Billion-Dollar Weather and Climate Disasters: Time Series | Hacker News Reader

Billion-Dollar Weather and Climate Disasters: Time Series

ncdc.noaa.gov

103 pointsperfunctory49 comments

1. The graph is apparently adjusted for CPI, but not house prices. House prices especially in big cities has mostly gone up much more than CPI over time. [1]

2. Including only billion dollars climate events does seem misleading and especially prone to statistical bias. Both bias towards local weather around cities, but also for modern days, partly because of 1.

3. Population has gone up and cities expanded so there's now a much larger area to make billion dollars devastation on.

Using percentage of GDP and not setting an arbitrary threshold of a billion dollars would rectify some of these problems. An alternative graph is provided here [2], and on that it's much harder to spot a trend.

[1] https://www.bankingstrategist.com/housing-prices-hpi-vs-cpi

[2] https://www.forbes.com/sites/rogerpielke/2019/11/07/everythi...

Florida is one of the places I look for when it comes to this. According to the NOAA [1] they have not really been able to detect an increase in hurricanes (the natural disaster that semi-regularly hits Florida) due to climate change. Yet, there has been growth in GDP and population. So, the increase cost in Florida is, at least in part, due to the increase in what people have.

If we have more and more expensive things (especially nice buildings) it's going to cost more even when the weather stays relatively the same.

Disclaimer, I'm not making a statement on climate change. I use FL as an example to highlight a different point. In the reference you'll see there is talk of places where observations of change have occurred.

[1] https://www.gfdl.noaa.gov/global-warming-and-hurricanes/

There has not been an increase in frequency of hurricanes, but there has been an increase in the energy carried by hurricanes and hurricanes intensify faster than they used to. It is wrong to imply all of the increase in damages is due to economic growth.

https://e360.yale.edu/digest/atlantic-hurricanes-are-becomin...

> they have not really been able to detect an increase in hurricanes (the natural disaster that semi-regularly hits Florida) due to climate change

Correct, not a significant increase in the number of hurricanes, but they are projecting an increase in destructive power.

> The IPCC AR5 presents a strong body of scientific evidence that most of the global warming observed over the past half century is very likely due to human-caused greenhouse gas emissions.

> it is likely that greenhouse warming will cause hurricanes in the coming century to be more intense globally and have higher rainfall rates than present-day hurricanes.

> the lifetime maximum intensity of Atlantic hurricanes will increase by about 5% during the 21st century

> We also conclude that it is likely that climate warming will cause Atlantic hurricanes in the coming century have higher rainfall rates than present-day hurricanes, and medium confidence that they will be more intense (higher peak winds and lower central pressures) on average.

> they are projecting an increase in destructive power.

Is this being measured by wind forces, tidal surges, etc?

Mostly higher rainfall / flooding (due to warmer waters).

For wind, it sounds modest. They project the strongest hurricanes to be 5% stronger (windier) by 2100 than the strongest hurricanes today.

Of course there are other factors too like sea level rise which isn't factored in I don't think. If your city is already flooded and a hurricane hits, I'd imagine the damage would be worse or last for longer.

The original article was about cost of damage from storms. It was looking at the past. If the atlantic basin has not yet seen an increase in number of hurricanes or intensity of them than that's important when looking at the reason for storm damage costs. We can't yet say the increase in cost is due to hurricanes.

Projections on future increases (which concern me) do not have an effect on already observed costs. I know places where this point is lost.

Seeing as the cause of these disasters is economic in nature I fail to see how this is misleading.

One interpretation implies the weather is getting worse(equally costly stuff). The other implies people just have more costly stuff(same weather pattern).

Both interpretations would be wrong because it's not either or. It's both.

Well, if you want to know the total economic loss due to weather, you might also want to e.g. include $900 million events.

If you want to know whether there is more disastrous weather in general, you probably don't want to completely ignore the vast majority of land. (Where destruction amounting to $1 billion is fairly unlikely)

For almost all interesting questions around changing trends in catastrophic weather's effect on the economy, I'd want a different measurement. What would you think, you could use this for?

Tools like these, while perhaps only answering very specific and limited questions, can be really useful when explaining phenomena to decision makers. More often than not (at least in America), the financial impact carries the most weight, particularly to those who might otherwise not care at all about climate change.

But why not make the data as unbiased as possible? Why not show total cost for weather events as a percentage of GDP?

Or if only large-scale catastrophes have increased, the distribution of weather events?

