Hacker News Reader

Comment by bdamm | Hacker News Reader

NASA testified to the US Senate in 1988 that "higher temperatures can now be attributed to a long-expected global warming trend linked to pollution". [1]

So, with more than 30 years of taking the issue seriously, what do you suppose we should do, from a risk-management perspective?

[1] https://www.nytimes.com/1988/06/24/us/global-warming-has-beg...

View on HN

There's one big issue inhibiting any sort of productive change on this issue.

"A senior U.N. environmental official says entire nations could be wiped off the face of the Earth by rising sea levels if the global warming trend is not reversed by the year 2030. Coastal flooding and crop failures would create an exodus of ″eco- refugees,′ ′ threatening political chaos, said Noel Brown, director of the New York office of the U.N. Environment Program, or UNEP. He said governments have a 10-year window of opportunity to solve the greenhouse effect before it goes beyond human control.

...

The most conservative scientific estimate that the Earth’s temperature will rise 1 to 7 degrees in the next 30 years, said Brown."

All of that is accurately quoted with one exception. That article was published in 1989, the 'point of no return' was 2000, not 2030. And incidentally this year makes 30 years since that article was published. In spite of the "the most conservative estimate" showing an expected temperature rise of 1 to 7 degrees in these 30 years, we've seen the temperature increase by 0.62 degrees since 1989. [2] And global CO2 emissions have spiked far higher and faster than anybody could have imagined due in large part to China's ultra-rapid industrialization.

Imagine you could go back in time 30 years ago armed with all the knowledge of today, and had the power to do whatever you could to try to enact action on climate change. What would you do? I think the most logical thing to do would be to put out pinpoint accurate predictions. When you do this you'd establish a substantial degree of reliability in the field. And then when the predictions start looking increasingly dire, you'd have decades of accuracy to justify action.

And I think we can start this today. We need to focus on accuracy above any and everything else. When you want to enact global change you first need to establish yourself as a reliable authority. Then what we can do is look back and emphasize how accurate our predictions have been. But without reliability, we cannot expect predictions to actually drive any change that would be in any way inconvenient.

[1] - https://apnews.com/bd45c372caf118ec99964ea547880cd0

[2] - https://climate.nasa.gov/vital-signs/global-temperature/

> And I think we can start this today. We need to focus on accuracy above any and everything else. When you want to enact global change you first need to establish yourself as a reliable authority.

Is anyone making more accurate predictions? Most people that disagree with NASA seem to saying anything from it's fake news to blaming other sources. How do you explain the observed warming?

> The most conservative scientific estimate that the Earth’s temperature will rise 1 to 7 degrees in the next 30 years, said Brown."

Hard to tell without units but this line leads me to believe he was talking in degrees Fahrenheit and your comparing it to degrees celcius:

> The difference may seem slight, he said, but the planet is only 9 degrees warmer now than during the 8,000-year Ice Age that ended 10,000 years ago.

No, nobody is making accurate predictions right now and that's the fundamental problem. We want to enact change on future predictions, but the models we're using to predict things have a history that's shown them to be unreliable. Each time a prediction fails to pass it ends up undermining the entire system of climate modeling more and more. We can't expect national bodies to engage in anything other convenient changes based on predictions with a poor track record. This is why you need accuracy above all.

As for the temperature increase, I don't picture things in such a polarized fashion. I think there's little doubt that humans are contributing to climate change. But what really confounds this issue is that climate, even absent humanity, is a very unstable system. These [1] are the data from the ice cores over the past 800k years, so literally before humanity existed. And you'll note a regular and often quite extreme shift in temperatures. For instance check out the end of the ice age before our most recent one - about 130k years ago. In an extremely brief period of time the temperatures skyrocketed something like 14 degrees celcius. We then gradually made our way out of that ice age hitting unprecedented lows dropping some 15 degrees celcius before once again starting an extremely rapid rise about 12,000 years ago which we are now in. And we've yet to hit the 'normal' highs of times past. So even if humans did not exist, we would almost certainly be seeing rapid warming.

