Sustainability Videos & Lecture Series

David Lobell: It’s a pleasure to be here. I realize this is a diverse audience which is always a bit of a fun challenge. I want to leave plenty of time for questions and discussions. I guess my goal today is to give you a sense of what it’s like to work on adaption to climate change and in particular in the area of food security.

Realizing that there’s a lot of undergrads here and graduate students I’ll take a little bit of time to give you some background both on what we understand in general about climate adaptation and in particular about food security.

Bear with me those of you who are [distorted audio] or have seen some of this. I’ll get to some more research results later on in the talk. I was teasing, Matt, so does anybody know what this crop is here; sort of a basic introductory point here.

This is the crop that’s grown on the most area in all of the world. Maybe it’s not a very good picture. Sorry about that. This is actually a wheat crop where I work in Mexico. This shows you actually a typical wheat experimental field after flowering, towards the end of the cycle. You see a kind of very rectangular pattern and you’ll see kind of different colors, different heights.

These are all different varieties of wheat being grown. The breeders and the geologists are going in and selecting varieties based on different aspects. We’ll get to that a little bit later.

I’m going talk in particularly about development of the next generation of crops; which is a necessary but not sufficient part of the overall food security picture. At the Center of Food Security at Stanford we’re very much interested in feeding people into the future. How do we do that also at the same time preserving the environment?

What I’ll talk about today is a little bit narrower in terms of the research which is a lot of what we do is on crop productivity. I’ll try and explain the connection between the two before I dive into it.

Now by way of background I’m not gonna go into climate change. I think that most of you are familiar with global warming. It’s not just some idea that, Al Gore, came up with. It’s a real thing that scientists are almost unanimously behind the basic consensus that humans are emitting greenhouse gases. The greenhouse gases lead to warming and it’s not only expected to happen, but it’s already happening. We see for example, last year was the hottest summer.

I’m gonna assume that you’re all familiar with that at least, or at least accept it for the purposes of this talk. What’s I think less commonly understood by people is in the last 5 to 10 years; especially, in the last 5 years we’ve really come a long way in understanding how sensitive our society actually is to temperatures.

I’ll give you some in depth today into agricultural aspects, but I wanted to give you a little bit more broader background on some of the interesting connections that I’ve been seeing in the literature lately.

One, since the start of the baseball season I thought I would mention, one, is the affect of temperatures on baseball. I think there’s one obvious one which people probably very familiar with which is that balls go longer distances in hotter weather, and so you get more homeruns on hot days.

There’s more effects than just that and in particular what I’m showing here is a player getting hit by a pitch. People have studied this believe it or not, and actually the chances of getting hit by a pitch go up quite a bit, about 50 percent when you’re in a very high temperature environment than when you’re in a cool. This controls for what park you tend to play in or what time of year you’re in.

You’ll notice the two most important factors are how many of your own players have been hit by—how many of the opposing players were hit by your own pitchers; which is a big affect, but also the temperature is a big affect. Anybody know why this might be? Yeah.

Audience: [Inaudible].

David Lobell: Yeah. It’s basically one of the reasons they study aggression in sports is ‘cuz there’s all this good data set, but this is an indicator or aggression. Actually there’s indications now from lots of fields that aggression is really affected by temperature.

This is a little bit hard to see. This is a working paper by Ranson recently looking at records of crime across the US and how it fluctuates with temperatures. Now when we study these kind of things empirically you can image all of the mistakes you might make by thinking it’s temperature and it’s actually something else.

You’ll have to take my word that these are smart people controlling for all sorts of variables, and even with those controls the black lines here are showing you the affect, say here, of temperatures on a simple assault, on aggravated assault, on robbery.

You tend to see much lower rates of cool temperatures and for many of these you see much higher rates at high temperatures, so this explain maybe why Canadians are always so calm and why I always feel so aggressive when I come to Arizona. I thought it was the either the fact that my in-laws live here, or because of high temperatures. [Laughter].

Okay so that’s a lot of evidence now, and there’s also more evidence, at least suggestive, of the risk of civil wars breaking out, other things related to violence and aggression.

All of which, you know simple assault it’s not like you’re gonna be walking down the street and all the sudden it’s hot and you’re gonna decide to assault somebody, but these are risk factors that combine with others. I hope, I hope.

Another one more related to performance, and this maybe isn’t so surprising to you, is the winning times of marathons are actually the average times of marathons are very affected by temperature. This shows you a plot of temperatures during the Boston Marathon, and the trend corrected winning times.

You can see that the difference between it being between 20 and 30 degrees Celsius is about a 5 minute difference here in speed. This ones a little more obvious maybe that when it gets hotter out you notice people sweat more, their bodies fatigue faster, but it’s another clear empirical indication.

You can see that the empirical studies first of all, they’re both very recent and they tend to be in prominent journals because we’re, I think, it’s a very new field of understanding really directly the affects of temperature on these crops of wheat.

Here’s a really interesting one it doesn’t look super convincing but, this is showing you trends in temperature over the last 30 years with trends and economic growth. Economic growth is incredibly complex due to all sorts of factors, including some of the civil wars and things that I alluded to before, but a lot of it is just the productivity of different sectors of the economy.

Agriculture being one, but also the productivity of workers very temperature sensitive this is why we’re sitting in an air conditioned office here. It’s obviously more comfortable for you to be working in air conditioned but you’re also much more productive. That’s why businesses cool their factories.

