Ecosystem Collapse and Sustainable Development
Andrew Dobson, Professor of Ecology and Evolutionary Biology, Princeton University

Roberta Balstad: Thanks very much, Parker, that was great. I had to check my notes again to make sure you didn't have a third degree in the social sciences. Maybe that's next.

The final speaker on this panel is Andrew Dobson, who is professor in the department of ecology and evolutionary biology at Princeton University. He is also, like so many people on this panel, associated with another institution and another set of disciplines. He's an adjunct professor at the Woodrow Wilson School of Public Policy at Princeton. He's director of undergraduate studies, he's a senior fellow at Butler College, and serves on a number of Princeton University committees. In addition, he's associated with a number of organizations in the UK, including a member of the Common Room of Wolfson College, University of Oxford, the British Society for Parasitology, the British Ecological Society, and a number of US organizations as well. His research is focused on population etiology of infectious diseases and the conservation of endangered and threatened species. Andrew.

Andrew Dobson: It's a great honor to be here today, thank you very much for inviting me. For travel reasons I missed yesterday. I spent quite a lot of yesterday at Belize Airport and then an even more gritty time at Houston Airport, so my apologies to yesterday's speakers.

I spent the weekend talking to fishermen and farmers in Belize, talking about what they did, and how they interact with the environment. And in fact it's their interaction with the environment that determines the economies of their lives. On Monday I'll go through the same horror of airports and go to Tanzania and spend time talking to the people who live around the Serengeti. Their lives are also dependent on the natural world that they live in.

So what I'm going to talk about today is how are human lives and the sustainable development of those lives dependent on the natural environment, because as the introduction told you I'm an ecologist concerned with how is the one voting species dependent on the viability of the non-voting species, the other ten to maybe fifty million other species? So the theme is going to be ecosystem services in sustainable development.

Essentially it comes from this idea that the services that we all rely on are supplied by natural ecosystems, they have an input into the human economy which we often forget to quantify. When we do try and quantify it we tend to underestimate its value, and we don't do enough experiments to try and quantify that value. But those services are crucial, it's things we use every day like the air we breathe, the water we get, even though it comes in little plastic bottles, things like flood control, food supply, recreation, aesthetic values right through to disease control.

It's also crucial to understand that in that slide we saw this morning those bottom four-fifths of the world, people who live in poverty, their lives are intimately linked to the environment they live in. They've more dependent upon ecosystem services than most of us, although ultimately as an ecologist you realize that the lives of all species are dependent on services supplied by other species. And the sort of bottom line question when we're worried about mass extinctions is how do the extinctions of some species lead to extinctions of the other? And of course it's one wonderfully naïve perception that the worry with the current extinction crisis is that all these other species are going to extinct and humans will be left. Isn't that going to be terrible? That's not only terrible, it's dumb. If you were betting we humans we go extinct halfway down the list, because we're very dependent on those services. So one of the key questions we need to work on is to understand better how are humans and how are other species dependent on ecosystem services? So I'm going to sketch out some ideas that relate to that today.

So one of the major exercises that's been going on for the last five years was the Millennium Ecosystem Assessment. It tried in many ways to ask this question, how do we classify ecosystem services, because if we can classify them we might be able to quantify them. And indeed the MA two extreme positions, resilient services which are ones where many of the species providing those services could go extinct but the services would keep working. There will be other ones that we would think of as brittle, where as soon as you've lost the species you've lost the service. Now it would be nice to convert that into a more biological or mechanistic classification, and that's partly what I'll lean toward today.

It also gets us towards this bigger question I think, which is relatively unaddressed, that as we lose species when does the whole ecosystem begin to collapse, because the interplay of services have been lost? And there are very few ecological studies of that. You would hope Halliburton or somebody like that would help us sponsor things like that, but that's yet to happen.

