The Holocene Deniers

[thaumaturgy]

Any one who believes one can get a 1 degree signal out of a 4 deg SD is also making a freshman level error. And that is what you are doing.

Let me introduce (again) the confidence interval.

The Confidence Interval based on the z-statistic (for large sample sizes normally distributed) has the following form:

95% CI = 1.96* s/sqrt(N)

s = standard deviation
N = number of data points

(Confidence Intervals)

Would you like an example?

Here's one:

If I have 1000 samples and the standard deviation of the mean of the data is 1 and the standard deviation is 4 the 95% confidence interval is:

1+ 1.96* (4/10) = 1+ 0.8

That means that with repeated sampling there is a 95% chance that the true mean will be within plus or minus 0.8 units.

NOTE: The confidence interval is only 20% the size of one standard deviation.

I know I've gone over this before.

Indeed if we have a heavy tailed distribution the standard deviation is a bit over estimated than if it were a nice gaussian distribution, but indeed the 95% confidence interval can be smaller than a standard deviation.

I am not saying it is signal. I am saying it is mostly noise--haven't you been paying attention?

I have been and I clearly disagree. With a narrow distribution with heavy tails it pretty much means that while there may be large outliers there are still a huge amount of signal within a reasonable limit of error.

With sufficient data you can even estimate the true mean rather narrower using the confidence interval.

This is pretty much intro stats.

If we are really measuring temperature correctly we shouldn't have many 2 degree differences in temperaure.

Are you serious on this point? Honestly? These are two themometers located 20 miles apart read by some guy (or guys) who drive out to the site every day at about the same time, but they are just instruments and they are just humans, and you expect that the temperature, let's say on a day when the average is 50degF to differ by no more than 4% relative error?

And you won't allow that there might be, on occasion, actual differences in temperature?

But further you keep studiously avoiding the really important fact that the trends are what is most important in this discussion.

Even if one thermometer read 5 degrees above the other consistently as long as they both moved in the same direction and with the same relative magnitude the trend is still preserved.

That is why one usually resorts to measuring the trend rather than the absolute.

This is also why, in chemistry, it is often easier to measure the change in energy in a system than the "absolute energy" of the system.

No, you have been too busy ignoring physics.

If you look at every single blip of noise in any data set and you try to assign a real physical effect you are interpretting noise as signal.

Let me ask you, when you flip around on the AM radio dial do you stop on the static stations and try to figure out if it's really Rush Limbaugh talking to you? Or do you keep scanning until you find a real signal? Because right now it sounds like you might sit and listen to the static trying to figure out if this is just a new form of language.

Then it was in appropriate of you to claim that you couldn't see the offset.

No, quite the contrary. The key point was to show that there was a lot of points above and a lot below and they seemed relatively randomly distributed. Not like for one whole year the whole data set was above zero and one year below zero. I will grant there does seem to be a consistent offset in the 365 point running median of about 0.2 degrees F.

In the chart you post now, you can learly see the bias in the data, and it varies over time. For years at a time it is biased one direction and then for years at a time it is biased another. By your clipping the axes, you were engaging in propaganda, not scientific analysis.

No, I was making a point. If one cannot make a point here without being accused of "propoganda" or lying I am unsure how to procede.

Man, you really don't listen at all do you. Let's start with simple physics. If you set two thermometers on a golf course far from a building, you would expect them to read the same temperature.

No, I would expect them to average to about the same temperature. There is noise in every system.

Let's put it this way: if my technician came to me with a pile of data that was supposedly repeated measurements and they all agreed perfectly, I'd ask them to do it over again because that doesn't look real to me.

Maybe that just comes form about 30 years in the experimental sciences, I don't know.

If, over time, they varied wildly, you would think something is wrong. Same thing with cities 20 miles apart. I dont' expect them to vary much more than a degree on the daily mean, yet regularly they vary by 4-5 degrees.

So why do you think these have outliers? Have you ever made a mistake in measuring something? Ever written a number down wrong? How about times when you didn't realize you'd made a mistake?

That is why the difference is important. We can't check up on the temperature reading in a single city series, but we can compare it to a town that should have the same temperature. This little fact has gone over your head in both of our debates.

Sorry, just accusing me of ignorance is not an effective strategy. Especially when I'm the one who understands the difference between a confidence interval and a standard deviation.

Yes I am accusing you of that, and maybe worse. You are the one who clipped the axes cutting off any chance to see the bias in the data and then you pounded your chest and acted as if it was idiotic to think that there was a bias. That shows either incompetence or worse.

Hmm.

If they were all within a degree, I would have no case to make. But as can be seen, they most assuredly are not within a degree. They aren't even within 2 degrees often.

No, the median difference over about a century is less than 1 degree. But that seems to be escaping you.

You continue to claim that the thermometer data is good in spite of bad gauges. Is it because you BELIEVE and your science be damned?

I am so far the only one dealing with statistics at more than the high school level.

If you really believed that the gauges were bad, you wouldn't keep trying to make them OK as you are doing.

