The role of mathematical modelling

Christina Pagel and Kit Yates have an article in the British Medical Journal (BMJ) on the Role of mathematical modelling in future pandemic response policy. It’s part of a series in the BMJ on the UK’s covid-19 inquiry. I have had some concerns about how an inquiry might reflect on the role of mathematical modelling in our response to the pandemic and I think this article makes some important points that I really hope are considered when people assess the role of mathematical modelling, both in the response to the current pandemic, and the role it might play in future pandemics.

Mathematical models are really just representations of reality that will always include assumptions and simplifications. They allow us to consider various possible scenarios that can then be used to inform policy making. Models are also continually evolving as more information becomes available, as our understanding of the system improves, and as techniques and computational resources evolve. They’re clearly not perfect, but they are an extremely useful tool when trying to understand what might happen.

Models can also be wrong. Sometimes this is a natural part of the process of development, sometimes it’s because there isn’t really enough data to constrain the model, and sometimes it is because modellers make mistakes. Modellers can also sometimes have more confidence in their models, and model output, than is actually warranted. Modellers should be willing to admit when models are wrong, or when they make mistakes, but – like most humans – sometimes find it difficult to do so.

The article also stressed that communication is key. This is partly to make clear the strengths, and limitations, of the models, but also to try and ensure that people understand in what way the models were being used. Models are often used to consider numerous scenarios, few of which will be close to what actually materialises. For example, if a model considers some kind of worst-case scenario and the results are then used to inform policy so that we avoid this scenario, then the model wasn’t somehow wrong.

Similarly, models might consider scenarios that cover a range of possible policy pathways. Again, that we don’t end up following pathways close to many of these scenarios doesn’t make these models wrong. In some sense, a model is never strictly right or wrong. What matters is how well it does when the pathway we actually follow is close to one of those considered by the model.

Of course, even if communication is taken seriously and done well, it’s still worth being aware that there are some who engage in bad faith and who will use supposed model failures to promote their agendas. There is little that can be done to avoid this, but this mostly highlights the importance of trying to communicate clearly about model strengths and weaknesses, the motivation behind the modelling, what assumptions were made, and which results we should regard as reliable.

One final thing I was going to say is that it can be important to highlight the different kind of systems that can be modelled. As the article says, one issue with epidemiological modelling is the intrinsic inability of most models to capture important facets of human behaviour. This can limit how far into the future one can realistically model. There are other systems for which this is less of a problem, and so one should careful of thinking that a limitation that applies to one modelling situation applies to all situations.

As usual, I’ve said too much, so I encourage those who are interested to read Christina and Kit’s article.

Links:

Role of mathematical modelling in future pandemic response – BMJ article by Christina Pagel and Kit Yates
Covid Inquiry – A series of articles in the BMJ about the UK’s covid-19 inquiry.

Some news

Royal Observatory Edinburgh

One reason things have been quite quiet here is that I’ve taken on a new role at work. I’ve just started as Head of the Institute for Astronomy, which is part of the School of Physics and Astronomy, University of Edinburgh, but is based at the Royal Observatory, Edinburgh (right).

As you can imagine, it’s taking up quite a bit of my time. I haven’t really had much time to think about possible blog posts, let alone write any. I’m also helping to organise the UK Exoplanet Community Meeting, which is in Edinburgh and starts on Monday, and am trying to prepare the first section of an updated Astrophysics course, which I have to start teaching a week on Monday.

So, a rather busy time and will probably remain so for quite some time. I may still find time to write some posts, but it will probably be rather sporadic.

The importance of science communication

A view from the Cuillin hills.

I’m just back from teaching at a summer school on the Dynamics of Exoplanetary and Solar System Bodies. It started in Inverness and then moved to Skye, which was a lovely place to visit.

I gave a couple of lectures mostly listened to the other lectures, or spent time talking with the “students”, who were PhD students and early career postdoctoral researchers. We did get one day off, which I used to go walking in the Cuillin hills.

One evening event involved a discussion about science communication, or public outreach. Most people agreed that it was important and something that researchers should aim to do. There was a clear recognition that it was important to be aware of the intended audience, and that we should think of innovative ways to communicate. Particular examples were engaging with artists, or musicians, as a way of making science more accessible. It was clear that virtually everyone recognised that effective science communication was tricky and that engaging with the broader public wasn’t as simple as just explaining the science.

