|
Warming 99.999999999999999999999999999999% Certain!
Grant wrote in message ...
Roger Coppock wrote: Then please take us step-by-step through the lag-one auto-correlation of the data set and produce your estimate of the number of degrees of freedom. The URL to the data set was listed in my original post, the part you snipped. Why me? I'm not the one asserting that global warming is 99.999999999999999999999999999999% Certain! No, you have asserted that there is a "lag-one auto-correlation" which produces your a number of degrees freedom which is different from the N-2=122 that I and others on this thread used. To back up that assertion, you should explain it. I've explained my assertion, and other people in this thread have understood my explaination and taken time to verify it. Personally, I don't think it should make much difference to reasonable people whether it's 99.999999999999999999999999999999% certain or merely 99% certain. And the latter claim will be far easier to prove without the help of questionable assumptions. Then, explain your assumptions to the groop and calculate your version for the certainty number. |
Warming 99.999999999999999999999999999999% Certain!
Roger Coppock wrote:
Why me? I'm not the one asserting that global warming is 99.999999999999999999999999999999% Certain! No, you have asserted that there is a "lag-one auto-correlation" which produces your a number of degrees freedom which is different from the N-2=122 that I and others on this thread used. To back up that assertion, you should explain it. I've explained my assertion, and other people in this thread have understood my explaination and taken time to verify it. The N-2 assumes that all N of the measurements are statistically independent of one another. That is, it assumes that anomalies in one data point are uncorrelated with those in the next (or previous) data point in the time series. This is a poor assumption for most geophysical time series. In general, atmospheric records contain both low-frequency and high-frequency variability. For example, over the course of a year, the temperature on one day is highly correlated with that on the next day. This correlation tends to decay for pairs of days that are more widely separated. The point is, if you took (for example) the max temperature for 124 straight days, subtracted the average annual cycle, and fit a straight line to your resulting anomalies, your computed slope (as a measure of long term trend) would not have the small uncertainty that would be implied by 124 independent samples. In fact, your slope would have almost no meaning whatsoever for long-term trends in temperature anomalies, because you'd almost certainly be fitting your straight line to part of a longer term (low-frequency) fluctuation that is not related to a long-term trend. For annual rather than daily temperatures, the problem with low-frequency variability is less severe but not negligible. There's no perfect way to account for the effect of low frequency fluctuations on apparent long-term trends, but using the autocorrelation of your data set to estimate a reduced number of degrees of freedom will allow you to attach a somewhat more realistic uncertainty to your aparent trend. Personally, I don't think it should make much difference to reasonable people whether it's 99.999999999999999999999999999999% certain or merely 99% certain. And the latter claim will be far easier to prove without the help of questionable assumptions. Then, explain your assumptions to the groop and calculate your version for the certainty number. Sorry, I'm not interested enough in the question to take the time to download your data set, look up the exact formulas, and write the required program to read the data set and do the calculation. I can say that the relevant formula can be found in Schaum's Outline of Statistics, because I'm pretty sure that's where I found it last time I needed it some years ago. As I said earlier, I already believe it is *likely* that we are seeing global warming. I just believe that it cannot be shown with anywhere near the confidence you claim, for reasons that are only partly related to statistics. |
Warming 99.999999999999999999999999999999% Certain!
Grant wrote:
Sorry, I'm not interested enough in the question to take the time to download your data set, The data set is only 14K bytes long. That shouldn't take more than a small fraction of a minute to download. You only need the January to December annual mean, which is in the 14th column, after the 13 columns that contain the year and 12 months. look up the exact formulas, That could take some time, because the formulas may only exist in your imagination. We are here to learn from each other, Grant. If you don't have the time to contribute something positive, then don't come here in the first place. You could, at least, publish the formulas clearly. -- "One who joyfully guards his mind And fears his own confusion Can not fall. He has found his way to peace." -- Buddha, in the "Pali Dhammapada," ~5th century BCE -.-. --.- Roger Coppock ) -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Warming 99.999999999999999999999999999999% Certain!
