NOAA&NASA: 2014 Warmest Year on Record

[cyptomcat]

Ok, overall I understand most of your problems with the data, and it's an improvement that you only used 'fraud' in reference to Mann in your reply and not NASA or someone else. It is important to note that other data sets do fit reasonably with the two data sets discussed in the OP. One example is provided in the Hausfather link (from Curry or me). That is part of my trust in the data set, in addition to lack of direct evidence of any intentional wrongdoing.

There are several specific points I disagree with:

I then simply reported that Goddard believed the same sort of biasing of the data may be occurring on global data sets.

That evidence by Goddard is incredibly small number of samples. That's not scientific analysis, and the effect has been explained for by non-NASA sources. For example Hausfather and Shollenberger describe this in detail. Former's analysis of the adjustment magnitudes show that the adjustments do not change the overall trend.

In that post #9, I linked to a Judy Curry article where she was taking seriously the questions being raised, in part by Goddard, about the US data. So, my point focused on the issue being taken seriously not on one particular guy, Goddard, raising it. That's what makes your last sentence in this post so laughable.

While she took the claim seriously, she also confirmed Goddard's analysis errors in the link you posted and cited or reported so many other scientists or fact-checkers confirming the analysis errors. This was many months ago. Later Hausfather explained for the errors on Curry's own blog. That such wrong analysis is still not fixed by Goddard and being propagated by him and others is anti-science. Nothing laughable about that.

As I said in that post, once you move away from the raw data, you are moving into the realm of interpretation of facts,

As Hausfather explains, raw data contains a lot of biases. I am not comfortable saying raw data is 'fact' since we know for sure raw data is wrong because of the many biases. In this case, we do not know the fact, that's what science is out to find out. Adjustments are science's best effort to actually get to the fact by removing the biases.

I've also more recently linked to a longer report by Watts and D'Aleo which considers the data set questionable:

I have only been through some of the beginning parts, but there is plenty of debatable or outright wrong claims in this document after reading Hausfather and Shollenberger among others. I don't think a messageboard post is the place for me to write a 100-page post about what's wrong with a 200+ page document especially when Hausfather and Shollenberger in addition to other sources (NASA/NOAA/BEST) have already addressed most of it.

 

[AE 87]

@cyptomcat fair enough.

You still over-weight my ref to Goddard. I wasn't using him as proof but simply data to be considered.

Your statement about not changing the overall trend is wrong from some of the data sets I've seen.

I may have time to look at the Hausfather v Watts later in the week/month.

For me, I don't claim the expertise to pass judgment quickly, so it remains an open question. That's been my point throughout.

 

[Animal02]

The links in AE's post do not make me doubt temperature records and in no way add up to fraud as others have said.

The bloomberg link says there was a mistake made. When it was pointed out, NASA confirmed the mistake. Mistakes happen and this is what one would expect the process to look like.
The NASA link is a follow up on the same issue.
Yes, the Goddard link says a lot of things, but the Judith Curry link also says that she disagrees with Goddards criticisms.

Unless your standard is that absolute accuracy is demanded before any science can be believed, this should just be seen as part of the natural process of science.

The fraud is using the surface temperature data as a basis for the claim.

 

[AE 87]

Freeman Dyson's take, fwiw:
http://www.independent.co.uk/enviro...-sceptic-professor-freeman-dyson-2224912.html

 

[GTNavyNuke]

I still like ocean level rise as an indicator of global warming; it collaborates air temperature and sea temperature rise. Given the massive advances in measurement technology and sheer number of measurements (we didn't used to care about world wide temperature much), there is little wonder that there are inconsistencies in the data and how it is processed.

When sea level stops going up, I'll think global warming has stopped.

But it's somewhat academic (tic) since Asian countries won't take care of a far more pressing health concern in air pollution ..... http://www.cnn.com/2015/01/27/asia/asia-air-pollution-haze/index.html .

 

[Animal02]

I still like ocean level rise as an indicator of global warming; it collaborates air temperature and sea temperature rise. Given the massive advances in measurement technology and sheer number of measurements (we didn't used to care about world wide temperature much), there is little wonder that there are inconsistencies in the data and how it is processed.

When sea level stops going up, I'll think global warming has stopped.

But it's somewhat academic (tic) since Asian countries won't take care of a far more pressing health concern in air pollution ..... http://www.cnn.com/2015/01/27/asia/asia-air-pollution-haze/index.html .

