Can We Detect Tipping Points in the Biosphere?

: Hillebrand: This entire concept of tipping points, of course, is only useful in terms of creating safe operating spaces if you can predict them and if you can actually detect them while or better, even before they are happening. But if you only retrospectively can say, oh, now we have crossed it, then of course it's rather useless. So the main message is not that there aren't tipping points. We can't say that. What we can say is we will have very strong difficulties in finding precursors or even concurrent evidence, which allows us to predict where these thresholds might be.

00:00:47: Hahn: Welcome to Inside Biodiversity. This podcast is hosted by iDiv, the German Centre for Integrative Biodiversity Research. My name is Volker Hahn. I'm head of the communications unit at iDiv and my guest today is Helmut Hillebrand. Helmut is a professor at the University of Oldenburg, and he's the director of the Helmholtz Institute for Functional Marine Biodiversity. Helmut is also a frequent guest and collaborator at iDiv. Today, we will discuss whether or not it's possible to detect thresholds or tipping points in ecosystems. We will talk about the Planetary Boundaries framework and discuss how valuable it is for policy and management. Enjoy this episode of Inside Biodiversity. Okay, Helmut. Very glad to have you on the show today. My first question is how did you become an ecologist and what is your motivation to do ecological research?

00:01:47: Hillebrand: Yeah, I started studying biology because I was actually interested in genetics. But then I realised that ecology is much cooler. So already at the end of my pre diploma, I was quite convinced that ecology is what I would like to do. Then my main courses were mainly in terrestrial, freshwater and marine ecology. I still think that the incredible diversity of how organisms interact with each other, and how they form entities with imagined properties is super fascinating.

00:02:29: Hahn: So those are very complex systems we investigate in ecology. And you decided to research ecological thresholds, which are also known as tipping points. And they're also mentioned in the context of planetary boundaries, safe operating spaces, regime shifts. Some of us may have heard of these terms. What made you look into thresholds in ecology? Were you inspired by these, let's say, public but also scientific discussions on this topic?

00:03:08: Hillebrand: A little bit more fundamentally, I started to realise that we can't play the game that we are looking at, at pristine conditions anywhere on this planet. Ecology has this, um, very strong scientific basis, but we are actually moving into a phase of, um, you could call it global change ecology or, um, post normal ecology, where everything we are studying is already transitioning is changing, and the nature of this change has massive consequences on how we understand these systems, but also how we predict the future of different types of ecosystems, how we Foresee human well-being that is relying on these ecosystems. So understanding the nature of this change. Understanding of how our systems affected in a rapidly changing environment became a fundamental aspect of of my research. And sooner or later, if you're dealing with these types of changes, you come across this debate about shifting regimes or planetary boundaries and so on. And initially, I wanted to know how often do these critical transitions happen, under what conditions that motivated me to to to look at this. I became a little bit worried of how this this entire discussion has been framed and what the political and policy takeaways from this discussion are, which also made me think about potential ways of analysing these kind of critical transitions.

00:05:01: Hahn: Yeah you're right. These concepts have a strong impact in in policy. And they play a big role in in these discussions. But let's as you suggest first talk about the scientific basis. And I think we need to start with defining this term threshold or tipping point in the context of biodiversity. What is a threshold. How is it defined.

00:05:29: Hillebrand: There is no one definition that would be subscribed by everybody. And I think that is very important. As a first message, I think we really have to define for each study we are doing in this context of what exactly we are meaning when we when we use these words. So for me and for the way we have been using this, I often think about an x and y axis. The x axis is a driver that can be simply time, but it can also be warming or another environmental axis. And the y axis is a phenomenon we are interested in. That could be something like a certain emergent property like yield or biomass production or something like that. It could also be a property like the number of species in an assemblage or so. So a threshold for me is a point on the x axis on this driver that leads to a transition in the y axis. So where we are moving from a gradual change in a system property into something that is happening disproportionately large scales. The classical Ethical view of this, for example, could be the point at which a regime shift occurs. That would be a classical example for such a threshold. So the threshold is the point of a driver or a temporal sequence at which something tips. Something changes into a new area of conditions.

00:07:15: Hahn: Can you give one very concrete example of an ecosystem where we have seen such a regime shift, such a tipping point where the system or the regime changed?

