 So, I just wanted to start by thanking Justin for organizing this day. It's already been interesting and I'm sure it'll continue to be. And certainly, nice of him to... Raise it up a little bit. Nice of him to coordinate this to match my visit and to flatter me with a keynote. I should say I don't hold any illusions that I deserve a keynote in a room like this with the other speakers and the happy audience who's probably forgotten more on these topics than I already know. I'll accept the keynote on the basis I've traveled before this to be here at least officially I'm facing the Stanford and now in Brisbane. So I am at sabbatical, I'm pretty much in the middle of my sabbatical so the talk today I think will have a bit of a sabbatical feel which is to say kind of half-baked ideas that I've been having some of which may strike you as completely obvious and some of which may strike you as completely wrong and maybe there'll be something in there that's neither wrong nor obvious and I don't know which parts those are but you can look for those. And really, when Justin told me about what he wanted to accomplish today and the kind of discussions he wanted to start I pulled together some thoughts I've been having and tried to draw lessons from the Australian experience for the types of things I've worked on. I don't really work very much on Australia although I've been participating in some projects. But certainly I came here with the idea that I would learn a lot that would be relevant for a lot of systems around the world and I think there's lots of lessons that are held here for others and I'll try to at least explain my thinking on that and why I see it as an important source of expertise and technology for the rest of the world. And I was getting in Brisbane and you become like I said any Australian city you become a sports fanatic. So I think that influenced my title a little bit but I'll talk in general about what I mean by Aussie rules of agriculture and why I think the rest of the world is essentially becoming more Australian over time or has Australian characteristics. In terms of Australian characteristics this is probably a plot many of you are familiar with this is something I use in teaching quite a bit because there's lots of good lessons to be seen in the history of agriculture in Australia. Certainly in the history of wheat this is a plot of wheat yields over time and as I said there's lots of lessons to be learned according to different phases for example you can make a lot of analysis here with the current situation in Africa back in the early days of Australian settlement but in particular what I'll talk about today is what we've seen recently in Australia which is really two things from my mind tremendous progress on the one hand in wheat productivity, average productivity essentially going up by double over this time period even in the last 20 years we're going up by very significant amounts so that's on the one hand. On the other hand what we see is remaining very high levels of variability and even potentially arguably higher levels of variability than we've ever seen before and this idea of pushing forward the yield frontier but at the same time being prone to big setbacks is something that I think in Australia is very well understood that agriculture is a very risky business. In the rest of the world as I'll try to argue that agriculture is becoming a riskier business for various reasons at the same time that we try to push up yield potential. So just to reiterate what I've been talking about Aussie rules agriculture, Australian agriculture essentially I'm thinking of frequently water-limited crops Australia is very low amounts of irrigation which is true in parts of the world but even for European parts of the world they're becoming increasingly water limited. And the second characteristic is that variability of yields is already high but is also seemingly increasing over time and essentially this is not so much that we're getting worse at dealing with adverse weather but we're getting so good at taking advantage of good weather conditions and so in a sense it's a success story it's a relative success being slower at the partial conditions than in really good conditions and that leads to a situation where you have increased variability with all sorts of strategies then to deal with variability including some of the work that's been done here to try to breed varieties that maybe trade off a little bit of yield potential for much greater resilience or stability whatever work you'd like to do. Now why do I think these are becoming common around the world where I think it's actually a bunch of things actually one of the things I won't really emphasize today is climate becoming more variable I think that is a component in some places but I think even without that issue of is climate becoming more variable are we having bigger swings in temperature or are we having bigger swings in rainfall even without that component I think lots of regions are set to see more variability. Why? Well, a couple of reasons in terms of water stress and the average state of water stress I think that this is becoming more common will become more common around the world partly because of rise increasing temperatures this is a summary plot from the I'll just use the mouse here rather than try to turn a point I'm not blocking the screen even then not that tall. So this is a plot out of the latest working group 1 report of the IPCC again this should be familiar to most of you but just to make sure historical changes in global mean temperature up until about 2010 future changes projected for different scenarios of emissions so this would be more like a business as usual this would be a very optimistic mitigation scenario but as sure as anything in this world we can expect warming over the next few decades especially in the 50 years the probability of not warming is extremely small and the degree of warming will potentially be different but that warming is really baked into the energy and climate system and it's something that we know very well will raise the saturated vapor pressure of the atmosphere this is Clausius Clapper on its base in physics it's something that drives a lot of what we see in terms of the climate projections but it also has this effect