I agree about the pollution....I was in Beijing in 2006 and it reminded me of US cities from the 70's. I understand it has only gotten worse.

 

[AE 87]

@cyptomcat
I'd like your response to this, in response to the link you posted, if you have time.

http://www.breitbart.com/london/201...e-this-global-warming-scandal-is-much-bigger/

 

[cyptomcat]

@cyptomcat
I'd like your response to this, in response to the link you posted, if you have time.

http://www.breitbart.com/london/201...e-this-global-warming-scandal-is-much-bigger/

Those animated graphs by Homewood/Goddard are horrible, especially when the limits of the axes change between different snaps of the animation. They should instead plot the old/new on the same graph with the same axes. They should also stop quoting evidence from a few cases, and do an analysis with ALL readings globally. Similar to someone looking at football stats from all GT games instead of 3 GT games. We are lucky, Hausfather has done that, so let's move on to Delingpole's claims on that:

How he goes from this:

The explanations he offers for the basic principles of temperature adjustments are plausible enough. They include things like the Urban Heat Island effect; weather stations which have moved locations; weather stations which appear to give false readings which need to be adjusted in line with their neighbours; changes in measuring equipment; changes in the time of day measurements are taken (formerly in the afternoon, now more usually in the morning,) and so on.

to this:

In other words it’s a case of “move along. Nothing to see here” and “trust the Experts. They know best.”

unbelievable. Someone reads the bold text, and is familiar with how temperature works: (Atl tomorrow for example)

should think "OMG, how much of an adjustment I am going to have to do for this! This is a big deal..." Instead we get "Nothing to see here." Unbelievable...

You’d expect, as a result of this, that recent (ie late 20th century) raw temperature readings from urban areas would be adjusted downwards in order to make them more accurate. Rarely though, is this the case.

Pretty sure this statement is wrong. You can read Hausfather et al. 2013 for that.

 

[AE 87]

Those animated graphs by Homewood/Goddard are horrible, especially when the limits of the axes change between different snaps of the animation. They should instead plot the old/new on the same graph with the same axes. They should also stop quoting evidence from a few cases, and do an analysis with ALL readings globally. Similar to someone looking at football stats from all GT games instead of 3 GT games. We are lucky, Hausfather has done that, so let's move on to Delingpole's claims on that:

How he goes from this:

to this:


unbelievable. Someone reads the bold text, and is familiar with how temperature works: (Atl tomorrow for example)

should think "OMG, how much of an adjustment I am going to have to do for this! This is a big deal..." Instead we get "Nothing to see here." Unbelievable...

Pretty sure this statement is wrong. You can read Hausfather et al. 2013 for that.

Thanks.

Do you know the collection times for each location or are you just assuming that this in part explains the data in these cases?

Why are rural temps at the beginning reported as cooler than measured iyo?

 

[cyptomcat]

Thanks.

Do you know the collection times for each location or are you just assuming that this in part explains the data in these cases?

Why are rural temps at the beginning reported as cooler than measured iyo?

I didn't say it was the TOB that had to do with these specific cases. My point is that looking at the global analysis, TOB is a significant issue along with other issues, and it's more than 'nothing to see here' in the context of a global analysis. I don't think Hausfather's analysis had to do with answering these specific cases, so Delingpole should read it as a global analysis, and it's a lot more than 'nothing to see here' as such.

I took a peek at the cherry picked data, and it seems like the stations they picked were stations with a lot of trouble that had bunch of station moves. No wonder their data got adjusted (usually adjusted according to the regional data from more reliable regional stations according to the algorithms that are publicly available and published.)

 

[AE 87]

I didn't say it was the TOB that had to do with these specific cases. My point is that looking at the global analysis, TOB is a significant issue along with other issues, and it's more than 'nothing to see here' in the context of a global analysis. I don't think Hausfather's analysis had to do with answering these specific cases, so Delingpole should read it as a global analysis, and it's a lot more than 'nothing to see here' as such.

I took a peek at the cherry picked data, and it seems like the stations they picked were stations with a lot of trouble that had bunch of station moves. No wonder their data got adjusted (usually adjusted according to the regional data from more reliable regional stations according to the algorithms that are publicly available and published.)

Thanks.

Just to be clear, I posted this link in response to your claim in post #61 above, "Former's analysis of the adjustment magnitudes show that the adjustments do not change the overall trend" and my reply in post #62.

That is to say, the point was that the adjustments did affect the overall trend of these data sets being adjusted, contrary to your assertion.