00:07:30: Hillebrand: So the most common used examples consider systems that move, for example, from a phytoplankton dominant in shallow lakes to a macro fight dominance or vice versa. So with increasing nutrients, a macro fight dominated lake shifts into a phytoplankton dominate dominated lake, and both of these system states have have positive feedback mechanisms that allow them to to stay in this different state. There's a lot of debate about the transition from coral dominated reefs to algal dominated reefs. That is also along a nutrient axis. And there is also a very classical alternative stable states that is often mentioned between savanna and forests, which is a rainfall gradient at which this transition, of course.

00:08:24: Hahn: So the idea is not only that the transition then is or the response to a global change driver is from the tipping point on, is non-linear, but it also means that it doesn't go back into the old state easily, but it's kind of reinforced through, uh, through positive feedbacks. Is that right?

00:08:47: Hillebrand: That is when we talk about regime shifts and alternative stable states. They are one potential scenario in which a tipping element occurs. There are other examples where this criterion must not necessarily be fulfilled, where it's simply a disproportional change, that is, that is characterising the tipping. And there might well be a way back. As I said, these terms have not a unique and very clear definition, and they have unfortunately also been used quite interchangeably between different studies and in different fields. I think that is that is an inconsistency that ecology has to live with. But as I said in the beginning, I think it's important to have that very clear in the way you are analysing your data as an ecologist.

00:09:41: Hahn: So but just to have this clear again, in your study, the threshold describes a point Where the response of a variable to a driver and an environmental pressure is non-linear, right?

00:10:00: Hillebrand: So you have to be non-linear. It is disproportionately large. So an exponential curve is also non-linear. But it's actually a gradual change. Right.

00:10:09: Hahn: Yeah. Yeah. So there's more of an abrupt. So we can imagine it as an abrupt change.

00:10:14: Hillebrand: A switch into a condition or a type of response that hasn't been there before. Um, we often talk about something like critical transition, or we also sometimes use the change point detection as, as a method where we look at at which point on the x axis does the response change its behaviour and stops being kind of part of a monotonic, gradual change. That is what we had as a definition for this Nature Ecology Evolution paper, where we where we did this meta analysis of meta analysis.

00:10:49: Hahn: Let's talk about this paper. You published it in 2020. It's called thresholds for ecological responses to global change do not emerge from empirical data. You publish that with a couple of colleagues, including Stan Harper from iDiv. And in that paper, you do acknowledge that there are ecosystems where ecological thresholds have been observed. So they they do exist in principle. But you say that they are not ubiquitous. And that was a result of this, uh, meta meta analysis that you did. What exactly did you do in that study?

00:11:29: Hillebrand: So we used 36 meta analyses. All of them have the same driver on the x axis in the same type of response on the y axis. And we looked at the magnitude of this response to the driver Across the studies in each of these meta analysis. Our null hypothesis was that if we have gradual or monotonic shifts in these responses, then along this gradient there will be a gradual increase in the mean and perhaps also the variance of these responses. In the case of a threshold, we would at some point see a new class of responses the higher response magnitude, either to the positive or to the negative. So we devised a number of statistical tests that first of all, told us something about whether the driver is important for the response magnitude. That was obviously the case in most of these meta analysis, and that is why they have been done in the in the first place. And then we had two tests that looked for the advent of new of new response types along this pressure gradient, and one of them never found them, and the other one found it in only a few cases. If you look at these cases a little bit more carefully, you realise that some of them actually found the opposite of what we would expect, and that is that responses became more predictable and smaller with increasing pressure. So that didn't fall into this category either. So at the end. Out of 36 meta analysis, three showed a sign that indeed, at some pressure moment, new types of responses emerged. That, of course, was a highly surprising result. We went on to extend our team included a PhD student who was very good in doing simulation studies. He simulated these meta analysis for us, but with the knowledge of whether there is or there isn't a threshold in these data. So we simulated different cases, and we realised that our statistical tax tests worked extremely nicely. They detected thresholds when they were present. However, even marginal noise in the scattering of the data made this detection fail. So the main message of this paper is we can't really say of how common tipping point type of responses are in this global change context. However, what we can say is we will have a very hard time to detect them before they happen and that of course, has a massive consequence of how this concept is used in global change policy debate, because this entire concept of tipping points, of course, is only useful in terms of creating safe operating spaces if you can predict them and if you can actually detect them while or better, even before they are happening. But if you only retrospectively can say, oh, now we have crossed it, then of course it's rather useless. So the main message is not that there aren't tipping points. We can't say that the absence of a proof is not the proof of absence. What we can say is we will have very strong difficulties in finding precursors or even concurrent evidence, which allows us to predict where these thresholds might be. There's a need to debate this concept in terms of environmental policy and how we envision global change responses.