that it will increase this factor at least by itself will tend to increase the vapor pressure gradient between the internal leaf and the atmosphere causing water to be lost at the fast rate so this in general increases water deficits a second factor which is I think something not as appreciated something I didn't really appreciate until about the last year is that relative humidity is also declining over many major agricultural regions this is another figure out of the work of the one report showing you for different seasons I'm used to calling this winter and summer but now I will not because I've learned better this is December through February we'll call it summer and June through August we'll call it winter what you can see the story of the global scale has always been relative humidity doesn't change much these were very early projections based on just simple models that you would probably maintain relative humidity you would increase specific humidity or absolute humidity but in general you'd rate maintain relative humidity that's true with the global scale but actually what it's mainly true is why it's mainly true is because there's some increase in relative humidity over the oceans which is compensated by which is counteractive by decreases in total land and in particular if you look at some of the key regions that we're going to grow food and what I've done is actually try to highlight that here by taking the projections just four places in the world where we grow a significant amount of food and then looking at the main growing season so this would be June through August in the northern hemisphere or December through February say in the southern hemisphere this shows you the mean projected change in relative humidity for a 50 year period 2060 relative to 2010 and you can see that in some of the bread baskets like in the US and throughout your and in eastern China you're seeing decrease in relative humidity expected on average to be on the order of 10% 5% to 10% say Australia is kind of in the middle ground there are parts for example in India where we don't expect very much now I shouldn't reiterate that the physics of this is actually quite well understood I just didn't appreciate it at least in the sense that most moisture in the atmosphere is last sort of evaporated over the ocean surface and because with climate change what happens is the ocean warms less quickly than the land you're tending to increase the saturation pressure over land more quickly than increasing the water vapor content in the air and that leads to a decrease in relative humidity over land especially in continental regions or intercontinental regions another interesting aspect of this is that depending on which model you look at you obviously get different answers but the change you get in temperature is very much highly correlated with the change you get in relative humidity and again the physics of this is well understood related to the amount of evaporated versus sensible late versus sensible heat flux but this is showing on the bottom here the correlation between out of 30 climate models a model's projected change in the daytime temperature versus the projected change in humidity and it's extremely negative high significant correlations throughout so what this means is that not only on average are we getting warmer and less humid but the models that get the warmest are actually the models getting the driest and for a cropping system that has very big implications because you're really ramping up the saturation pressure at the same time you're really turning down the humidity so this creates this huge vapor pressure deficit which could potentially drive a lot of water stress another factor separate from this so I'm going to start with the factors contributing to greater variability, greater water stress then I'll acknowledge some of the factors that may be working in the opposite direction and I should have made it for now and then I'll talk about essentially why this is an empirical question which one wins out and why I think it's going to be more the and so again I think this audience will be well aware that water resources in many key irrigated regions are stressed and especially in places like part of the Central US not our main rainbow but a significant amount of agricultural grade in North Texas and that area certainly in South Asia this is showing that groundwater depletion rates estimated by recent study really very very high rates on the water so you can see here the scales which is millimetre per year very very high rates in the North China Plain as well so this is going to lead to just generally more water stress a separate factor which is I think one again not why we appreciate although people who work in raising their potential very much recognizes is that any increase in biomass that you can achieve is going to come with an increased amount of transportation that there is a very fundamental link between how much carbon you can take in and how much water you get out and that depends a lot on vapor pressure deficit but for a given vapor pressure deficit this seems to be a very well conserved property of a given flow of synthetic type so Maze will have a particular coefficient we will have a very good coefficient but it's a very fixed quantity so that as you try to get higher and higher yields, higher and higher you're going to require higher and higher amounts of water this is a summary of that relationship from one of Thompson-Pair's recent papers and he's been aware of this for a very long time but certainly you have more efficient water use in Maze than the rice or soybean but the point here is that as you try to push yields higher the amount of water that you require is going to necessarily be higher which as you require more water the chances of you being short of that requirement is going to grow ok now the factors that we would think to be making the world less Australian or less water stressed certainly higher CO2 is a big deal it's again something that some people have appreciated for a long time some people are a little slower to recognize but there still are some interesting questions about how important this effect will be this is just a cartoon from a review paper by Andy Leakey just kind of giving you the sense that the idea is with