What was interesting, though, was that few seemed to have considered the potential pitfalls. During the early months of the pandemic, the UK government would often claim to be “following the science”. This was clearly intended to imply that government policy was being informed by the scientific evidence, but it also gives politicians a potential scapegoat (the scientists) when things don’t go as well as might have been hoped.

Similarly, the reason climate change hasn’t been effectively addressed, the reason some refuse vaccines, or believe the Earth is flat, is not because of poor science communication. This isn’t to suggest that science communication couldn’t have been better, but that there are so many factors that influence what people are willing to accept, or do, that science communication is unlikely to be a dominant determinant.

So, even though I think it is crucial that scientists engage with the broader public, and with policy makers, I also think it’s important to recognise the limitations of what science communication can, and should, achieve. The goal of science communication is mostly to inform, not to directly influence public opinion, or behaviour. The responsibility for the latter lies with those who have a mandate to do so.

I do think that most scientists recognise the latter, but I’m not sure that many realise that a failure of science communication could easily become the mantra if, or when, we don’t deal with various societal problems as well as we had hoped to do so. Even though it may be difficult to completely avoid such a narrative developing, it’s worth being aware of this possibility and doing our best to avoid it.

Considering Catastrophe

There’s been quite a lot of recent coverage of a paper suggesting that climate endgames, such as global societal collapse or human extinction, have been dangerously unexplored. For those who recall the contentious RCP8.5 debate, this may seem a surprising suggestion.

I’ve now had a chance to read the paper, and it mostly seems reasonable. There are possible feedbacks that haven’t necessarily been studied in as much detail as maybe they should have been. There are possible tipping points that we may want to understand better. Plus, we can’t rule out that climate sensitivity may actually end up being on the high side of the range. I agree that these are things that we should be looking at. I’d be surprised, though, if we significantly reduced the uncertainty about these outcomes any time soon.

They then discuss how climate change can impact our societies, suggesting that climate change could exacerbate vulnerabilities and cause multiple, indirect stresses (such as economic damage, loss of land, and water and food insecurity) that coalesce into system-wide synchronous failures. Again, it does seem worth studying these potential outcomes.

A couple of general comments, though, that I may not explain as well as clearly as I’d like. How the impact of climate affects societies depends on many factors, some of which we can influence. Hence, the outcome isn’t deterministic in the same way as it might be for a physical system. Of course, this doesn’t mean that one shouldn’t consider these type of scenarios, but they’ll almost always be conditional. So, we probably have to be careful of suggesting that climate change will cause some kind of societal response, rather than it might do so if we fail to take appropriate action.

The other issue is that we already largely know what we should be doing to reduce the risks associated with climate change. Essentially, collectively we need to limit how much is emitted and we need to invest in developing resilience and reducing vulnerabilities. Nothing wrong, of course, with considering what might happen if we fail to do so, or do so badly. Of course, if we’re already failing to do what should probably be done, it’s hard to see how more (uncertain) information is going to do much to change that.

This Realclimate post probably sums it up pretty well.

To get to the worst cases, two things have to happen – we have to be incredibly stupid and incredibly unlucky.

We can’t really do much about the unlucky part, other than trying to better constrain things that are uncertain. We can try to not be stupid and maybe highlighting what might happen if we are will help to avoid that. On the other hand, as someone who was involved in the RCP8.5 debate about a supposed focus on worst-case scenarios, I’m not convinced that it will have the desired effect.

Links:

Climate Endgame: Exploring catastrophic climate change scenarios, by Kemp, Xu, Depledge, and Lenton.
The never-ending RCP8.5 debate – one of my posts about a contentious discussion about worst-case scenarios.
The best case for worst case scenarios – Realclimate post by Gavin.

FAIL Better

Recently AT reminded us of FLICC, a taxonomy of contrarian tactics introduced in 2007 by MarkH. His bro’s Deck still shines. As a first post of a science blog, it does the job. As a permanent classifier, it deserves some love. I propose my own model – FAIL. First, some light criticism. Mark lists these items: conspiracy, selectivity (cherry-picking), fake experts, impossible expectations (also known as moving goalposts), and general fallacies of logic. Conspiracy does not resemble a tactic. Moving the goalposts doesn’t imply an unrealistic demand, it only switches criteria in a line of argument {1}. Finally, the last rubric could comprise others. Thus it needs specificity and homogeneity.