Countries that have been warming faster have actually seen slower growth than countries that have not been warming. You can look at this in various ways not just by trends, compared to trends but, when you see trends compared to trends it’s a pretty or what economists would call ‘differences in differences’, it’s a pretty compelling piece of evidence in that it’s unlikely other things are likely to be because.

Okay now here’s another example now closer to home for agriculture. This is showing you here a plot produced by Wolfram Schlenker and Mike Roberts, two colleagues, in 2009 showing you temperature on the X axis here and the affects on those temperatures on the productivity of corn or soybeans. A left and right.

The green on the bottom just shows you the typical amount of time in a given season that the country spends if you average over all areas of the countries at different temperatures. What you see is that corn and soybeans both are fairly indifferent as you get up to 29 or 30 degrees C. That a little bit of warming helps not that all that much, but then you get past 29, 30 C, so 30 C is about 86 Fahrenheit for those trying to do the conversion in your head, crops really take a hit, and actually they take such a hit.

These are in logged units that if you’re spending a lot of time out here, so 1 degree day, as we like to measure it, out of 35 could cost as much as, 5 percent penalty on your final productivity.

This is a very non-linear responsive temperature at a daily scale. I was thinking recently this is also a good description of if you were to plot my age here and the amount of hair on my head. There’s lots of things that follow this kind of trajectory. It’s a very non-linear response. You’re okay for awhile and then things sort of fall off.

The take home here from this initial tour of the literature is that there’s a lot of correlations out there between heat and outcomes that we care about, more than I think people realize. Often times we don’t realize them because we’re sitting in air conditioned offices but, also because lots of other things are changing with temperatures.

Hot years tend to be dry years, less rainfall, so people would be likely to say it’s low rainfall and it’s true some of it is the low rainfall, but you can actually piece apart these things if you have enough data. There’s clear evidence of temperature affect in lots of places.

In many of these, as I said, we’re pretty confident that there’s a cause in relationship. In other words, a batter getting hit by a baseball doesn’t cause the temperature to get warmer.

A lot of times in economics you’re worried about is it economic growth really cause conflict or does conflict cause economic growth to slow, but with temperature and rainfall economists like these things ‘cuz they’re exogonists. They’re not gonna be affected.

We can even make claims of causality but we can’t do in many of these cases with some exceptions like where we understand human anatomy, is to really understand the mechanisms. I would say in general a lot of the literature on heat impacts is really towards understanding mechanisms.

It’s not to say nobody has been looking mechanisms. The Australians’ have been looking at this for quite some time but, lots of fields are looking towards the understanding the mechanisms. You might ask, “Why would we care about the mechanisms as well?”

You don’t always care about the mechanisms. You might get a sense here. I’m pretty good at bad analogies [laughter] and I tend not to remember the ones that work well. If you’re walking down the street and you see your girlfriend with another guy there’s some mechanisms going on that make you feel upset, but you don’t really care about those mechanisms all you care about is the impact on you.

Similarly, if we’re seeing temperatures causing these impacts we’re concerned because know temperatures are increasing, we know that these impacts may become more frequent, but we do often want to know more than that because we want to adapt to it. We want to say, okay not only is it sensitive to temperature but can we make it less sensitive? How do we make it less sensitive?

Now when we talk about food security; which is what I’ll focus on today, let me just give you a little background on we think about food security, and then we’ll start talking about how do you make it less sensitive to temperature.

Food security is not that complicated as a concept. It tends to be complicated to measure it, to understand how it responds to things. As a concept it’s just simply that you have reliable access to food; that you’re not worried about where your next meal is coming from.

Most of us in this room are food secure. Hopefully, all of us are food secure. Most of the people in the country are food secure; not all of us.

It really comes down to three things. It comes down to basically if you think of yourself when you go to the market, how much does the food cost? What is your budget for buying the food? What is your household income and how is it distributed among the household? Do you have any diseases that prevent you from absorbing the nutrients that you require you to eat so much more to be able to absorb the [fading voice]?

That is actually a big deal in a lot of places with poor sanitation. As you get diseases of the gut that prevent you from really absorbing even if you have the food.

These are the three fillers of food security. You have the food exists at some reasonable price. You have some reasonable income and you don’t have a disease preventing you from absorbing those nutrients.

It’s a complicated subject. It’s broad. I’m not gonna try and get into all the details; although, questions are welcome at the end, but some key numbers to keep in mind. One; is that a billion people currently in the world, so about one in seven people, are not meeting the standards of food security that the international, like the UN, puts out.

Which is essentially, enough energy in your food to maintain a very modest amount of activity, so this is not if you’re an athlete or if you’re working in the field this is not that level of energy this is just minimal like 2,000 calories a day type of energy. A billion people don’t have secure access to that much food.

This gives you a sense of it. If you look here, this is the average percent of income spent on food. People I think, I don’t want to go off on too much of a tangent here but, people often I think, wrongly simplify this down to say it’s all a problem of distribution. We have enough food in the world it’s a problem with distribution.

It’s really about the ratio of your incomes to prices. If prices are high typically consumers are hurt. If your incomes are high typically consumers are helped. It’s useful to think about what’s the ratio?

One way of thinking about it is what’s the fraction of your income spent on food? Here we are in the US about six percent of our income goes to food. A very small fraction; especially, when you consider that most of that expenditure on food isn’t actually on the food itself. It’s on all the services that come with the food.