By the way, I agree with you and would have preferred if they had kept the y-axis the same in all the plots. I suspect that the problem raised by the failure to do this is ameliorated for most by the size of the y-axis annotations. It seems they did it to emphasize the shift from the cooling trend to the warming trend.

Sorry for my confusion, I got the impression from your previous post that you were citing time of day in which measurements were taken as an explanation for the data actually under discussion.

From what you say here, it seems that you just didn't understand the point the Breitbart author was making. He wasn't saying that the reasons listed by Hausfather were not legitimate. In fact, as you rightly quoted, he begins by conceding that they are "plausible enough." So, you were right, your conclusion that he didn't understand temperature changes over the day was unbelievable and you shouldn't have believed it.

So, let me clarify the point he was making. First, he conceded that all the explanations for adjustments were plausible. However, his point was that when you give plausible explanations of the sorts of adjustments that have to be made but don't concretely explain how they apply to particular data sets under dispute, you are saying, "move along," "nothing to see here," and "trust the experts."

And it's possible that this "trust the experts" strategy works. For, I get the impression, rightly or wrongly, that it's worked on you. You've cited Hausfather and Shollenberger (I'm assuming that you're referring to your earlier links to their blog posts) as responsive to these plots but they do not address the specific data sets.

So, could you give me a link or a source identifying these stations as "stations with a lot of trouble that had a bunch of station moves" or whatever you found? They are listed as at the same basic geographical location. I'm assuming that you are saying that they were each moved from a warmer place to a cooler place, resulting in the earlier temperatures having to be made cooler. Is that right?

Again, thanks for the more direct conversation.

 

[cyptomcat]

Just to be clear, I posted this link in response to your claim in post #61 above, "Former's analysis of the adjustment magnitudes show that the adjustments do not change the overall trend" and my reply in post #62.

That is to say, the point was that the adjustments did affect the overall trend of these data sets being adjusted, contrary to your assertion.

Seems like we had some confusion about those statements, probably misunderstanding on my part. My statement was about the overall global trend referring to Goddard's original analysis (but seems like you were referring to Goddard's specific cases), not a specific location. For individual locations, it's expected that things will change.

So, let me clarify the point he was making. First, he conceded that all the explanations for adjustments were plausible. However, his point was that when you give plausible explanations of the sorts of adjustments that have to be made but don't concretely explain how they apply to particular data sets under dispute, you are saying, "move along," "nothing to see here," and "trust the experts."

And it's possible that this "trust the experts" strategy works. For, I get the impression, rightly or wrongly, that it's worked on you. You've cited Hausfather and Shollenberger (I'm assuming that you're referring to your earlier links to their blog posts) as responsive to these plots but they do not address the specific data sets.

I think you are confused about this one. Judith Curry link you shared in this thread starts with a link to Goddard. That link contains Goddard's analysis of global temperatures. It's this analysis that Hausfather and Shollenberger responded to. Not the individual cases. That's the Goddard analysis that had numerous errors.

So, could you give me a link or a source identifying these stations as "stations with a lot of trouble that had a bunch of station moves" or whatever you found? They are listed as at the same basic geographical location. I'm assuming that you are saying that they were each moved from a warmer place to a cooler place, resulting in the earlier temperatures having to be made cooler. Is that right?

I am not sure that's how the process works. Read the Hausfather link for more details on this. Here is one of the stations moving multiple times:

That's from here:
http://berkeleyearth.lbl.gov/stations/157453

I didn't save the links for all the cited stations last night, so you can find the other stations by browsing the link above.

 

[AE 87]

Seems like we had some confusion about those statements, probably misunderstanding on my part. My statement was about the overall global trend referring to Goddard's original analysis (but seems like you were referring to Goddard's specific cases), not a specific location. For individual locations, it's expected that things will change.

I think you are confused about this one. Judith Curry link you shared in this thread starts with a link to Goddard. That link contains Goddard's analysis of global temperatures. It's this analysis that Hausfather and Shollenberger responded to. Not the individual cases. That's the Goddard analysis that had numerous errors.

I am not sure that's how the process works. Read the Hausfather link for more details on this. Here is one of the stations moving multiple times:

That's from here:
http://berkeleyearth.lbl.gov/stations/157453

I didn't save the links for all the cited stations last night, so you can find the other stations by browsing the link above.

Thanks for the quick response. I agree we may have been talking past each other on that first point. I hope that's been clarified.