That can be accomplished by organizing the tactics along strategies, or as Mark himself observes, as means to frustrate legitimate discussion. How to break exchanges? A classic answer stipulates that communication relies on cooperation. Accordingly, four principles guide our contributions: we try to make them informative, truthful, relevant, and perspicuous. Generalizing or rebranding, we get FAIL:

How to FAIL at Climateball
(Click on the image to zoom in)

In this model, an exchange FAILs when factuality, authenticity, importance or lucidity breaks down. All well and good, now comes the fun part – sorting tactics. To that end I cataloged tricks and determined how they succeed in FAILing. See the tree diagram above. Three characteristics underlie each principle. Connected to them, leaves (the white bubbles) exemplify frustrating tricks {2}. Most terms are commonplace. Before explaining the others, let’s put the model into words:

A contribution is factual with true support, a reality-based conclusion, and under reliable judgment. It is authentic if its author is truthful, committed to the position advanced, and habilitated to propound it. It has import if it addresses the question being discussed, implies a functional call to action, and abides by social conventions. It is lucid if the position is correctly specified (think WYSIWYG), to the point, and not self-defeating. This interprets the cooperative part, shown in colored bubbles. The leaves illustrate obstructive tactics. I offer pairs not for completeness but for contrast. Let’s explain the first row.

Mind probes are as easy to presume as hard to verify; the target of tilting at windmills fails to exist or apply; two sorts of unsoundness. Card stacking is one-sided, the BS Artist says stuff: two clashing ways to lack authenticity. The Peddler exploits cues to hook a recurrent pet topic whatever the exchange, whereas squirreling is the art of deflecting in any other direction than the actual topic. Parsomatics refers to overanalysing words so much that nothing of what is said makes sense anymore; the Soapboxer epilogues and sloganeers with vague platitudes. Two conflicting tactics that drown the exchange in gibberish.

You surely know about armwaving, punting, backpedaling, astroturf, charlatanism, double binds, playing the ref, galloping, soapboxing, racehorsing, and moving the goalposts. Incredibilism has been covered years ago at MT’s, peddling here at AT’s, and Eli introduced parsomatics in his aptly named Law Blogging. Four neologisms beg for a short description. Super Leap denotes any inferential overextension (H/T AlbertC). Void Operation does the exact opposite: nil, with self-sealing, circular or question-begging arguments. Infinite Replays echoes endless quests for Climateball perfection, seldom worth it. Fist Pumping represents excessive celebration, like in SpeedoScience.

Labels and examples ought to help develop a feel for the game. They do not caution to play Fallacy Man: it often bends acceptability {3}. They rather show how to spot what breaks. An unsound argument fails because it lacks support. Astroturfing fails because it fakes support. A double bind fails because nothing can satisfy the support requested. A double standard fails to support anything consistently.

Climateball players do not need any model to understand this. You already know the drill: say what you mean, mean what you say, keep on topic, and try not to hog the mic. Cool kids express their critical skills by meming the eternal questions: (1) u sure, (2) who dis, (3) y tho, and (4) wut. So you got no excuse. Nevertheless, the exercise allowed me to get a better grasp of the tactics for my Manual.

Ever tried. Ever failed. No matter. Try again. FAIL again. FAIL better.

Notes

{1} An expectation can be unreasonable without being impossible to meet it. There are many other dimensions to consider: cost, time, overkill, etc. A point-refuted-a-thousand-times comes to mind.

{2} There is an infinity of rhetorical tricks and more ways to classify them. To capture each one in an exhaustive list would be a fool’s errand. I contend that each breaks at least one of the four axes of FAIL.

{3} Fallacy Man fails to realize that his stance falls to the fallacy fallacy. In fairness, the fallacy fallacy succumbs to the fallacy fallacy fallacy, in turn countered by the fallacy fallacy fallacy fallacy. In short: don’t be the Fallacy Guy.

{4} Here is a simple recipe to recognize a trick. Ask yourself: how does it hinder the ball from moving forward? In a way, armchair contrarians know how to run out the clock.

Revisiting causality using stochastics

The Proceedings of the Royal Society A has just published two papers by Koutsoyiannis et al. on revisiting causality using stochatics, the first being the theory paper and the second presenting some case studies. One of the case studies in the second paper considered Atmospheric temperature and carbon dioxide concentration, with their results suggesting that

the results in figure 14 suggest a (mono-directional) potentially causal system with T as the cause and [CO₂] as the effect.

