones.

The reference? FAO’s training manual on Building on Gender, Agrobiodiversity and Local Knowledge. Where there is this:

But also this:

More than 90 percent of crop varieties have disappeared from farmers’ fields; half of the breeds of many domestic animals have been lost. In fisheries, all the world’s 17 main fishing grounds are now being fished at or above their sustainable limits, with many fish populations effectively becoming extinct. Loss of forest cover, coastal wetlands, other ‘wild’ uncultivated areas, and the destruction of the aquatic environment exacerbate the genetic erosion of agrobiodiversity.

Which is a bit confusing. The reference for that box? This. From 1999.

Oh well.

{ 22 comments… read them below or add one }

Cary May 30, 2017 at 2:27 am

If you follow the threads provided in this and earlier posts, you come to the 75% figure in Shattering, a book by Pat Mooney and me. Let me take the story one step back. The 75% figure in Shattering originated with a prediction made to us by Erna Bennett (of FAO) that if the Common Catalog came into effect in Europe with many varieties being de-listed for sale, the result would eventually be a loss of three-quarters of the varieties there. Several times during my own later tenure at FAO, the 75% figure was used more “expansively” as in the quote in the box. Some instances even originated in the office of the Director General at FAO. I recall, as a senior technical officer, contacting that office to explain the inaccuracy of this use, and indeed the impossibility (as you note) of having any precise estimate. Alas, it was no use. The figure continued to be used. It was true because the DG said it and that meant it was true. (I was told just that once, even though I knew the DG was just signing on to something someone on his staff had written.) Now, as you show in Box 5 (above), one can even find references to the “fact” that 75% of the diversity (not just the varieties) has been lost. Globally!

FAO has long needed to correct the record. Yes, it is like something generated from a game of “Chinese Whispers” as mentioned in one of your earlier posts. But hey, maybe it’s become a meme by now! What would Richard Dawkins have to say? I’m guessing he would wish you good luck in reeling this back toward reality, whatever that is.


Tim May 31, 2017 at 6:13 pm

Then there is this now famous infographic from National Geographic. This is based off a study that compared a USDA census of US varieties in the early 1903 with what was in the USDA collection in 1983. The infographic correctly explains this, but the summary of the infographic incorrectly interprets and then generalizes these results to: “The survey…found that about 93 percent of the varieties had gone extinct…” That statement, that 93% of crop diversity has gone extinct is repeated over and over again.


Cary June 1, 2017 at 12:02 am

I guess I’m missing something. The summary refers to varieties, not diversity, lost. So, it seems consistent.


Luigi Guarino June 5, 2017 at 12:07 pm

Cary, what is the relationship, if any, between your study in Shattering, and the National Geographic study?


Cary June 11, 2017 at 9:43 am

I believe the latter was based on the former.


Tim June 5, 2017 at 4:08 pm

It wasn’t that they were lost, just that they were not present in the USDA collection at that time.


Cary June 11, 2017 at 9:56 am

I suspect that in most cases, “lost” and “not present” were/are just about the same. Of course, one can never prove the negative, so it’s impossible to prove that something that’s not in the collection no longer exists. And, in fact, I know of a few apple varieties, for instance, that were not in collections in the ’80s that have seen been identified. But, I’m not aware of substantial numbers of lost “American” heirlooms entering the USDA collection in recent years – at least nothing on the scale that would dramatically change the statistics – so I presume that documented varieties from the old days that are not in the collections now are for the most part “functionally extinct.” Even if we were to stipulate that 10% (that’s generous) of the varieties thought to be lost still exist somewhere, I’m not sure that’s a difference that makes a difference. The fact would remain that many varieties are no longer with us. Of course, I’m talking about the loss of varieties, not necessarily the loss of genetic diversity. The two are not synonymous. That’s a difference that does make a difference.


Colin June 2, 2017 at 11:33 pm

People like and hold onto numbers because they provide a sense of validity. We in this tiny insignificant community of crop diversity folk know these numbers to be incorrect, perhaps wildly off the mark, and that its essentially impossible to come up with a true single global number.

