Mapping the 1970 corn blight

Here are my 2 maps ((Quick & dirty, without cross checking the numbers, but I think the maps speak for themselves.)) for this discussion. I used linear regression to predict corn yield for each county in the US, using time (year) as the independent variable. I used the years 1950 to 1969 to create the model, and to predict corn yield in 1970. This should be a reasonable estimate of the ‘expected yield’ for 1970 for each county, if it had been a ‘normal year’.

I then computed the difference between the expected yield and the yield obtained by farmers, and expressed that as the percentage of the expected yield. Negative numbers mean that yields were lower than expected in a county, positive numbers mean that they were higher than expected. Counties with data for less than 9 years were excluded.

1970 corn yields were indeed much lower than expected in the southeast. Corn blight hit very hard. But also note that yield was stable or up in the north and in the west, and look were US corn was grown in 1970. The map below expresses corn area as the percentage of the total area of a county.

Most corn is grown in the corn-belt. The southern parts of it were much affected by the disease (The Illinois Secretary of Agriculture’s estimate that, by August, 25 percent of his state’s corn crop had been lost to the blight may have been spot on). But 1970 was a normal or good year for corn yield in the northern and western parts of the corn belt, and that compensated for the losses incurred elsewhere. If you sum it all up, corn production was about 15% lower than what could have been expected. That is whole lot of corn — but perhaps not that exceptional as far as bad years go.

Here is a table of estimated corn yield by state, as percentage of the expected yield for 1970, and the corn area, as percentage of the national area (only for states with more than 1% of the national corn area in the counties data set).

State Yield Area   State Yield Area
Florida -36 1   Minnesota -12 8
Georgia -33 3   Missouri -11 5
Illinois -31 18   Nebraska -9 9
Indiana -27 9   North Carolina -5 2
Iowa -26 18   Ohio -1 5
Kansas -24 2   Pennsylvania 0 2
Kentucky -22 2   South Dakota 6 4
Michigan -12 3   Wisconsin 15 3

Landscape-agro-ecology

The February issue of BioScience has an article about the connectivity of the agricultural landscape in the USA. Margaret Margosian and colleagues used a graph-theoretical approach to characterize the ‘resistance’ of the American landscape against the spread of crop pests and diseases ((Margaret L. Margosian, Karen A. Garrett, J. M. Shawn Hutchinson, and Kimberly A. With, 2009. Connectivity of the American agricultural landscape: assessing the national risk of crop pest and disease spread. BioScience 59: 141–151)). The idea is that the resistance of the landscape to, say the spread of a maize disease, is higher if there is less maize planted in a region. The authors show that wheat, grown in distinct and poorly connected regions, is less vulnerable than soybean, which is grown in a single contiguous region. They suggest that this approach should be helpful for risk assessment and responding to newly introduced diseases. So far, the results are at the level of a general characterization. It would be great if they could validate their predictions with observed disease data. That will be hard, as there are many other factors, like local weather, that come into play.

But it should be possible. Landscape effects on pest abundance were recently quantified by Douglas Landis and coworkers ((Douglas A. Landis, Mary M. Gardiner, Wopke van der Werf, and Scott M. Swinton, 2008. Increasing corn for biofuel production reduces biocontrol services in agricultural landscapes. PNAS 105(51):20552–20557)). They found that:

Recent biofuel-driven growth in corn planting results in lower landscape diversity, altering the supply of aphid natural enemies to soybean fields and reducing biocontrol services by 24%. This loss of biocontrol services cost soybean producers in these states an estimated $58 million per year in reduced yield and increased pesticide use.

Now think of the work by Claire Kremen and co-workers showing how landscape pattern influences the ‘pollination ecosystem service’ by wild bees ((Kremen, C., Williams, N. M. and R. W. Thorp. 2002a. Crop pollination from native bees at risk from agricultural intensification. PNAS 99:16812-16816)).

And the call for ecological engineering of landscapes to avoid outbreaks of rice pest. And conservationists that work on shade trees in coffee fields, to help birds and other wild organisms — and get high quality coffee.

I think we are witnessing the coming of age of landscape-agro-ecology. The study of agriculture and its biodiversity beyond the field scale.