They then concluded that

[h]ence, the common perception that increasing [CO₂] causes increased T can be excluded as it violates the necessary condition for this causality direction

As regular readers will probably realise, this is clearly nonsense. There is overwhelming evidence that the increase in atmospheric CO₂ is due to human-caused CO₂ emissions, and that this increase is the dominant cause of the increase in global surface temperatures.

There probably isn’t much point in going through all the reasons why what is suggested in this paper is almost certainly wrong. I thought I might simply highlight that I started a PubPeer thread about this paper and Gavin Cawley has already posted a couple of useful comments. A PubPeer thread about the Zharkova et al. paper produced quite an extensive comment thread and probably played a role in it being retracted.

You might argue that a paper shouldn’t be retracted just because one of the case studies produces results that are almost certainly wrong. On the other hand, you might also argue that if one of their case studies produces such results that it rather undermines their whole method. You might also argue that it’s rather embarassing that one of the Royal Society’s journals could publish a paper with what is, these days, a very obviously wrong result.

Limits to Growth?

Tom Murphy, who is a physics professor at UC San Diego, runs a blog called Do the Math. Just over 10 years ago, he had a popular blog post asking can economic growth can last?, which I discussed in one of my a blog posts. A couple of days ago, he published his main arguments in a Nature Physics comment titled Limits to Growth, the full text of which you can also access here.

He makes two main arguments. One relates to waste heat, and the other to physical resource limits. When we utilise energy on the Earth’s surface it eventually turns into waste heat that needs to be radiated into space. At the moment this is negligible. However, if energy usage grows at 2.3% per year, it will increase by an order of magnitude every century and this waste heat would become significant within a few centuries. His argument isn’t that this is likely to happen. It is simply an illustration of a real limit. Admittedly, as pointed out on Twitter, he also doesn’t really make clear that this only applies to non-renewable sources of energy, but there are also limits to how much renewable energy we could use.

His other argument relates to physical resources. There must be some limit to our use of physical resources. This doesn’t necessarily limit economic growth because not all economic activity relies on the usage of physical resources. You could imagine a scenario where we reach a stage where economic growth decouples from the use of physical resources. In other words, we reach the limit of physical resources, the usage of which is now fixed, but economic growth continues through activities that don’t utilise these resources.

The problem that Tom Murphy highlights is that this means that economic activity associated with these physical resources will become an ever decreasing fraction of total economic activity. However, these would be activities that are crucial to our survival, such as food and energy. He suggests that this would be ludicrous. For example, if they become effectively free, how do you stop someone from buying it all and raising the price?

What I would be very interested to hear are mainstream rebuttals to the above argument. It could be that mainstream economics just doesn’t consider century long timescales. If so, does that imply that Tom Murphy’s arguments have validity over these longer timescales? Another could be, as the paper suggests, that the finite physical resources will always be associated with a non-negligible fraction of total economic activity. Hence, they essentially act to limit how much overall economic growth is possible. Alternatively, maybe there are ways to have continued economic growth even if those sectors associated with physical resources become a vanishingly small part of the global economy.

Alternatively, if no mainstream economics are willing to discuss the arguments in Tom Murphy’s paper, maybe my commenters can let me know what they think 🙂

FLICC

I should probably be writing about the UK recording temperatures above 40^oC for the first time, but it’s been covered pretty extensively elsewhere. Instead I thought I might briefly mention something that I’ve become more interested in and have been thinking about quite a lot recently. Even though I think science communication is important, I also think it’s becoming much more important to identify, and potentially counter, the spread of misinformation, or disinformation.

In that vein, I thought I would promote Skeptical Science’s FLICC taxonomy. These are techniques of science denial, and the acronym stands for Fake experts, Logical fallacies, Impossible expectations, Cherry-picking, and Conspiracy theories. I suspect many who have followed the climate debate can identify examples of each of these techniques.

For example, promoting the credentials of contrarians even though they don’t really have relevant expertise (F). Claiming that climate change today must be natural because it’s changed before (L). Demanding unrealistic levels of certainty before taking any action (I). Selecting a very short period of a dataset that appears to support your arguments, while ignoring all the data that doesn’t (C). Suggesting that climate scientists are somehow incentivised to promote alarming possibilities (C).