Now one way we could quantify services, and this is something that goes back to some work that Osvaldo Sala did and inevitably a book by Hal Mooney, what I'm talking about is the relationship between the number of species times their abundance, which we call biodiversity, and ecosystem function, the efficiency with which some function is undertaken. So if we were talking to these Belizean fishermen ecotourism might be very brittle. If you have an intact ecosystem and everything's present then you're going to get highly efficient, lots of maximum revenue from ecotourism. As you begin to lose charismatic species that service might collapse.

In contrast, you might have a more resilient function that might be something like fishing or agriculture where you can lose lots of species and it might even get better because you're trading off the sort of natural function for some altered function.

Now one thing the Millennium Ecosystem Assessment tried to do was to take a list, and don't write all these down, it's just to illustrate that there a huge number of things we could classify as ecosystem services, the millennium ecosystem function [?] illustrated what happens when you do this by committee, because there are twelve different ways that ecosystems provide functional cultural services, but only two in which they contribute to food, which I find a bit odd because wishing I was the Italian I find that there are twelve different cultural ways of perceiving food at least, but anyway it gets confused like that.

What we then tried to do was to take this list of services and say could we classify them in this spectrum from resilient through to brittle, where essentially we could try and make a little diagram for different types of ecosystems and different types of services to say which ones tend to be brittle, and notice across ecosystems some consistently tend to be resilient, others consistently tend to be brittle. And in fact if you could read this, what tends to happen is that services performed by charismatic species that are high in the trophic level, predators, things like pandas, bald eagles, tend to be brittle services. You go to Belize to see a manatee. If that manatee's not there then that service is gone. You go to Yellowstone to see a grizzly bear or a wolf, you don't go there to see the nemitoads [?[ in the soil. In contrast, services that are photosynthesis, cleaning air and water, those are performed by the boring species that aren't charismatic, but are mighty dear to me, the nemitoads, the bacteria. You can lose lots of those species but those services are still undertaken. So there's some relatively coarse mapping of where you are trophically, at the bottom of the food web you tend to be resilient, at the top of the food web you tend to be brittle. But there will be exceptions to that, but that's a coarse mapping we might be able to make.

So roughly as we go from purification of air, purification of water, carbon sinks, we tend to be at low trophic levels, pollination of local agriculture, recreational and spiritual value, which deeply offending the sociologists I condense down to one thing, because they all tend to require you to have an intact ecosystem, unless you're a golfer when you go to lower trophic level you're aesthetically happy at that level. So we roughly have a mapping, aesthetic goods and services, some food and ?? fish, top trophic levels, plants, oxygen production, CO2, more resilient, and they vary linearly with area, most resilient to the things at the bottom.

Now we could then do the ugly thing, because the ugly part of my personality is a desperate person who wants to be a mathematician to say could we convert these into some functions and do something useful which might tell us about how the services collapse as we begin to lose species. So I have to sort of have economics envy and do some algebra. To do that I can set up a relatively simple function. Let's have something that as we lose species from an ecosystem we will tend to lose the top species at a faster rate than the things they eat, at a faster rate than the plants, and the things that we'll lose at the slowest rate will be the decomposers. And I can sort of try and parameterize that for various ecosystems. I can then take the well known sort of species area relationship, though some people can think of it as a specious area relationship, between the number of species composition of a habitat and its area, and say that as the area declines we would expect to lose a certain proportion of species, and the sort of MacArthur and Wilson rule of thumb for that is if we lose 90% of the area we'd lose 50% of the species, but I'm going to say when we lost 50% of the species we have a much higher probability of losing species at the top of the trophic web than at lower. Or I could discard that trophic assumption and just say that I've lost 90% of the area it's much more likely I've lost the brittle services than the resilient ones. So essentially that would give me that as I lose area, I've converted area into species, and said that very quickly I lose top trophic levels, the other levels go at slower rates, whereas the net reduction of species is on the solid line.