If you go up to the forced end of the spectrum and lots of African countries, Pakistan also this is the average household that’s spending half their income on food. These are basic commodities they’re buying. They’re buying basically bulk maize grain or bulk [fading voice] gram; half of their income on average, so the poorest are spending 70, 80 percent of their income on food. This is where the food insecurity is really high. It’s in south Asia and then in [fading voice].

The other thing to understand about food security and this is really the link to today’s talk is that both income and prices depend on productivity of the agricultural sector. Now for some this may be obvious for some not so obvious, but I think the least obvious one is why is income so dependent on agricultural activity?

This is the story of the green evolution is that the poorest of the world are unusually dependent on agriculture. This is the history of almost every country including ours. If you look at, this is the World Bank figure, you look at poorest countries to richest countries on the X axis here, and look at the share of the labor; which is shown in red or the share of the DEP, the total economic activity in our country, it goes up as you go to the poorest areas.

Labor being almost 100 percent of people in agricultural labor, meaning they either working their own lands or working somebody else’s land. Over half of the economic activity is in agriculture, so if you improve the productivity of half of your economy that’s gonna be a big deal for incomes for the poorest of the poor.

Also, what I’m not showing here is that within a country the poorest tend to be in the rural areas, so that’s even disproportionately important to their income.

One the price side here’s a simple way to think of it; which is if anyone has read Lewis Carroll’s ‘Through the Looking Glass’ like the Alice in Wonderland companion. We call this often The Red Queen Problem; which is that you have to run super fast just to stay in place. This is when Alice is running super fast she’s not going anywhere and the Red Queen is as well, you have to run twice as fast if you want to get anywhere.

This is the story of agriculture. You’re trying to keep up with constantly shifting demands. This is a simplified supply and demand curve. I see, Michael, in the audience this is his bread and butter. Basically people will demand more as the price goes down, and people will supply more as the price goes up.

Typically, the actual realized quantities are where you intersect these, but the demand curve is constantly shifting out to the right because of population growth, because of income growth, bio-fuels now. The question is “How do you push supply out?” How do you keep pushing supply out? Historically we’ve done that and I’ll talk about that in a little bit.

How do you keep it going? If you have lower productivity you potentially get big increases in prices. Increased productivity is generally gonna be price reducing and also income improving in the poorest areas of the economy. Agriculture productivity is very important for both of those reasons.

Okay and now the obvious link in here into climate adaptation is that all these aspects are sensitive to warming. I’ve shown you, I mean I don’t think the baseball example directly impacts any of these three, but all of the other ones in some way or another you can links into disease burdens, you can imagine links into food prices, into household incomes.

What I’m gonna really talk about is the effects on agriculture productivity; which as I just told you is really clear on both of these aspects. When we think about the impact of heat on food security I want you to think more broadly than just what I’m gonna talk about today.

There’s all sorts of reasons to think, and we’re still trying to figure these out I would say, but within each of these areas where we’re trying to figure out how much does it matter we’re at the same time trying to figure out what are the mechanisms?

What I’m gonna talk about today is agriculture productivity and the mechanisms of impact. Here we are at the outline for today. I’m gonna try to give you just a little bit more in depth about the trajectories of productivity and why we feel like we need to adapt to climate change.

Then I’ll talk about adaptation and what in particular the community that’s trying to develop new crops what do they really want to know about the climate change and how is that gonna help them adapt.

Then I’ll try to provide two case studies. We might not get through all of them. That’s fine. I added some more introductory stuff today when, Matt, told me a little more about the audience, so we’ll see how far we get.

I’ve been working recently with wheat in India and also corn in the United States both of which provide interesting data sets to ask these questions of what are the mechanisms, and also how well are we doing in adapting; which is obviously something else that people would like know.

Okay so the need to adapt crop fields. Here so I can in one slide how I think of this when I think of what kind of research we’re doing and why it’s important.

What I’m showing you here again is a picture of wheat; a different picture. This is the history of wheat productivity. The field is measured typically in tons produced per hector of land over the last 40 years, actually 50 years.

What you can see is the productivity of wheat has basically tripled over this time in terms of land, how much land you need to grow a given amount of food. We now need one-third the amount of land to produce as much food as that same amount of land would’ve produced 50 years ago.

This is a complicated story but, it really comes down to three major innovations. One was the new prevalence of irrigation throughout the world. A lot of countries invested heavily in new infrastructure in South Asia and East Asia.

Water as you well know living in Arizona, is a very essentially part of crop growth, so that really raised the potential for how much you could produce.

At the same time, scientists were developing new varieties of these crops that could really reach that potential and not for example, fall over because they were getting so heavy and tall. A lot of the mutant Japanese varieties of wheat actually were very useful in developing the modern.

Pretty much every product you have that has wheat in it, whether it’s the Baklava or whatever else is over there with wheat in it, are basically derived now from these dwarf wheat varieties like I’m showing you here. Much shorter than the old ones so that they can get fatter and still now fall over.

Finally, and not least important, is the prevalence of a lot of chemical additions to agriculture; including synthetic fertilizers that came out of World War II. A lot of industrial activity in producing synthetic fertilizers from abstract nitrogen hugely important for plants is the major macro-nutrients.

