Tuesday, November 23, 2010

Global Crop Yield Map

I found the above map in a recent paper in PNAS.  It shows average yield of all crops (in dry tons/hectare) across the globe.  It makes fascinatingly clear where food comes from. You can click for a large version of this and the other maps below.

The most obvious high-level point to be taken from this map is that high crop yields are associated with development (the US, western Europe, Japan) .  There is some question about how correlation runs here.  Certainly, yields have increased enormously in the twentieth century with mechanization and agricultural chemistry, so development increases crop yields.  However, it's likely also true that development has historically proceeded furthest and fastest in places particularly congenial to agriculture, which now thus show both high yields and high levels of development.

Overall one would guess that there is significant scope for improvements in yields in places like India and China as they continue to develop and urbanize.

It's interesting to compare the map above with one of NPP (net primary productivity - how much carbon plants fix in total).  That map (from here) looks like this:

Clearly, the places where most of the NPP happens are in the tropics, whereas most of the crop yield occurs in temperate regions of moderate plant productivity.  Humans have still not figured out how to exploit tropical areas for food production on anything like the scale of temperate regions, even though the former is where most of the plant growth occurs.  It also makes it clear that trends in global NPP and trends in global food production may not necessarily correlate, since they are really centered in quite different places.

It's also interesting to compare the crop yield map above with the first principal component of the Palmer Drought Severity Index, which would be my guess for how the emerging trend of global warming induced drying/wetting will look:

Unfortunately, there is a lot of overlap between where humanity currently produces high-yield crops, and the places that appear to be getting drier (including southern Europe, western North America, and eastern China).

Presumably that is going to create a lot of pressure to more effectively exploit areas that we currently don't utilize very much.

Note: This post is part of the Future of Drought Series on Early Warning.


jewishfarmer said...

It is going to point us to places we don't exploit much, but what we're going to find there is that soil matters a great deal, Just as we can't produce on tropical soils the same kind of production that we make in temperate areas, we're also going to find that boreal and tundra soils aren't an equal exchange for the topsoils we've been using.

Stuart Staniford said...


WRT to the tropics, it's clear that attempts to impose temperate agricultural practices there don't work very well. But that's not the same thing as there being no way for people to exploit them in the future with a different choice of crop and agrosystem.

One of the things I was struck by in comparing the maps is that Malaysia/Indonesia is one region where you see both high crop yields and high NPP. I'd like to understand what's happening there better.

WRT the boreal regions - are there quantitative indicators that soil scientists use to assess agricultural potential of soils?

Unknown said...

I guess my thoughts here are that Western techniques are not only heavily dependent upon mechanization, but they are also heavily dependent upon fertilizers and pesticides.

Where does that stuff come from? Pesticides are basically petrochemicals.

Fertilizer mainly consists of 3 components. Nitrogen (in the form of ammonia), potassium, and phosphorous. The ammonia in general is made from natural gas via the Haber process. And P and K are mined.

There may well be ways to produce some of these things (ammonia and the pesticides) using renewables of some sort or another. P and K, not so much. If you don't mine it from the earth somewhere, you don't have it.

So the question is whether Western methods of agricultural production are anything even close to sustainable given the inputs that are required.

Mike Aucott said...

NPP is essentially a measure of carbohydrate production. One of the key aspects of agricultural productivity is protein production. To produce protein efficiently, especially in the absence of high inputs of industrially-fixed nitrogen, one needs to be able to grow legumes, which need calcium. The high crop-yielding areas are, for the most part, those areas with calcareous soils, or, like Japan, rich volcanic soils that support the growth of legumes such as clover and alfalfa.

kjmclark said...

I'm not sure that first map is showing quite what you'd think. Most food crops are measured in bushels, not tons. It's usually forages that are measured in tons. Take corn for instance. That's usually measured in bushels of grain. The stalks and cobs are basically a form of waste, and aren't usually measured at all. So if you're just looking at the weight of the economic yield, you can get some odd values when some regions harvest forages and others harvest grain.

The NPP covers that, so looking at the two, the yield data may be skewed by forages. Look at Wisconsin vs. Kansas. Wisconsin is a part of the grain belt, but they're more know for dairy products than grain. Kansas is a major grain producer, and they don't grow as much forage. So if you're measuring the yield in tons, you're probably only counting the grain in Kansas, but you're counting most of the above-ground plant in Wisconsin. That would make Wisconsin look like a major producer - by tons per hectare - when in reality it's just producing a lot of forage to feed cows.

kjmclark said...

