Global warming--the Data, and serious debate

[thaumaturgy]

Interesting note here:

Note that changes in borehole temperatures (Section 2.3.2), the recession of the glaciers (Section 2.2.5.4), and changes in marine temperature (Section 2.2.2.2), which are not subject to urbanisation, agree well with the instrumental estimates of surface warming over the last century.(SOURCE )

Here's some graphs of Sea Surface Temperature anomalies:

Figure 2.5: (a) Annual anomalies of global SST (bars and solid curve) and global night marine air temperature (NMAT, dotted curve), 1861 to 2000, relative to 1961 to 1990 (°C) from UK Met Office analyses (NMAT updated from Parker et al., 1995). Smoothed curves were created using a 21-point binomial filter to give near-decadal averages. Also shown are the equivalent SST anomalies from the SAR - dashed curve. (b) Smoothed annual global SST (°C), 1861 to 2000, relative to 1961 to 1990, from USA National Climate Data Centre, Quayle et al. (1999) (thin dashed line, includes satellite data); USA National Centres for Environmental Prediction, Reynolds and Smith (1994) and Smith et al. (1996) (thin solid line, includes satellite data, to 1999 only), and UK Met Office (Jones et al., 2001) (thick line). (c) UKMO SST and NMAT anomaly time-series from a 1961 to 1990 average for the Northern Hemisphere. (d) As (c) but for the Southern Hemisphere. Both for 1861 to 2000.

(Ibid)

Now certainly there is room for some possible gauge errors on these as well, owing to heat from ship-based collection, but here's a bit of a detailed explanation (emphases added where appropriate):

Box 2.2: Adjustments and Corrections to Marine Observations. ​

The SST data used here comprise over 80 million observations from the UK Main Marine Data Bank, the United States Comprehensive Ocean Atmosphere Data Set (COADS) and recent information telecommunicated from ships and buoys from the World Weather Watch. These observations have been carefully checked for homogeneity and carefully corrected for the use of uninsulated wooden and canvas buckets for collecting seawater prior to 1942. However, corrections prior to about 1900 are less well known because of uncertainties in the mix of wooden and canvas buckets. Nevertheless, Figure 2.4 provides good evidence that even in the 1870s, SST was little biased relative to land-surface air temperatures globally. Since 1941, observations mainly come from ship engine intake measurements, better insulated buckets and, latterly, from buoys. SST anomalies (from a 1961 to 1990 average) are first averaged into 1° latitude by 1° longitude boxes for five-day periods; the anomaly for a given observation is calculated from a 1° box climatology that changes each day throughout the year. The five-day 1° box anomalies are then aggregated into 5° boxes for the whole month with outlying values rejected, and monthly average anomalies calculated. Further adjustments are made to monthly SST anomalies for the varying numbers of observations in each 5° box because when observations are few, random errors tend to increase the variance of the monthly mean. NMAT data are treated similarly and have quite similar characteristics. However, a variance adjustment to NMAT data is not yet made. NMAT data are also corrected for the progressive increase in the height of thermometer screens on ships above the ocean surface, though no corrections have been made since 1930. Because there are only about half as many NMAT as SST data and NMAT have smaller temporal persistence, monthly NMAT anomalies may be less representative than SST anomalies even on quite large space scales. On longer time-scales, and over the majority of large ocean regions in the 20th century, there is good agreement between NMAT and SST. 19th century NMAT anomaly time-series should be viewed cautiously because of the sparse character of the constituent observations, and regionally varying biases, only some of which have been corrected. (ibid)

So you see, it isn't a matter of drooling semi-moron bureaucrats attempting to bias the data. We have honest, hardworking scientists making great effort to eliminate bias in the data.

And besides, an unmanned bouy in the middle of the ocean will have a tough time encountering an AC vent or hot parking lot.

 

[grmorton]

Funny you should say that. Exactly yesterday they finally installed the 3kW solar Photovoltaic system on the roof of my house!

It will cover 99.9% of the electricity my wife and I utilize.

Of course, we do live in Southern California where that is possible, it isn't always so efficient elsewhere in the world.

You are lucky and get a tax subsidy (at least that is my understanding). Texas doesn 't give any tax subsidies.

 

[grmorton]

Yes, but note the deviation in the last 30 years when the global temps are climbing at a higher rate and solar irradiance is flat.

As I said, there are other reasons for the rise on land. Did you notice the methodology the weather service uses for correcting 'bad' stations? They change the trend. That means, you make the trend match what you expect.

THOMAS C. PETERSON, “EXAMINATION OF POTENTIAL[/font]
BIASES IN AIR TEMPERATURECAUSED BY POOR STATION
LOCATIONS,” American Meteorological Society, Aug, 2006, p. 1078 fig 2]
“The homogeneity adjustments applied to the stations with poor siting
makes their trend very similar to the trend at the stations with good
siting.”

I am afraid this point is becoming far too important. Indeed, in any model input system there are flawed gauges. Are you now going to state that the majority of the temperature gauges that feed into the model are likewise set in a systematic way such that there is a measurable positive bias?

I have looked at the data in many many different states of the US and in 3 different countries. The temperature records all have these lunacies. I posted on Watersville, Wa. Tonight lets look at Walden, NY See the temperatue record there. Notice the unbelievable drop in temperature. I also subtract this record from West Point, just about 20 miles away. Can you tell me that there were several years when there was 10-12 degrees difference in temperature over this short of a distance? (I can keep showing you this kind of data for longer than you can protest it).

That will be somewhat more difficult when assessing sea surface temperature measurement and satellite measurements which are usually not near an ac unit.

Bias is, as you no doubt know, often measurable in a specific direction.

Yeah, I just did a spatial map of the temperature trend over the past 30 years. Land has heated up faster than the oceans, significantly faster. CO2 can't do that unless you say that CO2 only hovers over land.