higher CO2 plants are able to improve that trade-off between carbon gain and water loss by closing their storm out of that you have essentially for a given amount of photosynthesis or potentially for a C3 crops and even greater amount of photosynthesis you lose less amounts of water that's again you can see this in any sort of experiment that's wrong with high CO2 over and over again and so yes it's true that higher CO2 it's not necessarily as big of a factor as we in some places it will be a huge factor in other places it won't be as big of a factor as some of these other changes I've already talked about and as an example of that I'm just showing here some simulations we're doing with Graham Hammer and others looking at main systems throughout the world and what the projected changes in temperature and humidity and CO2 imply for the amount of water stress which is a very close predictor of the amount of yields that you get so I'll take a minute just to explain here you have just looking at the temperature changes predicted by the miles over a 50 year period you have a projected in this particular site say in Nebraska which is a very dry site you really exacerbate water stress you have an average yield loss of about 20% projected by this model now if you look at the humidity changes on top of temperature the humidity changes again in crop studies you actually increase that to about 30% and you also significantly widen because as I said there's a negative correlation at play here where the hotter models are also much drier so you're actually making the downside projections much worse now if you then introduce the CO2 changes we expect corresponding to these climb scenarios you certainly have a beneficial effect relative to say temperature only you reduce the losses but you reduce them by about the same amount that you would increase them if you were accounting for humidity so the net effect of accounting for both humidity and CO2 is something that looks a little bit on average like what you have before but with a wider distribution again because of this correlation between humidity and temperature now it depends on site so there are some places and I haven't fully unpacked why there are places in France for example where the humidity changes although they're significant but the CO2 is certainly enough so CO2 here is actually driving this particular system in France say to a less water stress situation even though temperature is going up and humidity on average is going down but throughout the corn belt of the US the net situation is negative in other words CO2 is not enough to overcome these driving factors towards more water I should say that I think one of the reasons I say these CO2 issues are still slightly on the top is that there's a lot of interesting interactions between chemistry temperature and CO2 obviously as you change the transpiration rates of the canopy you change the temperature of the canopy as you change the temperature of the canopy you change the vapor pressure surrounding the leaves as you change that so there's a lot of intricacies that are not necessarily capturing the experimental data or in the models and if anything these are going to push the CO2 effect to being smaller in terms of the net season there's also increases in the leaf area that again most of this audience is well aware of and this is maybe beyond the simulations I showed you why the models can capture and the reasons we don't think CO2 will be able to help as much as might appear in the context of humidity changes this would be an additional factor that's not in this model the other thing I think counteracting a lot of these other trends in terms of more water stress is that autonomy continues to change and often cases become more water efficient around the world so in the US we've had a big introduction of no-till lots of improved weed control so there's less evaporative losses or transpiration losses associated with weeds you have better earlier ground cover which limits evaporative losses so these are all going to in general make a system more able to capture rainwater and make it into useful production less likely then to fall short of the water you need to meet your target okay so this was all by way of kind of explaining the types of things that will drive systems to be more or less water stressed and more or less variable in terms of yields given the variation in water stress that happens from mid-year but I haven't yet sort of made a firm conclusion I showed you some examples from the US and the West from simulations but I think that overall this is an interesting question that you have factors pushing both ways and how does it all play out and so what I would like to do is explain a little bit of some recent results we've had and I think that in deciding what to talk about there's always the opportunity to keep it very broad but I'm going to use this chance to present some research results which may seem a little now for a little while but I'll bring it back at the end and hopefully this is not in the category of being either obvious or wrong because these are recent and peer reviewed results so in the US we grow a lot of corn and we grow it throughout the country but especially it's in the central United States mostly rain fed systems very high yield and very productive very deep soils and very high yields except when you get a year like 2012 so this was a big drought in the US now I've heard Australians kind of stop at the drought that we had in 2012 as you know we're a bunch of whips and we call that a real drought but in the US it was a big deal it was this is a picture taken in July which is this is not what corn should look like in July this is right around when it should be flowered and this is a nice picture because you know ironically there's a lot of interplays between supply and manufacturers that we were talking about this morning ok but corn largely in the grain fed and in the corn belt of the US one thing that we have in the US which doesn't exist in Australia is subsidized insurance and one of the benefits of that so we can talk about the downsides of that but one of the benefits of insurance in the US these records are now made public and so what this is showing you here is out of the three big producers and three biggest producers