As for his comments referencing temperature, I'm also not entirely sure. His comment was somewhat unusual since the last ice age obviously did not only last 8,000 years. It also didn't end 10,000 years ago (though that's not too far off). And it would also be quite unusual for a UN speaker to use Fahrenheit. In this case, I think we can dodge this issue since 0.68C = 1.2F and that's given not only a failure to reverse carbon emissions but an unprecedented spike in them, so the point holds true in either case. Extreme CO2 increase = bottom or below the bottom of "most conservative scientific estimates" in temperature increases? We can't expect nations to engage in urgent action based on data like this.

[1] - https://en.wikipedia.org/wiki/Global_temperature_record#/med...

[2] - https://en.wikipedia.org/wiki/4.2_kiloyear_event

I don't understand why you are doubling down on your misunderstandings here. The actual increase that you yourself quoted falls entirely within the error bounds of the IPCC estimates. It doesn't matter that it is towards the bottom of those error bounds - the whole point of an error bound is that it is a range that captures the uncertainty. To argue about where in the range it falls is to fundamentally misunderstand the purpose of an error bound in the first place.

And the 'conservative estimate' is again a quote from the same person who made the dubious unverifiable 1-7F claim, so is also not relevant compared to the IPCC data. The IPCC range is their best estimate, so it is not a conservative range according to the IPCC report, which is what matters here.

It is sad that you are focusing on the words of one person who was not even directly involved in the study, while ignoring or misunderstanding the published work of hundreds of climate scientists who contributed to the actual numbers which have proved 30 years later to be accurate predictions.

This is a misunderstanding of ranges. Imagine you flip a coin 100 times and you only get 41 heads. Is this coin fair? 41 heads falls just within the 95% confidence interval for a fair coin over 100 flips, so you can answer yes in that regard. But things are not binary. It's all probabilistic. In other words, it's not like 40 heads = rigged coin, 41 = legit.

Your assumed interval is going to be a normal distribution. That means, if your assumption of it being a fair coin is correct, that you'll hit close to the perfect spot (50 in this case) disproportionately more often than you'll hit at the tails ends of your distribution. By contrast, now let's assume the coin was rigged to hit tails about 60% of the time. Now you'd be hitting extremely near that 41 constantly.

Climate predictions are annoying because we can only get one sample at a time on a time frame that generally requires years to decades, so it's really really hard to determine whether we're right or wrong. But what you can safely say is that the further you fall away from your expected value - the less probable it becomes that your hypothesis (of said expected value being true) was accurate. In other words hitting 1 on e.g. a range of 1 to 7 with an expected value of 3.5 is a much less likely scenario than hitting 1 on a range of 0.5 to 1.5 with an expected value of 1.

Dumb question, but how can you assume the ranges offered are a normal distribution?

Having spent several years doing a PhD that was primarily statistical analysis, I can assure you I understand ranges very well. I certainly did not assume it is a normal distribution. There are going to be multiple sources of uncertainty including statistical and systematic errors which have different distributions. Taking all that into account they came up with an overall bound, and the actual number is within that range. That means the data are consistent with the prediction.

And you need to stop quoting that 7F number. That number is not in the IPCC report, which is the thing we need to be comparing against - it is an unverifiable number quoted from one person who appears to have made multiple other errors, as you yourself pointed out (eg ice ages timeframes).

Then of course you know what I said does not strictly require a normal distribution. That the distribution is not normal is perfectly well indicated by the fact that their variance is imbalanced. They have some sort of high skew distribution in expected outcomes. This leads to two points. The first is that there is no model where you would argue that falling at a tail end of your distribution, repeatedly, is supportive evidence of your hypothesis. That's invariably going to be strong evidence for the null hypothesis.