These three together the point here is that at first they had a big affect. Second is that they all three were pretty big innovations in agriculture. It’s very hard to imagine things that’ll be that big. We try to imagine them but, it’s gonna be hard to maintain this trajectory. What I’m showing here is just a fifth of the last 30 years of data, and what it would mean to maintain that trajectory.

Now to give you some context, here I’m showing you the 60 to 100 percent increase in production that we anticipated needed because of demand growth. This is an aggregate number across all grain crops, but I’m showing it here on the wheat.

Again, we need that it’s a variety factor is one. It’s largely developing countries getting richer wanting to eat more, wanting to eat higher on the food chain. It’s population growth and it’s bio energy production.

We want to maintain that. I’m gonna skip over a lot of science here, but the best we understand the heat is gonna have a negative affect on productivity of crops like wheat.

Here I’m trying to show in magnitude what we think it is. The blue is showing here what we think heat would do to the medium trajectory of yield, and the red is showing you what if everything is on the lower end of what we would like it to be if say things warm faster than we think or if crops are a little more sensitive than we think it could be as much as this.

It’s a huge increase in the short fall from what we were barely meeting before to here. I should have said before that basically any difference in this you can think of as being made up by a combination of either reduced consumption because prices will rise and the poorest won’t be able to eat. It’s always the poorest where the elasticity’s are. It’s not gonna the richest.

It could be you expand the areas, and that’s typically the response of poor countries is that if they have forests they’re gonna expand [fading voice].

Okay so I’m taking too long to go through this. Anyway but maize is a similar story. Corn, maize we use interchangeably. I’ll probably switch back and forth. The proper name is maize, but obviously you know it as corn. That’s a picture of a corn plant.

How do we adapt? A lot of conversations on climate adaptation, and I know a lot of you are probably interested in this from totally different angles from agriculture. When I think of that in agriculture, I think, it’s really important to start with what kinds of decisions are thinking about that.

What are actually decisions that you might make differently because climate is changing. That you wouldn’t have made differently if climate had not been [fading voice], and so that’s what this says here.

In agriculture, I think, people all too often think of a farmer. That’s not because, I’m not saying, “all too often” because farmers aren’t critical or they’re not great, but it’s not the only decision made in agriculture.

The farmer’s are say users of technology and all sorts of decisions went into who made those technology, what they decided to work on. That farmers are all interested in policies; those are all influenced by decisions. The farmers are all influenced by the infrastructure around them. I mentioned irrigation; those are all influenced by decisions of [fading voice].

The point here is that in any area of interest like whether it’s food or water, urban systems I think there’s lots of different decisions being made and it’s important at the outset to be explicit about what kinds of decisions you have in mind and what the scale of interests are.

I think that’s very useful because different scales often have very different implications for what we can say about climate impact, and what we can say about the robustness of switching a decision and how likely that is to help you or not.

My own work in agriculture tends to think about broader spatial scales, broader time scales exactly because I don’t think that we have all that much to tell people working in the individual field about what next year or the year after is gonna to be like. There’s obviously people working on forecast but that’s different than what I think of as adaptation to climate change. The reason I’m talking about crop development today or developing new varieties is because of that choice.

When did we start? 12:10 or so? Okay.

Let me explain to you a little bit about how the decision maker, in the sense of like, a crop breeder thing. Now breeding it’s actually a lot like trial and error; anything that you do is trial and error. This is another picture from India this time of a breeding field. You can make out the different partitions. You can see it’s actually some bags over the grains because when they’re trying to make crosses they have all these paper bags covering the anther.

The idea is you throw a bunch of stuff out there. Presumably, you pick that stuff from some rationale because you can’t grow everything in the world there’s thousands of varieties out there. They’re trying to pick which ones will be the next big thing. I think a good analogy, maybe not a good analogy, but an analogy is if I were in charge say entering somebody in a race, and this was a running race around the world and I had to pick one of you in this room, and multiply you 1,000 times and send you off to that race.

How would I do that? Well, pretend you’re a plant and I can’t just ask, “Who’s a really good runner?” and you would raise your hand. I’d have to have in mind a little bit about what race you’re gonna run right? Are you gonna a marathon or you gonna run a 100 meter dash? That would help if I knew something about that.

Then I would also have to have in mind what kind of traits am I looking for in a runner? Do I look for somebody who is younger or older? Somebody who’s got more hair or less hair for the aerodynamics? Somebody who’s thinner or fatter? That’s the idea is you take that; I’ll call it a bad analogy now, and bring it to the world of crops.

They’re trying to think—let’s add to that analogy before we leave it. Let’s say if I wanted to take you and take you out to the track and run a certain race that has some cost associated with it. I’d have to buy you lunch or something and I have a budget. I can’t take every possible one of you and test you.

I have to first screen based on my expectation on how different [fading voice] work, and then I’ll maybe test a handful of you. Then I’ll select one, multiply it, and send it out. That’s kind of the process of crop breeding, obviously with plants, and focusing on target environments that vary a lot.

There’s different breeding programs designed for different target environments. It’s a very empirical exercise you can’t just say, “Oh this plant has this trait and that trait, so I’m gonna take it and throw it in that environment” you always test them. You always find surprises. You always have to further test it, throw it in lots of environments.

It’s like March Madness. Why don’t we just say Louisville is the champion? Because first of all, we want to be entertained, but second of all, we don’t know. It has to play itself out. Okay I’m mixing bad analogies now, very bad.