00:15:48: Hahn: There was a response paper to your paper by Lade et al. They acknowledged your findings. They didn't question your findings, but they said that safe operating spaces, which can be used in management and policy, can still be defined even without knowing a tipping point in the system. Do you agree with that statement, or would the positioning of this safe operating space be completely arbitrary?

00:16:26: Hillebrand: First of all, let me say I'm extremely happy about this comment. It's actually one of two things that happened after this paper. And as we were quite aware of the fact that we were kind of a little bit away from the mainstream argumentation in that field. With our paper, we were actually super happy also to see how constructive and collegial this this interaction went. We also wrote a reply to this comment, which we were very grateful for, for because some of the points we are trying to make in our original paper, we could actually explain a little bit better in our reply to the Steve Lade's paper and his colleagues. I see where they are coming from. I still have the big problem that if we can't estimate where this threshold actually is, we have a very high risk of exerting too much pressure on ecosystems. One of the unfortunate truisms in in ecological management is that if you put up a boundary somewhere, there's always the tendency to use this boundary completely. It's a little bit like speed limits. If there is a sign saying you can drive 80km/h, you will find very few car drivers that only use 70 or 60. The same is true with boundaries such as fishery quotas, for example. They are in the political process, negotiated higher than they originally have been recommended, and then this higher quota also completely used. So having a boundary somewhere for a certain environmental driver will always lead to the complete use of that space. And if we then are not sure how far away from this potential threshold we are, we risk highly negative consequences. The second thing that I fear is that this, this kind of myopic focus on, on tipping points keeps the gradual changes out of focus. So we are having shifting baselines in all of our ecosystems. They are greatly affected by by human actions. So it's a little bit like waiting for this big bang which is not coming. But ignoring that we are on a shifting trajectory all the time.

00:19:09: Hahn: So that means even even while we're below this boundary, we're still impacting ecosystems. And we can have negative responses, even though they're not abrupt, but they're gradual.

00:19:22: Hillebrand: Exactly. And that is if you let people draw a safe operating boundary graph, you often see a very level thing below the safe operating boundary, and then something bad happens and then we are in a different state. But in almost all studies we are doing, we are seeing that even small environmental changes already impact a system. And they these gradual changes accumulate over time and over space. So we have a problem if we if we envision good impacts that are below this boundary and bad impacts that are behind this boundary. This becomes especially important if you deal with what I'm mainly dealing with and that is biodiversity. So what is what is a safe operating space for biodiversity? Is it can we lose some species because they are not so important and we are still below a certain boundary or what? What is how how would we envision a safe operating space for for biodiversity? That's something that we published a essay paper on as kind of a follow up to this, where we kind of try to conceptualise our criticism on this, this type of tipping narrative in biology space.

00:20:49: Hahn: So we're already mixing the scientific results with consequences for management. To summarise that you don't find ubiquitous tipping points in ecosystems in your analysis, and then you also argue that setting such boundaries and safe operating spaces is problematic, because the incentives for management might underestimate the impacts happening at low levels, but that would also apply to systems where there are tipping points. Right. Because even systems with tipping points would have gradual changes below that tipping point. And even then you would argue that this concept is problematic for for policies that. Correct.

00:21:36: Hillebrand: Yeah. And you're right. We should perhaps be very careful about what we are mixing here. Um, so on a general level, I would say thresholds, for example, have some advantages from, for example, from a legal perspective. Right. You can set a certain threshold that is allowed and not allowed. And um, it might make some environmental regulations more binding and more impactful if thresholds are formulated.

00:22:08: Hahn: But but what we what we said is the target the policy target. And we we base it on a scientific finding. If we know that there is a tipping point and we can quantify it.

00:22:25: Hillebrand: Yeah. So the the danger in using tipping points as a central element Ment of environmental policy is one we are not able to predict them for many aspects. There's certainly a great way of predicting when a glacier will be beyond its way of of surviving. But for many biological systems, that is obviously not not easy. As we have shown in our analysis, the second thing is, even in a system which has a tipping element, this kind of expectation of everything below this tipping element is fine is something that I think is not true. And then and that's the third level and that is the more specific level. I am really worried about the applicability of this concept in terms of biodiversity or biological composition. I think that creates a number of scientific, ethical and also political problems that are, for me, very hard to envision of how that could lead to good results.