I cannot stress this enough. The point of clipping the axes was to show that the data was distributed around zero.

the land data is crap. I will say it again. You can only assess the same trend if the airconditioner isn't running on either station.

Are there air conditioners at one of the two Mississippi stations? Please show me the information. Were they there in 1915?

Honestly you can run what you want, but I am not likely to trust what you say.

And I would dearly hope you don't! I wouldn't want anyone to believe what I say just because I say it. That is why I obsessively (and I do mean obsessively) provide links to my sources, and the calculations and links to information behind the formulae I use.

You are thus not required to believe a single word I say.

You are the guy who clipped the axes and who depended upon a borehole study which dumped 95% of the data in order to achieve a result. You hadn't done enough research to even know that the previous borehole study existed. And if you had, then you failed to mention a very important piece of contradictory data--that too is anti-science.

As you wish.

 

[grmorton]

I wonder if part of some of these issues is around "time of observation". I think it might be possible to download the STATION HISTORY and get that information out of it. This might explain some of the errors (but perhaps only some of the errors). Thankfully apparently much of this is documented.

Here's what the folks who provide all this data have to say about the data (I've added emphasis where appropriate)



I would be curious if some of the offsets might be due to subtle differences in time of observation.

4 degrees????? hardly. There are too many of them. You would have to have cold fronts running around quite often. Once again, physics, the thing you ignore would rule that out. I am not through calculating but it looks like you would need a cold front, a very very strong cold front 1/5 of the days to explain the distribution. That is why I think it is noise, not signal. But if it is noise, the temperature record is crap. You don't have strong cold fronts 1/5 of the time

But at least now you are acknowledging the offsets, something you ridiculed me for earlier. That is a step in the right direction.

 

[thaumaturgy]

Honestly you can run what you want, but I am not likely to trust what you say. You are the guy who clipped the axes.

That is why I posted the expanded version. I will admit when I plotted the running 365 point average there were extended times when it was above and below the line. That is why I went back to the entire data set point-by-point.

I'll gladly cop to whatever the data shows even the median bias of one degree in those cases where I've found that. I'll even gladly admit that when you look at the Mississippi data there are stretches where there's a positive bias of a degree or two in one direction and then maybe a long period where it's flipped negative.

I simply don't see that as damning to the entire data set. The picture I posted is what it is. I didn't "hide" the y-axis, it was easy to see. I later posted the entire data set, and I ran both the mean and median of the 365 point running data.

I have presented all the data and it is easy to check out. You can run the exact same things I have run. As I said I verified for myself the 365 point running mean and how it swung back and forth as you showed in your graph. Again the picture isn't quite so jarring when you look at the raw data itself.

And, again, means are far more succeptible to deviation due to outliers than the median. (Another one of those intro stats things). So in the case of heavy-tailed data or data with a small number of extreme outliers I tend to prefer the median.

Of course that is my attempt to be honest with the data and not bias it by using a less appropriate method (like the mean).


(By-the-by: when I did clip the axes you are obsessing on I thought long and hard about how to do it without clipping but still show the relative point-to-point distribution around zero. I though perhaps if I just shifted the whole data set so it was all positive and then put it on a log scale it might compress it without making it unreadable. But because I'd have to go to a lot of trouble to explain what I had done there I opted for the expedient of making my point that in any given stretch of the data from a point-to-point basis there were no clearly visible cases where the difference was consistently strongly positive for years at a time with no negative values, which was my point from the beginning.)

Just an FYI. Please resume the accusations. I can take it. I only care about the data.

 

[thaumaturgy]

something you ridiculed me for earlier. That is a step in the right direction.

I don't recall ridiculing you for the offsets. Please refresh my memory.

 

[thaumaturgy]

Any one who believes one can get a 1 degree signal out of a 4 deg SD is also making a freshman level error. And that is what you are doing.

Well, to be fair to statistics that is not necessarily accurate to my understanding.

Here's an example:

I went through and generated a data set where I generated 1000 random numbers according to a "random normal" generator with a mean = 2 and a standard deviation = 6 and then I numbered each data point 1-1000.

Then I constructed a "trend" in the data by multiplying the ROW # of the random data (1 to 1000) by 0.01 and then adding the random number to it so it would generate a line with a slope ~ 0.01 but the data itself would have a normal distribution around the trend line with a standard deviation ~ 6

Then using statistics I tested the data to see if I could find the trend and find a statistical assessment of how valid that trend was.

Here's the result:

Note the top graph. See that "red zone"? That's the 95% confidence interval on the fit .

The data is noisy, in fact the fit itself only explains 9% of the variation. But a trend, in fact the trend I originally generated to test, was, indeed, "findable" by the statistics.

The final data came out where it automatically found the slope of 0.009 (which was pretty durn close to what I built it to have), and the lower graph shows the histogram of the residuals (the difference between each individual data point and the "fit line". It is normally distributed (the normal quantile plot proves that) AND it has a standard deviation of about 8 (not too far off, considering these are computer random generated numbers).

But here's the key take away:

I could find a statistically significant trend (that's what that F-test shows me...it is statistically significantly "not-zero") and that trend was what I expected it to be.