What the crop development community wants to know when you talk them is basically, they want to know what is it about the temperature stress say that is really causing yields to go down? Because if we can understand that mechanism then we can look or do we need plants that transpire quickly? Do we need plants with really strong roots? Do we need plants that stay green for longer?

These are all different traits that they could look for and that they could expose to certain stresses when they screen them. They’re different ways of screening. I could take you out and run a 100 meter dash. I could you out and run three miles. You design your screening based on what you think the mechanisms and the target environment are looking for.

The other thing they want to know is well do we really have to do anything? Aren’t we already adapting? Because each year we’re selecting and temperatures are going up and CO2 is going up, I think we’re just doing it anyway. You can give us the adaptation funds but, we don’t actually have to change what we’re doing do we?

That’s a legitimate question. I actually was at a conference recently where the COO of the big food company said, “We’re not so interested. We think we’ll adapt to climate change because our breeding cycle is 10 years and we’ll keep up with change.” They were more interested in what the policies are gonna be and how they might help to mitigate carbon and things like that.

I think that’s an empirical question. We really don’t know, and if anything I’ll maybe talk about at the end, all the evidence is that when you target stresses you make a lot faster progress than if you’re just relying on them to occur more frequently and then hope that you select.

Okay so this is what they want to know. What I’m gonna talk to you about is the evidence along both of these lines. Maybe as one piece of background is to understand that this seems like the kind of thing we should already know. People have been growing plants for 10,000 years. Scientists have been looking at plants for well over 100 years. Why don’t we already know these things?

Isn’t temperature just an affect? Can’t you just look that up in a text book? The fact is that temperature has lots of different affects on plants, just as it has lots of affects on people as I showed before. It’s really hard to know the relative importance of what’s happening.

The way I teach this to undergrads and graduate students is basically there’s five big areas that I think of as separatable effects of temperature. There’re photosynthesis and respiration directly affected by temperature. There’s how fast the crop develops. Actually, the life cycle is very much dependent on temperature.

There’s direct damage, so if you get extremely cold, you get a frost, things die. You probably know that from some of the frost you’ve had recently. If you get extremely hot things die right? Again, you probably have some familiarity with that.

Water stress is a very important way that temperatures can affect plants, and I’ll talk more about this later but, essentially, again you know this as well, the hotter you get the more water plants need, the more water air can hold this is a Clausius Clapeyron but, maybe the reverse way of the way some of the students are used to thinking about it in a climate model.

Finally, pests and diseases are also responsive to temperature and those indirectly affect lots of what we grow. There’s estimates that a third of what we could produce is lost due to pests and diseases every year.

Okay and then the reason we can’t just say it’s 20 percent this, 30 percent this is that it’s gonna depend on where you are, what temperature you’re starting from. Obviously, in Canada the frost is more important. In Indonesia the frost is not really such a concern etcetera.

What we started to do in my group in part is to try and think about clever ways to get evidence on these particular aspects of the problems. For example, we don’t do experiments but, that’s clearly one way to do it. You could try to say take a crop expose it to a heat shock at a certain time of the year where we think there’s a mechanisms involved, like right at flowering where maybe the plant actually aborts of its kernels because of the really high temperatures, and you could do that, and you learn something, but it’s typically in controlled environment, you don’t necessarily know how it translates to the field and it’s very expensive to actually do this enough combinations.

While there’s important work going on there it tends to be slow progress and it tends to be I would say fairly inconclusive. There’s exceptions when experiments are really designed well, but for example, when you do a heat treatment you actually change things about the humidity of the environment, you change things about the radiation regime.

You typically often have ‘em in a pot and so their roots aren’t actually behaving as if—there’s lots of concerns sometimes about experiments and you’d like to have independent ways of looking at it.

A second way, which we’ve started to do, is well pretend you’re a doctor and look for symptoms of particular mechanisms that you might expect to happen. This is where, Matt mentioned the remote sensing, this is where we try to use remote sensing as a way of monitoring croplings. We’re looking for symptoms that we think are associated with different mechanisms.

Now at long last we’ve reached a research example. We’ll talk about wheat in India. This is a picture of India; Google Earth. The green areas there show the major zones of wheat production in India. It’s basically a northern crop. This is the Indo-Gangetic Plain. It’s heavily in the northwest, but it’s an important crop throughout the north.

What we’d like to know with wheat in particularly there is how important is a mechanism which we refer to as ‘Accelerated Synethics’. What this means is that you have crop growing along, doing okay and then a really hot day happens or a really couple of hot days happen, and even though the crop, you would think, might be able to recover what it does is it goes into a shut down mode.

It starts senescing, the leaves start turning brown and it’s basically trying to ensure its survival. If you think about it, I’m gonna do another terrible thing which is to give the crops some human features here, but if the crop is thinking, “Oh no it’s really getting hot. We must be entering the brutally hot part of the season. I better ensure that my offspring survive I’m just gonna shut down and focus on preserving, reaching maturity, preserving my seeds.”

It’s a conservative strategy and it makes a lot of sense if you’re out in the wild, but what happens in agriculture is you start senescing you stop absorbing radiation, you stop growing and your yields get hurt.

This is some experimental evidence on this. I mentioned before how experiments can sometimes be inconclusive. You see the exposure to days about 34 degrees Celsius and the amount of yield loss with 100 being full yield and loss. You can see that there tends to be lower yields and there tends to be—you know you can draw a line through that.

