5 ms·

"Scientists look for inconsistencies and potential flaws, and try to understand how significant they are, if they can be controlled for or if they completely in

"Scientists look for inconsistencies and potential flaws, and try to understand how significant they are, if they can be controlled for or if they completely invalidate the research they're reviewing" That's giving academics a bit more credit than they deserve. This is a theoretical idea of how they should operate, rather than what they actually do. The problem of urban heat distorting weather station readings over time is very real. It has affected the US weather station network in the past. The problems were completely ignored for decades until climate "denialists" (actually: people outside academia who care about science) physically visited many of the stations, took photos and then wrote a big report which caught the attention of Republicans in Congress. They documented thermometers that had once been in open fields but were now next to air conditioner exhausts, transformers, mobile phone towers, main roads etc. An especially big problem is weather stations that were once in farmland and are now at airports, the biggest heat sources you can imagine! In fact over time there has been a lot of dropout of stations and now many of the remaining stations that report reliably are sited at airports. They also did some work where they found stations that were still properly sited (the US has far more weather stations than needed, so this was possible), and looked at what happened when you used readings only from them. A huge amount of the warming trend simply vanished. In the wake of that a couple of things happened: 1. Some of the stations the report highlighted were shut down. 2. Congress can't agree on much but, spending money is an exception. They funded a brand new weather station network with ideal siting, the latest sensors etc. Really state of the art, and so well done that even the non-academics who wrote the "denialist" report were satisfied with it. This is called the US Climate Reference Network (CRN). So, happy ending right? Unfortunately no. There are several problems that persist: 1. The US CRN output isn't actually used by climate researchers. Instead they have kept on using the old broken network which they now adjust via models to, they claim, fix the problems. This is problematic because these are the very same people who blatantly ignored their own severe data quality issues for decades until bloggers noticed what was happening. 2. The old network was never actually fixed; the government just shut down the bad stations that were highlighted in the original report. This year the people who did that original report did another one where they once again visited weather stations, and they found that basically none of them were in compliance with the siting rules (which are objective, this isn't a matter of opinion). In other words, the claims that standards were raised were deceptive, some rules were written down but aren't being enforced. 3. The US CRN shows no warming trend whatsoever. This is for the entire USA since 2005. This data appears on its face to be incompatible with the claims of runaway warming and an impending crisis/emergency. Take a look and see for yourself. Go here: https://www.ncei.noaa.gov/access/monitoring/national-temperature-index/time-series/anom-tavg/1/0 https://www.ncei.noaa.gov/access/monitoring/national-tempera... and then change the graph to plot "all months" and unselect nClimDiv leaving CRN. You'll see that the temperature trend is flat. "A more mainstream sort of climate change denial you will see is right-wing news outlets saying something along the lines of "it is snowing, so much for global warming" during a snowstorm." I think you've been left behind by recent changes. Attributing single weather events to climate change in either direction is indeed not supposed to be valid, but the new field of "attribution studies" does exactly that by claiming that individual events can be directly connected to climate change. The rules changed and now saying e.g. "it is snowing, that's evidence of climate change" is not only not right wing denial, it's now considered denial to object to such claims! We're constantly being asked to trust that scientists of all stripes are trustworthy, honest, and behave as model scientists should. But when people dig into the details they constantly find non-optimal behavior along with a culture that doesn't seem to care. That's the source of the replication crisis, the source of constantly shifting claims about COVID, etc. Trust is easy to lose: when people notice that scientists aren't holding themselves to the highest standards, but demand us to act as if they do, how is trust restored?