The 95% confidence interval on the fit is a special calculation that you can learn more about

HERE
or
HERE

Now I will stress I am not a statistician. I am merely a very interested scientist who uses statistics and in order to use them better is constantly learning. So take it for what it is worth. If you have any disagreement with this, please present it not as "gut feelings" or accusations of ignorance but rather in mathematical/statistical formalisms and possibly a few actual citations of how my stats were in error.

(I might have to go "radio silent" for a while as I'm off to a mini-vacation with the wife in Colorado for the weekend. But I'll make sure to jump back on as soon as I am able.)

 

[grmorton]

That is why I posted the expanded version. I will admit when I plotted the running 365 point average there were extended times when it was above and below the line. That is why I went back to the entire data set point-by-point.

That is good. I am going to tell another story. Years ago I worked as a consultant for a very small company. This company invested its own money. They didn't go out and get investor money as many small companies do. I was their only geophysicist. There were two geologists. Our boss who worked for the owner had a guy (I will call him Trusty) bring a prospect in for us to look at. There was a Texaco well that had produced 3 billion cubic feet of natural gas before watering out. But the seismic that Trusty had showed that the Texaco well was at the foot of a huge anticline. My boss asked me if I thought we could get up dip from that well (a good thing). I said yes.

While Trusty was in the office I borrowed his seismic, made a map and approved the location. My boss was happy. I was happy. I had kept a small zerox of the seismic for while the well was drilled--with Trusty's permission as he said it was his seismic.

We drilled the well and came in lower than the Texaco well. In fact, there was a huge fault between us and the Texaco well and we were on the wrong side of the fault. I pulled out my zerox looked at it wondering how we could be low when the seismic showed we should be high. It was then that I noticed that the seismic consisted of a very carefully done cut and paste job. Ole Trusty had cut out the original seismic and inserted something from somewhere else--scissor photoshop.

So, I told our the geologists and our boss that we had been had. OUr boss asked Trusty to come into the office, I insisted that he bring the seismic data. he did. Only he brought entirely different data. What he had looked nothing like what he had shown us a few weeks back. I pulled out the zerox, which ol Trusty had forgotten about. He then claimed that one was 'field processing' and what he had now was the final processing. That was absolutely ridiculous--there was no field processing in that day in age.

Our boss agreed with Trusty and indicated in no uncertain terms that I was not to pursue the issue. The next morning, I told the president of the company that I thought we had been had. He said "We will get to the bottom of it". Then nothing happened. Nothing. No one cared that we had just spent a million dollars on a fraudulent dry hole. That was money thrown down the drain.

I needed a job but I wasn't going to throw out my ethics for a place so I started looking. The company bought more Trusty dusters and the louder I yelled that the next well was going to be dry hole the more of their own money they poured into the well. It was crazy. Science didn't matter, data didn't matter. They weren't stealing so far as i could see cause I wasn't selling any of these to investors. The family was using their own money.

It took me years to figure out why the family didn't care about the science. The company I worked for was in a horrible legal struggle for ownership between two parties. The family was apparently giving ol Trusty money for wells he had sold to other investors, and when the well was dry, Trusty gave them back money under the table. (At least that is the only thing that makes sense. Effectively they were draining the resources of the company I worked for and the louder I screamed that it was a dry hole a comin' the more of their money they were willing to risk with Trusty.

I left the company as quickly as I could.

I tell this to show you why I am extremely sensitive to data manipulation. It is an awful thing to do. As to caring about the data, I do care about the data, Thau which is why I am sensitive to propagandistic data manipulation. I have been burned too many times by people who don't treat the data with care.

Now you asked about if time of observation would account for the bias. It won't. Let us say that Bob in Brookhaven City measures the temperature at 3pm and Mary in Monticello measures it at 5 pm. In between a cold front comes through so there is a 4 degree difference. That will work on an occasional day but lets look at Sept Oct 1963. To explain this with time of observation you would have to have a strong cold front per day. Remember that a 4 degree difference creates a deg/mile temperature gradient greater than a strong cold front.

http://home.entouch.net/dmd/weatherMSBrookhavenCityMonticelloSep1963.jpg Picture will be re-added tonight.

If you would but think in terms of physics, you would understand why I think your statistical red herrings are meaningless.

 

[thaumaturgy]

I can't wait til we get back to dealing with the data.

(Please in the future don't post such big pictures, it tends to throw off the formatting on my computer screen. Thanks!)

 

[grmorton]

I can't wait til we get back to dealing with the data.

(Please in the future don't post such big pictures, it tends to throw off the formatting on my computer screen. Thanks!)

I apologize, and will try not to.

I will upload a similar 2 month plot for Brookhaven MS and Monticello MS for 1985. It shows a strong bias for months going the other way. This is why I raised my objection to your median is zero thing. The median being zero doesn't ensure that everything is ok. We have in this town a long period of time where Brookhaven is degrees hotter than Monticello just 20 miles away. Then we have another period of time at a later date when Monticello is degrees hotter than Brookhave, just 20 miles away. Unless you are ready to say that this is a new meteorological phenomenon, there is something very very wrong with the temperature measuring system set up by the weather service and used by the climatologists. Why you deny this over and over is way way beyond me.