We will disagree on whether or not Hansen's claim is a strawman. He does this to get funding. Remember, the best way to get someone to fund you is to scare them into thinking that if they don't bad things will happen to you. It is one reason I don't like governmental science. Al Gore is getting rich by scaring people into buying carbon credits from his company. Neat scam

Based on limited time I am still poring over the specific data. I too took some of the various temperature trends recently. I was putting together an example for a statistics training and used temperatures from an historic station in Massachusetts and correlated the data to look at both the R[sup]2[/sup] as well as the F-statistic of the model. Significant in one example, insiginificant F-test in another, both on the continent.

But, again, I am unwilling to draw major conclusions from two data points which disagree.

Fair enough. I will keep posting pictures of crap data. It is not hard to find. I suspect that before this is over, I will have filled up my alloted space for pictures and have to erase them to make room to keep going.

As you know from your years of modelling that if one looks at any two given points in the entirety of the data one is merely looking at the relative noise. Not the trend.

I also know from modeling that models are only as good as the assumptions they incorporate.

Actually in my years as a carbon scientist I did work on hydrogen storage and was in attendance at a conference on the topic in Norway several years ago. I used to be a believer in H2, but much less so today. I agree H2 is not the way.

H2 is a big scam as well. Often what is out there in the popular consensus is wrong. People make fortunes by not going with the crowd.

What I note in your post is that:

1. Carbon based energy is the only reasonable energy source.

2. Carbon based energy is a limited resource

3. We are all going to see, within the next 50-100 years the complete collapse of an energy-based economy

That leaves me with the only conclusion to draw that we are doomed. Is that it?

Sigh, this thing got submitted by accident before I could finish and upload the pictures, I will do that next reply

Humanity will live, but have you ever heard of the reindeer on St. Matthews Island? It is an interesting story.

When we use up the fuel we will be in trouble.

Ultimately we will be "weaning" ourselves, not stopping cold turkey.

I am in favor of weaning, but that isn't what I see people doing.

Answered already. I find it appalling that anyone would set a gauge in a bad place. I also note that this is the nature of any large-scale analytical study. It must be proven out that the majority of these gauges are thus set and inducing a measurable positive bias.

The surveys say that 53% of the stations can have a 2 deg F bias and 13% a 5+ deg bias. I posted that chart, I think

Give me time, man! I will readily agree that you and I can take any study and find "bad gauges" employed.

I am unwilling to assess single data points as if they, by themselves, have specific meaning. That is my statistics training.

Fair enough. I have studied the issue from a raw data perspective and you haven't. Thus, my arguments are going to be new to you and others.

Please, I beg you, I spend my days dealing with data. I've seen bad and good data and I've seen horrid conclusions drawn from anecdotal data. Let me assess the responses.

I work 12 hours per day 4 days per week, then 2 days on my ranch, and have about 2 hours per day to debate--I like debate so I can pour it out. I find it relaxing and educational. When people find something I have done wrong, I learn and am better the next day.

But again, please do not assume anecdotal data will be impressive. I've got about 350 JMP stats software files on my computer right now that, were I to take two data points right next to each other on the x-axis I'd see dramatically different y-values. That's the nature of statistical data. Two neighboring points mean nothing if the overall trend is significant.

It is the hard reality of where R[sup]2[/sup] meets p-value.

Claiming something is anecdotal is far too often used as a cop out to ignore data, so I won't make this anecdotal if you won't make it a cop out. Fair?

 

[grmorton]

Well, that tiny bit of net warming does appear to be statistically significant (assuming I plotted the proper things, I plotted the "Global" column against the Year column. I came out with an F-test p-value of <0.0001
(F-statistic = 134.97, n=357, df 1,356. Now granted the adjR^2 value kinda sucked at a palty 27%, but clearly the trend was significantly "non-zero".

Well, it's off to "meeting town" again. Sorry, back later!

You didn't seem to notice that the ending temperature, in 2007 was about the same as the starting temperature in 1979. That is my point. I have no doubt that the temperature goes up and down, but a trend? not necessarily because today's low is not significantly higher than that of 30 years ago.The fact is that we have more CO2 in the atmosphere today than 30 years ago and we don't have a higher tropospheric temperature. Indeed, it has been cooling down over the past few years, just as the oceans have been cooling down (but the land heating.) And once again, you should know that the late 1990s and early 2000s were a high solar output, one of the highest because we had an exceptionally high number of sunspots and solar flux. Below is a solar flux plot which matches the high on the tropospheric temperature. When the sun's output went down, so did the temperature, back to nearly the starting point. No net gain.

I will now post the pictures I couldn't post earlier about Walden NY and West Point minus Walden. I doubt you will defend the idea that Walden spent years being 12 deg colder than West POint just 16 miles apart. Walden spent time in a glacial age in the 1960s while West Point just 16 miles away was a balmy upper 40's. Does that make sense to you? By the way, I have looked at every single month at Walden for those years. It is in the monthly data.

 

[grmorton]

I will grant the annualized average numbers are problematic. Again, a "gauge" issue. Presumably the IPCC did not rely solely on the data from these stations.

I appreciate that. There is so much of that in the US data and I haven't even yet delved into the Chinese data. I lived in China for a year and a half, speak Mandarin, read a little. I got their data, and as I expected, it is much worse than in the US. I expected this because I know how China works. If the boss says Dec. is the coldest month, it WILL be the coldest month. If the boss in the next town doesn't care or decides January is the coldest month, either reality will reign or January WILL become the coldest month. I will show some Chinese data shortly after going through some more of the US data. Asia is one of the areas that is rising in temperature rapidly, and it is the place where since 1980, they have laid down more cement than any other place on earth. The urban heat island effect is rising, causing the temperatures to rise, and that temperature rise is interpreted as global warming, not interpreted for what it is--more cement on the ground.

But further, not Peterson's point:



Now the treatment of the data should allow for better assessment of "specious" data points. In addition reasonable outliers can be "tested against" (a "Q-test" as I recall).

