Is corn becoming more or less tolerant to drought and extreme heat?

At Bioenergy Camp in Perry Iowa, I presented a paper with Wolfram Schlenker on the evolution of heat tolerance in corn.  I blogged about this paper awhile back.  The paper will be a chapter in a forthcoming NBER volume, Climate Change: Past and Present.

My seminars sometimes provoke a lot questions.  I'd like to think this a good sign that my talk is interesting.  Maybe I just need to work on being clearer  This talk, however, provoked a particularly sharp response from Ted Crosbie, Vice President of Global Plant Breading for Monsanto.  He didn't say why, but Dr. Crosbie claimed our results were wrong.

Our main finding is that extreme heat tolerance in Indiana increased from 1940 to 1960 and then declined between 1960 and the present.  The turning points--1940 and 1960--coincide nicely with adoption of double-cross and single-cross hybrid corn.  Our estimates show relative tolerance to extreme heat at the end of our sample (2005) to be close to than in 1940.  That's the part Dr. Crosbie didn't like.

I don't think our results are wrong, even though there is a lot of uncertainty.  But allow me to speculate a little about the challenges here.

There is a lot of chatter in the business about plants being more drought tolerant than they had been in the past.  But it's hard to find persuasive evidence to back that claim.  One problem is that there are two closely related but different factors: drought resistance and extreme heat resistance.  When it's really hot, corn doesn't tassel, and this can kill seed formation and yield.  Also, when it's really hot, evapotranspiration increases, and there tends to be less precipitation to replenish soil moisture.  So, statistically, it's hard to separate stress coming from lack of moisture from stress stemming from extreme heat, including associated issues with flowering and seed formation (which tends to happen in the hot month of July).  All of this is complicated further by the fact that data on precipitation is poorer than data on temperature, which means temperature is likely to pick up effects from lack of moisture.  (We find it's hard to predict precipitation between weather stations but can predict temperatures well.)

So, what we find is that drought tolerance has in fact increased steadily over time.  That is, the detrimental effects of too little precipitation has attenuated.  But the effects of extreme heat (measured by degree days above 30C), conditional on precipitation, became steadily less damaging between 1940 and 1960, and then became steadily worse between 1960 and 2005.  Now it is possible that drought effects are somewhat confounded by extreme heat effects and vice versa.  After all, these are almost two sides of the same coin.  But temperature measures have the stronger association, probably because temperature is measured more accurately than is precipitation.

Things are complicated further by the fact that, despite global warming, weather has been good--a little cooler--during Midwestern summers in recent years.  If you're a farmer in Iowa you may think your high yields are due to the fact that your corn stalks can better take the heat.  But it may actually be (and to some extent almost surely is) because it's been little less hot than it used to be. We haven't reported these statistics, but when we run regressions without weather variables, the trend comes in about 10 percent steeper between 1950 and 2005.  Technically, we need to be clear about difference between the size of the dose and the size of the dose response.  The "dose" in the case being extreme heat. 

The good recent weather also raises some interesting questions about why it's been so cool, and whether we should expect that to continue with climate change.  All the climate models we've looked at predict serious warming in the Midwestern summers.  John Mirankowski says the climate modelers at Iowa State predict it will be warmer in Winter but not in summer.  With the amount of warming generally anticipated for the latitudes of the corn belt, it's hard for me to believe there won't be a fair amount of warming in summertime.  But it's certainly something to quiz the climate scientists about...

A focus on recent changes in heat tolerance is a little askew from this paper's focus on the long sweep of history.  Our sample stops too early and uncertainty too is large to discern with any degree of accuracy what's been going on in the last 10 years or so.  And we focused on the the relatively temperate state of Indiana because it has the best weather records in the first half of the last century.  We needed a number of years of good data before and after first adoption of hybrid corn.  Our main goal was to get good weather data from the terrible dust bowl years. 

Perhaps a better perspective of the more recent years can be found in our earlier PNAS article (check out the supplement, that's where the meat is) that examines the whole country from 1950-2005.  What that paper shows is that, looking broadly over time and geography, the basic relationship with yields is remarkably robust and stable. Looking over time, however, we simply split the sample in two.  While this would also obscure recent changes, the stability of the results suggests that if there were big improvements in heat tolerance that happened steadily over time we would have seen at least some indication that. I think this means that if heat tolerance has grown, it has probably been very recent and has probably been fairly modest.

The thing to do is update our fine-scale weather data and look more closely at evolution of heat tolerance in recent history over all states combined.  It should be easy enough to do once we download all the weather station data since 2005.