I thought I would briefly mention a paper, by Simon Sharpe, that is currently under review and that discusses how we should present climate risks. It was highlighted by Steve Forden on Twitter and suggests that climate risks should be plotted as probability over time. The suggestion is that what we mostly see are impacts (such as global surface temperature) plotted against time. The paper argues that
The ideas, as I understand it, is to consider some thresholds beyond which we might expect some impacts to materialise. We can then present the probability of us crossing these thresholds and how this probability will change with time under various scenarios.
a more appropriate means of assessing and communicating the risks of climate change would be to produce assessments of the likelihood of crossing non-arbitrary thresholds of impact, as a function of time
For example, as the figure on the right illustrates, if we cross 2oC we have a high chance of losing coral reefs, and a reasonable chance of losing Apline glaciers and summer Arctic sea ice. If we warm further, then other outcomes become possible, and become increasing likely if we continue to warm. If we could quantify the probability of these various outcomes, and how this changes with time, then we could inform what we would need to do if we wanted a good chance of avoiding some, or all, of these outcomes. Similarly, we could highlight at what point some of these outcomes become very likely if we do not take action to avoid them materialising.
Although I do think it is useful to consider things in this way, I can see some complications. The probability of various outcomes depends largely on the emission pathway we actually follow. Although we have some control over future emission pathways, it’s not straightforward to determine the actual probability of actually following one. Hence, it’s not clear how we can really quantify the likelihood of crossing non-arbitrary thresholds.
We could, of course, present conditional probabilities (i.e., how the probability of crossing some threshold varies with time given some emission pathway) but this suggests that this framing isn’t quite as straightforward as it may seem. I do still think, though, that this is a reasonable way to consider this. It may be that there is some straightforward way to present this, but I suspect that this is a topic about which nothing is ever quite straightforward.