Thursday, February 17, 2011

Do Mediterranean Crop Yields Show Climate Stress?

Yesterday, we were discussing global crop yields (focussing on cereals as the most widespread and nutritionally significant crops).  I was pointing out that in the overall global data, there just really isn't room to see a climate change signal.

But, if we were in the early stages of climate change related agricultural problems, you might expect the signature to show up in particularly impacted regions long before it shows up in the global data.  Readers who remember the discussion in the Future of Drought series will recall that the Mediterranean tends to show up at the top of any list of particularly strongly impacted climate regions.  It was fairly dry to begin with, there are clear drying trends in recent decades, and climate models show it getting worse still in the future.  So if anywhere shows a climate change signal in the yield data, this region would be it.

For example, this figure shows the first principal component in the global data for the Palmer Drought Severity Index

If you don't feel like clicking the link above to figure out what that means, you can just think that the orange/pink/red regions are places that are tending to get drier over time (in an agriculturally significant sense).  As you can see, the western Mediterranean is pretty much all pink or red.

Borrowing this map from Google Maps,

I selected Portugal, Spain, France, Italy, Greece, Morocco, Algeria, Tunisia, and Libya as the sample here.  I excluded Egypt, Turkey, and points east since the drying trend is much more equivocal there.  I excluded the Balkans due to potential confounding effects from the political instability in those countries in recent decades.

So then I went to the UN Food and Agricultural Organization again, for country-specific yield data.  Here is the raw data (together with the world average from yesterday):

Since that's rather noisy, I applied a seven year centered moving average just to remove some of the short term weather noise and make the graph easier to read:

I think for the most part, this raises cause for concern.  The biggest and most developed agricultural economies (France, Italy, Spain, Greece) all show very clear signs of a slowdown in yield gains in the 1980s or 1990s.  The data for the less developed countries are harder to interpret, I think.  For example, in Portugal, you see clear signs of the transition to democracy in the late 1970s resulting in a rapid rate of economic development, and some catchup in agricultural yields to those of its neighbors.  It's going to be hard to see a climate signal in that line.

The most obvious other explanation is that France, Spain, Italy, and Greece are running into some kind of technological limitations in raising yields.  I think the best way of looking at that would be to compare trends with some regions that aren't encountering drought stress, and I'll try to look at that soon.

Update: An alternative explanation for the shape of the yield curves in the developed countries here is that Europe effectively banned GM crops in 1998. So one possibility is that most of the yield improvement elsewhere has been coming from crop genetic improvements, and Europe has cut itself off from those.


Paul Yarbles said...

France seems to have flat-lined for the past 15 years even though their population increased over 8% during that time! I assume they are probably the most technologically advanced nation in your sample. On the other hand, maybe France is just unwilling go 100% pedal to the metal when it comes to currently more productive industrial agricultural and this accounts for the near zero increase and not the climate. France not using available agricultural technologies wouldn't explain the slow-downs in the others in your sample of course.

Anyway, some related questions are:

How much have the increasing yields due to improving agricultural techniques and greater land use masked the effects of climate change?

How much of whatever growth there is in agricultural production is sustainable without certain resources the world seems to be running low on?

How will the agricultural techniques used over the past decades decrease or increase(!?) future yields that will not be able to use the same techniques?

Mike "Pops" Black said...

I'm wondering how much short term impact we would expect to see in a Mediterranean climate of mild wet winters and hot dry summers where much ag is already dependent on irrigation?

Or are those areas as dependent as, say, the Central Valley of CA?

kjmclark said...

OK, so first, FAO's "yield" is just trunc(production/harvested area). I checked the world and the countries you selected, and it gets the same numbers. Harvested area is ***much*** more interesting than the production or yield series, especially for the world cereals harvested area series. Harvested area declined from a peak of 727 million hectares in 1981 to 660 million hectares in 2002, then shot up to over 700 million hectares again by 2009. A good part of that is the decline in the former Soviet Union (from about 120 million hectares in 1981 to 79 million in 2002 & back to 83 million in 2009).

I think your argument so far can be summed up as whether or not climate change has had an impact, farmers have been more than able to adapt, probably through better seed, equipment, and techniques. The question would be whether climate is getting in their way, and what kind of production they could have had if they weren't having to adjust for climate changes. I think you're making a convincing enough argument that price changes aren't primarily due to climate change, but not convincing enough from these data to say that climate change isn't having an impact.

France, Spain and many other of these countries had a sizable drop in harvested acres from about 1993-1996. It shows up in their production as a significant dip as well, but gets papered over in the yield. Don't know what happened those years, but if something like that were due to a multi-year drought, it may not show up well in a 'yield on harvested acres' measurement.

Alexander Ac said...


I think the crucial question here is if the current yields/hectare are sustainable. If the soils are depleted, I do not think so, if the subsidy is fossil fuel, I do not think so...etc.


Lars-Eric Bjerke said...


The agricultural production in the EU as well as the US is also dependent on subsidiaries. One of the explanations for the curves is that EU the last years base the subsidiaries on area of farmed land and not on the production, in order to promote less intensive farming and limit the use of fertilizers.

Stuart Staniford said...

Interesting point Lars - I'll look into that.

Douglas said...

So I dug a bit into that FAO data from yesterday, and found something unexpected. Looking at tonnage of primary crops, there is a steady increase from 1990 to 2009,as Stuart's data indicates.

Then I separated out all the categories marked "fodder, forage, or silage (which accounts for a substantial chunk, about 26%, of the total).

Now the odd part: With all the talk about increased meat consumption in developing countries, I was expecting an increase in these categories, but instead there was a strong decrease in the 90s, leveling off in the early 2000s, and then continued decrease from 2005 on.

Maybe I'm reading the data wrong, but this seems quite odd to me. Maybe the "increasing meat consumption" theory is a myth?

Paleodoctor said...

Is Peak Water playing a role?

Bytesmiths said...

It's not clear from the UN link provided what crops are being plotted. Grain? Annuals only?

One anomaly of much of the world -- including the Med -- is the conversion of forest to annual crops. This can produce an increasing trend line that is, in fact, an overshoot of capacity, since annual crops generally deplete the soil, and highly-mechanized annual crops are more CO2 intensive than the forest and perennial plants they replace.

In the Med, 200-year-old olive and cork trees are being pulled up to plant dryland wheat. Not good.