Or maybe the percentage of houses destroyed or the total acres of land rendered temporarily useless?

Or number of weather-related casualties per capita?

While none of those are perfect or complete measurements, they'd all dodge several issues discussed here.

But what conclusion will decisionmakers draw? That climate change is something we need to address urgently, or that our population densities have grown and maybe we need to consider that in disaster planning?

Side note, because of the way the page is constructed the data isn't part of the backup kept by the Internet Archive.

Programmer but not a web dev here - why not? View Source tells me it's pulling data from https://www.ncdc.noaa.gov/billions/disaster-mugl.xml. It's formatted by the fancy JS graph stuff, and displayed by:

    <div class="multigraph" data-src="/billions/disaster-mugl.xml?disasters[]=drought..

with links to the time series data below the graph. In what way does this not make the data available to the Internet Archive? How could this page (or a hypothetical similar page on my personal blog) have been constructed to (1) make the data available to IA and (2) make it not vulnerable to an HN hug-of-death (it looks like just some static assets with a user-side dynamic Javascript graph that would be trivial to serve to my untrained eyes).

The server seems to be overloaded.

Yeah. The graphic in the article is loaded with ajax so doesn't work from Google's cache either.

It loads on Wayback Machine after a minute.

https://web.archive.org/web/20200105124509/https://www.ncdc....

It's certainly malfunctioning, I'm getting an 'invalid upstream proxy response' error and a 502 Bad Gateway on the error handler. This at around 1500GMT 6 Jan 2020.

Even though the methodology is iffy, the evidence for CO2 and mean temperature increases is solid enough to take it seriously. And if the data is valid and the trend is real, the next decade or two should be interesting.

Solid enough? Its cast iron.

I agree. And it just keeps getting better and better.

Or worse and worse, if we're considering the impacts.

Climate news brings out all the nit picking. We will never reach 100.0000000000% certainty, and even then you can just argue for more zeros.

https://en.wikipedia.org/wiki/Significant_figures

This tells nothing about weather or climate and everything about how rich we've become and where we are able to build today.

It's simply misleading to put it up this way. Even if the weather was getting less extreme you could experience an increase in cost.

This is all the way up there with the claims of oil subsidies by measuring only arbitrary negative externalities with not talk about the positive ones (and no they are not factored into the price)

Future impacts are not simply based on the weather but also people’s building habits. Any prediction ignoring such factors is going to be meaningless.

Edit: But, that does not mean the current impact of climate change on these numbers are meaningless nor that further climate change can’t make the situation worse.

Thats what I am saying.

Not quite, saying the economy is the most important part of this graph is not enough to dismiss the impact of climate change.

What I am saying is that the reason the cost can be bigger even if there were less extreme weather is that we have more stuff and we build places we didn't before. We have more stuff to be destroyed.

You would have so show the same amount of stuff, extreme weather and that this extreme weather is because of climate change to make a case for this to be climate related.

It's the usual sloppy catastrophic thinking that we unfortunately see in this highly policial debate.

Trends generally have multiple causes.

The link between extreme events and high costs plus the link between extreme events and climate change is enough to say climate change is a factor rather than the only factor.

Further, we don’t need to care about the causes to look into mitigation strategies. Banning construction at significant risk from storm surges is simply a good idea even if climate change where a non issue.

Edit: I avoided saying numbers of events because type and intensity could be more important than overall numbers.

It's non-sensical to use climate change as an explanation when it comes to cost for reasons I have already pointed out.

There are multiple obvious reasons for why the cost is going up since the is more to get destroyed before we even get to "climate change".

But there is not a single piece of evidence anywhere that it has anything to do with "climate change" which itself is a completely meaningless and sloppy expression that really lumps all sorts of things together that should be separated.

I am willing to bet that you can't actually find a single concrete piece of evidence that the rising cost have anything to do with climate change.

And using your logic, Holland shouldn't exist at all.

We are becoming better and better at living places that used to be inhabitable for humans. That has a cost because it's more exposed, but over time this will be countered by improved mitigation processes.

I would love to see any concrete demonstrated evidence that climate is an issue here. There is a reason why they call it weather and climate disasters. Highly manipulative.

If you want any evidence:

https://www.nature.com/articles/nclimate1452

“Here, we review the evidence and argue that for some types of extreme — notably heatwaves, but also precipitation extremes — there is now strong evidence linking specific events or an increase in their numbers to the human influence on climate. For other types of extreme, such as storms, the available evidence is less conclusive, but based on observed trends and basic physical concepts it is nevertheless plausible to expect an increase.”