Now that suggests, that creates a hypothesis that said if species map onto trophic levels then as natural ecosystems are eroded we would expect to see ecosystem services not lost at random, but in some hierarchical fashion, losing the aesthetic ones first, ultimately losing the things that all life is dependent on last. The key crunch is to say well how can we put that into models that look at how habitat is converted, and how could go out and do the Halliburton experiment, how could we quantify that, when do we know whether these services are going to break down? And that's the big hard question for ecology which we've probably got to answer by the end of next week, on a realistic time scale, because these are quite urgent things we need to know.

Now our big worries with the world is that we are messing it up. We've heard lots about global climate change. The other major thing destroying the world is global habitat change. When we talk about global it's conversion of habitats from land to agriculture that's a major force of change. And indeed if we're worried about biodiversity 90% of the threat to biodiversity is habitat change in the tropics, climate change is a much smaller threat, and then mainly to species that live in the cooler climates, and indeed the next place I'm going to go when I go from Tanzania is to go up with the Inuit communities up in Alaska. There they're seeing slight losses of biodiversity, but huge breakdown in human societies because climate change there means that the older people in the society can no longer predict what the weather tells them about where the food they eat, the caribou and the musk ox, where that's going to be. So again if we're worried about climate change, it's going to come in from the polls to the equator, if we're worried about biodiversity loss, that's going to go from the equator out to the poles. So let's think about habitat loss and biodiversity loss.

This is habitat loss in action. This is about Trinidad about ten years ago. It's what happens when you used to take slides and you try and convert them to PowerPoints, but this is sort of swidden agriculture. And indeed we can think of agriculture as a process that immensely complicated, and we've seen examples of that. I as a sort of ecologist tend to think of it as relatively simple. It's actually we have pristine habitat, we convert it to agriculture, it lasts as agriculture for a period of time, in which case it becomes degraded, and then succession, the sort of recolonization by natural species slowly coverts it back to natural habitat.

And of course what we're worried about when we're worried about poverty anytime or human population is how much land do we have agriculture and how many people does that support? And indeed how many people living in the agricultural lands and the urban areas does that support? Now that's a relatively simple framework, but it's well that's well understood in a completely different context. Because essentially one of the best set of understood models in biology, you can pretend it's in economics, but it's really in biology, are models for infectious diseases where you have susceptible hosts who become infected, they then recover and remain immune, or lose their immunity and become susceptible. Now we could think of habitat conversion in exactly the same way, it's just you assume humans are some sort of up market fungus, which sort of throws away the intelligent design assumption, but essentially what's happening is we have pristine land, we add humans who convert the land to agriculture, it's then abandoned and may then return to pristine land, or you could grab it before it gets infected and set it aside as a nature reserve. And succession is essentially birth and restoration. And that means we have a model whose dynamics we understand. I'll bring a real nerd now, write it down as a set of really quite simple equations which we can solve, and you can show that its dynamics are just like a disease. You take a forest, add some people, they convert it to agriculture, it's lost to agriculture because it degrades, so it becomes degraded land, which slowly goes back to forest, but the system settles to some equilibrium with a human population sustained by the land under agriculture. Now that's a very grim scenario. It's assumed all those things are relatively independent of each other to begin with. It resembles the types of dynamics we see. This is just data for the Philippines where forest is converted to agricultural land, it's then abandoned, so we get this increase in abandoned land. Similar things in the Pacific coast of the US, and indeed if we come back home to New England we see the agricultural land is abandoned, returns to forest, and is now being reconverted into sort of housing lots for people. So different landscapes around the world are essentially going through this process.

What we can do is then say where will it finally settle out? And where it settles out can be determined simply the time the land lasts under agriculture, the speed with which it recovers. And it's very gritty. If we were under ?? agriculture, land that's converted and abandoned really quite quickly, then a relatively high proportion, about 10% of the landscape, will be protected for other species that supply ecosystem services. The longer the land lasts for agriculture the more of it gets converted to agriculture and the less is left an unconverted land. That's a bit worrying, that means we either have to be very active setting aside land for nature, or we have to recognize that perhaps the agricultural land, or the land where humans live, is dependent on services provided by the other habitat.