smcl 4y ago
When you plot "all months" you get a very noisy graph which is hard to visually discern any patterns, I guess it's why you say it is "flat". We can't really eyeball trends like that. However despite the noise you can still see that USCRN generally agrees with the ClimDiv. You dedicated most of your post to painting ClimDiv as a fatally flawed network of weather stations which didn't ever get fixed. But it's returning the same kind of results as the new one, so ... > I think you've been left behind by recent changes No. There may be people like you who exist outside academia, but still take an interest into networks of weather stations and the like, but the vast majority of climate change deniers takes are along the lines I described - "it's snowing, ergo suck it libs" This may sound dismissive, but "look at this graph of this measurement you never think about but ignore this dataset, ok and only look at these years, ok now squint your eyes... does THAT look like climate change!?" feels like you're trying to dupe me, not convince me. Combined with the vaguely anti-science conspiratorial tone ("don't believe the scientists, they seem to care, believe me!" - um no?) and allusions to shenanigans over COVID... I'm afraid that I'm not on board.
- origin_path 4y ago
  Re: Trends. You can compute it if you like. People have done that. I don't recall the results but they were close enough to zero to be equivalent. Re: ClimDiv. Yes, but from my comment: "they have kept on using the old broken network which they now adjust via models" So indeed you'd expect them to follow each other closely now. There's no attempt to dupe you - a full treatment of the problems with temperature data is a book sized topic, that's all. The reason for the selection of years is simply because that's all the data that has been collected by the new network. By all means, treat this as a good faith attempt to provoke questions and further research. The CRN is just an example - the problem I'm getting at here is what it reveals about the methodologies and culture. Specifically, how do we know that people who claim to be scientists are actually using the scientific method? That's the topic I responded to. For decades the US data collection system was degrading in ways that created false warming, but it wasn't the people who are paid to study climate who raised the alarm. It was volunteers. That should be a serious problem for any rational person who wants to know true things; scientists are meant to care about measurement error and take steps to minimize it. So you'd want at least some sort of post-mortem process to identify the root causes of this collective institutional failure, but, it didn't happen. Instead data accuracy became politicized and now researchers refuse to use the CRN even though it was built specifically to resolve disputes over data quality! The problem here is actually more severe than this one case. Climatologists were well aware of other problems with temperature data! Data quality problems are common in science so that's not inherently an issue. But the scientific approach is to quantify and then communicate the flaws in data through CIs/error bars, then propagate those through any statistical derivation you do, and finally be honest about what the CIs are telling you even if that's "we can't say much". That's the point you were making originally and we agree. Climatologists don't do this. Instead they think up reasons the data must be flawed, "fix" it via post-processing and then issue new dataset versions without error bars. They don't only do this for data collected long ago. Once they opened that pandora's box they couldn't stop and now temperature readings collected even just 5 years ago are subject to large scale revisions post-hoc. These revisions don't just change individual data points. Entire trends appear or disappear due to the revisions. https://retractionwatch.com/2021/08/16/will-the-real-hottest-month-on-record-please-stand-up/ https://retractionwatch.com/2021/08/16/will-the-real-hottest... https://www.nature.com/articles/nature.2015.17700 https://www.nature.com/articles/nature.2015.17700 http://www.climate4you.com/images/NCDC%20Jan1915%20and%20Jan2000.gif http://www.climate4you.com/images/NCDC%20Jan1915%20and%20Jan... That behavior is considered fraud in other fields and for good reason: the temptation to change the data to confirm your theory is too great if you allow that. Climatologists though, don't consider it fraud, they think it's normal and fine. Sure enough, there are problems like temperatures as measured by satellites and weather balloons not matching the modified temperature timeseries from ground based weather stations. This is the sort of thing you'd expect to see if the ground station data was corrupted but pointing out this divergence is considered "denialism" and verboten. All this opens up lots of meta-scientific questions like, is the scientific method objective? Do the institutions attempt to enforce the usage of it? What even is the scientific method anyway, can you write it down in a way nobody will object to? Clearly here we have a major disconnect: if you did what climatologists do in a physics experiment, your papers would be retracted and you'd be booted out in disgrace. Yet both fields claim to be scientific. the vast majority of climate change deniers takes are along the lines I described - "it's snowing, ergo suck it libs" Sure, and that's indeed not a useful rebuttal to claims of long term trends. I don't know about "vast majority" because I don't know of any surveys that could be used to check that claim. But unfortunately it's also no longer particularly meaningful to check. Climatologists, governments and media are now all too happy to attribute individual forest fires, cold years, hot years etc to climate change. If they can do it then "it's snowing ergo suck it libs" is in fact a valid argument by the scientific institution's own standards. Hence my point: are they really scientific?