There are, of course, many other examples, and each technique can also be considered in more detail, as illustrated by the figure below. I do think it’s worth being more aware of this and trying to work out how to identify when someone, or a group, is promoting misinformation. I would add that it’s probably worth being careful of assuming something is misinformation just because it appears to satisfy one of the techniques highlighted below. It’s probably better to consider if there is a pattern of spreading misinformation, rather than being too quick to judge on the basis of what might be an individual example.

A bit of a holiday

Lindisfarne castle

You may have noticed that it’s been a little quiet here lately. That’s partly because I’ve had little to say, but also because I’ve been away on holiday with the family. We rented a cottage and spent a week on the Holy Island of Lindisfarne. In case you don’t know, it’s a island off the coast of Northumberland which you access via a causeyway that you can only cross when the tide is low enough.

There’s not a great deal to do on the island, but there is a castle (picture on right) an old priory, a few pubs and cafes, and some really nice walks. You can also cross to the mainland along the Pilgrim’s Way, which is marked by poles and is also only accessible when the tide is low enough.

What was nice was that it would be very quiet either before people could cross the causeyway, or after they had left in order to avoid being trapped for the next four to five hours. I’d mostly use these times to go for a cycle without having to dodge all the other visitors. There was a headland where you could sit and watch the sea birds, look out for dolphins (which I did see on one of the days) and even watch the lifeboat crews try to free a yacht that had managed to get some rope wrapped around its rudder when it sailed over some buoys that were marking lobster pots.

I was also going to comment about the state of the discourse which, as usual, I find somewhat disappointing. However, I’ve probably written enough, so may expand on this at a later stage. I’ll just highlight a tweet from Mark McCaughrean that mostly illustrates the issue that I have. Essentially, I find it difficult to know who to trust when it comes to topics I’m not that familiar with, such as how to implement climate policy, or what is realistic when it comes to energy policy.

It almost feels that everyone is engaging in rhetoric styles that are more aimed at appealing to certain audiences than in providing reliable information about complex issues. Maybe this is simply unavoidable, but I do find that it makes it difficult to sift through things to try and work out what to trust and what to be more dubious about. Maybe this is unavoidable, but I do find it frustrating at times.

Climate change and social justice

The Institute for Strategic Dialogue (ISD) recently released a called Deny, Deceive, Delay: Documenting and responding to climate disinformation at COP26 and beyond. It highlighted a number of people who will be familiar to those who have followed the public climate debate, and used a taxonomy for Discourses of delay presented by Lamb et al. These discourses of delay are various narratives that can be used to argue against and, hence, delay effective cimate action.

There are a number of different climate delaying discourses, but one of them is an appeal to social justice. Essentially, arguing that climate action will have large costs that will pre-dominantly impact the most vulnerable. As highighted in the ISD report, this has led to environmentalism becoming a new front in the culture wars.

This issue is something I have pondered from time to time, but have never quite seemed to express my thoughts as clearly as maybe I should. I also worry that maybe it’s a form of just asking questions, so will acknowledge this in advance. I should also acknowledge that this falls well outside my area of expertise, so some of my terminology may be, un-intentionally, not ideal.

It seems clear that there are social justice issues associated with climate change. Some groups will be more severely impacted than others, and it seems likely that those who’ve contributed least will suffer most. So, it seems that if we want to develop climate policy that is fair, then these kind of social justice issues should be taken into account.

However, the more we focus on these kind of issues, the more we would seem to run the risk of falling into the culture wars and, potentially, validating what are probably disingenuous social justice arguments. For example, those (such as Alex Epstein) who seem to argue that we should expand the use of fossil fuels so as to deal with some of these issues. On the other hand, there probably are perfectly valid arguments for expanded fossil fuel use in some circumstances. Also, I certainly don’t think that those of us who have benefitted from the use of fossil fuels should be telling those who haven’t what they should do.

I’m not sure if I’ve expressed my concern all that clearly. I’m certainly not arguing against highlighting the importance of social justice when thinking about how to deal with climate change. Mostly I’m wondering how you do so without it ending up being counter-productive. Maybe one option is to highlight discourses of delay and identify who is spreading disinformation, and how they’re doing so. Maybe we just should just make the strongest arguments we can and shouldn’t really care about bad-faith actors. On the other hand, being aware can at least help to identify easily avoidable pitfalls.

I should probably stop there, as I’m not really sure what else to say. I’m actually on the train down to Cambridge for a few days, so am going to get out my book and relax for a while. However, if anyone does have any thoughts on this, I would be interested to hear them.