These two towns are exactly in the situation of 2 thermometers set close together, say 20 ft apart in which thermometer A measures say 5 degrees hotter than thermometer B for a full year. Then the next year, a leap year, thermometer B measures 5 degrees hotter than thermometer A except for leap day when the temperature difference is zero. The median is zero, and the average is zero and you think there is no problem. That is, of course, absolutely ridiculous. Something is very very wrong. But your lovely statistics make you happy.

Like my auditor friend who thought the travel expense account was OK because all the numbers added up and thus over looked the fact that the guy never traveled, you overlook the fact that this is a ridiculously, laughably awful situation. And it is data like this upon which global warming conclusions are drawn. And you are the one thinking it is ok. I would never argue that way on data this bad. But then, I am not a true believer.

Lurkers, compare the uploaded picture here with the oversized picture above. You can see both pictures on Brookhaven at

The Migrant Mind: Brookhaven City MS vs. Monticello MS

Tell me, just tell me that this is OK. Doing so, loses one one's credibility.

 

[grmorton]

I will try to downsize the uploaded photo tonight when I get home.

But I want to raise another set of towns whose error for at least a short time is 10-15 deg F. Below is from my blog back in May

Lets look at Perry Oklahoma and the two months it spent in the refrigerator.

Stillwater, Enid and Guthrie have about the same temperature But Perry, about 15 miles from Stillwater, 30 from Enid and Guthrie, was 10 to 25 deg F colder than the other 3 cities, if we are to beleive the temperature record. When one subtracts Enid from Perry, Perry is 10 to 15 deg F colder. The same result happens if you subtract Stillwater from Perry's temperature profile and Guthrie from Perry. This means that if we believe the temperature record, for two months Perry, Oklahoma was 10-25 deg F colder, living in a layer of cold air not experienced by any of the cities surrounding it. That should have caused a 2 month long thunderstorm and winds blowing INTO Perry, but of course, there is no record of that. The raw data is crap and this is the crap upon which we base our belief in global warming. Dave who thinks this is good data, has never actually bothered to examine the data he believes supports his viewpoint. What he doesn't know is that the big secret of the USHCN is how statistically bad the data is.

The red line in the picture is the global warming over the past 100 years but the raw data has as much as 25 degrees F of error. Anyone who deals with signal to noise issues will clearly see that there is no way such a small signal can be brought out of the noise of this partly raw dataset.
**end of blog**

Ok, should we believe the statistically OK Perry, that it spent 2 months in the refrigerator, or is something wrong? Your call Thau.

 

[thaumaturgy]

there is something very very wrong with the temperature measuring system set up by the weather service and used by the climatologists. Why you deny this over and over is way way beyond me.

So, barring the complete elimination of all weather data ever in the past, how would you propose to improve the system?

Remember, these stations, going back to 1907 were not put there to support some "global warming conspiracy". They were put there to measure temperature. It seems to me that the posting of "average temperatures" has been quite common for decades, even before any "hysteria" around global warming. Presumaby this data was based on the horrible, horrible flawed trash we are debating today.

Yet I don't recall any moves to strike from the record these terrible data stations until it became a political football for some folks.

I will gladly admit that indeed there are stretches sometimes for several days at a time when there is a consistent positive offset one way or the other. Usually it's only a degree or two from what I could see. But again, I didn't see year after year of consistently one side of the "zero" versus another.

These two towns are exactly in the situation of 2 thermometers set close together, say 20 ft apart in which thermometer A measures say 5 degrees hotter than thermometer B for a full year. Then the next year, a leap year, thermometer B measures 5 degrees hotter than thermometer A except for leap day when the temperature difference is zero.

I will treat your example with more respect than you treated mine. Let's take a look at this "data" scenario of yours.

I "made up" two stations worth of data, called Station A and Station B. I then generated 100 "observations" using a "random normal" generator.

Then I added an "offset" to A of any where from 0 to +4 units and then added the same, but opposite sign offset to those same "days" to B and I had these run for days at a time. I then took these two "offset data sets" and "induced" a trend as I did in the last example. I set them so they would both have a slope of about 0.1. Take a look at the results:

The top two graphs are the two trends. Now remember, these things have a median difference of 1.53 (lower left hand graph) and up to +15 degrees difference at times! If you look at the lower right hand graph you can see the difference going up and down.

But the key point is: look at the slopes of those two top graphs: both are about 0.1 which is just about the trend I "induced" in the data so it would have a trend.

Again, we have data that has a broader spread than the slope. The data generation algorithm has a standard deviation of 4 and a mean of 2. (Also the F-test on the slopes was significant, ie statistically significant non-zero, the red zones are, again, the 95% confidence interval on the fit.)

The median is zero, and the average is zero and you think there is no problem. That is, of course, absolutely ridiculous.

EXCEPT when we are more interested in the TRENDS of the data. Which I have just shown can be very close regardless of swings back and forth of offset.