Now looking at Peterson's Table 1 it does not look like the data is too alarming.

He lists the 'mean annual maximum temperature', which I believe is taking the maximum temperature throughout the year and then averaging that temperature from all the years. I calculated the average of the yearly averages in deg C, and it is no where near the values he posts. So, what he has done is take the average of the highest temperatures each year. That may not show some of the problems that exist in the data from a global warming perspective.

If you run the numbers on the means the means taken together have a standard deviation of less than 1degC which amounts to about a 5% relative standard deviation on the Mean Max for all stations. Standard deviation for the Mean Min is about 10% relative standard deviation.

I dont' think I see a massive problem with these stations.

Oops gotta run, more later.
[/size][/font][/size][/font]

You are only looking at the data in his table. I just did the standard deviation for Eads, Co. It is 1.97 deg C, or 3.5 deg F. I think you are mistaking what his table represents. The average of the annual temperatures is 9.6 C, not 19.8 C.

 

[grmorton]

Indeed. And even if there were a massive problem with them, we'd still have satellite data and other signatures of warming like sea ice coverage and glacier melt.

Well, lets look at the southern Hemisphere Sept ice extent again. Note the upward trend. That seems to be missed by too many people thinking that the ice is all melting down there.

I would also note that since the sun's output dropped, Alaskan glaciers have started to grow again--according to the US Geological Survey.

Alaska[/COLOR] glaciers
MASS BALANCE: For decades, summer snow loss has exceeded winter snowfall." Anchorage Daily News, October 13th, 2008 11:08 PM ]
Two hundred years of glacial shrinkage in Alaska, and then came the winter and summer of 2007-2008
]Unusually large amounts of winter snow were followed by unusually chill temperatures in June, July and August.
"In mid-June, I was surprised to see snow still at sea level in Prince William Sound," said U.S. Geological Survey glaciologist Bruce Molnia. "On the Juneau Icefield, there was still 20 feet of new snow on the surface of the Taku Glacier in late July. At Bering Glacier, a landslide I am studying, located at about 1,500 feet elevation, did not become snow free until early August.

"In general, the weather this summer was the worst I have seen in at least 20 years."

Never before in the history of a research project dating back to 1946 had the Juneau Icefield witnessed the kind of snow buildup that came this year. It was similar on a lot of other glaciers too.

http://www.adn.com/news/environment/story/555283.html

We have more CO2 in the atmosphere than at any time since 1700 and yet, the glaciers are growing. Where is the effect of CO2? Shouldn't it stop those glaciers?

 

[thaumaturgy]

You are lucky and get a tax subsidy (at least that is my understanding). Texas doesn 't give any tax subsidies.

Actually we get a tax subsidy and my employer provides an additional subsidy, but in point of fact SoCal is a good place for solar despite the tax break. It's one of the few places that there is significant sunshine all year long.

But the tax break isn't all that great when you consider how much the system costs. I do it partly for the economics but partly because it's the right and reasonable thing to do.

Part of my "money where my mouth is" attitude.

 

[grmorton]

I just went onto US HCN and found the data for EADS and HOLLY stations (# 52446 and 54076 respectively). Am I doing something incorrect? The annual average graph I got for these two stations look like this:

I may be graphing something incorrectly but I dont' see any annual averages at or near 20deg F.

Please let me know where the data you got is from and if I am missing something.

Thanks,

His chart is average annual MAXIMUM temperature, not annual average temperature


I would also note one thing about Petersons, approach. If we are trying to determine the temperature trend, then we dare not change trends we deem as bad. Peterson's methodology is to tilt the trends of bad stations to match what he deems as good stations. That is what we call cheatin' down here in Texas.

He admits it twice in that article

&#65533;“The homogeneity adjustments applied to the stations with poor siting
makes their trend very similar to the trend at the stations with good
siting.” THOMAS C. PETERSON, “EXAMINATION OF POTENTIAL
BIASES IN AIR TEMPERATURECAUSED BY POOR STATION
LOCATIONS,” American Meteorological Society, Aug, 2006, p. 1078 fig 2


“Again, the homogeneity adjustments applied to the stations
with poor siting make their trend very similar to the trend at the
stations with good siting.” THOMAS C. PETERSON, “EXAMINATION OF POTENTIAL
BIASES IN AIR TEMPERATURECAUSED BY POOR STATION
LOCATIONS,” American Meteorological Society, Aug, 2006, p. 1078 fig 3.

I think it is time to show another air conditioner thermometer. This one from Oregon. I will also show the temperature record of Perry, OK, subtracted from Stillwater Oklahoma. Note that these are YEARLY averages. I want to emphasize that these cities are only 16 miles apart. Note that in Stillwater OK was warmer than Perry by 1 deg from 1899, declining to zero in 1917, and then Perry is warmer than Stillwater by more than 2 deg F for about 3 years. Lets stop here and consider the implications.

Temperature differences drive the winds, and drive thunderstorms. If there was a temperature difference of 1 deg over this short distance, it represents a higher thermal gradient than exists from the equator to the pole in January. That is 0.0268 deg/mile. Now, in 1899 Stillwater to Perry gradient was about .0625 deg/mile. But between 1917 to 1920, the gradient reversed and it became .125 deg per mile the other direction--about 5 to 6 times the average temperature gradient from the equator to the pole. And that gradient was there for an entire year!!!!!!!!!

What would be the meteorological effects of such a temperature difference? Well, winds should blow from the hotter to the colder city, warming it, and these winds would have had to have blown that direction for 3 years from 1917 to 1920. Then the winds must reverse for decades because for 1921 to 1938 Stillwater was hotter than Perry. But then for most of the next 3 decades, the wind should have blown from Perry to Stillwater up until 1974. Then once again the temperature reversed and up until 1992 that case stood, when in 1992 to 1995 or so, Perry once again became hotter.