Turning to your paragraph beginning, "I think you are confused about this one," it seems that we may again be talking past each other. I admit that some of this arises from a lack of clarity, hopefully not confusion, on my part.

Hausfather was responding generally to the issues raised by Goddard with respect to the need for adjustments in station temperature data. He was not responding specifically about either Goddard's data or the data-sets in Delingpole's article. I get that. Delingpole cites Hausfather as "[a]bout as close as we've got to an attempted justification" for the adjustments in the data set reflected in his charts.

After conceding the plausibility of Hausfather's principles etc, Delingpole says, "The problem with Hausfather's explanations is that though their fine on the theory they don't seem to bear much relation to the actuality of the adjustments that have been made around the world." I think I accurately glossed this point as:

when you give plausible explanations of the sorts of adjustments that have to be made but don't concretely explain how they apply to particular data sets under dispute, you are saying, "move along," "nothing to see here," and "trust the experts."

Delingpole rightly contextualized it within the debate which had been going on for years before Hausfather's post while I bracketed that context as assumed.

Looking at the Concepcion data since you had the data for that one. It does appear that a major change occurs with the shift of the station in about 1978 after which the adjusted data seems to track the raw data. However, the adjusted data from ~1960 to ~1978 still seems odd to me. Any ideas on that?

 

[cyptomcat]

Thanks for the quick response. I agree we may have been talking past each other on that first point. I hope that's been clarified.

Turning to your paragraph beginning, "I think you are confused about this one," it seems that we may again be talking past each other. I admit that some of this arises from a lack of clarity, hopefully not confusion, on my part.

Hausfather was responding generally to the issues raised by Goddard with respect to the need for adjustments in station temperature data. He was not responding specifically about either Goddard's data or the data-sets in Delingpole's article. I get that. Delingpole cites Hausfather as "[a]bout as close as we've got to an attempted justification" for the adjustments in the data set reflected in his charts.

After conceding the plausibility of Hausfather's principles etc, Delingpole says, "The problem with Hausfather's explanations is that though their fine on the theory they don't seem to bear much relation to the actuality of the adjustments that have been made around the world." I think I accurately glossed this point as:


Delingpole rightly contextualized it within the debate which had been going on for years before Hausfather's post while I bracketed that context as assumed.

Looking at the Concepcion data since you had the data for that one. It does appear that a major change occurs with the shift of the station in about 1978 after which the adjusted data seems to track the raw data. However, the adjusted data from ~1960 to ~1978 still seems odd to me. Any ideas on that?

I can look at the old/raw Concepción data if there is a non-irrittating graph you can link somewhere. It's annoying and difficult to look at the animation.

Of course, at the end of the day, we would need to dig through the code, because the algorithms make use of other data besides the raw temperature data (for example IIRC illumination data from satellites, described by Hausfather and others.) Code is public, but it is pretty large.

 

[AE 87]

I can look at the old/raw Concepción data if there is a non-irrittating graph you can link somewhere. It's annoying and difficult to look at the animation.

Of course, at the end of the day, we would need to dig through the code, because the algorithms make use of other data besides the raw temperature data (for example IIRC illumination data from satellites, described by Hausfather and others.) Code is public, but it is pretty large.

No worries. Click on the graph while it's morphing and you should be able to get it on a separate page. Otherwise, it's at the GISS website, iirc. Don't worry about digging through the code, I thought you might know what was going on since you seem so sure that it's right. Correct me if I'm wrong, but the fact that you don't actually know what's happening with the adjustments to the particular station temperatures but believe that the adjustments must be reasonable reflects that you've been persuaded by Hausfather et al. that "there's nothing to see here," "move along," etc.

 

[cyptomcat]

No worries. Click on the graph while it's morphing and you should be able to get it on a separate page. Otherwise, it's at the GISS website, iirc. Don't worry about digging through the code, I thought you might know what was going on since you seem so sure that it's right. Correct me if I'm wrong, but the fact that you don't actually know what's happening with the adjustments to the particular station temperatures but believe that the adjustments must be reasonable reflects that you've been persuaded by Hausfather et al. that "there's nothing to see here," "move along," etc.