 

[grmorton]

So, barring the complete elimination of all weather data ever in the past, how would you propose to improve the system?

Sigh, why do you make that limitation? If the past data is bad, your wishing it to be useful is not going to suddenly make it useful. Thermometers should be moved from cities and out into the countryside where they are far from buildings and only the effect of CO2 can be seen. Unfortunately the modern mmts stations require electricity and that has moved many stations closer to heat-emitting buildings over the past few years. One has to eliminate NOISE in the system.

Remember, these stations, going back to 1907 were not put there to support some "global warming conspiracy". They were put there to measure temperature. It seems to me that the posting of "average temperatures" has been quite common for decades, even before any "hysteria" around global warming. Presumaby this data was based on the horrible, horrible flawed trash we are debating today.

YOu always seem never to understand the criticism. I have criticized modern thermometers for being next to air conditioners and heat sources. 100 years ago no one had air conditioners they were quite rare, and now they do. and that puts a positive bias on the MODERN stations not the ancient ones. So, you seem to intentionally distort my criticism.

Yet I don't recall any moves to strike from the record these terrible data stations until it became a political football for some folks.

I like the 1907 stations and so do the folks at good ole GISS. They don't 'correct' those temperatures but they sure do add heat to the modern stations, even above the heat cities give those modern stations.
Here is that picture again.

I will gladly admit that indeed there are stretches sometimes for several days at a time when there is a consistent positive offset one way or the other. Usually it's only a degree or two from what I could see. But again, I didn't see year after year of consistently one side of the "zero" versus another.

DAYS??? Months and years are more like it Thau, but you seem never to acknowledge that. Getting you to acknowledge this is like pulling teeth. Go look at the first picture on

The Migrant Mind: Iowans can't measure the temperature

to see some offests.

Go look at the first picture here to see the same

The Migrant Mind: May 2009

I will treat your example with more respect than you treated mine. Let's take a look at this "data" scenario of yours.

I "made up" two stations worth of data, called Station A and Station B. I then generated 100 "observations" using a "random normal" generator.

Then I added an "offset" to A of any where from 0 to +4 units and then added the same, but opposite sign offset to those same "days" to B and I had these run for days at a time. I then took these two "offset data sets" and "induced" a trend as I did in the last example. I set them so they would both have a slope of about 0.1. Take a look at the results:

The top two graphs are the two trends. Now remember, these things have a median difference of 1.53 (lower left hand graph) and up to +15 degrees difference at times! If you look at the lower right hand graph you can see the difference going up and down.

But the key point is: look at the slopes of those two top graphs: both are about 0.1 which is just about the trend I "induced" in the data so it would have a trend.

Sorry, what a silly example you have. You are basically treating temperature as if it is a random number grab out of a deep dark bag. The reality is that when Brookhaven City is freezing, there is a very high probabilty that Monticello MS is also freezing. Yet in your silly graph, there is no correlation like this at all. So, you seem not to know that temperature is supposed to be measuring something REAL rather than just being a random number generator. That isn't what temperature is all about.

SEcondly IF temperature is as you model, then it confirms what I say. I have said that the noise, the SD is far greater than the supposed global warming signal. If the SD is 4 deg it makes no sense to say that the world has warmed by .84 +/- 4 deg C because in fact it might have cooled by 3 deg plus. This is something that in neither of our two debates have you been able to grasp. You can only claim the world is warming IF and only IF the noise is less than the signal--0.84 deg C. You can say that the world has warmed by 0.84 deg C +/- .5 deg C and that means that the world might have warmed by 1.34 deg or as little as .34 deg C. But it has warmed (highly probable). But with the noise of 4 deg in the system you can't say it has definitely warmed. But for all your self-proclaimed expertise in statistics, you never seem to acknowledge that little fact.

Again, we have data that has a broader spread than the slope. The data generation algorithm has a standard deviation of 4 and a mean of 2. (Also the F-test on the slopes was significant, ie statistically significant non-zero, the red zones are, again, the 95% confidence interval on the fit.)

Once again, like my auditor friend, you keep looking at the receipts rather than at the fact that the guy writing the expense account wasn't supposed to be traveling. What silliness you bring to this. INdeed, Thau, I am amazed that anyone takes you seriously in this and suspect that only the blind belief of your friends, not to mention their cruelty at not stopping their friend from arguing what you are arguing, keeps them from telling you that .84 +/- 4 deg C doesn't mean that the world has definitively warmed!

EXCEPT when we are more interested in the TRENDS of the data. Which I have just shown can be very close regardless of swings back and forth of offset.

Trends created by having 1907 thermometers lacking their air conditioner pets and 8% of 2009 thermometers having air conditioner pets? Yeah there is a real scientific trend.