But, having something to say about this important issue, and making it widely known to the larger community is important. Is it too idealistic to imagine our community collaboratively coming up with a better global one sentence statement, and offering it to the world? Perhaps with no numbers at all? Perhaps only that “in the past century, the number of crops and their varieties, and the genetic diversity within them, has declined as a general trend in farmers’ fields around the world”.

I realize that with putting this quote forth in this venue, I’m about to be slaughtered by the exactoids in the crowd. But I’d really like someone to show me science that disproves the above statement. Case studies of genetic diversity levels of improved varieties within farmer’s fields after the Green Revolution are disqualified by default of not being able to deal with the real time period of greatest lost.


Colin June 2, 2017 at 11:39 pm

Or, to put it more succinctly, stop complaining and propose something better!


Tim June 5, 2017 at 4:16 pm

Colin, I think we need to start with ideal then and work towards what can be worked out and if, by the time we have worked out what CAN be measured, we are still getting data that meaningfully answers the question we intended to answer. I think that the ideal is to compare the genetic diversity of original genepools that contributed to the crops we are growing and to see how much we have in our genebanks (ideally over time, as this will change and we are after all talking about TRENDS) and how much we have in our fields (again, ideally over time). I realize that is a monumental, and likely impossible data set to construct. For me the question is, how close can we get to that dataset and will it still be valid? Variety names have been used as a proxy for genetic diversity in the public eye for entirely too long. New varieties are added every year, but the addition of an “Improved” cultivar that is simply a more inbred version of a previously released cultivar with a few small new tricks is not a large net gain of diversity, right? And yet, to translate this work into easily digestible bits that we can use to communicate with real people often means talking about varieties because people understand what that is.


Colin June 5, 2017 at 6:45 pm

Hi Tim and all — I very much agree that new studies specifically directed at generating real tangible data on the genetic erosion issue would super useful (at least to our little community). For me, such studies would use the best of cutting edge genetic tools to assess what can be assessed in comparing past to present. As Tim notes, its impossible to fully assess the past. Actually its impossible to fully assess the present as well. But we could do some case studies preferably across a number of different crops (a couple of cereals, legumes, vegetables, fruits, oil crops). Livestock too. I also think its really important to think through whether we are assessing genetic diversity in fields, or overall genetic diversity (wherever it is, including in genebanks). I’d prefer those get pulled apart and treated differently; in any case the basic idea that spatial diversity (traditional systems) has been replaced by temporal diversity (modern systems, with new varieties every few years) needs to be grappled with. Can they be compared at all? If so, if a modern system is less diverse at a single point in time, is that an issue? Where’s Mark van de Wouw? I think we need his thinking here.

Just a note — here at the USDA national seedbank numerous colleagues are interested in this question, perhaps specifically by revisiting those earlier studies about the USDA genebank system, but with a molecular approach. We’ve been keeping our eyes peeled and ears open for PhD students or postdocs that might want to work on this, and any funding opportunities that we might be able to apply to. Would you send anything our way? Thanks.


Dave Wood June 5, 2017 at 4:54 pm

“The Seed Vault has the capacity to store 4.5 million varieties of crops”. Do they actually mean samples? This `variety’ terminology seems to be more of a problems than the FAO 75% figure. Or is that “talking about varieties because people understand what that is”? The interesting thing is that varieties was what the system was all about in the days of Fairchild and the rest of the USDA collectors. They would make specific steamboat journeys just to get the very best variety of a crop from a country – dates, lemons, just name it. Worked a treat.


Paul Heisey June 5, 2017 at 7:46 pm

Well, there’s this:

Hyten, D.L. et al. (2006) “Impacts of genetic bottlenecks on soybean genome diversity.” PNAS 103(45):16666-16671.

Does anyone know of other similar detailed studies at the molecular level?


Luigi Guarino June 6, 2017 at 1:26 pm

Interesting, of course, but this doesn’t say whether soybean diversity has disappeared from farmers’ fields of late, for whatever reasons.


Clem June 6, 2017 at 11:16 pm

Thanks Paul for pointing to the Hyten et al. paper.