Yet, we don't see these effects. I chose this area because it is just the area that I grew up in. I know it well. I can assure you that no such meteorological effects are seen there. Smaller temperature differences than this, cause huge thunderstorms and tornadoes in Oklahoma--a very tornado prone area. Thunderstorms lasting for a year or for decades are not seen up there.

I can only draw one conclusion.The data is crap and is not fit for knowing what is happening climatologically.

As I said, I can keep posting data like this from the raw data longer than you can keep believing in global warming. My point is t hat t he data is not good enough to know what the temperature was 100 years ago. Remember the entire warming over the past century is 1.1 degree F. There is far more noise in the raw data than 1.1 deg F.

Are you familiar with signal to noise analysis? If the signal isn't 2x the noise, it is really hard to see the signal. As it is, the noise is much greater than that.

Also note from 1920 to 1929 there is a total of 5 deg F temperature swing from Perry being hotter to Stillwater being hotter. Is that realistic for two towns merely 16 miles apart?

Can we say that the weather bureau is competent when they don't see these problems?

 

[thaumaturgy]

Yeah, I just did a spatial map of the temperature trend over the past 30 years. Land has heated up faster than the oceans, significantly faster. CO2 can't do that unless you say that CO2 only hovers over land.

Correct, but my limited understanding is that the land is where most of the anthropogenic CO2 is coming from. Again, that is merely an aside.

We will disagree on whether or not Hansen's claim is a strawman.

Actually I was referring to your point as a "strawman". Surely Hansen a trained climate expert doesn't believe humanity is capable of controlling weather. He would be perfectly correct in assuming mankind can control mankinds' impact on the weather. That being likely the actual point he was making. Albeit maybe a bit hyperbolically.

Fair enough. I will keep posting pictures of crap data.

Again, you are missing my focus on the statistics. You can point out bad data til the cows come home (literally in your case, perhaps), but the key is that I can take any given graph and find lots of neighboring points that are not on the "best fit line". In fact, most of them won't be on the best fit line.

Rather than merely finding anecdotal data that is bad, please provide me with a systematic analysis. In other words: "more than 50% of the temperature stations are badly located" or some substantiated claim of a systematic bias.

I also know from modeling that models are only as good as the assumptions they incorporate.

Speaking of "cop outs". Sadly in my work on statistics this is a very easy place to hide. Thankfully statisticians have provided us with many ways to determine bias and test against Type I or Type II errors. Knowing we cannot minimize both.

Fair enough. I have studied the issue from a raw data perspective and you haven't. Thus, my arguments are going to be new to you and others.

Grant me the possibility that your argument points will not necessarily be new to me. What is new to me is to have to field countless anecdotal data points and address them.

Again, as a statistically somewhat savvy individual that is a drag.

I work 12 hours per day 4 days per week, then 2 days on my ranch, and have about 2 hours per day to debate

Excellent. I work 10-11 hours a day 5 days a week and debate when I can squeeze it in. It is alarmingly easy for me to find data and respond to it, but it isn't instaneous.

Claiming something is anecdotal is far too often used as a cop out to ignore data, so I won't make this anecdotal if you won't make it a cop out. Fair?

I don't believe I have taken it as a "cop out" yet. You have yet to do more than merely find examples, numerous though they may be, of single data stations showing data that does not fit along the same line.

Here's an example, from some unknown source just to make a point:


The spread around the data in the top graph is rather large which means I can find a point in which the X-value is 200 and the Y is about 10 and then move over to an x-value of 210 and a y of about 230 or so. The graph trend doesn't move "negatively" but indeed these two data points would indicate that. That is why it is important to know the overall statistics. In the top graph the correlation coefficient is 70%.

It is wholly anecdotal if I were to simply focus on a series of points that lie along a straight line from left to right. The fact that the ensemble of points shows a trend of some statistical significance and measurable p-value of the F-test indicates that the trend is likely real.

I don't mean to be pedantic or speak down to you on this, but what we are dealing with is the ensemble of data.

While individual points are problematic and it is always a good thing to poke around and eliminate the bad gauges, when the data is in the graph one has to look at the overall picture.

It is possible to easily find 35% of the data points on that top graph to be not on the line. Do we go through and query each one?

 

[thaumaturgy]

His chart is average annual MAXIMUM temperature, not annual average temperature

That still seems strange because in neither Holly nor Eads do I find a preponderance of 20degF temperatures, which would surely have biased the data downwards, rather than keep it up to 50 to 60deg F as in the graph I made. Did I miss the months with 130degF temperatures that offset these?

I would also note one thing about Petersons, approach.

I hope we come back to the Eads and Holly graphs since they were in response to your earlier graphs of the data set. Either you and I were working off different sets or our Excel spreadsheets performed quite differently.

I think it is time to show another air conditioner thermometer.

They have become the highlight of my day.

As I said, I can keep posting data like this from the raw data longer than you can keep believing in global warming.

And you are perfectly free to do so. Until I see that you have posted a substantive amount indicating a systematic bias in the overall data then you will indeed have to clear the caches several times.

 

[thaumaturgy]

You didn't seem to notice that the ending temperature, in 2007 was about the same as the starting temperature in 1979. That is my point.

Did I miss something? The data indicates a statistically significant "non-zero" line. The p-value says I am 99.999% likely to not make a Type I error in rejecting the null hypothesis.

While I am not happy with the correlation coefficient, it surely indicates that the temperature trend is not flat from 1979 to 2007.

That's why the statistics is important. Please address the statistics and not the individual points.

I have no doubt that the temperature goes up and down, but a trend? not necessarily because today's low is not significantly higher than that of 30 years ago.

Oh my. You are going to have to take that one up with Fisher and the statisticians. What part of the F-test do you disagree with? The fundamentals? Or the specific application? Did the stats software I use mess up a sum or squares term or something? The software is from SAS and is a highly recognized and respected organization.

