Allan Hsiao’s paper is short (literally five pages plus references), but it’s for the American Economic Association, so it also has its share of equations. The killer table is below, though, and Klement summarises it like this:

Analysing satellite data, he can rank the 30 largest cities by the percentage of infrastructure at risk from sea level rise. The table below shows that for a sea level rise of 1m, Osaka, Jakarta, and Tokyo face the largest possible loss of infrastructure. But once you go to 2m or 3m sea level rise, the cities that are much more at risk are Chinese and Southeast Asian cities, such as Bangkok and Shanghai, which are particularly affected.

Source: Allan Hsaio

The 1.5m tipping point

The notes say that the percentage shows the proportion of infrastructure that is submerged at different levels of Sea Level Rise (SLR). The tipping point seems to be around 1.5 metres, which sounds fine in theory since the base case IPCC projections are for around a one metre increase by 2100, but that doesn’t allow for subsidence (often caused by water extraction), or the possibility of cascading climate change. Hsaoi summarises his conclusions at the start of the article:

How exposed are cities to the threat of sea level rise? I quantify this exposure worldwide with a focus on urban infrastructure. I document three facts. First, Asian cities are highly exposed. Second, poorer cities and neighborhoods are less exposed. Third, exposure accelerates as sea level rise passes 1.5 meters.

The actual method is pretty complicated, and involves a huge amount of data—this is one of those studies that would have been unimaginable in the days before computers. The dataset is also available separately on his website.

Sea level risk

But broadly, he starts with the IPCC sea level rise projections, and uses multiple maps and satellite data to break cities into ‘cells’ of 30 square metres. A set of rules is applied to each cell to decide if it is inundated, and if so, what the impact is on adjacent cells.

The model captures permanent risk under SLR (sea level rise). SLR is a relative quantity: Inundation depends on land elevation above sea level, and so it occurs whether the sea rises or the land sinks. Cities thus experience SLR as the sum of global mean SLR and local land subsidence. ­ Fast-subsiding cities experience each SLR scenario before slow-subsiding cities do.

But the model doesn’t capture risk from storm surges or tidal surges—the modelling for this would be much more complex. But that does mean that you have to allow for the possibility that in practice these maps might be worse.

Infrastructure risk

And here’s a couple of his maps, for the worst affected cities at one metre sea level rise and two metre. On the left: Osaka at a one metre rise in sea level, on the right Bangkok at a sea level rise of two metres.

(Source: Allan Hsaio)

As well as just assessing impact, Hsaio is also interested in the social equity of all of this, and that requires more analysis:

For a given SLR scenario, I compute the percentage of education, health, and transport infrastructure at risk of inundation in each city. Education infrastructure is schools, health is hospitals and clinics, and transport is highways and primary roads… I define infrastructure exposure as the average of these education,health, and transport percentages.

Being a finance guy, Klement wonders about this from the perspective of insurance companies, since, as he points out, there are large chunks of Miami that are largely uninsurable already. Hsaio’s more interested in policy response.

Moral hazard

Hsiao has also done more specific work on Jakarta, which experiences frequent flooding. There are some interesting conclusions here, broadly that a strong government commitment to sea defences (in Jakarta’s case a sea wall) creates a moral hazard, because it

attracts coastal residents, slows inland migration, and lowers the incentives for inland development. The consequence is continued spending on coastal defense and large damages should it fail.

Insurance doesn’t work because places that don’t flood don’t want to pool with places that do (“Zurich doesn’t want to pool with Jakarta”). Alternatives that might place more of the financial burden on people who choose to live in coastal areas might work, but is open to political lobbying. And once you’ve decided to go with sea defences, and people decide to live behind them, you face political pressure to keep on strengthening the sea defences.

Options

Taking the bigger, global picture, Joachim Klement thinks this leaves us with three options.

1. We do nothing and hope that our models are correct in forecasting sea level rise of 1m or less (I call this the ‘what could possibly go wrong’ choice).
2. We continue to accelerate the decarbonisation of our economy to ensure that we have a margin of safety in place, in case our forecasts are overly optimistic (but that would require the coordinated efforts of all countries in the world, including the US).
3. We increase our efforts in climate change adaptation to make sure that when sea levels rise, our infrastructure doesn’t drown.

He seems to think we’ll go for option 1. And judging by our current responses to climate change, he’s probably right.

The normalisation of climate change

But it is maybe also noticing this as part of a wider change—the normalisation of climate change effects in our policy discourse. Similarly, The Conversation had an article by Paul Behrens in the last week about the impact of climate change on food prices. (H/t to Ian Christie for the link). Just a quick couple of extracts from that here:

Climate change is pushing up the prices of the food that we buy and therefore changing what we eat. One-third of UK food price increases in 2023 resulted from climate change, according to research by agricultural economists. This extra cost contributed to food price inflation and the UK’s cost-of-living crisis.

This trend, he says, is likely to continue. But we can make changes to our diets that reduce the impact:

The scientific consensus shows that the biggest opportunity we have for reducing food’s environmental impacts across many countries is increasing the amount of plants we eatand reducing meat and dairy intake… The plant-rich diet we investigated isn’t vegan. It’s not even vegetarian, although it does include a reasonable (and healthier) amount of meat and dairy.

Breaking the silos

The wider point here is that we’ve now reached a point where climate change has become a fact of life. One of the problems with this is that people in different silos are used to thinking of their climate change problems as being different from other people’s, rather than connected.

Another one—this isn’t my point, but it came up in a work meeting and I can’t credit it, at least yet—is that climate change people are used to thinking of themselves as insurgents, or, in Three Horizons language, as the Horizon 2 innovators. But all of this is now firmly in the domain of the people who maintain our Horizon 1 systems, and this requires a whole new language.

—

Thumbnail image of Shirahama, Japan, by Wikimedia contributor Gpwitteveen, CC BY-SA 4.0

A version of this article is also published on my Just Two Things Newsletter.