in the US a plot of field level yields that we have records for and the color is an indication of the level of yield so red is high yielding green would be sort of average yielding and purple would be low yield and this is for a particular year in 2012 through 2010 we have these records going back to the mid 90s excuse me so what you can see here and I talked to you yesterday a lot about satellite based appreciation of heterogeneity let's say and you can here see this an insurance based appreciation of the tremendous amount of heterogeneity yet within a system even within say an individual county you'll get a lot of variation in yield that will surprise people who study these things but it's a very nice example of how variable productivity is even in a very productive very well managed system now this is a movie showing you all the data and I like this I like this because it took me like an hour to make it so I want to make sure it's easy but I like it because it really demonstrates the the difference in the spatial patterns from your years so there's not a single spatial pattern that repeats itself the amount of temporal variation you get depends a lot on where you are and from an analytical standpoint this is really great because we can do a lot of rigorous statistical analyses without worrying that we are confusing cause and effect or at least confusing one factor for our left factor and so what we've done is now pair this data with lots of really high resolution weather data that exists in the region and ask questions such as what are the key factors environmentally that drive yield variations and then back to this key question are we getting any better or worse over time at dealing with drought in the US and this is the question I want to talk about today and this is an issue that somebody raised this morning about needing to rely on I think this is Eric needing to rely on really good data not an anecdote so in the US we are a wash in anecdotes of how clever farmers are and how great seed companies are and how better at dealing with drought now I don't necessarily disagree with any of those statements I mean seed companies are clever farmers are great and we've gotten better as I said at dealing with bad years but we've gotten potentially even better at dealing with good years and so this is an empirical question of how is variability how is sensitivity to bad conditions changing over time okay so first what we do is we link this with all sorts of other temperature data so that was supposed to be an animation but it didn't translate this is just two snapshots of two different years and I'm highlighting here a variable that seems to be quite important which is daytime temperature in the flower room or 60 to 90 degrees after so and actually what's even better than Tmax but this slide I just had is Tmax is the vapor pressure deficit so there's a high correlation between daytime temperatures because again the daytime temperature is driving the saturation and empirically that correlation is well about 0.9 and so we can build very I would say models that are maybe half as sophisticated as some of the genomic models that you guys are using but that are able to deal with lots of potential variables lots of data avoid overfitting and we can actually ask the question what are the key drivers of variability and it turns out that it is vapor pressure deficit in the ASCII window and even more than rainfall variation what drives water stress in the U.S. is high vapor pressure deficit this is different than Australia my understanding of Australia is the soils are not very deep there's a lot of in-season rainfall there's a very high variability in that in-season rainfall and so rainfall is driving a lot of variations but the U.S. is actually a vapor pressure deficit story which has been a hard story to tell to the people who are used to talking about rainfall but I think it's borne out in lots of different ways because the rainfall isn't important but above a certain threshold the rainfall doesn't matter nearly as much as the high vapor pressure deficit okay so what we do is we use all this data we use all this weather data we use this technique called Mars or multivariate adaptive regression splines to try to figure out what the two relationships are we're going to use that all to define what the exposure of each individual place in each end and year what was the exposure to drought for that particular environment and then ask how are they performing the indifferent drought conditions over time okay so this is a summary of for corn we've done this for soybean as well of what variables emerge is really important and as I said the most important is vapor pressure deficit in the window of 1690s after selling this is the relationship it picks up so it's able to pick up non-linearities it essentially fits piecewise linear functions and what you see is a declining yield with high vapor pressure deficit and then a really quickly declining yield with very high vapor pressure deficit penalties associated with late sowing in the US this is something that's well known and it emerges as very important and then again as I said we have some role of precipitation but it's actually not a very significant factor in corn because we have these deep soils that are pretty much starting at saturation in most years okay so what we do then is for each place and year as I said we know what the weather was so we can actually say according to our model is this a good or bad situation for maize in this case or for soybean and this just shows you for each place and year whether we think it was particularly bad which would be a red dot or particularly good which would be a blue dot and you can see again variation over time you can see 2012 is a particularly bad year and this is years like 2004 and so the idea is quite simple I think which is that well we have in every year at least we have a few well maybe not every year but most years we have at least a few or say at least a hundred fields that were exposed to a particular type of drought stress and we can say okay how did it do under drought stress in 1995, 1996 and we can do that for different levels of drought stress so even though this year was very heavily droughted there would be some locations that didn't have very severe drought