You are free to calculate the numbers for yourself. What F-statistic for this trend do you come up with?

And once again, you should know that the late 1990s and early 2000s were a high solar output

Didn't we go over the Max Plank data earlier? I thought there was no significant increase solar irradiance since about 1950 or so.

Yup, here it is again:

But again, there is reason to believe this number is hard to pin down, so I might be willing to acquiesce that the question is still open, not that it is a "decided" issue.

You can, I suppose, take it up with the MPI.

I will now post the pictures I couldn't post earlier about Walden NY and West Point minus Walden. I doubt you will defend the idea that Walden spent years being 12 deg colder than West POint just 16 miles apart.

Now do me a favor and take the individual data points and do a quick R^2 and a quick F-test on them. Is the trend flat (ie p>0.05)? What is the y-intercept of that trend? 2degF? 3degF?

Again, I am indeed bothered by the spike in the early 60's. And if the IPCC had based their global warming analyses on the Westpoint data station I would be bothered, especially if they started collecting data in 1960 and stopped in 1963.

Walden spent time in a glacial age in the 1960s while West Point just 16 miles away was a balmy upper 40's. Does that make sense to you? By the way, I have looked at every single month at Walden for those years. It is in the monthly data.

Sounds like a bum "gauge". I guess someone should do a full gauge R&R study on it.

You know I have a pH meter in the lab that started giving me bad data the other day. Turns out the probe finally crapped out. I had to change it out. Get a new one. Things break. Gauges go bad.

The best we can do is be mindful of this. That's why statisticians and "Statistical Process Control" people are out there.

 

[thaumaturgy]

Yeah, I just did a spatial map of the temperature trend over the past 30 years. Land has heated up faster than the oceans, significantly faster. CO2 can't do that unless you say that CO2 only hovers over land.

Just to balance this point a bit, here's a graphic of a map generated by a NASA satellite showing temperature anomalies:

(SOURCE)

Looks like there's significant warming across water bodies as well. Perhaps less or more depending on circulation of aircurrents and buffering capacity of seawater (Revelle Buffer Factor).

 

[grmorton]

Can we look at more than merely single data points? Statistics works precisely because it is an ensemble rather than an analysis of individual data points.

I will readily grant that there is no perfect gauge and the potential for human error is immense. Clearly we can easily find bother "outlier data points" as well as bad gauges and incompetent scientists and technicians.

The key is can we find a systematic bias. In other words, how many badly placed stations are there in relation to the whole? How does this affect the massive efforts that were undertaken to avoid ship-contamination (heat) in sea-surface temps and nowadays the autonimous and unmanned (ie not-ship-bound) data bouys and probes?

69% of the US stations are sited in ways that would cause more than a 2 deg F bias. The entire global warming is 1.1 over the last 100 years. Noise and bias are greater than signal.

In all efforts at full disclosure, I used to be a volunteer for a local California initiative for a single payer healthcare system. Not a government run medical facility but a government funded (ie single payer) system staffed by non-government doctors.

In the effort of full disclosure, I am dying of a cancer which could have been caught early, but wasn't because I lived in the UK during those years and lived under national health care. The government decided that the PSA test was too expensive so they didn't do it. If I had lived here in the US, it would have been caught early and maybe I wouldn't now be dying because of socialized medicine. If you haven't lived under it, you don't have a clue how bad it really is. If you are well, it is fantastic. If you need a heart bypass, you get to wait an entire year in the UK. If you need a neurologist for your child, you get to wait an entire year before seeing one. I got into a cab one day in Aberdeen Scotland and the cabbie said it was his first day back after his heart bypass. I laughed and said that that should have taken a year. He said, "no, it took two years to get the bypass". Shocked I asked what happened. He said he was on the waiting list for a surgeon and 3 weeks before his surgery, his surgeon died. The government socialized medicine saw fit to put him to the back of the line for another surgeon.

My neighbor, another American expat, had a cough. The national health care wouldn't do an xray--too expensive. They gave her cough medicine. By the time she flew to the states to get the cough checked out, it was too late to cure her lung cancer. She died.

They don't do regular mammograms on women over 65. It is like they say, you have lived long enough, go die. I have British friends who are convinced that the NHS put their elderly parents to death. They were doing well in the hospital and died for reasons totally mysterious, unknown and undiscussed.

When you go to an NHS dentist, they don't take xrays of your teeth. They only polish the top front four teeth.

I hope you like what you campaign for. I will forever vote against anyone wanting national health care. It ain't worth a bucket of warm spit.

Just so ya know.

Likewise.



Anyone who takes a single data point and tries to make a statistically valid comment about global trends should be ridiculed. I am all for that.

Is that what has occured here? Again we need a better assessment of the systematic bias.

I can pull up probably more graphs of good correlations in my personal data sets here around coatings chemistry that have noise upwards of 10% (adjusted R[sup]2[/sup]) and that's OK by me. A 90% correlation coefficient is a happy thing. That means that any single given data point could be wildly off the curve and it's neighbor wildly off in the opposite direction from the line of best fit.

Again, if we have a systematic bias (in the present discussion, certainly a quantification of how many stations are abysmally set and how that weights the data).
[/quote]

Below is a survey chart of a volunteer evalluation of temperature stations around the US. you can find more info at www.surfacestations.org

Most stations are sited such that the siting handbook I mentioned on the first page of this thread say would have more than a 2 deg F error. Once again, the entire rise in global temperature is only 1.1 deg F. Bias is greater than the signal. 100 years ago we were not a cemented over country. IN 1900, Dallas County Texas had more chickens than people Today you can't hardly find a live chicken in the county.

I will. This is not good. Again, so what? Statistics works on ensemble data and is therefore more robust than a mere single station.

We must strive to make sure the data is good and eliminate bad "gauges", this is simple statistical common sense.