If you look at figure 5 in Monfreda et al (https://www.gtap.agecon.purdue.edu/resources/download/2981.pdf) you can see this. This is the paper that West, et al is referring to in the caption for figure 1. Notice that the scale for cereals (first map) goes to a high of 12 tons/ha, but the scale for forage crops (third map) goes to a high of 70 tons/ha. Also notice that there's a good deal of forage production in the northern US and western Europe.

There is more cereal production in Wisconsin/Minnesota than I thought, but at 12 ton/ha, it's overwhelmed by the moderate production of forages in those same areas, at about 35 tons/ha.

Phil said...

How, exactly, is yield per acre a meaningful measure of anything of value?

It might have been once, pre fossil fuels, when it was a measure of the efficiency of conversion of solar energy (when compared to your neighbour's farm, at any rate).

rjs said...

there are a lot of things that bother me about that pnas map...

ie., high yields in mountainous areas, & it visually conflicts with the EIU data, which has china #1 in the world in every food category except coarse grains...so unless they're into eating corn & sorghum, the US high tonnage doesnt feed a lot of people...

Stuart Staniford said...

Kjm: southern Wisconsin, at least, has very high corn yields, significantly better than Kansas - see here for a map. More on your other points later.

jpbenney said...

On thing omitted from this discussion is how, as Gordon Orians and Antoni Milewski have shown in ‘Ecology of Australia: The Effect of Nutrient-Poor Soils and Intense Fires’, Australia’s soils – which look more than reasonable on the maps above – are in fact much less fertile even than the “famously poor” tropical soils.

What Orians and Milewski point out critically is that Australia soils are uniquely deficient vis-à-vis any other extant continent in not only phosphorus and sulfur – critical macronutrients – but also more interestingly in a large number of volatile chalcophile (also called thiophile) micronutrients like copper, zinc and selenium. This has excluded large herbivores except those of a fecundity too low to tolerate even the most minimal human presence, and has made fire the almost exclusive consumer of the primary productivity. In fact, Australian secondary productivity is so low as to be orders of magnitude smaller than any other extant continent, although the limited paleopedological record can easily be interpreted as saying Australian pedological conditions were globally normal before the Oligocene when the Alpine Orogeny and Antarctic Ice Sheet first developed.

This is consistent with their extreme age – very few new soils have been formed anywhere in Australia since the Carboniferous/Permian glaciations occurring 300,000,000 years ago. In other words, Australian soils are around 30,000 times older than even the oldest soils in most of Europe, North America, North Asia, East Asia, Central America, Andean South America, New Zealand and/or extratropical South America. Indeed, the factor of 30,000 may be an understatement as to how weathered Australian soils are vis-à-vis soils of the northern and western hemispheres. Most of the period between the Carbo-Permian glaciations and today was globally hot and wet, with tropical cyclones extending even to high latitudes and no permanent ice anywhere in the globe. Under these climatic conditions, although the continent then lay at high latitude Australia’s soils were subject to extremely intense weathering, so that almost all of them show the effects of formation in a warm to hot and wet climate even under (seasonally) arid meteorological patterns.

Thus, the maps above do not cover Australia adequately – one would really need a much darker red covering virtually the whole continent for some degree of accuracy. Interesting and importantly, the very area of Australia where rainfall has increased most – due to polar ozone depletion and greenhouse gas emissions (in which Australia is undoubtedly the planet’s single worst offender) – is the region with the very worst soils of all. In this region, deficiencies in nutrients are off the map compared to soils in any other extant continent (e.g. available P is typically under 0.1 part per million), and exclude even the low-energy, low-fecundity Macropodidae grazers or the most extensive cattle rearing. Even if this region changed from arid to even sub-humid due to a deeper monsoon, there would be no possibility for agriculture of any sort.

jpbenney said...

With respect to the boreal and polar regions, warming could in theory improve their soils a great deal if it releases nutrients from the permanently frozen inactive layer where most nutrients are stored in the Arctic and Subarctic. Inactive layers in permafrost contain much higher nutrient levels the the shallow and easily leached active layers that do thaw during the summer. How much this would help in practice I am not certain, because the transition from no permafrost to continuous permafrost (where nutrient depletion in the active layer would be most severe) is spatially very slow and areas of discontinuous polar permafrost corresponding to mean annual mean temperatures between -5˚C and 0˚C (23˚F and 32˚F) tend to have the same problems as continuous permafrost regions.

High yields in mountainous areas can actually be realistic because the young, fresh soils are chemically exceedingly fertile.