and vice versa and this is what we see in the data so this is plotting now the average fields four fields exposed to a particular set of conditions over time so red would be the most stressful conditions blue would be the least stressful conditions and as you'd expect the more stressful conditions have lower yields and this tells us that our model is sensible that if you fit a trend to these different sets of conditions or these different levels of stress you see yield gains throughout this 18 year period for all levels of stress but you see yield gains that are faster at the higher yield levels or the lower levels of stress then at the lower yield levels or the higher levels of stress you can see 2012 really it's a severe yield loss but even if you remove that yield from the analysis you retain this kind of spreading of the difference between a good and a bad condition and you can see that a little bit more clearly if we take some of the key individual components of stress so for example the key one for drought would be high vapor pressure deficit this is what I showed you before and I've split it up into quintiles of this so these colors indicate the quintiles of vapor pressure deficit in region and you can look at yield trends associated with high vapor pressure deficit and again we have lots of yield progress in cool conditions or low vapor pressure deficit conditions but we have essentially no progress under very high vapor pressure which is the more stressful condition and interestingly for sowing dates it's kind of the reverse story where actually the less advantageous sowing days or the late sowing days we've actually been making quite a good progress there so there's a couple of ways we've tried to make sure that this makes sense the point was to take the apps in the model many of you are familiar with very good at simulated water stress compared to most models at least in my opinion and actually simulated some of the management changes that have been going on so what I should say is that this story of greater increases of yields overall is very much a story that has been associated with increased planting density of maize in the US this is kind of a well known thing that over time the planting density has increased in the US what's surprising to me or what was surprising to us people is that even over this very recent period there's been about a 25% increase in sowing density of corn so what we did was take an option and run it for a 25% increase in sowing density and what you see is that you get higher average yields but so these green lines of a higher sowing density relative to the control of higher average yields but you get actually bigger drops so in a dry year you're not able to produce anymore sometimes you can even produce less if you're over stressing individual plants and they move though and their harvest index goes way down this just shows you a scatter of vapor pressure deficit and interestingly versus yielding interestingly the increase of slope from about 15% to 28% so almost a doubling of the sensitivity to high vapor pressure deficit matches very well what we see in the empirical observations and another thing we did as a test was we just each year independently fit a regression let's go back here if that's alright fit a regression between the basically the high amount of vapor pressure deficit areas for that particular year there's lots of reasons you wouldn't want to rely on that as your estimate of sensitivity to vapor pressure because we know these different places for example southern Illinois has very poor soils has high vapor pressure deficits in this year you don't want to attribute all of those yield drops to the high vapor pressure deficit but what you can do is do this over time and see is there a trend in the spatial relationship so you're not now relying on looking at temporal trends and yields for different levels you're just looking at each individual year and sure enough what we see is some variation in that inferred sensitivity but a steady decline and again you get that very strongly if you include 2012 but you also get the stash one shows you it's very statistically significant these red dots are showing you what absent showed as the inferred sensitivity so if anything the empirical data shows something more extreme than what absent shows now one implication of this in relation to my original sort of thesis I guess is that you have a higher sensitivity to drop or in this case a higher sensitivity of vapor pressure deficit which is the main driver of drop in the US and what's shown here is a projection of vapor pressure deficit in this region over time driven again both by higher temperatures and by lower humidity so we expect average vapor pressure deficits for example in 2050 to be almost as high on average as they were in 2012 this is a dot here this isn't an unusual feat this is actually a dot hidden by a poor placement of the but you can see that there's been a historically if you include 2012 historically certainly a trend and that trend is expected to continue the bands here show the span from the 25th to 75th percent out of the 5th to 95th percent out of the mouth so there's some uncertainty associated with individual projections for this region but certainly the way this towards higher vapor pressure deficit now again the implication of being more sensitive to this is that this increased frequency of drop will actually be a bigger deal than we thought because rather than being more adapted and much better dealing with these drop conditions we're actually becoming more and more sensitive to these drop conditions so if you project out the implications of this for yield using current sensitivity you get something like a 15 percent yield loss but using the trend you get something like a 30 percent now this is not a prediction this is not to say we think this historical increase in sensitivity will definitely continue over time of course we would hope to do something about that but just to kind of drive home that this trend is not a small trend this is a very almost a doubling of the sensitivity over time in these last 20 years so another way to think about this and this is again I think it's going to happen in a lot of places and I haven't again this is not something that's fully fleshed out