At this place you can see a contented MMTS thermometer system and its pet air conditioner.Wellsburg_Closeup_MMTS_View_W

 

[grmorton]

People don't lookk at the data, they just simply believe. Below is a picture of the temperature difference between Brookhaven City MS and Monticello MS for the years 1962-1963. There is clearly a bias towards hotter temperature at Brookhaven City. But there is also each winter a phenomenon in which Brookhaven City gets even warmer. Do you think there was a heater next to Brookhaven City's thermometer? Is that why winters are warmer at Brookhaven City, 20 miles away from Monticello, or do you think the climatologists are measuring the temperature correctly?

Please tell me Thau,, that you think that this is ok because the median is zero when you look at the entire record of these two towns. PLEASE tell me that this is ok.

Believers it is time to speak up.

A question for all of you that I often ask of young-earth creationists: Is there anything that would falsify global warming? Any data at all? Or is everything set in stone?

 

[grmorton]

Another wonderful MMTS thermometer with its pet air conditioner. During the 20th century more and more thermometers felt the need for the warmth and friendship of a pet. They chose air conditioners about 8% of the time. But hot heat emitting buildings were the most popular. This one is in West Virginia, and he has two pets--the air conditioner and the building. But don't worry, Thau will show that the median is zero and we can all sleep peacefully.


Isn't this a fun way to warm the globe and scare all the kiddies into giving you more funding so you can scare them again next year at budget time?

 

[thaumaturgy]

Sorry, what a silly example you have. You are basically treating temperature as if it is a random number grab out of a deep dark bag. The reality is that when Brookhaven City is freezing, there is a very high probabilty that Monticello MS is also freezing.

I rather assumed you understood something about this data set. Let me repeat: each data set was set to be random numbers in a normal distribution around the same mean (2) with the same standard deviation (4), ergo this was set up precisely to show you that even with alternating offsets of up to 8 degrees or more that switched back and forth and lasted several data point in a row, consistent trends were derivable from this.

If this is a "silly example" then your whole debate point is, by definition, "silly" as well.

I merely show you how the statistics works.

Look again at the graph in the lower right hand corner. It shows data points going up and down around the "zero" line, not unlike your running averages.

Yet both data sets are still capable of showing approximately the same trend (both at 0.1 slope to 1 significant digit).

I cannot make it more clear how this all works.

The two data sets are not "coupled" in any way other than that they are both trending the same and they are both "seeded" to be the same mean and standard deviation at any given point before the trend is added. And the same trend is added into both.

SEcondly IF temperature is as you model, then it confirms what I say. I have said that the noise, the SD is far greater than the supposed global warming signal. If the SD is 4 deg it makes no sense to say that the world has warmed by .84 +/- 4 deg C because

I believe I have addressed the difference between the confidence intervals and standard deviation now about 3 times at least. Please review it and let me know what you are missing or what I am missing. If you like there are a number of cited references that describe the difference between these two measures.

in fact it might have cooled by 3 deg plus. This is something that in neither of our two debates have you been able to grasp.

Please do me a great favor and look at just ONE of the things I've posted about the confidence intervals. Thanks.

 

[grmorton]

I rather assumed you understood something about this data set. Let me repeat: each data set was set to be random numbers in a normal distribution around the same mean (2) with the same standard deviation (4), ergo this was set up precisely to show you that even with alternating offsets of up to 8 degrees or more that switched back and forth and lasted several data point in a row, consistent trends were derivable from this.

Look, just because you can make a model it doesn't mean that it actually matches reality. I criticised the fact that your model doesn't match reality because there is no correlation between the points. Therefore, any 'conclusions' you draw are about your model, not about reality.

If this is a "silly example" then your whole debate point is, by definition, "silly" as well.

I don't really care to debate. I am here to show my data, which rarely comment on, preferring instead the building of models that don't match reality. Would you care to comment on the validity of the temperatures measured at Brookhaven City and Monticello which I posted this morning? Do you think those stations, with their winter time issues are really measuring temperature correctly? Do you? An answer to that question doesn't require another fruitless exercise in statistics. It is physics that is important. Why do you think the spread in temperature difference is more in the winter than in the summer? Do you think there are heaters involved?

All those questions, you avoid.

Do you think that the MMTS I showed with its pet air conditioner is a good thing? (I know you have said you don't think it should be next to an air conditioner, but you never let that shake any belief you have in the validity of the climatologists to measure the temperature.

I merely show you how the statistics works.

Which is utterly irrelevant to the issue of whether or not the physical system should be behaving in tis fashion.

Please do me a great favor and look at just ONE of the things I've posted about the confidence intervals. Thanks.

No, I am not getting sucked into irrelevancies. That is your forte in all the discussions I have had with you. By your own admission your first or second post in this thread was a huge irrelevancy that cost me time.

As to your confidence intervals, you haven't ever responded to my post a few days ago about why comparing two towns is the only way to know the error in the measurement and you want me to answer things for you that are irrelevant.

You admitted that the temperature could be biased for a few days. But have failed to acknowledge that it can be biased for years. Below is a picture of Columbus and Seymour Indiana. I took the temperature difference between the two towns and it is biased for 2 years. And the bias isn't constant. Comment on that. Do you think something is wrong with the temperature collection system?