The second is that the IPCC paper was published in 1990. I was using the heat data starting from 1989, as that's when the article I referenced was published. Here are the correct temperature deviations for testing the IPCC hypothesis which starts with 1990 as the baseline:

---

1990 = 0.44

2000 = 0.40

2010 = 0.7

current = 0.8

---

Your argument, which I believe is flawed, does not speak positively of the IPCC data even if we accept it to be true. From 1990 to 2000 the temperature dropped 0.04 degrees, which is of course well outside their 0.2 to 0.5 bound. From 2000 to 2010 it increased by 0.3 degrees, which does hit square on their expected value. From 2010 to present we've seen an increase of 0.1 degree. Unless we see a substantial increase over the next couple of years, we're likely to once again completely fall completely outside the IPCC predictions.

If you'd like to aggregate the data we see a 0.36 increase over 28 years = a decadal increase of less than 0.13, once again completely outside the IPCC range. And of course as we increase the time frames involved, we should be able to tighten that range as variance is reduced, which means we're even further outside their expected results. However, I'd also emphasize that this isn't really radically different than if we hit a decadal increase of 0.2. We're looking at seeing a tail distribution over and over, or something a bit below the tail distribution over and over.

----------------

Okay, but that is all an aside. The reason I am focusing on the director's comments is because politicians and other individuals are not debating the minutia of distribution mathematics, variance, and other such things. They are listening to these well established individuals and their predictions. And so why wasn't the individual who claimed a 1 to 7 degree increase immediately remanded and demoted by the UN? At the minimum why was the AP not immediately contacted by the UN to issue a correction?

In my opinion the answer is because people are being intentionally hyperbolic, mostly with good intentions. The idea is that people won't do anything if you tell them that what we expect is somewhat hotter days and maybe an increase in extreme weather events. And that is probably a correct assumption. So instead they resort to hyperbole preaching apocalyptic scenarios to come, as our UN Director did. But this is an even worse idea since when these 'prophecies' fail to come true it breeds skepticism towards further predictions. Climate progress plays out the tale of the boy who cried wolf!

Though I do have to say even when we just look at the models, something seems to be going wrong. This [1] is an image from the 2000 IPCC report indicating expected temperature changes from 1990 to 2030. We're currently about 75% of the way through that time frame and sitting on 0.36. It's possible that next decade will see an unprecedented ultra-rapid rise in heating, but if it doesn't then we're again looking at yet another complete failure of prediction. I think there is a clear hypothesis to explain what is going wrong, but this is already an excessively verbose post!

[1] - https://archive.ipcc.ch/ipccreports/tar/wg1/fig9-15.htm

He was definitely using Fahrenheit because he was talking in New York, and most of the conversations on global warming in the American sphere was conducted in Fahrenheit until about 10 years ago.

We don't have time machines. Predictions can only be made with models, and the models aren't perfect.

> And I think we can start this today. We need to focus on accuracy above any and everything else

I'm not sure why the onus for absolute accuracy has to be solely on the side that's urging action against climate change. The opposition has no such regard for accuracy or even basic truthfulness - and they've been brutally effective so far in squashing action. Are you really saying numbers are the reason we haven't won this argument yet? Most people don't go by numbers, they go by their gut.

At least the climate change predictions have been in the right direction - temps have actually increased, just not by as much as they had previously thought. EDIT: It seems like the numbers weren't even that far off - other replies have pointed out the discrepancy is largely due to Fahrenheit - Celsius conversion. And climate around the world really has worsened.

> we've seen the temperature increase by 0.62 degrees since 1989.

Let's count ourselves fortunate, and not waste the time we've been given.

> The most conservative scientific estimate that the Earth’s temperature will rise 1 to 7 degrees in the next 30 years, said Brown."

This quote seems suspicious. The only place I see it is in this AP article, as a quote from this one person, and from a bunch of denialist websites that link to this article. I couldn't find any original source for the number (or any indication whether it is Fahrenheit or Centigrade, which of course makes a large difference.)