but if you look at a lot of places like in Iowa what's happening is yield potential isn't increasing in this case often driven by an interaction between the cultivars and an increased sowing density so you have very high yield potential and increase over time and what happens now is if you look at average yields of 10 tons per hectare or more the water demands of that crop are approaching the amount of water available in the system in the range of the system so what I'm showing here so again there's very strict relationships between carbon uptake and water loss which are governed essentially by the type of crop which in this case is corn and by the vapor pressure deficit which depends a lot on temperature so you can look at the simulations and see very clearly this type of relationship in terms of efficiency and so what I'm showing you here is if you take the amount of rainfall that falls in this county and you look at the efficiency that you can expect the maximal efficiency of all well actually kind of an upper limit is a rule of thumb at least according to Scott when I was talking with him and I still have to track this down a little further but say that almost you'd expect 75% of the annual rainfall to be consumable by the crop and you can only do so much of making sure the water gets in the soil and doesn't evaporate and doesn't go out for a few weeks and in the US for example we have a lot of runoff under the hiring minutes which Australians know something about as well so if you figure you can get 75% of the rainfall to the crop at best and there's a certain maximal efficiency of converting that to biomass and there's a certain maximum efficiency of maximum fraction of biomass you can get to yield then what you get is sort of a maximum possible yield for a given amount of rainfall and this is showing you the lines for the historical 5th percentile of annual rainfall and the historical medium of annual rainfall and so what I think the story is here is that first of all we start lowering our efficiency of water use and that drives water stress because for a given amount of biomass you need more water so that's the shift from they say the dash lines which are representing an average temperature of 28 degrees to the 29 degrees and at the same time what you're getting is yields pushing up against what the dry years can support and again this is the Australian story that the dry years can only support so much biomass as you can imagine for that amount of rainfall and so you wouldn't expect the bad years to get much better than they are right now actually don't have up to 2012 which was a little bit worse in terms of combinations of rainfall and temperature and eventually you can continue to increase under good years but even eventually you're going to hit some wall in terms of advocates because of the water bill now of course there's interventions that can be made on the water side and there's already reports of increased irrigation in the Midwest which is a novel but I think that there's a lot of indications that what you'd expect to happen in the US is something very similar to the last 20 years of what's happened in Australia continued progress but increased variability and again this is without even considering changes in climate variability and for that matter without even considering the increase in DPD just even if you state at a constant temperature you're already very close to what you can expect out of a dry year now I haven't I haven't got the data or the time so I'll blame the time but I haven't gotten to talk about other regions but for example we have a study going on in the Chinese systems which shows that as a combination of better yields over time higher yield potentials and also climate change which are increasing again in vapor pressure deficits that essentially the ratio of rain-fed to irrigated yields which you can consider kind of a measure of the importance of water stress has been declining over time but we're now in this situation in this region where it's becoming much more frequent in water stress not necessarily because water well it's partly because the temperature changes and mostly because of the yield potential increases not necessarily because rainfall itself is changing that much rain-fed European systems I would say as I showed you in the beginning have this very large humidity signal especially in southern Europe decline in rainfall this is one of the areas where decline in rainfall is quite robust one of the few areas in the world where we have a general expectation of decline in rainfall higher temperatures and higher vapor pressure deficit or lower humidities and on top of that we already know relative humidities yield potentials are quite high so again that system is set up for not being able to improve much in the four years and maybe all the gains will have to be associated with increased variability and then as I said this is not even counting the issue of increased rainfall variability increased temperature variability which will drive that further okay so that's sort of my argument or my thought process for well partially for coming to Australia besides the cricket and the football that I wanted to see but in terms of trying to learn about how how people here have thought about these issues of how to cope with increased high variability and how to cope with water stress and from my perspective this has always been a mecca of dealing with of crop growth under stress and continues to be although yesterday he said I should never have come here to destroy my great visions of camera but it's still re-examining after periods and in Brisbane as well I'm learning a lot about this but certainly there's a lot of lessons to be held I think for the research side of things but also for the farm management side and for the farm policy side of things and I don't know exactly what those are I'm still thinking through those but I think that there are a few and one is simply that we may make that not have this mindset which exists in the US and I think exists in a lot of systems that I've worked in which is that farmers will think about a new technology if there's no trade-offs associated with yield potential because they make most of their money in a really good year and that's sort of been the starting point for a lot