00:23:41: Hahn: Is the problem maybe that we try to find physical indicators for what we want to have as humans for something that should actually be based on more on our values, and that we should discuss more. What what do we value about the biosphere, about biodiversity? Because what we want to conserve can be different things depending on what your values are. Isn't that true?

00:24:10: Hillebrand: That is certainly true. And biodiversity or emotions as as part of how we deal with biodiversity is certainly also a super important topic. But from a very, very first principle type of thinking, how would you define a threshold or a tipping element in in biodiversity? In principle, you can look at it from two perspectives. One is biodiversity is what is changing and what you are interested in. So what is on your y axis is some emergent property like yield or primary production or something like that. And then you can say, how much diversity can you lose or change before this emergent properties switches into something different? And I think if you have one take home message from this entire biodiversity ecosystem, functioning literature of like three and a half decades, it is that this thinking does not work because all of these species are entangled in multiple emergent properties. And as soon as you think about the multi functionality of these systems, there is no diversity to be lost. This classical type of redundancy that you see, if you focus only on one specific process or one specific property, goes away into something much more gradual, and where much higher percentage of the biodiversity is needed to maintain this, this property or this process. So this idea that we can lose some of the biodiversity without affecting the emergent properties of the system is potentially simply wrong. But even if it would be true, who makes this decision? So even though I'm a marine ecologist, I'm incredibly fond of orchids. So is there anyone who can say we don't need these orchids in these, like, um, calcareous meadows in Europe? Who makes this decision about importance? And do we even have the ethical stance to to actually do that. So that leads to the other question. And that is how much change can there be before biodiversity is tipping into something else? And then you have this. These examples we talked about in the beginning macro feeds the phytoplankton corals to algae, forest to savannah. These are some remarkable cases. And they often rely on the response of so-called foundation species, species that have a habitat defining role in these systems. So what you are essentially saying is, if we are stressing these foundation species to an extent that they are not performing very well anymore, then the system is going into something else. In most other systems where you don't have this foundation type of of species, what you see along a gradient of pressure is. Reorganisation is a change in the identity of species, the type of interactions they have. And for all of these systems, there is no tipping. We have we do a lot of experimental ecology, and we stress the hell out of the organisms we are working with. And they reorganise. They gradually shift into new conditions. So also from that perspective, I think the tipping narrative has no real meaning in this, in this type of context. And that is why we were quite critical about how using it in this biodiversity context, as I said in other contexts, it might be fine, in other contexts it might also have its very strong need, but especially in dealing with biodiversity, I think it is a concept which is a little bit wide.

00:28:26: Hahn: Until now, we have mostly talked about ecosystems on the local and regional level. Now there's also the concept of planetary boundaries. And in this framework there are nine boundaries. One of them is Functional Biosphere Integrity. Or I want to elaborate with you whether we can scale up from your findings to a global level, because one of the fathers of this planetary boundaries framework, Johan Rockström, does acknowledge that there is no scientific evidence for planetary scale tipping points. But still he upholds the boundary for Functional Biosphere Integrity. And the way it's measured is through a control variable that's called Human Appropriation of Net Primary Production, which describes the amount of energy that we extract from ecosystems as food or feed or timber. My question to you is if the threshold and boundary concept is already problematic or difficult to establish on a local level, can we make sense of having such a boundary on the global level?

00:29:47: Hillebrand: Yeah, that has been one of the major criticisms about this planetary boundary framework that they put biodiversity and the biosphere in there. What would this, again, this boundary look like? They will also under different conditions, there will be life on Earth, right. So what is what is actually biosphere integrity? It is very hard to grasp this. And there has been quite harsh criticism around this, this biosphere component of the planetary boundary discussion especially, Um, which has been much more contested than others, like temperature or nutrient cycles or things like that. One of the criticisms has been exactly this planetary scale. What does that mean? A planetary biosphere boundary is that between life and no life? It's very hard to envision, and it's also not really aligning with the fact that most of our impact on, on biodiversity is actually local and regional so far. It will potentially change with accelerating climate change, but pollution, land use change, cuz all of these things are happening at sub planetary scales, at regions, at localities. I don't really see the connection to what I see in my research, or what I think would be good ways of thinking about our impact on the biosphere. The use of something like net primary production is kind of a master variable to to define the integrity of the biosphere is also something that is um, if you if you connect this to this entire nature contribution to people (NCP) framework. Right. It is one aspect of one category of NCPs, whereas many of the other aspects like climate mitigation, uh, water purification, or also the more cultural NCPs, sense of place and so on. All of these things are not not in there. So for me, this biosphere aspect of the planetary boundary debate is still super opaque. It doesn't influence my thinking very strongly, to be honest.