Not a whole lot of evidence on what happens to wheat at these really high temperatures. Again 34 is 93, I think, about Fahrenheit so pretty hot. Nothing for you guys I know, but for most of the world it’s pretty hot.

Wheat is a cool season crop. It tends to grow in the winter season in most tropical areas and where it’s a very important crop you have a lot growing, across the border in northern Mexico for example, and it’s grown November to March [fading voice].

What we do in India, and one of the nice things about India, is it’s wall to wall wheat. I should have put in the picture of flying over some of these areas. It’s just incredibly humongous during the winter season. With satellites you can look down and you’re pretty sure you’re looking at wheat and then we can look at tracking the greenness of the crop over time with a satellite instrument called MOTUS.

What we’ve done is take MOTUS is this area and I’m just showing you an example here, so you can look at MOTUS over time. This is what a filtered time series looks like. I’ll just point with my hand since it’s not working. You can see the green up, so this is a measure of greenness, so wheat is planted say, greens up, peaks, senesces, gets harvested. Anybody know what this is? [Pause].

The second crop in northern India is rice typically. So wheat, rice, wheat, rice, wheat, rice it’s a very intensive system. You get the wheat out, put the rice in, get the rice out, put the wheat in. It’s been going on for a remarkably long time, so anybody who talks about sustainability of agriculture there’s few better examples, at least in terms of a rotational system, of wheat and rice. Obviously, they’ve intensified it a lot over the last 30 years, and they’ve drawn down the ground water and [fading voice] like that.

Here’s, I’m just showing you that, kind of get a sense of the rhythm. We can go in with the wrong sensing and say, “Okay when did it green up this year? When did it green up? When did it senesce?” and we can plot these things.

This is the average green up day. You tend to sow a little bit earlier here than over here because the monsoon sets the calendar in rice. Then you harvest and that sets the calendar of wheat. Then season life you tend to grow shorter varieties here because as I said it’s well known that wheat doesn’t like the hot times of year, so you want to get the wheat out. If you’re sowing later you have to plant shorter variety.

We can look at this and yes, okay it works. I’m such an amateur when it comes to using animation. It’s embarrassing to say but.

What we’re looking at here is now we’re tracking the phenology year after year, and we can actually look or we could look at the exposure to high temperatures. The key thing here is that the exposure is not in the same place every year, so we again be sure that it’s actually the very high temperatures.

You can see this is the extreme heat measure here. Sometimes it comes in here. Sometimes it comes in over here. There’s enough diversity in the pattern that we can do a pretty good empirical job of separating out what could be other factors and what exists.

Let me just explain this figure a little bit. Here it’s plotting the average, basically the average degree days over a season. Each point on that map and each year is serving as it’s own observation, so we have thousand and thousands; almost a million observations of green up to green down.

Now what we do is plot that as a function of the average temperature. I mentioned one of the reasons temperature matters is because it affects the development rate. Plants develop more quickly in high temperatures.

Now you can see by this downward sloping line this is season length against temperatures, downward sloping. What is interesting here is we’re separating out now, you have the same exposure to average temperature but some fields had more of these very hot days, and some just had a lot of pretty hot days.

The red line shows you what the average season length for the ones that had say 25 hundred degree days of average, accumulation of average temperature, but some very hot days. The blue is the same amount of average temperature, but less hot days.

The point is that these lines are separated; that the seasons are much shorter when you’re getting exposed to very hot days. This is a confirmation that the hot days are doing something special in terms of the development. It’s not just the fact that the season on average was warmer; it was the extra hot days that caused things to accelerate.

On the right here it’s just showing you what we predict, based on this data, what we predict in these red bars would happen if you warmed up the region by two degrees; which is what we expect over the next few decades.

The blue and green are showing what the crop models typically used in the region [fading voice]. The crop models are seeing much less, or depending on what sowing day you talk about, because different sowing dates see different amounts of heat. They see less shortening of the season than we would predict from this data because they don’t have this accelerated senescing.

Okay I think we’re doing okay. How else do we look at mechanisms? Well, another way that we’ve been trying to use a lot is to not only look at data but to simultaneously look at models and compare the two, and by understanding when they agree and when they don’t agree it can point to certain mechanisms. I’ll explain that a little bit.

Here’s an example, a nice example from a group out of Europe, looking at heat stress and antithesis. This is a particular mechanism, as I mentioned before, of basically the flowering that at the critical time of the plant when it’s flowering, the male’s flowering the female doesn’t go well because it’s hot.

There’s reasons why; you can think of it as like the pollen dries out or it’s more likely to dry out before it gets to the female and if it’s really hot or dry, so you get unsuccessful pollination.

Most models don’t include that affect they’re just concerned with the basics of photosynthesis and development. What these guys did was they said, “Okay let’s put it into the model and see how much better it agrees with data” so they’re showing here that with each stress, these solid lines, you get better agreement with the data.

That’s an indication because this model agreed better than this model with the data. It’s an indication that that one added mechanism is actually important in reality.

What we’re gonna do is look at corn in the US. This is a cornfield in Illinois from last year, in July. This is not what corn is supposed to look like in July in Illinois. We had a very hot year. You probably saw a lot of news stories about this. A very dry year; it was kind of unprecedented both on the rainfall side and the heat side; which is why it was such a spectacularly bad year for production.

