I’m Christopher Neill, a senior scientist at the Woods Hole Research Center in Woods Hole Massachusetts. In a previous lecture, I talked about how deforestation, fragmentation, fire, the rapid rise of crop agriculture, and changes to regional climate influence the dynamics of the Amazon forest and the potential future dynamics of that forest. In this talk, I want to talk about specific work that we are doing at the Amazon’s agroindustrial frontier — that is, the southern boundary of the evergreen Amazon forest, where crop agriculture is expanding, and where climate and rainfall are just high enough to sustain evergreen forest. So, it’s a zone of both rapid human change and potentially rapid ecological change. What I want to do is talk about how all of these factors interact to potentially shape the current dynamics of the forest, the future of agriculture, and the potential future trajectories of the remaining forest. Our work is focused on a very large farm that we call Fazenda, or Ranch, Tanguro. It’s an 80000 hectare farm that’s situated along the southern boundary of the evergreen forest, in that orange zone on this map, which is the zone of highly deforested land. This farm is an interesting place to do research because it contains about half of its land area in large blocks of remaining forest that both incorporate areas of forest, but entire watersheds. And then half of this farm, the pink area on this right-hand inset map, are croplands. This is what that farm looks like from the air. In the background, you can see these large areas of forest. You also see some protected riparian zones, although they, in this case, in the foreground, they’re highly degraded because they don’t have much forest left, but these zones are also important parts of this agriculture landscape, which I’ll talk about. At Tanguro, we think it’s an excellent place to look at the interaction of land cover and land use, how fires interact with forests in this agricultural landscape. We can look at land-atmosphere exchanges, such as how the rise of crop agriculture and the expanding use of fertilizers influences the releases of greenhouse gases to the atmosphere. And we can also look at interactions of soils and waters, and that is, how does rainfall move from forests or upland cropland, through the soil profile, into streams and waters, and what those ultimate effects will be. So, I’m going to cover all of those points, because they’re all part of our interacting research at Fazenda Tanguro. And, of course, the big overlay is how all of those interact with climate, and how all of those might play out in the future under a climate that’s much more likely to be warmer and with longer dry seasons. I want to start by talking about a unique prescribed burn experiment that our group conducted over more than a decade at Fazenda Tanguro. This is the basic layout of the main experiment. It contains a control area on the right, and this is a half a square kilometer, a large block — a kilometer by a half a kilometer. In the middle is a plot that was burned once every 3 years, on a schedule, during the dry season, when fires typically happen. And the plot on the left was burned every year. And we’re comparing how those fire regimes affect the structure of the forest and the likely potential that that forest would be flammable enough to burn in a repeated way. So, I’m going to show you a little bit of those results. Like I described before, fires in the Amazon forest, because the forest often doesn’t get particularly dry, often occur in the dry season and as fires in the litter layer, in the understory. So, these are fairly low intensity fires to start, but low intensity doesn’t necessarily mean that they don’t have important ecological effects. So, here’s a plot that compares tree mortality and biomass in response to those fire regimes. On the top panel we’ve got a control forest, which is the line on the bottom of that panel with the open circles, and it shows fairly low, but some, tree mortality over six years of this experiment. But what shows up in those top lines, in the square panels at the top, is that the cumulative tree mortality reaches 50% in the plot that was burned every 3 years. And the same change in live biomass, and the same change in leaf area, which is simply the… sort of the area of leaf surface that occurs in this forest… the leaf area is highest and the live tree biomass is highest in the control, and it takes a real nose-dive in that second 3-year burn. So, what happens is that one fire sets this forest up to burn again, and to burn more severely, because one fire results in some mortality of small trees, increasingly dry conditions during severe dry seasons, that allows that second fire to be much more damaging. And if you burn every year, you’re using up fuel. But if you’re burning only every few years, in those very dry years that come in the Amazon with El Nino conditions, for example, every several years, you’re really setting up that forest to burn and burn in a way that transforms its structure. So, I want to show you how that structure is transformed, just by showing you some pictures from this experiment. On the upper left is an intact forest — that’s the control site. The upper right is the plot that was burned every 3 years after the first burn. And then in the lower left, after the second burn. And then 3 years after that second burn, what you see is that this site has been invaded by grass. That’s very important because that grass dries out, becomes fuel during the dry season. So, this change in structure, as it plays out over time, results in a much more flammable ecosystem, and you can see that lower right-hand panel indicates that this forest is well on its way to being transformed into a