Then there are some hints of optimism, and indeed we see some of this in the MA. There's a beautiful set in the assessment section of the Millennium Assessment that looks at what's happened to land conversion, both globally and in Asia. Essentially what this shows you in blue is the amount of land under cropland in 1961. In purple is the amount of land under crops in 2004. But in yellow is the amount of land that would be under crops if we were still producing crops as efficiently as we were in 1961. So the green revolution has been majorly successful in slowing down that rate of land conversion. And indeed if we were to go to southeast Asia you'd notice that if we didn't have the green revolution we'd need more land for agriculture to feed the people of southeast Asia than there is land. So again, lots of that comes from the green revolution. More of it may come from the genetic revolution, suggesting that many environmentalists when they blame the green revolution for progress are actually barking up the wrong genetically modified beanstalk. Nevertheless, this is just changing the transient, it doesn't change the end conditions. We have to do something else for the end conditions, and there we can also look at the green revolution, because the green revolution is very dependent on water. We've heard a lot about water this afternoon, and indeed water and the presence of water may well be dependent on pristine habitat. So if we go to Ranomafana in Madagascar, there's a wonderful dam there. We tend to be negative about dams. This dam has had a major influence on biodiversity. It may be elegant because the French designed it, I guess. It supplies the hydroelectric power for the middle third of Madagascar. Originally they set up the woods, the forests that surround it for utilitarian reasons, they needed that forest so the dam works so they have electricity. They retrospectively discovered it was one of the best places in the world to see lemurs in the wild, there are thirteen species of lemurs there. So it then becomes an ecotourist place, but initially it's set up for water preservation.

So let's take this idea that we set aside lands for water because we need water for agriculture, that we could do the same for other ecosystem services, and put it back into our disease model. So we'll make the productivity of agriculture and the number of humans it can support a function of the ratio of the land under forest and the land under agriculture. So essentially we'll have a really weird sort of disease where how sick you are is dependent on how many healthy people there are there, which isn't too bad. But again we could do the nerdy bits. What I've actually done is to say the proportion of land forested will determine the agricultural productivity. Relatively simple Michaelis-Menton type curve. If I do that and look at where that curve crosses, so how dependent is agriculture on the amount of land forested? As we go here it's more and more dependent. If I didn't have that dependence this would be the net land that would be left pristine as forest. As I increase agriculture's dependence on forest then I need more and more forest to sustain agriculture, though I don't actually see a huge reduction of land under agriculture. Now that assumes a relatively resilient dependence of the agriculture on water. I could make it more brittle by just putting in a function like this, and again the higher up here that this crosses here, this is a more and more brittle service. I could put that back into that framework and I get an even stronger dependence. I have to protect more and more forest to have my functional agriculture. Now that's quite intriguing. It says that once we acknowledge the dependence of converted land on unconverted land, there's a much larger incentive to save it.

And it's slightly scarier than that, because we could also work out what's the size of the human population that this settles down to? Well there we find something that at first sight initially counterintuitive. In the absence of that dependence we'd have a very large human population living in abject poverty. The more we recognize their dependence then we have people living at a better standard, possibly because they're getting both the benefits of the ecosystem services and the benefits of the agriculture services, but a much lower population density. The worrying thing here is the later you realize your dependence on those ecosystem services, then to get back to where you want to be you've got to go back down the slope, just lower population growth so the resources you have are divided amongst more people and you've set up the situation that Partha Disgupta talked about, you've got to start discounting the future negatively to do the restoration to get back where you want to be.

Okay, what have I tried to say? What I've tried to say is essentially encapsulated that we need more interactions between ecologists, economists, sociologists, etcetera. We particularly need to understand better how do ecosystems collapse, and a better understanding of what's the relationship between habitat conversion and ecosystem services? It might be that there's a nice mapping of trophic level diversity onto the services of diversity provided, but we need more information about that. We do need to do the Halliburton experiments of collapsing ecosystems, or restoring them and looking at how ecosystem services change as we do that. But most of all we need to realize that the time we have to do these problems is really short.

Thank you.