We’re gonna look at it throughout the Midwest here; which is the heart of the Corn Belt. This is showing you the average productivity of corn. What we’re gonna do is use a model called ABSAM. This idea with ABSAM is that it has a lot of mechanisms related to development, photosynthesis and water stress but, it doesn’t have these direct heat damage during flowering. It also doesn’t have pests and diseases during flowering.

What we wanted to see was ABSAM do a good job of recreating the observulationships between temperatures and yield. Some of the relationships I showed you between Wolfram Schlenker and Mike Roberts slides way back. How well does it do of recreating those? If it does well, we think well, maybe these mechanisms are driving it, and if it doesn’t do well we think maybe these may or maybe if it doesn’t does well, it’s these mechanisms that the model doesn’t treat the mechanisms properly.

As I said, we’re gonna look at the Midwest, particularly three sites here. Here is kind of jumping to the results. This is the observed in black, the up start in green. The observed yield versus, and these are de-trended, the versus the accumulation of very high temperatures.

If you think back to that non-linear curve it’s basically the amount of time you spend on that steeply sloping part of the curve, and negative slope. ABSAM actually does a very good job. This is not with calibration or anything. It does a surprisingly good job of recreating that sensitivity to extreme heat. I’m showing you for one of these sites but, we’ve done this in multiple sites and it looks similar.

What does that tell us? Well, it tells that, if you follow my logic from before that it’s probably what’s going on in reality is probably something that the model has in it; otherwise it’s just dumb luck. We don’t expect things to work for just dumb luck. Okay we always have key values and things for dumb luck, but it’s very unlikely that this would happen year after year in different areas.

We’re looking at a couple mechanisms within ABSAM. One is, I mentioned, is photosynthesis response to temperature. Well, if you look at daily growth directly in response to temperature all these points are individual days within the simulation over the time, but the black line is the average. It’s like a smooth average to all these points because you can’t really—it’s too many points to really the tendency.

Yes, temperature affects daily growth and you have some optimum for maize around 30 degrees Celsius, but you see a tremendous amount of scatter. It’s not the main thing going on with temperature.

In contrast what you see is growth is very closely related, with again some scatter, to the ratio of water supply it demands. This is a measure stress in the model. When we talk about ‘water supply’ that means basically, how much water can it pull up from the soil and ‘demand’ is how much water does it need to transpire to maintain growth.

I realize I didn’t talk about that as background, but plants transpire a lot of water; that’s why they need water. The temperatures again, Clausius Clapeyron, higher temperatures mean you need more water or you lose more water if your cellmates are of a certain size. Cellmates are how plants lose water.

It seems to be more so water stress than just direct affect of temperatures on the crops. When we look at water stress you can see that the water stress, the supply, ratio supply to demand is fairly sensitive to, this is not showing you daily temperature, maximum temperature versus daily water stress.

There is a fairly linear relationship between the two. To get a sense of how important temperature is if you go from 25 to 35 Celsius you more than double amount of water needed to maintain a given level of growth.

Again, I’m talking to a group in Arizona, I don’t think that this is a surprise, but as I said there’s lots of mechanisms going on. In this case it happens to be that this mechanism is dominating the other mechanisms.

Now it’s even more than that because that was showing you at the daily time scale, but of course, what happens is if you’re hot on a day you’re actually using more water which sets you up for moisture stress the following day or the following week or the month.

If you look at a monthly time scale actually the affects of high temperatures at the monthly time scale is that much more because it not only affects its demand, but it’s affecting the supply of water; how much you’re using, how much you’re running out.

This plot, let’s just focus on this, this shows you the individual terms, so the demand, the supply. Here we can run the model adjusting just the temperatures or adjusting just the rainfall to get a sense of how important temperature is as a driver of demand. One would be no stress. That would be the ratio is just one, and so you’re able to meet all of your demand.

As you go below one, you have more stress, so I should have marked that but, more stress in this direction. The orange bars here are showing you what happens as you warm by two degrees. The blue bars are showing you, sorry the green bars are showing you what happens when you have current climate.

This is the average amount of stress in current climate, the average amount of stress when you warm by two degrees and the average amount of stress when you reduce rainfall by 20 percent. You can just see, this gives you a sense of how important two degrees is for the water stress, and compared to a 20 percent rainfall drop it’s much more important. Why I focus on because that’s a few months of corn growth [fading voice] you can see a similar thing applies in [fading voice] months.

Now in the five minutes or so I have left, so I’m almost done, I want to address the second question that I said the development, crop development can be one; which is, “Do we really have to do anything different?” Aren’t we just doing this anyway but autonomously adapting as we go?

How do we look at this? Well, this is the fun part of doing science is you try to think about what is the evidence. How do we know it was happening if it was happening? This is a very tough question with adaptation. You think any field of adaptation whether it’s infrastructure or water how do you know when you see it?

Now here’s a couple of options that we find. One is to look at actually the breeding trials that are going on and compare how the breeding trials are progressing in hot conditions versus cool conditions. When they send out these seeds to grow in 100 places every year to test them not all those places are seeing the exact same climate, so you can compare how well they’re doing in the hot areas versus cool areas.

The other one is to look at actual farmer yields and compare it in good conditions and bad conditions and see are the yields in fact getting better faster in the bad conditions as the good conditions, and that gives you a sense of are we becoming more or less sensitive to bad conditions; in this case adverse climate.