savannah ecosystem. I want to talk about a little bit of work we’re doing in these riparian forest fragments. These areas, as occur in the middle of this picture, are forests that are along streams. Now, Brazilian farmers, by law, are mandated to leave buffer zones along stream channels. These forests, even though they occupy a fairly small total portion of the area, are disproportionately important, I think, because they protect stream water, they provide shade and cover and protect the aquatic ecosystem. We’re interested in how the dynamics of these forests play out over time, but they’re subject to many of the same dynamics that fragments of upland, or what we call terra firme, forests are subject to, that I talked about last time. One of the things that we see happening, and we’ve documented, is that these riparian zones are narrow. They also can be invaded by grass, and it looks to some extent like their structure is changing over time, simply because they’re fragments, but also because they are subject to disturbance because they dry out and they are invaded by grasses. So, these are an important additional fragment dynamic in this landscape. Now, let’s take a step back at the larger scale — this removal of forest cover caused by deforestation and agriculture, actually, is changing the regional climate. This is a graph, a diagram, of results of a model that come from remote sensing, that show that this large green area in this central part of this image, the land cover on the left… the green area is remaining forest, and on the right you see a graph of temperature. Now, the temperature is low in the big forest area, but it can be 5-7 degrees higher out in that cropland landscape. Well, what does this mean for forest dynamics? It means that higher temperatures are occurring along these edges, just where this change in structure is playing out. So, this flammability is driven by both the changes to forest structure, but these large changes to climate that are driven by the land conversion. Another thing that our group is doing is trying to understand, how do these changes in land cover and cropping influence runoff from streams and the chemistry of streams? We’re interested in how water moves from landscapes into streams and rivers, and one way we get at that is we look at the infiltrability of the soils in a landscape that looks like this image behind me. We do this by putting a column of water over the soil and measuring the rate at which soil… the soil can infiltrate that water, and we get information such as the following, and I’ll explain why I think this is important. This is simply a graph of the rate at which water infiltrates into surface soils in the forest and in soybean fields. So, there’s a clear result that the water infiltrates less rapidly when you have agriculture. These forests infiltrate a large amount of water; they can absorb a meter or more of water in an hour. It never rains that hard. This dotted line along the lower portion of the graph is the maximum 5-minute rain intensity. So, even when this landscape is soybean fields, it can absorb the maximum… the maximally intense rainfalls. So, what that means is that you almost never get lateral flows — you don’t get a lot of erosion off of this landscape — and that’s an important thing, because if farming can be sustained to maintain soils that maintain this high infiltrability, it means that the risk of erosion and runoff that brings sediment and surface nutrients from land to streams remains fairly low. We also looked at how this land cover changes the discharge of streams. So, this is, again, this map of Fazenda Tanguro, and each of these little gridded areas, the darker areas, are entire watersheds that are either in soybean fields or in forest. And we log water flow out of all these watersheds using level loggers and stream rating curves, and we can come up with a measure of the total amount of water and the timing of that water that left these streams — a very standard hydrological measure. And when we did this across multiple watersheds at Fazenda Tanguro, we got a very interesting result. And that was that the overall discharge from streams that were 100% — almost — in soybean cover was 4-5 times larger, in total, than the runoff from the forest stream. So, this plot… the bars on this plot show rainfall… rainfall is highly seasonal, the maximum rainfall occurs from November through March, and the dotted line at the top is soybean watersheds, and the solid line at the bottom is forest watersheds. And what’s, I find, absolutely remarkable is that, despite the intense seasonality of rainfall, streamflows hardly vary over the year at all. And that derives from this very high infiltrability, right? Rainfall falls, it’s absorbed into the ground, it goes into the groundwater, and that water just moves out of the groundwater in a very regular fashion. In the soybean fields, there’s some indication that there’s a little more rapid flower, but the big point is there’s much more water coming out in those soybean streams. And why is that? Because you don’t have trees that are grabbing water, using deep roots that are down deep in the soil, and pumping that water back into the atmosphere. Without trees, croplands put much less water back into the atmosphere. And that’s extremely important, because as that plays out over very large areas, that’s going to influence the future of the Amazon climate. So, here’s a picture of those very deep but highly permeable soils. These soils are absolutely remarkable. They can be up to 30 meters or more deep, they’re very uniform and they’re highly permeable. They don’t have the typical layering that we think of… that we temperature zone biologists usually are familiar with. These are what we