of discussions about adaptation climate change because you don't want to sacrifice what you can do in really good conditions but I think to some extent in the story of Drysdale and other things in Australia as I understand was very much making a trade-off now I'm happy to be correct about that but there is certainly a point where a trade-off makes sense when you look at the value of more stability as opposed to higher potential certainly diversifying activities within a farm I think that this might be a topic of Bob's talk later I'm not actually sure but there is certainly you see in Australia a tremendous amount of diversification within farm that we don't have in the US or a lot of systems around the world certainly more grain storage will be a rational response to more variability at this aggregate scale and I think the beginnings of that can be seen already and I think one of the big things obviously in Australia there are huge land holds whether that is you know there's lots of reasons for that it's not just dealing with variability but certainly having large land holds is one way to do a very high variability and then the only thing now Australia is I think leading the edge on which holds a lot of lessons is actually decision support systems for farmers getting on the the leading edge of seasonal forecasting and being able to inform decisions and one of the responses to high variability is to essentially either become very, let's say conservative in your input because you're not sure what are your record conditions or to become just very inefficient because you're applying so much so I think there's as I look at the landscape there's lots of lessons and lots of technology being developed in Australia that are going to be and I think you know above all it's the at least from my perspective what I'm interested in learning throughout this Chicago it's about the development and the management of drought-powered crops and what that means and how much does it mean I started off with a picture of grain sorghum how much does it mean switching out of the crop into something like grain sorghum and we talked a little bit about this this morning maybe wholesale shifts in the clocking system are the I'm going to steal Mark's words here and I'm not even sure I'm going to use it correctly but this would be a more transformative change than just trying to be incremental adjustments to the systems that exist maybe that's the strategy to deal with with the prospects of much higher variability is to switch out of a given system that's exposing it to that so nothing else I didn't go over time and hopefully I stimulate a little bit of thought and I'm happy to receive my time for Mark because I know he tends to go on and on where I can take questions it's up to him I've got a few mics here John Evans, probably Graham Fakwa I'm done with this here, would be asking this question but I don't know whether you're aware of the papers that he's published with my project on the decline in pan-evaporation which is somewhat counter to the story of the decrease in relative humidity equate to drying so in Graham and my projects experience in fact it's not supported the long term observations of evaporation do not support any dry world with warming yeah, good question I'm aware of the work I have to say that I'm still wrapping my head around the differences between well, I know I have the differences between pan-evaporation and transpiration efficiency on the other hand the differences between kind of an absolute view of the world which I've been maybe falling into which is that BPD is what drives transpiration and there is no consideration of sort of boundary layer effects of greater wind speeds just in general the effects of wind speeds so it is it is a puzzle in my mind of why you don't see observational that you might expect it but over the time we have observations there's lots of explanations that you can encode for why you haven't seen it even though it's what you expect but you know I would hope to resolve that in my mind at least it may well be resolved in the notion that looking at humidity and temperature is not going to tell you the whole story from a plant's perspective just like it will tell you from a pan-evaporation that on that point we haven't seen big changes on the bottom can you hear me now we haven't seen big changes in BPD in Australia over the last 30 years I think it's fair to say the reason for the reduction in evaporative demand has been a mixture of reduction in wind speed as we discussed yesterday and then some places dimming as well this is my quadric alongside me backwards I wanted to ask on the diagram you showed us running into a limitation by water in the May sites in the Midwest that horizontal red line is the upper bound the upper bound was horizontal yet it must increase with increasing CO2 certainly it does that was a lazy way of just showing a reference point for current CO2 levels and for the 5% rainfall where would you expect that line to fall but you're right it should be slanted with a slope according to the situation yeah and at the risk can I pass this to Mike to finish the question I was just interested in the numbers in one of your slides the one where you had a BPD of 2.2 kPa so could you go to that slide it was about 10 bank was it this one or I think one bank yeah this one here it's probably been that one it's the one also I could rate sure well this is the same panel so you have this one yeah so I'm wondering how that is going to be calculated okay the yeah that one this one here yeah so in absent the way that BPD sort of the average daily BPD is calculated is you look at the BPD at the at the peak of the daytime maximum and you look at the BPD at the minimum which is typically very small or even zero and then you take the you take 75% of the difference I'm just trying to be sure so essentially it's kind of a weighted average of the peak and the daily it's not a sort of an hourly calculation of BPD it's something that's so what's the temperature the roughly ballpark those numbers if I look correctly you're going from 2.2 to 2.6 kPa on though this is yield here this is log yield this is BPD on the x axis here so these are average daily BPDs over a 30 day period and what are those numbers these numbers go from about 1 to 2 kPa