While preparing some slides for a presentation the other day I came across some interesting data supplied by the Ag Statistics Service of the United States Department of Agriculture (USDA). The data was presented in long columns of numbers organized by year. This was data presenting production numbers for US wheat beginning in 1866 (the year after the conclusion of the US Civil War) and running through 2008. Quite an extensive data set.

At first look the data didn't seem all that interesting and patterns, if present, were difficult to detect. So I decided to graph the data by year and see what showed up. The first data set (see below) graphed as a scatter plot with years running along the x-axis and bushels along the y-axis shows the number of bushels produced over the time period given above.
I think most of us would agree that the results aren't too terribly surprising: the amount of wheat produced year on year has gradually, but consistently, risen over time.

An interesting feature of the data set once graphed is the gradual increase in year-on-year variability over time. Look on the far left. You see that the number of bushels produced each year is slightly more than the previous, and that the increase is gradual and consistent. On the right side you find the number of bushels produced each year can vary widely from the previous year. Not only that, but the points seems to exhibit something of a cyclical nature. The production will gradually increase, then hit a peak and fall for a few years in a row before rebounding.

My own interpretation of this (and this is backed by nothing other than reasoning over the data) is that once we reach the later half of the 20th century farmers have become increasingly "market aware". Meaning that increasingly the farmers are using market forces to determine what crops they should plant. An increase in production will generally result in a decrease in price for that particular crop. One can assume that the peaks form once a low enough price has been reached forcing the farmer to plant another, higher value crop in its place. Over a few seasons, supply retreats and prices go back up pushing farmers back into that grain again.

Next, let's have a look at another different but related data set: Acres of Wheat Planted. See chart below.
Here we see the number of acres planted to wheat over the same time period. Again, not all the surprising. Thinking about the United States just after the Civil War there was much land to be opened up and made available to wheat production. This expanded for a while but, inevitably, most of the cultivatable land was turned over to agricultural production, and being a finite resource, its expansion leveled off. And this can be clearly seen in the data set. Again, we see the cyclical trends start to show up in the later half of the 20th century and the year on year swings in planted area are quite pronounced.

It is interesting to think about these two data sets in relation to one another. In the first graph we see, clearly that the amount of grain produced has be steadily increasing over the 150+ years covered by the data. In 1866 a mere 170 million bushels of wheat were produced. Contrast that amount with 2008: 2.4 billion bushels. But then, put that in the context of the second data set which describes a clearly finite resource: land available for wheat production. We find a well defined increase in wheat production and a leveling off of acres planted. How does one justify the two trends?

The answer lies in the following graph:

To me this is extremely interesting. Notice how the yield per acre (bushels of wheat produced per acre) remains relatively constant from 1866 until just about the year after World War II (1946): 11 bushels per acre compared with 16.4 bushels per acre. Then an amazing thing begins to happen: the yield per acre abruptly starts to climb rapidly, going from 16.4 bushels per acre to the 43.5 bushels per acre achieved in 2008. So, in this way we can see how the wheat produced has climbed steadily year on year: at first it was due to expanding areas of cultivation followed by increasing yields per acre.

But the question then must enter one's mind: what could possibly have sparked just a dramatic increase in our ability to continually push the yields up starting in 1946. The answer is quite simple: Beginning in 1946 and beyond the widespread use of chemical fertilizers, pesticides and mechanical cultivation began. This technology spread all throughout the farms of the US and the world in what has come to be known as the Green Revolution. In 2009 this use of the word "green" sounds at odds with our current use of the word where we use it to refer to environmentally friendly practices. In 1940's and 1950's "green" meant the greening of the planet through increased agricultural production brought about through the use of fertilizers, pesticides, and petroleum powered farm equipment.

Where will the these graphs extend in the future is an interesting question. Obviously arable land is shrinking around the world through over use of the resources (water and nutrients) resulting in desertification. Secondly, how much more grain can we squeeze out of an acre of land. I would have to guess that we are nearing the limit given the current technologies in use (ie traditional breeding programs and fertilizers used). This doesn't bode well for a planet with an expanding population. If the production is going to continue to increase to meet increasing world demand then the next threshold to be crossed is, mostly likely, the introduction of genetically modified grain species. And that is something that will have to continued to be debated as we weigh our competing needs: increasing food production to feed a growing population while keeping our food supply safe.