term oxisols. They’re very, very old, highly weathered, but extremely deep. And, despite high clay content, aggregated clays make them very permeable to water, and water can infiltrate quite rapidly. So, one result at a slightly larger scale of these changes to the amount of water that’s leaving because we don’t have trees under… over cropland, is that we can measure the total amount of water in a large… in a deep soil profile over a fairly large area using a technique of resistivity, which measures water content in soils simply by placing electrodes and measuring electrical resistance. So, what you get when you do that is a two-dimensional picture like this, that actually is soil moisture down to almost 15 meters deep. And we run a transect out from a soybean field, on the left, into the forest, on the right, and what you see absolutely clearly in this image is that the forest is drawing water from deep down, and you get warmer colors, that is, drier conditions, on the right-hand side, under forests, deep down, than you do under soybean fields. There’s just a lot of water… wetter soils under soybean fields and that manifests itself in this greater overall runoff. We also are interested in the impacts on aquatic systems, of this intensive agriculture. If you think about what happens in intensively agricultural areas of the Northern Hemisphere, such as the “bread basket” area of the middle of North America, we see very large impacts of runoff of nutrients from farm fields, movement of nitrogen, particularly, into rivers, and when that nitrogen hits the coast, such as in the coast of Louisiana, just as an example shown here, you get this anoxic zones in estuarine regions that are very, very, potentially, damaging for aquatic life. So, we’re interested… you know, do you get these changes in this intensive Amazon cropland that you would get in, potentially, the same intensity of agriculture in North America? And so we took our watersheds and we actually added up, by measuring the concentrations of solutes — these nutrients, nitrogen and phosphorus — in the water, and comparing between forest and soybean watersheds, we calculated the total export of material — and we do this as kilograms of nitrogen or kilograms of phosphorus per hectare of land area — and we can compare export from forest to export from soybeans across multiple watersheds, a very powerful methodology. And here’s what we find. For nitrate, we find that the export from soybean watersheds if four times greater than it was for forests, but there was absolutely no increase in the nitrate concentration — quite a remarkable and surprising result. The difference of four is simply that there’s four times more water going out the soybean stream, carrying water with the same concentration of nitrate. The same was true for ammonium and the same, basically, was true for phosphate, that there’s a small increase in export, but it’s driven completely by the fact that there’s more water leaving the watershed, and not by any transport or excess amount of nutrients leaving that landscape. So, this is a pretty surprising and interesting result, especially when you compare it to the result of intensive agriculture in the temperate zone. And these are very, very low numbers — export of less than 1 kilogram/hectare of nitrate, or barely 1 kilogram/hectare of ammonium, or very, very low export, compared with tens of kilograms of nitrogen/hectare potentially going off of cropland that has seen intensive agriculture in other parts of the world. Another factor that’s happening in this zone is cropping is further intensifying, right? So, under soybeans, farmers don’t use nitrogen fertilizers. Soybean is a legume — it fixes nitrogen. So, while phosphorus fertilizer and potassium fertilizer are applied, there is very little or no application of nitrogen fertilizer. That game completely changes when cropland intensifies from the single cropping of soybean to the double-cropping of soybean followed by corn or followed by, in fewer cases, cotton, where the second crop does require a substantial amount of nitrogen fertilizer. So, this is a picture of Fazenda Tanguro and what’s happening right now at the farm is the spread of a double-cropping regime, and the replacement of single-cropped soybeans, which we see has very little impact on the solute concentration in streams, to the double-cropping of soybeans with corn, with potentially unknown but potentially serious consequences, because there’s a large use of nitrogen fertilizer. So, we did another experiment to look at this. This just shows the extent of that across the entire landscape of the state of Mato Grosso, which is a Texas+Oklahoma-sized area where the intensive cropping is occurring. And the left-hand panel, from 2001, shows mostly green cropland, that is, soybean-only cropland, and on the right, the yellow colors shows… this is soybean/corn rotation, in red is soybean/cotton rotation. So, within little more than a decade, almost the entire landscape changed from single-cropping of soybean to double-cropping of soybean with another crop that needs nitrogen fertilizer. So, a very dramatic change occurring in the Amazon over a very short period of time. We conducted a field fertilizer experiment to try to look at the fate of the nitrogen that would be applied in the corn, that second fertilized phase of this intensifying agriculture. We did this by enlisting the farm and the planters — this is the size of the planters that operate on this farm — and we created an experiment that compared 0, 80, 120, 160, and 200 kilograms of nitrogen. Now, this farm is using itself, on corn, about 80 kilograms of nitrogen/hectare. We wanted to know, what was the limit of intensification, and what could you expect if this