I have a friend with whom I have published some statistical papers--I was his lab rat. I am trying to get him interested in the bad bad state of the data. He is interested but thinking about it. Luckily he is retired and arguing againts GW won't destroy his career as it normally would in a university. He is a professor.

It is not religion. Believe me, I suspect I spend as much time as you and other modelers in assessing the value of data. That's why I've expended so much time in my life over the past several years increasing my stats skills. I have a long way to go, but I am pretty sure it is equally invalid to draw statistically unreliable conclusions from "single points" or anecdotal evidence.

We are in a battle of Type I and Type II errors here.

As I said, I can post more bad stations for longer than you can expresss your belief that this is anecdotal. Lets post the Baltimore Maryland station. It is like a cat on a hot tin roof, near an air conditioner but also on a hot roof top. Have you ever done roofing? I have. It is very hot. It is a bad place to get a good temperature. That is an exhaust condensor on the other side of the thermometer. What a grand place to get a temperature!!!!! It helps professors and hansen scare our politicians into giving them money to study global warming. That is the real purpose of roof top thermometers.

Well, it's a good thing I condemn it then, isn't it?

I will tell you, few AGW advocates do that. I applaud you highly.

Well, to be fair, I'm sure you are assessing a motive without evidence. But more importantly I assume your work in exploration geology has kept you somewhat distant from manufacturing and industrial chemistry. Believe me, there are processes in which the gauge is ignored until something ships out the door that massively fails the spec. I should know, I once had to help figure out why a customer's multi-thousand gallon tank of our liquid product was suddenly a nice solid brown chunk filling the volume of their tank requiring several people go in with pick-axes to hammer it out.

Well, I have been a 3rd level manager. I would fire people who did such sloppy work, motive or not. In my business we invest hundreds of millions of dollars in a single well (yes, more than most movies cost) and if we find oil it can cost more than a billion dollars to build the offshore platform. Sloppy work means we have a higher chance of losing that money in a worthless dry hole or a faulty platform. In my business we pay attention to the guages. It is far too costly not to. So, I would respectfully disagree with your assessment of my QC abilities.

It is highly unlikely that the data was maliciously allowed to accumulate. The fact that you could find information about this individual station is testament to the fact that this is an open-source ideal.

I don't care if it is malicious or not. It is attrocious work. It is work that should be condemned (and I know you do) and it is work not worthy of support.

Just out of curiosity, how do you "know" when the data is bad?

See the post above where I describe what we should see in the way of meteorological effects. When the temperature gradients from town to town are greater than the average pole to equator gradients for an entire year, one must question the data. I will post the tempreature gradients for an area of Oklahoma about 35 x 25 miles across. Note that the annual average temperature gradients are much higher than the average equator to pole gradient. Such high temperature gradients should cause winds and thunderstorms, which are NOT OBSERVED!!!

That is an unevidence blanket statement. I assure you I could probably with enough time and effort dig up an incompetent geophysicist for you. I could dig up an incompetent chemist or two for you. Doesn't mean anything other than "people are people" and "scientists are people".

yes, there are incompetent geophysicists and geologists. In the oil industry, especially in the small companies, they get fired. In government, they don't. They are protected by civil service laws, so they don't care.

 

[grmorton]

What would anyone be doing that would require knowing the temperature to the 10th decimal place?

Your questions are irrelevant.

 

[grmorton]

Did I miss something? The data indicates a statistically significant "non-zero" line. The p-value says I am 99.999% likely to not make a Type I error in rejecting the null hypothesis.

The problem with linear regressions is what is wrong with most investment decisions. People think linearly, not cyclically. A linear regression is grand if the phenomenon is linear. if it is cyclical, it isn't so grand and goes up and down depending upon when one does the regression. Most investors think lineary--stock market going down now, therefore it will always go down. Stock market going up therefore it will always go up. With cyclical phenomenon, linear regressions are not good. Look at the number of ups and downs over the past 30 years and you use a linear regression. tsk.

While I am not happy with the correlation coefficient, it surely indicates that the temperature trend is not flat from 1979 to 2007.

That's why the statistics is important. Please address the statistics and not the individual points.

I will agree to a very very slight warming but it is not comparable to the quantity of CO2 rise in the atmosphere over that time frame. ONce again, I ask you, what is the sensitivity of temperature to a doubling of CO2???? The IPCC has used many different values for that number over the past 20 years. They don't know

Based on analysis of several leading climate models, the Intergovernmental Panel on Climate Change/Third Assessment Report (IPCC/TAR) estimated this climate sensitivity to be in the range of 1.7-4.2 oC (IPCC, 2001).
Steering Committee on Probabilistic Estimates of Climate Sensitivity, Estimating Climate Sensitivity, Washington: National Academies Press, 2003), p. 1

You know that is a range of over 2 fold in uncertainty.

Are you aware that during the little ice age, CO2 LAGGED temperature? When the oceans warm, they exsolve CO2. When they cool, they absorb it. So, if the world's temperature goes up because of more solar energy input or the Milankovitch cycles, the oceans give off CO2 and the atmospheric CO2 rises. Below is a chart.

Oh my. You are going to have to take that one up with Fisher and the statisticians. What part of the F-test do you disagree with? The fundamentals? Or the specific application? Did the stats software I use mess up a sum or squares term or something? The software is from SAS and is a highly recognized and respected organization.

You are free to calculate the numbers for yourself. What F-statistic for this trend do you come up with?

See above. I am going to have to go to bed. I get up really early and it is late here. You are in California where your evening is just starting. I will try to answer another one before I go to work tomorrow.

Didn't we go over the Max Plank data earlier? I thought there was no significant increase solar irradiance since about 1950 or so.

I thought I countered that point.

But again, there is reason to believe this number is hard to pin down, so I might be willing to acquiesce that the question is still open, not that it is a "decided" issue.

You can, I suppose, take it up with the MPI.



Now do me a favor and take the individual data points and do a quick R^2 and a quick F-test on them. Is the trend flat (ie p>0.05)? What is the y-intercept of that trend? 2degF? 3degF?