If the quote was from 1989 then it was presumably referring to the IPCC report of that year, for which Wikipedia quotes:

under [BAU] increase of global mean temperature during the [21st] century of about 0.3 oC per decade (with an uncertainty range of 0.2 to 0.5 oC per decade);

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

which would be 0.6 to 1.5C over 30 years, or roughly 1F-3F. So the 7 degrees seems to be nothing to do with the IPCC report, and it is unclear where it comes from (or even what unit it has, but since he is from NY it seems plausible that it is F rather than C). Perhaps he was misquoted? Or he was quoting extreme outer bounds of the probability distribution. Building a case of inaccuracy on one poorly attributed number seems excessive.

In reality, temperature trends have tracked IPCC predictions remarkably closely, and that fact is far more significant than a single poorly attributed quote.

Noel Brown was the director of the UN Environmental Program in North America. He's going to be sourcing whatever internal figures they're working with. And yeah, given that he is the directory of the program in North America, I expect you're right that he was referencing fahrenheit and not celsius. I assumed celsius as he was described simply as a "senior UN Environment official." Given he directed the program in North America and was likely being interviewed in this context for public consumption, I agree it was probably fahrenheit. This, however, does not really change the point. Hitting 1.2 when you claim that "1 to 7" is "the most conservative scientific estimate" indicates a general tendency of inflation.

The reason I chose this report was it was the earliest authoritative statement I could find on climate change that we could then test against reality. I'm sure you appreciate that this sort of hyperbole is not an isolated incident. Would it really change your mind if I dug up x more authoritative sources of predictions that were completely unrealistic? Even with the IPCC report we ended up barely squeeking by the bottom of their range which was 0.6.

So results are putting us at or below the bottom of most all estimates. Why is this? And going forward, why should we expect the results to be more accurate? I don't think these are particularly radical questions to ask. To be clear I'm completely in favor of mitigating against carbon emissions, even for reasons that have nothing to do with climate change. I enjoy the outdoors. Millions of internal combustion engines, fossil fuel driven refineries, etc do not do pleasant things to the outdoor experience. The only thing I'm not in favor of is people freaking out to the point that they think climate change is "the most probable factor for the possibility of a near-term extinction event." That's nonsense by any measurement and is counter productive to ever actually getting anything done, though you might help spur on a loony or two with a few screws loose.

[1] - https://archive.ipcc.ch/ipccreports/far/wg_I/ipcc_far_wg_I_s...

[2] - https://archive.ipcc.ch/publications_and_data/publications_i...

> Noel Brown was the director of the UN Environmental Program in North America.

He was the head of the New York Office. I'm not sure those two things are synonyms.

A quick search for 'Noel Brown unep director' will resolve any doubts, and incidentally also provides lots more information on him and other commentary he's provided over the decades.

What I'll take away from this: the estimates given in the early days of the debate -- the IPCC was formed in 1988 -- were surprisingly accurate.

Again, why are you focusing on this one person when there is an entire report backed up by hundreds of scientists that has the actual numbers, which are not 1-7F? This makes no sense at all. So one guy misspoke, or was misquoted, or exaggerated. His stuff on the ice age was also nonsense as you pointed out yourself. The important thing is that the actual IPCC prediction was accurate - that is the real "authoritative statement" and can be read at length on line - that is what you should be quoting.

> And global CO2 emissions have spiked far higher and faster than anybody could have imagined due in large part to China's ultra-rapid industrialization.

This is not true. If anything, CO2 emissions were actually slightly overestimated:

The IPCC’s First Assessment Report (FAR) in 1990 [...] featured business-as-usual (BAU) scenario assumed rapid growth of atmospheric CO2, reaching 418ppm CO2 in 2016, compared to 404ppm in observations.

https://www.carbonbrief.org/analysis-how-well-have-climate-m...

>I think the most logical thing to do would be to put out pinpoint accurate predictions. When you do this you'd establish a substantial degree of reliability in the field

they DID, dude. https://www.theguardian.com/environment/climate-consensus-97...

Reply on news.ycombinator.com