We looked at the first of these approaches with wheat the second with maize and I’ll go through them briefly. With wheat what we’re looking at is this international nursery run by SYMMIT. SYMMIT is the, going back to that green revolution side of the tremendous yield progress; SYMMIT was the main key player in developing the crop varieties.

They continue to be the main player. It’s a publicly funded, international group, produces varieties, distributes them at no cost to national programs around the world that then modify them if they want for their local environments and release them to the farmers.

SYMMIT runs trials every year; every year throughout the world. This shows you the location of the trials where they run wheat. It’s an incredibly well organized group of basically participating groups around the world that are coordinated by SYMMIT. SYMMIT is based in Mexico. They breed here and they send it out.

What we did was look at the progress in these environments over the last 30 years as SYMMIT has been keeping records. Luckily they’ve digitized some of these records or we had to help them digitize it.

What I’m showing here is now the trend over the 30 year period of wheat yields in these nurseries. Okay? You take the average or the top yields of a nursery and you look at how that changes over time. We’ve split it up into the coolest conditions that the wheat crop sees and the warmest conditions.

Every year it’ll see cool conditions at some sites, warm conditions in other sites. Over time because of warming you’re tending to see more warm conditions in recent years, but every year there’s some of these.

When you look here you see, well these two bars are intended to show the same thing. We try to correct for the fact that the locations switch over time, and also that within the 19 degree, so this is a cool bin below, 19.5 degrees during grain filling. It hasn’t steadily been the same average temperature over time, but whether or not you correct for those movements around time you tend to see pretty significant progress in wheat yields over time.

This is confirmation that they are actually doing what they say they’re doing; which is improving wheat yield’s potential over time. What’s interesting is if you look at now how they’re doing in the hotter sites it’s a much lower rate of progress, and in fact, not really significant statistically in the higher two temperature bins compared to what they’re achieving in the cooler bin.

This is an indication that there main breeding strategy, which is embodied in this data set, which they call The Elite Spring Wheat Trial, are actually not generating that much progress for the hot conditions. In fact, you could argue they’re increasing their heat sensitivity because they’re getting so much better in the cool conditions that the warm years are that much worse relative to good years.

Now I should say that they also have other nurseries that are targeted to stress. Right now these nurseries do operate separately, and the ones that are targeted to stress really don’t try to maintain high yields in these environments. Those you do see actual significant progress to the high temperatures, but what’s key is that you want varieties that do well in both good and bad conditions.

They have yet to really figure out how to, and it’s a difficult task to try to produce crops that are both less sensitive to bad weather but also equally good in good weather, and that’s more or less what I just said, so why I have this animated I don’t know.

Now here’s the story for US corn. We don’t have a public, international, openly accessible data sets for corn because it’s basically a private seed industry. What we do have and this is very recently is access to lots of farm level data on corn yields.

This gives you a snapshot of what it looks like for one year in 2000. These are individual points here. I’m just plotting the average yields, but you get a sense that each point here is an individual farm that we have data for, so thousands of records. For those of you who work with data you can start to see why we get excited about this sort of thing.

This gives you an indication that we have it now for multiple years, so you can look at; this is showing you for Iowa yields over time, you see nice beautiful patterns. Again, these red areas are higher yields, blue are low yields. We also have all the environmental data ‘cuz US has a pretty good weather network.

For example, this shows you the average vapor pressure deficit which is that metric of how much, how hot this temperature is how much water it can absorb. Actually, I like flashing ‘cuz you can get an indication of how important these temperatures are just by for example, looking back and forth.

You can see the hot areas tend to be the lower yield or conversely the redder yield areas you can look and you’ll see that they tend not to be the red areas on the temperature map. Oop. You’ll see that a couple times.

I promised I’d wrap up. What we do is each year we try to pick out where did these crops grow under good conditions, where do these crops grow under poor conditions? We’re defining good and poor so that they’re the same each year. This is just showing you a scatter plot of all the data and how we’re identifying good and poor.

Good is basically low vapor pressure deficit or low temperature and higher rainfall. Poor is lower rainfall, higher temperature. Each year it varies, so in really hot years you’ll have less of the good, but you’ll still have some.

This is how it’s looking. This is kind of a recent analysis, but the blue is showing you the average yields on all the fields that are high yielding, are good conditions, and the red average yields on all the poor conditions. You can see on average the good conditions are better than the poor conditions; although, there are some years where funny thing happen.

What we’re really interested in is this gap between good and poor, growing or shrinking over time. Shrinking would be an indication that you can believe these stories of how farmers were so well adapted to 2012 yields. Now it’s true that poor yields are rising, but they’re not rising fast. You can see here the difference is actually going, it’s basically flatter, it’s slightly negative.

It’s true that poor conditions are producing higher yields than they used to, but they’re not producing higher yields relative to good conditions and that’s really the key measure of the impact of the weather.

In summary, there’s lots of affects of temperatures. I used the baseball example as one that hopefully will stick in your memory. Lots of reasons that society responds to temperature, is impacted and needs to adapt and that it’s important to understand the mechanisms and lots of ways that we try to do this. There’s not really a one size fits all approach.

That in the recent stuff we’ve done on how are the crop development communities doing? It’s not really clear that they’re just autonomously gonna adapt to these hotter conditions. It seems like more targeted efforts are needed, more stress specific types of breeding approaches for example.

I’ll end there and hopefully I’ll have time for a few questions and hand it back to you. [Applause].

[End of Audio]