Not exactly. First of all, once again, Hausfather's explanation was in reply to the Goddard global analysis, but you are correct that reading that explanation teaches one how the adjustments work along with helping one make better judgment of any adjustment examples. With that knowledge, one can judge the adjustment examples. That's what I did with the 5-10 examples Delingpole provides. If I refused to look up actual data on these stations and moved on, that would have been "nothing to see here, move along". Instead, I looked for more information about the stations, then made a judgment when I had enough information (the breakpoints for the algorithm along with the regional mean data). In my opinion, the adjustments look plausible based on that data. Granted, I worked within the time budget I had, but it was still an evaluation based on the data.

Pausing the GIF is neat. I learned something new. I'll look at the time you referenced for Concepcion when I have time later.

 

[cyptomcat]

Looking at the Concepcion data since you had the data for that one. It does appear that a major change occurs with the shift of the station in about 1978 after which the adjusted data seems to track the raw data. However, the adjusted data from ~1960 to ~1978 still seems odd to me. Any ideas on that?

I had some time to take a peek at this.

Looking at the Berkeley Concepcion link, their algorithm identifies a break point in 1969. This isolates the period of 1960-1969 for this specific station as having data with a bias compared to stations around it (second figure at the link). Likely causes of this can be the unknown switching of equipment, unknown fixing of equipment, or some other cause that is not indicated on the station record. The algorithm calculates the bias from the available neighbors' data (regional) and applies it for the 1960-69 period. That’s why the peaks of 60s have come down from 'old' to 'new'. As a disclaimer, this is from the Berkeley algorithm, which is not the same exact code as NASA, but the math for adjustments are similar (and raw data mostly same), so it should apply to Homewood graphs (NASA GISS) and give us an insight on what happened here.

I wonder if Paraguayans thought that two guys on a GT football messageboard would be analyzing their adjusted temperature record from 50 years ago...

 

[Northeast Stinger]

I had some time to take a peek at this.

Looking at the Berkeley Concepcion link, their algorithm identifies a break point in 1969. This isolates the period of 1960-1969 for this specific station as having data with a bias compared to stations around it (second figure at the link). Likely causes of this can be the unknown switching of equipment, unknown fixing of equipment, or some other cause that is not indicated on the station record. The algorithm calculates the bias from the available neighbors' data (regional) and applies it for the 1960-69 period. That’s why the peaks of 60s have come down from 'old' to 'new'. As a disclaimer, this is from the Berkeley algorithm, which is not the same exact code as NASA, but the math for adjustments are similar (and raw data mostly same), so it should apply to Homewood graphs (NASA GISS) and give us an insight on what happened here.

I wonder if Paraguayans thought that two guys on a GT football messageboard would be analyzing their adjusted temperature record from 50 years ago...

Nerds, all of you.

 

[AE 87]

@cyptomcat
I really appreciate the conversation over the last few posts. Your posts have led me to some links and sources that I had not seen before.

While I think we actually talking more with each other than past each other now, it seems that we may have reached the point that highlights my current agnosticism on whether the data sets are available. You say:

This isolates the period of 1960-1969 for this specific station as having data with a bias compared to stations around it (second figure at the link). Likely causes of this can be the unknown switching of equipment, unknown fixing of equipment, or some other cause that is not indicated on the station record. The algorithm calculates the bias from the available neighbors' data (regional) and applies it for the 1960-69 period.

I find this methodologically disconcerting. I would prefer that you either determine and adjust a dataset based on known issues associated with the data set or just call it a bad data set. It just seems weird to me, and I'd have to think about it more. I think trust or distrust in the data set comes down to how one considers the trustworthiness of the adjustments and the methodologies for calculating them. I think I appreciate the logic of both sides (say, Hausfather and Watts) of this issue, and frankly can't decide on the basis of the descriptions I've read.

So, I guess what I'm saying is this: my take on our conversation is that you have been persuaded by the Hausfather et al arguments in light of your look at the Berkely algorithm etc, and I'm not yet persuaded (see my previous paragraph). I'm not sure that the data is available for what I'd like to see methodologically.

As I've said before, my hesitancy in simply accepting the official position has arisen from things like the climategate e-mails (I don't buy the so-called debunking) and the mainstream reaction to Mike Mann, i.e. that the hockey stick wasn't exposed as inaccurate by the mainstream climate scientists, as well as the propaganda nature of the press releases etc.

In other words, we may have hit an impasse in our conversation, the interpretation of data as data, past which we cannot get.

 

[00Burdell]

I have never understood the climate debate.

There have been many snowball earths and many hell-on-earths over the last billion years.

Is the point of the climate debate to figure out which one is up next because one of them surely is.