smcl 4y ago

When you plot "all months" you get a very noisy graph which is hard to visually discern any patterns, I guess it's why you say it is "flat". We can't really eyeball trends like that. However despite the noise you can still see that USCRN generally agrees with the ClimDiv. You dedicated most of your post to painting ClimDiv as a fatally flawed network of weather stations which didn't ever get fixed. But it's returning the same kind of results as the new one, so ... > I think you've been left behind by recent changes No. There may be people like you who exist outside academia, but still take an interest into networks of weather stations and the like, but the vast majority of climate change deniers takes are along the lines I described - "it's snowing, ergo suck it libs" This may sound dismissive, but "look at this graph of this measurement you never think about but ignore this dataset, ok and only look at these years, ok now squint your eyes... does THAT look like climate change!?" feels like you're trying to dupe me, not convince me. Combined with the vaguely anti-science conspiratorial tone ("don't believe the scientists, they seem to care, believe me!" - um no?) and allusions to shenanigans over COVID... I'm afraid that I'm not on board.

origin_path 4y ago
Re: Trends. You can compute it if you like. People have done that. I don't recall the results but they were close enough to zero to be equivalent. Re: ClimDiv. Yes, but from my comment: "they have kept on using the old broken network which they now adjust via models" So indeed you'd expect them to follow each other closely now. There's no attempt to dupe you - a full treatment of the problems with temperature data is a book sized topic, that's all. The reason for the selection of years is simply because that's all the data that has been collected by the new network. By all means, treat this as a good faith attempt to provoke questions and further research. The CRN is just an example - the problem I'm getting at here is what it reveals about the methodologies and culture. Specifically, how do we know that people who claim to be scientists are actually using the scientific method? That's the topic I responded to. For decades the US data collection system was degrading in ways that created false warming, but it wasn't the people who are paid to study climate who raised the alarm. It was volunteers. That should be a serious problem for any rational person who wants to know true things; scientists are meant to care about measurement error and take steps to minimize it. So you'd want at least some sort of post-mortem process to identify the root causes of this collective institutional failure, but, it didn't happen. Instead data accuracy became politicized and now researchers refuse to use the CRN even though it was built specifically to resolve disputes over data quality! The problem here is actually more severe than this one case. Climatologists were well aware of other problems with temperature data! Data quality problems are common in science so that's not inherently an issue. But the scientific approach is to quantify and then communicate the flaws in data through CIs/error bars, then propagate those through any statistical derivation you do, and finally be honest about what the CIs are telling you even if that's "we can't say much". That's the point you were making originally and we agree. Climatologists don't do this. Instead they think up reasons the data must be flawed, "fix" it via post-processing and then issue new dataset versions without error bars. They don't only do this for data collected long ago. Once they opened that pandora's box they couldn't stop and now temperature readings collected even just 5 years ago are subject to large scale revisions post-hoc. These revisions don't just change individual data points. Entire trends appear or disappear due to the revisions. https://retractionwatch.com/2021/08/16/will-the-real-hottest-month-on-record-please-stand-up/ https://retractionwatch.com/2021/08/16/will-the-real-hottest... https://www.nature.com/articles/nature.2015.17700 https://www.nature.com/articles/nature.2015.17700 http://www.climate4you.com/images/NCDC%20Jan1915%20and%20Jan2000.gif http://www.climate4you.com/images/NCDC%20Jan1915%20and%20Jan... That behavior is considered fraud in other fields and for good reason: the temptation to change the data to confirm your theory is too great if you allow that. Climatologists though, don't consider it fraud, they think it's normal and fine. Sure enough, there are problems like temperatures as measured by satellites and weather balloons not matching the modified temperature timeseries from ground based weather stations. This is the sort of thing you'd expect to see if the ground station data was corrupted but pointing out this divergence is considered "denialism" and verboten. All this opens up lots of meta-scientific questions like, is the scientific method objective? Do the institutions attempt to enforce the usage of it? What even is the scientific method anyway, can you write it down in a way nobody will object to? Clearly here we have a major disconnect: if you did what climatologists do in a physics experiment, your papers would be retracted and you'd be booted out in disgrace. Yet both fields claim to be scientific. the vast majority of climate change deniers takes are along the lines I described - "it's snowing, ergo suck it libs" Sure, and that's indeed not a useful rebuttal to claims of long term trends. I don't know about "vast majority" because I don't know of any surveys that could be used to check that claim. But unfortunately it's also no longer particularly meaningful to check. Climatologists, governments and media are now all too happy to attribute individual forest fires, cold years, hot years etc to climate change. If they can do it then "it's snowing ergo suck it libs" is in fact a valid argument by the scientific institution's own standards. Hence my point: are they really scientific?