As to climate change, the term is specifically used because it’s far from a uniform change in temperature. The poles warm far more than the equator with massive amounts of heat shifting between them. In some instances this can result in events where winter cold fronts stretching much further south on the same day warm fronts extend into the far north. The global temperature may be above average on the same day you get snow flurries in Texas.

Its used because its unspecific, you can lump any notion whether speculated or demonstrated under that category and just claim whatever you want.

Your example is as vague as they come which proves my point. No specifics and everything gets lumped under one to conclude what they want to. Thats not science thats scientism.

Pretty much my thoughts on the graph. Look at the trend of the 5-year mean: It spiked in 2005, and has been pretty steady since then.

Measuring these disasters with money is about as meaningful as measuring IQ with potatoes.

Why on earth would you say that? Arguments against attempting to avert these disasters will inevitably be based on cost. Presenting the consequences in monetary terms will help counter these arguments.

Because first there’s M0 inflation to take out, and then it’s well known that claims inflate for the same damage over time, and courts award bigger damages and so on.

Dollars are a problematic way to measure it.

They're a problematic way to measure anything, but often they're the best we've got, especially when what we're measuring is economic costs.

It depends on what we want to measure and for what reason.

Here I expect this time series will be used to illustrate consequences of climate change without taking into account that the US' population has grown by 45% since 1980 and that the US' real GDP has grown by 171%, which makes this time series a poor proxy metric especially when there are direct metrics (e.g. number of hurricanes, etc. per year) available.

This sounds interesting and, to my amateur ears, at least worth considering. Is there any possibility of correcting for those factors? Would that satisfy as an indicator of climate change (or lack thereof)?

The fact that more people are building in climate-vulnerable areas is a very good thing to know about — and that’s the kind of data that this kind of series will highlight. If you want historical hurricane data I’m sure you can get that from NOAA. However, that kind of data won’t give you insight into the human impact of that weather in the same way this series does.

number of billion dollar events per billion dollars real gdp should be easy enough to calculate once the page comes back online.

Excluding the "well known" claim, and passively agreeing on inflation, population density might increase underneath + (real) gdp growth per capita might occur underneath (+/-) we might be better or worse at surviving through various catastrophes now (think of wood only buildings in old huricanes). Kind of fascinating having to disentangle all the comingling elements.

For the evaluation of costs of prevention measures it is exactly the right way to measure it, isnt it?

Prevention measure cost dollars are identical to catastrophe damage cost dollars.

Likely because trying to measure using some nth order variable will leave you open to attacks on the quality of your data. Measuring cyclone landfall events, number of cyclones, and cyclone energy are all first order events, nearly impossible to argue with, and certainly impossible to argue against general trends.

But I'm starting to think throwing crap against the wall for people to argue about is now the point since it divides us into tribes. If all we discussed were incontrovertible facts, we'd quickly reach consensus.

You can measure almost anything with money, provided you're careful about it. That's how insurance companies evaluate risk without loosing money. In the case of disasters, it's really quite simple. Somebody has to pay for the damages (incl. life insurance, medical bills, etc.), and you can estimate these costs fairly accurately if you have enough data.

I'm not saying that all values can be compared to monetary value. There are (or may be, depending on your viewpoint) certain kinds of value that are hard to translate to monetary value, for example those that are based on merely ordinal comparisons, those that are highly subjective and for which at the same time there is no market and it makes no sense to assume one, or those that involve context-dependent preferences or other non-standard features like non-transitive preferences with semiorder or interval order representations. However, those rarely play a role in policy making where every implementation of a policy has a cost and different budgets need to be brought into some reasonable equilibrium.

It is meaningful but in a slightly confusing way as a mixed proxy for wealth of the area, density, and severity with the last potentially "rounded down to zero" by appropriate countermeasures.

It does ignore the human costs so one empty mansion getting wrecked vs an entire favella killing thousands for instance the previous looking worse.

Combine that with a rise in development making it look worse without completly mitigating counter measures - it models the fiscals of the disastee only. Important if it is your job to try to assess liabilities and baseline what is cost effective for prevention but not a measure directly of human cost or devestation.

As you can see by the replies, most people can only think in terms of potatoes. If we were smarter we'd do it a better way.

Reply on news.ycombinator.com