I don't deny that the regression shows a rise. What I am pointing out is that the starting and ending point are almost the same. this is a cyclical phenomenon, not a linear phenomenon. The sun, in my opinion is the biggest cause of the 2000-2005 rise in satellite temperature. The solar cycle is the cause, not CO2. And now that the sunspots have been at a 100 year low, the weather is getting colder (yall had snow in San Bernadino Co, May 31 this year, and the tropospheric temperature has dropped. CO2 is not involved very much in this.

Again, I am indeed bothered by the spike in the early 60's. And if the IPCC had based their global warming analyses on the Westpoint data station I would be bothered, especially if they started collecting data in 1960 and stopped in 1963.

So, tell me how many bad stations will make you uncomfortable. I don't want moving goal posts.

Sounds like a bum "gauge". I guess someone should do a full gauge R&R study on it.

You know I have a pH meter in the lab that started giving me bad data the other day. Turns out the probe finally crapped out. I had to change it out. Get a new one. Things break. Gauges go bad.

But the weather guys never seem to replace these things or even notice, for years or decades at a time. YOu are competent and fix things. They don't.

The best we can do is be mindful of this. That's why statisticians and "Statistical Process Control" people are out there.

I wish they would just look at where they are putting their thermometers.

 

[chaim]

I am sorry I have missed the earlier posts and they are a little too fragmented to follow, so I will start with this one below.

It is dishonest to try and diagnose a trend in a dense time series (such as the one here) by picking two points at random (or not so random if you are trying to bias the result). The analysis presented by thaumaturgy clearly shows a statistically significant trend globally averaged temperature. There is inherent noise in the climate signal. Based on a spectral analysis, the highest frequency noise is obviously the diurnal cycle. On top of this is a minor spectral peak at ~7 day periodicity in synoptic features as they circle the globe. Next would be the annual cycle. After that comes ENSO with a period of 3-8 years. Claiming that the previous 2 years of cooling is a sign of a global cooling trend is akin to claiming that winter is coming when it starts getting cooler at 7p on a June evening.

On longer scales there is the AMO with a period of ~60 years, however with a very weak amplitude. On longer scales still is the Milankovitch cycle and glaciaton. One of the reasons that 'global warming' is of such concern is that it is not in phase with any of the above cycles, but more importantly the rate of change of temperature is much much faster than that associated with any of these cycles (with the exception of the diurnal and annual cycle of course).

Just to answer a few of your other points. Sea surface temperature will change at a much slower rate than land surface for simple physical reasons not associated with the distribution of CO2 (which is well mixed). The primary reason that SST rises less slowly than lans druface temperature is that the ocean is a dynamically active fluid. As such warming of the surface is much more rapidly transmitted to the subdurface layers than on land (which doesn't mix or circulate too much). The scary thing is the excess heat from the surface is being seen deeper in the ocean:
http://www.sciencemag.org/cgi/content/abstract/1112418

You didn't seem to notice that the ending temperature, in 2007 was about the same as the starting temperature in 1979. That is my point. I have no doubt that the temperature goes up and down, but a trend? not necessarily because today's low is not significantly higher than that of 30 years ago.The fact is that we have more CO2 in the atmosphere today than 30 years ago and we don't have a higher tropospheric temperature. Indeed, it has been cooling down over the past few years, just as the oceans have been cooling down (but the land heating.) And once again, you should know that the late 1990s and early 2000s were a high solar output, one of the highest because we had an exceptionally high number of sunspots and solar flux. Below is a solar flux plot which matches the high on the tropospheric temperature. When the sun's output went down, so did the temperature, back to nearly the starting point. No net gain.

I will now post the pictures I couldn't post earlier about Walden NY and West Point minus Walden. I doubt you will defend the idea that Walden spent years being 12 deg colder than West POint just 16 miles apart. Walden spent time in a glacial age in the 1960s while West Point just 16 miles away was a balmy upper 40's. Does that make sense to you? By the way, I have looked at every single month at Walden for those years. It is in the monthly data.

 

[grmorton]

Just to balance this point a bit, here's a graphic of a map generated by a NASA satellite showing temperature anomalies:



(SOURCE)

Looks like there's significant warming across water bodies as well. Perhaps less or more depending on circulation of aircurrents and buffering capacity of seawater (Revelle Buffer Factor).

Well, My data is from NASA/NOAA as well, and it is temperature TRENDS, not temperature anomaly. The trend is more important than the anomaly. If the anomaly had been going up for years, but then began to turn around, there would be a time when the anomaly was still above the base line, but the TREND is downward. That is what I posted. We need to be clear about what an anomaly vs trend map is. Again below is the trend.

For those who might not know what a temperature anomaly is, they take the temperature, and subtract some arbitrary value from it. Normally it is a mean annual temperature from say 1950 to 1970, but it could be any value or yearly range. That leaves you a deviation from the norm. It isn't a temperature TREND

 

[chaim]

I am not going to get embroiled in the argument about land based temperature measurements as it is somewhat irrelevant due to the large number of other measurements that show exactly the same trend.

However I do want to elaborate on one of the points you brought up in the quoted post, which I bolded for clarity below. You are exactly right when you point out that CO2 lags temperature for the little ice age and for all the glaciations before that as well, generally by about 800 years wich coincides well with the over turning time for the ocean. You also are completely correct about the mechanism. The implication of this is to refute the claim that the current warming is non-anthropogenic. As you pointed out previous warmings are forced by changes in Total Solar Irradiance (TSI) either due to the sun or variations in the earths orbit. These changes drive CO2 from the ocean (due to reduced solubility) into the atmosphere, where it feeds back to amplify the initial warming. Now for the important bit - In the current warming, CO2 leads temperature, it does not lag. So clearly the mechanism for the current shift in climate is not the same as in previous shifts. You inadvertantly refuted your own point.