 

[thaumaturgy]

Look, just because you can make a model it doesn't mean that it actually matches reality. I criticised the fact that your model doesn't match reality because there is no correlation between the points. Therefore, any 'conclusions' you draw are about your model, not about reality.

In point of fact the picture in the lower right corner was showing exactly what you are complaining about. Systematic offsets that flip back and forth. It showed that even with that in play it is possible that two data sets can still behave in parallel to show equivalent trends

The fact that there is no necessary correlation between points actually makes my point stronger. But even so it really isn't a bad example because all the data points were seeded to be normally distributed about the SAME VALUE with the SAME STANDARD DEVIATION. Ergo, in a sense, the only "differentiators" before I threw in the "offsets" were due to the random distribution around that mean and standard deviation.

So, in fact, I have rather proven the point. It is also a direct analogue to the imaginary scenario you developed of two thermometers near each other but reading with "offsets".

What I developed here was a mathematically robust version of your example. The fact that you couldn't do that to make your point, I don't even begin to understand.

I don't really care to debate.

Unfortunately you've run up against someone who can debate somewhat knowledgably about how data is dealt with in a real-world scientific manner.

I am here to show my data, which rarely comment on, preferring instead the building of models that don't match reality.

Actually, in truth, Glenn, I have repeatedly and I must stress this, repeatedly commented on your data. I have repeatedly downloaded it and treated it with statistical robustness.

Unfortunately you cannot "bully" statistics. They are what they are.

 

[Frank Lovell]

Hi, Folks!

I have finally read all of the posts here on this thread, and I see that skeptics of the claim that anthropogenic global warming is politicize-worthy "settled science" are considered here as borderline (if not over-the-line) crackpots who deserve dismissal via pejorative innuendos (if not outright ad hominems) rather than respectfuly presented reasoned argument.

And that's fine, that is what I experience in most other dialogue forums groups too (especially as an atheist on Christian dialogue lists). And y'all have somewhat changed the tone of your replies to Glenn Morton somewhat from your initial um, "greetings" of him. I have long thought (from other Internet forums he and I have both participated in) that Glenn does a fine job of presenting an objective case for skepticism of the presently highly politicized claim that humans are causing (as distinct from contributing-to in some degree or other along with a whole and complex passel of interacting natural/non-human processes that also exert influence on global climate change and have in the pre-human past produced multiple MAJOR global warming and MAJOR global cooling).

Like Glenn, I too am also a skeptic of the claim of anthropogenic global warming (or AGW, as distinct from plain old-fashioned GW). I think the idea that human-CAUSED global warming (AGW) is presently politicize-worthy "settled science" is ludicrous, and that the intense political (and social) marginalizing of we skeptics of AGW (as distinct from natural global warming or GW) -- is downright criminal (with shades of the disastrous history in Russia of Lysenkoism during which premature politicization rendered the legislative OUTLAWING of expressions of skepticism about Lysenko's ideas). And of course if this means to y'all that I, too, must be some sort of a crackpot, well then, sobeit (for looooooong is the line of my intellectual antagonists, but there is always room for more) -- but before you aim your dismissive pejorative innuendos at me I ask you to do just one thing first, if you please will.

Please go to [...WHAT -- I CANNOT POST A LINK UNTIL I HAVE POSTED 50 UNLINKED POSTS??? AAARRRRGHHHHH!!!!] and give a thoughtful read to Paul Drallos' article, "Is Global Warming Caused By Human Activity?" And especially study carefully Drallos' Fig. 2 in which the last ~600 million years of Earth's (prehuman) global temperatures are plotted along with Earth's atmospheric CO2 concentrations -- see if you see the implications for Earth's future global climate changes which veritably LEAP off that chart to me -- and note that humanity's entire industrial revolution is thinner than the rightmost pixel of the rightmost margin-line of the chart shown in Drallos' Fig. 2. Note from that chart that over the last ~600 million years, more than 90% of the time Earth's global temperature has been either cooler or warmer (and mostly much warmer) than it is today -- and we think we're able and ready to tinker with and fine-tune colossally complex global climate dynamics and henceforth keep Earth globally at today's comfortable (but historically quite rare) temperature??? Holy Unintended Consequences, I sure HOPE not!!

Anyhow, please give a careful, thoughtful read of Drallos' well-articulated case against AGW (which harmonizes with my own but he articulates it so well I find no present need to re-invent an articulation of my own) -- and then, if after that you nonetheless feel I still deserve to be implicitly or explicitly accused of being a crackpot, fine, I will then happily plead guilty (but if you then also want to SAVE me from my crackpottery (and/or Glenn from his), I (as Glenn) will need to hear from you objective counter-evidence presented with valid logic employing only sound premises -- otherwise I will remain tragically lost).