00:32:07: Hahn: What I find problematic about this control variable, Human Appropriation of Net Primary Production is that at the moment, the definition is that if you are below 10% HANPP – that's the abbreviation for it – then you're in the safe zone, the safe operating space. But the the authors also acknowledge that we already crossed that boundary in the 19th century, when we had less than 2 billion people on the planet. And human appropriation means that we extract food from these ecosystems. And if we were to go below 10% again now with 8 billion people on the planet, I figure that would mean starvation on a global scale. So how can you define 0 to 10% HANPP as a safe zone if it's actually a zone of starvation? Or am I getting this wrong?

00:33:10: Hillebrand: Yeah, that is one problem. The other problem is that it also would actually mean that we are all vegan and vegetarian. The even stronger problem is, and that is one of the things that that is really problematic with this planetary scale of this discussion is that Western, the industrialised countries, the HANPP is higher than 100% because we are actually taking the primary production from somewhere else. So planetary wise we might be still below that boundary, but regional, that actually means that we are, uh, have to transport part of this net primary production from one place to another and all colonial, post-colonial aspects of the global trade market. Um, all of this is a consequence of this regional injustice. So what does it tell us when we are planetary wise, below this threshold, if it still is so extremely skewed in its geographic distribution that it leads to massive problems in justice, in the way people can live, and so on and so on. So it's this planetary scale. In this type of argumentation, makes no sense to me. Um.

00:34:39: Hahn: Okay. Helmut, this is a highly complex issue, obviously, because the Earth system is so complex and maybe we can't simply describe it with a handful of metrics on a planetary scale and define exact positions.

00:35:01: Hillebrand: I think, to to be kind of, um, a little bit differentiated. I think publishing this paper in the first, the Rockström paper in the first place, and also the follow up things have Strongly reinforced the discussion about things that the Club of Rome called limits to growth. Right? 50 years ago. So we are we are having these discussions about that. Our planet is not an endless entity, and that we have limited resources and that it's our responsibility to to deal with these resources in a better way than we are doing today. All of this is is something that I'm happy that this has been reinforced by, by this paper. Still. And that is often something that is that that is happening in these kinds of debates. It's not something you can take literally in the sense that it actually gives you a boundary. It doesn't give you a boundary where you can define your actions by. Spy for the German government or the Lower Saxony Agency for Coastal Protection or the European Union. These boundaries do not inform how they have to define their policies, and I think that is an acknowledgement that has to be made, that it is not something that is. It is a warning sign. It is a cry for action, but it is not. And it cannot be transferred into a practical solution for a problem. And that is something that we have to, um, work on together as society and scientists and policy makers in how to find these, um, solutions. We are living in a time where the political agenda is going into a completely different direction and finding ways of showing how important action towards maintenance of biodiversity and towards better ecological status of the environment is something super important. And in that case, it doesn't really help to think about the world as we are still in a safe operating space and somewhere else is a bad area, and we just have to avoid to go to this bad area. As you already indicated, for some of these things, we might have transgressed these boundaries quite a while ago, and we rather have to turn around and see how we can go back instead of still having time in front of the the threshold.

00:37:55: Hahn: I agree, and it's a very difficult exercise for so many reasons. We have to come to a finish. This was very interesting. Helmut, what would you like, very briefly, in a nutshell, what would you like the audience to remember from this conversation? What's your take home message?

00:38:17: Hillebrand: Biodiversity is a very unique and complex thing that defines how this planet works. This complexity allows biodiversity to adapt to new conditions and thinking about biodiversity in binary ways. About below and above a threshold is the wrong way of of thinking about how biodiversity works and how we rely on biodiversity.

00:38:53: Hahn: I hope you enjoyed listening to this episode of Inside Biodiversity. Be sure to follow iDiv on LinkedIn, on X, on Bluesky, and on your favourite podcast app. For feedback, ideas or questions, use #InsideBiodiversity or email us at podcast@idiv.de.