The problem with linear regressions is what is wrong with most investment decisions. People think linearly, not cyclically. A linear regression is grand if the phenomenon is linear. if it is cyclical, it isn't so grand and goes up and down depending upon when one does the regression. Most investors think lineary--stock market going down now, therefore it will always go down. Stock market going up therefore it will always go up. With cyclical phenomenon, linear regressions are not good. Look at the number of ups and downs over the past 30 years and you use a linear regression. tsk.



I will agree to a very very slight warming but it is not comparable to the quantity of CO2 rise in the atmosphere over that time frame. ONce again, I ask you, what is the sensitivity of temperature to a doubling of CO2???? The IPCC has used many different values for that number over the past 20 years. They don't know

Based on analysis of several leading climate models, the Intergovernmental Panel on Climate Change/Third Assessment Report (IPCC/TAR) estimated this climate sensitivity to be in the range of 1.7-4.2 oC (IPCC, 2001).
Steering Committee on Probabilistic Estimates of Climate Sensitivity, Estimating Climate Sensitivity, Washington: National Academies Press, 2003), p. 1

You know that is a range of over 2 fold in uncertainty.

Are you aware that during the little ice age, CO2 LAGGED temperature? When the oceans warm, they exsolve CO2. When they cool, they absorb it. So, if the world's temperature goes up because of more solar energy input or the Milankovitch cycles, the oceans give off CO2 and the atmospheric CO2 rises. Below is a chart.




See above. I am going to have to go to bed. I get up really early and it is late here. You are in California where your evening is just starting. I will try to answer another one before I go to work tomorrow.





I thought I countered that point.



I don't deny that the regression shows a rise. What I am pointing out is that the starting and ending point are almost the same. this is a cyclical phenomenon, not a linear phenomenon. The sun, in my opinion is the biggest cause of the 2000-2005 rise in satellite temperature. The solar cycle is the cause, not CO2. And now that the sunspots have been at a 100 year low, the weather is getting colder (yall had snow in San Bernadino Co, May 31 this year, and the tropospheric temperature has dropped. CO2 is not involved very much in this.



So, tell me how many bad stations will make you uncomfortable. I don't want moving goal posts.




But the weather guys never seem to replace these things or even notice, for years or decades at a time. YOu are competent and fix things. They don't.



I wish they would just look at where they are putting their thermometers.

 

[chaim]

Without more information your plot is impossible to evaluate. It appears that it is the difference in two monthly averaged global temperatures from two years (2004 and 2007). Is that correct?

Well, My data is from NASA/NOAA as well, and it is temperature TRENDS, not temperature anomaly. The trend is more important than the anomaly. If the anomaly had been going up for years, but then began to turn around, there would be a time when the anomaly was still above the base line, but the TREND is downward. That is what I posted. We need to be clear about what an anomaly vs trend map is. Again below is the trend.

For those who might not know what a temperature anomaly is, they take the temperature, and subtract some arbitrary value from it. Normally it is a mean annual temperature from say 1950 to 1970, but it could be any value or yearly range. That leaves you a deviation from the norm. It isn't a temperature TREND

 

[grmorton]

I am sorry I have missed the earlier posts and they are a little too fragmented to follow, so I will start with this one below.

It is dishonest to try and diagnose a trend in a dense time series (such as the one here) by picking two points at random (or not so random if you are trying to bias the result). The analysis presented by thaumaturgy clearly shows a statistically significant trend globally averaged temperature. There is inherent noise in the climate signal. Based on a spectral analysis, the highest frequency noise is obviously the diurnal cycle. On top of this is a minor spectral peak at ~7 day periodicity in synoptic features as they circle the globe. Next would be the annual cycle. After that comes ENSO with a period of 3-8 years. Claiming that the previous 2 years of cooling is a sign of a global cooling trend is akin to claiming that winter is coming when it starts getting cooler at 7p on a June evening.

On longer scales there is the AMO with a period of ~60 years, however with a very weak amplitude. On longer scales still is the Milankovitch cycle and glaciaton. One of the reasons that 'global warming' is of such concern is that it is not in phase with any of the above cycles, but more importantly the rate of change of temperature is much much faster than that associated with any of these cycles (with the exception of the diurnal and annual cycle of course).

Just to answer a few of your other points. Sea surface temperature will change at a much slower rate than land surface for simple physical reasons not associated with the distribution of CO2 (which is well mixed). The primary reason that SST rises less slowly than lans druface temperature is that the ocean is a dynamically active fluid. As such warming of the surface is much more rapidly transmitted to the subdurface layers than on land (which doesn't mix or circulate too much). The scary thing is the excess heat from the surface is being seen deeper in the ocean:
http://www.sciencemag.org/cgi/content/abstract/1112418

Sorry, but this all assumes that there are no long term trends or frequencies. If you take a 1000 year cycle and then take 50 years out of that time frame, you can get a linear regression that shows whatever you want. But you don't have a linear phenomenon. I repeat. Linear regressions are only great for noisy LINEAR phenomenon. If a phenomenon is cyclical with many different wavelengths, finding a linear trend is meaningless.

You can call this dishonest if you want. It is what I have come to expect from the GW advocates who would rather bully than actually discuss t he data.

Would you care to explain Watersville Washington? See above. You pick and chose what points you want to discuss if you don't want to discuss the raw data.

Another question. Do you think it is grand to put a thermometer next to an air conditioner vent??? Is that a sign of competence on the part of the weather burearu to approve those sites?

Please answer the above.

I will again upload the satellite data. From 1979 to 1997, there is no rise at all. Indeed, the zero line almost perfectly bifurcates the cyclical data. Then there is a bump with the very active solar cycles of the early 90s and early 2000s, then the temperature goes back to about where it was in 1979. I stand by this. This is NOT a linear phenomenon. The ups and down are NOT randomly distributed but cyclically distributed.

Please look at the data, not at your bias.