Thanks! -- Frank

PS: GLENN -- SEE ABOVE FOR WHAT THE SYSTEM HERE LIMITS ME TO POSTING; I WILL SEND YOU IN PRIVATE EMAIL THE LINK TO DRALLOS' ARTICLE TO WHICH MY POST ABOVE REFERS, WILL YOU THEN PLEASE REPOST IT HERE SO THAT WHATEVER FEW FOLKS HERE WHO ARE INTERESTED ION DOING SO CAN READ IT? THANKS!! IT WILL BE QUITE A WHILE BEFORE I POST 50 UNLINKED POSTS -- I WONDER WHAT OTHER HANDICAPS TO DIALOGING THE SYSTEM HERE WILL SADDLE ME WITH... -- FL

​

 

[Baggins]

Paul Drallos doesn't appear to be a climatologist but a theoretical physicist with a career in plasma physics and computer software manufacturing why should I be interested in what he has to say about climatology?

AGW is settled science amongst climatologists does it really matter if a plasma physicist doesn't like that?

Mr Morton is certainly not interested in objective evidence as his hand waving dismissal of statistics that undermine his argument shows, he rejects AGW for political, rather than scientific, reasons and the same is almost certainly true of you Mr Lovell and it is certainly true of Mr Drallos who is a contributer to the Republican Party.

 

[thaumaturgy]

Hi, Folks!

I have finally read all of the posts here on this thread, and I see that skeptics of the claim that anthropogenic global warming is politicize-worthy "settled science" are considered here as borderline (if not over-the-line) crackpots who deserve dismissal via pejorative innuendos (if not outright ad hominems) rather than respectfuly presented reasoned argument.

And that's fine, that is what I experience in most other dialogue forums groups too​

Glenn wanted a bare-bones, hard core discussion of data. I have provided that to him.

I will gladly grant I started off on the wrong foot but I have made a concerted effort over the past 15 or so pages of discussion to keep my aggression in check. I am not so sure it is reciprocated.

(especially as an atheist on Christian dialogue lists). And y'all have somewhat changed the tone of your replies to Glenn Morton somewhat from your initial um, "greetings" of him.

I do dearly wish people would get the whole story. In an earlier thread several months ago I engaged Glenn out of enormous respect. His reputation preceded him and I was pleased as punch to actually be talking to the famous Glenn Morton. As the thread went along Glenn got much more aggressive and I followed suit. It was generally unpleasant precisely because that is how Glenn wanted it. I got down and dirty.

That experience was precisely why I met this thread the way I did initially. I knew how Glenn operated at that time.

However, after a while I actually felt guilty for being the first on this thread to get snarky. So I stepped back and took a way out so I could come back and actually discuss the data.

To that end I have endeavored to keep the discussion related only to the data. In response I have been told my points are "silly" and that statistics (in a scientific debate) are a "red herring" and I specialize in "distraction" of the debate from the main point.

Like Glenn I am a professional scientist. I have years in academia, industrial and governmental research. So I think I know a few things about how data is dealt with in the real world. So I am curious as to why the debate seems to be pulled precisely in the opposite direction of how data is dealt with by professional scientists.

I have long thought (from other Internet forums he and I have both participated in) that Glenn does a fine job of presenting an objective case for skepticism of the presently highly politicized claim

Glenn does, indeed, do a fine job of presenting pictures and graphing the data. He does support his contention. However it seems to miss the overall point of how the data is actually used in the field under discussion. He focuses on individual surface stations which are only a small component of the overall scientific foundation of global climate change.

Focusing on individual stations is called "anecdotal data interpretation" first off. Focusing on the outliers and tails of the distribution of any data set is called "interpretation of noise as signal".

All data has noise. Ergo, that noise must be assessed and dealt with. That is why I bring it back to statistics. Further the data in question are used in gridded averges on continental scale and in parallel with numerous other methods of assessing temperature changes. And finally, rather than the absolute temperature being of importance the TRENDS of temperature are the key.

I have shown how trends can be assessed both from noisy data and from two "gauges" that don't read the same absolute value and in fact have alternating offsets one way and the other.

of the disastrous history in Russia of Lysenkoism during which premature politicization rendered the legislative OUTLAWING of expressions of skepticism about Lysenko's ideas). And of course if this means to y'all that I, too, must be some sort of a crackpot, well then, sobeit (for looooooong is the line of my intellectual antagonists

And science needs skeptics. It doesn't however, mean that all skepticism is ipso facto necessarily correct. Skepticism serves a valuable purpose. But that skepticism of science must deal with the data as the science deals with it.

​

 

[grmorton]

PS: GLENN -- SEE ABOVE FOR WHAT THE SYSTEM HERE LIMITS ME TO POSTING; I WILL SEND YOU IN PRIVATE EMAIL THE LINK TO DRALLOS' ARTICLE TO WHICH MY POST ABOVE REFERS, WILL YOU THEN PLEASE REPOST IT HERE SO THAT WHATEVER FEW FOLKS HERE WHO ARE INTERESTED ION DOING SO CAN READ IT? THANKS!! IT WILL BE QUITE A WHILE BEFORE I POST 50 UNLINKED POSTS -- I WONDER WHAT OTHER HANDICAPS TO DIALOGING THE SYSTEM HERE WILL SADDLE ME WITH... -- FL

​

They treat newbies like well, you know what until they have 50 posts.

Here is the link http://home.comcast.net/~pdrallos131681/CO2/co2.html