I can’t speak to whether soybean diversity has disappeared from farmers’ fields on a global level, but I would suggest that phenotypic diversity in the North American soy crop has actually increased. The loss in diversity measured in the Hyten paper compares ancestors to present day cultivars. Domestication alone accounts for the vast majority of the “loss”. But this framing of loss is a bit misplaced in my opinion. In the process of selection (whether by nature or by human agents) some alleles are selected and others are rejected. The rejected alleles are only lost if the germplasm harboring them is permanently discarded. The study mentioned here demonstrates that the other sources of alleles have not been discarded.

Preservation of diversity by farmers makes for a bucolic image, but hardly a commercially satisfying one. We can cuss and discuss the diversity issue among ourselves or in a wider forum till blue. But until there is either a political will with funding attached, or some serious failing in the commercial crop there likely won’t be much departure from the current state of crop improvement.

The current implosion of commercial seed sources could be held up as a potentially disastrous development. But the shrinking market returns to R&D that is driving the merger mania is exactly the sort of evidence I’m speaking to. Want more diversity? Someone has to pay for it.


Jay Bost June 6, 2017 at 9:28 pm

Something I frequently wonder about in the context of this discussion is diversity measured by varieties versus actual genetic diversity. I have an unfounded hunch that in say the case of tomatoes, while number of named heirloom varieites is lower now than the beginning of 20th century, that actual genetic diversity – due to introgression of genetic material from CWRs is higher now than in early 20th century. Heck, maybe the highest since domestication. So have we really “lost” anything important in this case? Maybe some unique alleles for taste or color have been lost, but the overall genetic diversity of the crop has actually increased?
For better or worse, as Colin mentions, the message to the public can only be so complex and nuanced, but crop conservation folks have to believe in what they are saying too.
We do either need more substaniated numbers or a general statement, lest we become purveyors of fake news


Jay Bost June 6, 2017 at 9:47 pm

Some interesting insights in here.

“Public germplasm are potential allele mining sources for crop improvement as illustrated by previous authors who sampled among US seed banks 30 accessions from the five continents. The study confirmed that history of crossing with wild tomato species and distribution among different environments across the world has spread allelic variation (Labate, Sheffer et al. 2011).”

“The recent discoveries of the molecular events shaping tomato fruit indicate that the germplasm is frequently more diverse phenotypically than the wild related germplasm but not necessarily showing a similar pattern at the molecular level. “The irony of all this,” says Steve Tanksley (geneticist at Cornell University, and precursor of all these studies) “is all that diversity of heirlooms can be accounted for by a handful of genes. There are probably no more than 10 mutant genes that create the diversity of heirlooms you see” (Borrell 2009).


Jacob van Etten June 17, 2017 at 9:44 am

An analysis of genetic diversity that will probably always be deficient because there are unknowable aspects. But one knowable aspect that is still underexploited in diversity analyses is that geographic diversity distribution patterns have some predictable properties because human seed dispersal is not random. A lot is still possible to estimate crop genetic diversity by building spatially-explicit models of crop dispersal (e.g. van Etten & Hijmans 2010).

But we should not fall into the mental trap that we should know all about the past to understand the present and deal with the future. We should not save the 0ld paper archives while forgetting to archive the Internet. We need to direct the main effort at building systems that monitor diversity in the field, linked to human processes that actively manage this diversity, such as breeding, variety introduction, seed sales, community seed banks, seed savers networks, etc.

“75% lost” is perhaps not a meme but a myth. It is not a neutral piece of information that rides the waves of noise. It is a strategic way to grab the attention. I wonder if we should be generating a better number for the myth, be trying to dismantle it (= the loss is in fact a gain). Science myths don’t die. We should get better at communicating the complex issues that are at stake.


Susan June 27, 2017 at 12:30 am

On another figure, the need to double production by 2050 to feed the planet. Has anyone done the same excavation of the origin of figure, the assumptions underlying the claim etc.? I may be missing well-known critiques, so point me in the right direction.


Susan June 27, 2017 at 12:35 am

Something that answers my question:
Tomlinson, I., Doubling food production to feed the 9 billion: A critical perspective on a key discourse of foodsecurity in the UK, Journal of Rural Studies (2011), doi:10.1016/j.jrurstud.2011.09.001


Luigi Guarino June 27, 2017 at 8:34 am

Well we talked about this a while back.


Susan June 28, 2017 at 9:19 pm



Leave a Comment