large landscape further intensified? To do this, we created a series of plots and we measured two important fates of nitrogen. One was the emission of nitrous oxide, a potent greenhouse gas, to the atmosphere. We expect that to increase following fertilizer application, if what happens here is consistent with what we know from other areas. And, second, we used soil water collectors to measure the concentration of nutrients in the soil that are leeching down and potentially out into streams and into watersheds. Now, I’m just going to summarize these results. This is work in press. Kathy-Jo Jankowski, who was at MBL, is leading this charge. And this table shows the fate of the fertilizer in these different fertilizer treatment levels, right? The 0-200 kilograms/hectare nitrogen. So, what this shows is a lot of the nitrogen that’s applied ends up into he biomass. We actually get more nitrogen going into the biomass then we put on in fertilizer, because we’re taking advantage of the fixed nitrogen from the soy phase of the cropping, the cropping cycle. But what’s remarkable here is that very, very little is leeched, even under high fertilizer. And the N2O release is actually very, very small, it’s in fact smaller that releases from a lot of other cropland — less than one kilogram of nitrogen/hectare. What’s surprising is this right-hand column, that there’s a lot of nitrogen that’s just stuck in these very deep soils, deep down. And this is really an interesting thing that we’re finding across these weathered tropical soils, both in Brazil and in places in Africa, where fertilizer use is increasing. And that is, these soils have the capacity to actually absorb nitrate at depth, and we think that this provides a remarkable capacity to buffer watersheds from the increased impacts of higher fertilizer use, when you have this high infiltrability and when you don’t generate these erosive surface flows. Now, there’s some big questions remaining about this, and we think that this reduces leaching to streams, but we don’t really know, what is the ultimate size of this capacity? Or will we, at some point, after decades of fertilizer use, exceed this capacity? Or are there some circumstances, like high flows, that will actually exceed this capacity? So, it looks like there’s… done right, this intensification can occur with fertilizer uses up to something like 120 or 160 kilos/hectare. Beyond that, the N2O fluxes appear to increase, but the leaching appears to be buffered, at least for now, but we haven’t looked at this over the decades in which this agriculture is likely to remain active in these Amazon croplands. Finally, this is probably important over very large scales. This is an interesting map that we created that shows Fazenda Tanguro, there, circled in black, and the dark green in this case is double-cropping, and the light orange color are these deep oxisols, or latosolos in the Brazilian classification. And what this shows is that, not only has double-cropping expanded dramatically on these deep soils, but there’s potential for it to expand a lot more, because there’s a lot of available soil, a lot of available area on which this could potentially happen. So, those really remain our very large questions. You know, what is the extent that you can intensify production on these deep soils? Brazil has done a very good job over the last decade of reducing deforestation. Here’s a graph from work by Marcia N. Macedo at the Woods Hole Research Center — it simply shows that, since 2006, deforestation has fallen, but cattle production and soybean production has actually continue to increase to some extent. So, what Brazil has done, through a series of policy initiatives and a moratorium on the selling of soybeans from land cleared after 2006, is it’s forced cropping and intensification of cropping onto already cleared land. And the questions really remain are… the questions that remain are, how intensive can that cropping be and with what impacts? So, this has been a success. It remains to be seen the extent to which it can remain a success, because as the landscape starts to look like this… right?… soybeans followed by corn, this looks very much like other parts of the world with intensive grain agriculture, we really still have some big unanswered questions about how intensive cropping can be, how much greenhouse gas emission, how much leaching, what are the limits to this intensification before we see environmental damage, especially to waters and to increases in greenhouse gases in the atmosphere? And finally, there’s an important question remaining… how much deforestation can this landscape sustain until the evapotranspiration through remaining forest becomes insufficient to sustain the rainfall regimes, which are necessary, both to sustain the remaining forest and to sustain the intensifying cropland that’s replacing it? And this is really, I think, the most challenging area of current Amazon climate and forest research. We don’t know the answer to that question, but it’s absolutely vital for Brazil, and other Amazon countries, to move forward to create a system that allows protection of forest, and agriculture at an intensive level that actually allows forest to remain standing in other places. This work at Fazenda Tanguro was made possible by a very large array of both collaborators and funding sources. I especially want to thank my colleagues at MBL, the Woods Hole Research Center, and the Institute for Amazon Environmental Research, and the University of Sao Paulo, on whose work this talk is based. Our funding comes from NSF, NASA, Gordon & Betty Moore Foundation, the Research Foundation for the State of Sao Paulo, the Brazil national funding agency CNPq, and the US Fulbright program. Thank you.