Roger Coppock wrote:
Grant wrote: Sorry, I'm not interested enough in the question to take the time to download your data set, The data set is only 14K bytes long. That shouldn't take more than a small fraction of a minute to download. You snipped out the other, vastly more important reasons. look up the exact formulas, That could take some time, because the formulas may only exist in your imagination. You have an interesting way of trying to elicit cooperation. We are here to learn from each other, Grant. Are you 99.999999999999999999999999999999% certain that that's why I'm here? For the record, about 50% or more of what I read in the sci.geo.meteorology newsgroup, especially with respect to global warming, ice meteors from space, and so forth, is written by people with only a superficial understanding of meteorology or climatology and surprisingly little recognition of what they *don't* know. One of my main reasons for coming to this newsgroup from time to time has been to try to inject the perspectives of a professional atmospheric scientist into discussions that seemed to me to be careening into the abyss of ignorant speculation. A few years ago, there were many other experts around. I think many or most of them decided at some point that they had better things to do than get drawn into "debates" with arrogant and argumentative laypeople who think they know better. If you don't have the time to contribute something positive, then don't come here in the first place. See above. You could, at least, publish the formulas clearly. At the risk of repeating myself, it would take me time that I don't have to find and post formulas that can probably be gotten from any reasonably complete textbook on statistical regression. Here's the operative topic: "estimating the number of degrees of freedom in a time series of non-independent (or correlated) measurements." If you're truly interested in accurately portraying the uncertainty in global warming trends, you'll do your own homework and not attack those who point out, however cursorily, the shortcomings in your arguments. On a side note, I completely agree with whoever pointed out (I saw only a response to that person's article) that the lag-one autocorrelation only allows you to get a *first-order* estimated correction to the number of degrees of freedom. That person may be better able than I to point to you to a source for more information. Though I would advise against *demanding* that he do so, as you have done with me. |
Warming 99.999999999999999999999999999999% Certain!
February 10, 2004
Grant wrote: At the risk of repeating myself, it would take me time that I don't have But you have enough time to post long, convoluted, content free, text arguments and baseless skepticism on the usenet, eh? Keep up the good work! Don't forget, we like citations and references. And we love http://www.google.com. Thomas Lee Elifritz http://elifritz.members.atlantic.net |
Warming 99.999999999999999999999999999999% Certain!
Thomas Lee Elifritz wrote:
February 10, 2004 Grant wrote: At the risk of repeating myself, it would take me time that I don't have But you have enough time to post long, convoluted, content free, text arguments and baseless skepticism on the usenet, eh? Keep up the good work! Don't forget, we like citations and references. Then you'll like http://www.uoguelph.ca/~rmckitri/research/surface.pdf And we love http://www.google.com. which I just found via Google using search terms "serial correlation" "degrees of freedom" trend Happy reading. |
Warming 99.999999999999999999999999999999% Certain!
Grant wrote:
Don't forget, we like citations and references. Then you'll like http://www.uoguelph.ca/~rmckitri/research/surface.pdf And we love http://www.google.com. which I just found via Google using search terms "serial correlation" "degrees of freedom" trend Oh, and here's another one from the SAS manual (same Google search): http://gsbwww.uchicago.edu/computing...hap8/sect3.htm |
Warming 99.999999999999999999999999999999% Certain!
Grant wrote:
Grant wrote: Don't forget, we like citations and references. Then you'll like http://www.uoguelph.ca/~rmckitri/research/surface.pdf And we love http://www.google.com. which I just found via Google using search terms "serial correlation" "degrees of freedom" trend Oh, and here's another one from the SAS manual (same Google search): http://gsbwww.uchicago.edu/computing...hap8/sect3.htm And one more (do you have a PostScript viewer?): http://www.cgd.ucar.edu/stats/manuscripts/trend.ps |
Warming 99.999999999999999999999999999999% Certain!
Grant wrote in message ...
At the risk of repeating myself, it would take me time that I don't have to find and post formulas that can probably be gotten from any reasonably complete textbook on statistical regression. Here's the operative topic: "estimating the number of degrees of freedom in a time series of non-independent (or correlated) measurements." If you're truly interested in accurately portraying the uncertainty in global warming trends, you'll do your own homework and not attack those who point out, however cursorily, the shortcomings in your arguments. You are wasting your time. I had a go at explaining the problem to Roger Coppock a few months ago, but he's too ignorant to even realise that there might be a problem. Frankly, I find the True Believers just as tiresome as the True Disbelievers, both of them seem to rely on revealed truth (ie their own prejudices) to the complete exclusion of any remotely competent or objective analysis of the problem. Indeed, as you say, most of them are too stupid to even realise how little they know. James |
| All times are GMT. The time now is 06:12 AM. |
Powered by vBulletin® Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
Copyright ©2004-2006 WeatherBanter.co.uk