It looks like you're using an Ad Blocker.
Please white-list or disable AboveTopSecret.com in your ad-blocking tool.
Thank you.
Some features of ATS will be disabled while you continue to use an ad-blocker.
originally posted by: Dfairlite
a reply to: TheRedneck
Excellent data. Cue the crazy lefties claiming "you're not a climate scientist so what you found is irrelevant!"
The only thing that one has to know in order to completely and utterly disprove CAGW (or even AGW) is the numbers.
The observed greenhouse gas effect of CO2 is ~1.5C for every doubling of CO2 in the atmosphere. In 1800 we were at 282ppm. So when we reach 564 we will have raised earths temperature by ~1.5C. We are currently at 408ppm, so in 220 years we have added 126ppm, we still have 156 to go. It will take us another 150-200 years to add that, and that's if we don't transition away from fossil fuels like we already are doing.
Furthermore, we wouldn't see another 1.5C increase until we hit 1128ppm of CO2 in the atmosphere. When would that happen? About a millennia from now.
All of this, of course, assumes that people are responsible for 100% of the increase of CO2 in the atmosphere. Which they are not. Probably more along the lines of 50% give or take.
originally posted by: Dfairlite
a reply to: TheRedneck
All of this, of course, assumes that people are responsible for 100% of the increase of CO2 in the atmosphere. Which they are not. Probably more along the lines of 50% give or take.
originally posted by: 1947boomer
OK, so you conducted your own independent research effort and now you are publishing the results. First of all, kudos for investing your time and energy.
However, I assume you are putting this out there for critical comment—basically peer review. So, I will respond in that spirit.
I don’t know what your education level or what your field(s) of expertise might be, but for the record, I have a BS in Physics and MS and Doctorate level degrees in Aerospace Engineering. My major fields of study in graduate school included a lot of classes and experience in topics like optimal estimation theory.
I don’t have any strong criticism of your data acquisition methods; the missing data might have been handled a little better, but I don't think it affected the outcome materially. I won't quibble with that.
What I do criticize, however is your process of fitting curves to the data set and the implications that you then draw from those curves.
When you fit a straight line to the data set from 1950 to the present you showed almost a two degree F increase over that time span (about 1.8 F, as best I can make out). (Parenthetically, that’s just about what the Intergovernmental Panel on Climate Change (IPCC) says is the current average global temperature increase since pre-industrial times. So, it looks like you are in agreement with their finding.) You didn’t say what method you used to generate that curve fit, but I assume you probably used the standard minimization of least-squares error criterion.
You mention that you were surprised to see this linear relationship between increasing time and increasing temperature even though (or perhaps because) it agrees with what such groups as the IPCC have been saying for some time. I’m sure you realized that if you simply extended this linear relationship into the future it would continue to predict what the IPCC predicts—about a 4 degree F average rise by the end of the century. So, I guess you looked around for a different answer, and settled on fitting a fourth degree polynomial. And sure enough, when you project that polynomial into the future, you see average temperature going down at some point.
So here are the problems with that. First, you give no physical basis whatever for choosing a higher order polynomial other than your hunch that a linear process shouldn’t continue into the future. You made the statement, “ Nature rarely follows linear anything—4th degree polynomial trends are typically more accurate.” This statement is totally unsupported by any facts and therefore represents nothing more than your confirmation bias. In fact, your statement is contradicted by a lot of known physical phenomena: Gravitational and electric forces, for example, obey an inverse-square relationship. Most solid materials obey a linear relationship between stress and strain. Radioactive materials decay exponentially, and so on. What you seem to be saying here is that somehow, Nature prefers 4th degree polynomial trends as a general principle and therefore you chose that for your model. That is just nonsense, Nature does not have any such preference.
Now it is true, that when confronted with a complex data set such as the temperature history of Huntsville, it is common for scientists to attempt to fit the data to various kinds of curves in order to try to deduce the underlying physical laws. When this is done however, it is important to be unbiased and accurate in the curve fitting, otherwise you can end up falsely confirming a pre-existing belief. For example, if you carefully measured the gravitational attraction force between two objects over a wide range of separation distances and attempted to fit various curves to the force data with no prior belief bias, you would find that a single inverse square term would fit the data perfectly, with essentially no error. Knowing this, you could infer Newton’s law of gravitation. But suppose you had the prior belief that gravity should obey a fourth degree polynomial relationship and you tried to fit the data to that polynomial, using a least squares error algorithm. The algorithm would probably return a series of polynomial coefficients for the curve fit. Does that mean that gravitational attraction is actually a 4th degree polynomial? No. It means that you chose the wrong mathematical model and got a curve fit, but a really crappy one.
Which brings me to the last point. When doing a least squares curve fit to a data set, standard algorithms attempt to minimize the sum of the differences squared between the proposed function and the empirical data (the residuals). Standard least-squares algorithms usually output this number to the user as an indicator of how good the fit is between the proposed function and the actual data. In the example I gave above of curve fitting gravitational data, if you chose an inverse square relationship, the sum of the residuals would be zero, and you would know you hit on the correct mathematical relationship. If you chose a 4th degree polynomial you would get a large residual error. I don’t know if you used this approach; if you did, you did not present or discuss what the error residuals were. For that reason, we have no idea whether a linear relationship or a 4th degree polynomial is actually a better fit.
Or, for that matter, whether some other curve entirely is a better fit. You point out that when you include only the data from 1960 to the present you get a steeper linear relationship than when you include the data from 1950 to the present. To the unbiased observer, that would suggest that whatever the actual curve is, it is getting steeper with time. In other words, the actual curve may be concave upward. That would suggest that a power curve of the form T = T0 ^k or an exponential curve of the form T = T0 exp(t/) would be a much better fit.
Try fitting curves other than those that reflect your confirmation bias and see which ones produce the smallest residual errors.
a reply to: TheRedneck
To the unbiased observer, that would suggest that whatever the actual curve is, it is getting steeper with time.
originally posted by: OccamsRazor04
a reply to: TheRedneck
I think this detailed post highlights the exact problem we have. Things are changing, but why? Our data set is so limited it's impossible to ascribe changes to any one particular reason. If it is cyclical we have not gone long enough to see the cycle.
originally posted by: ChaoticOrder
This chart shows Texas temps going back to 1900:
This shows the avg mean temp measured by all USHCN stations:
OK, so you conducted your own independent research effort and now you are publishing the results. First of all, kudos for investing your time and energy.
However, I assume you are putting this out there for critical comment—basically peer review. So, I will respond in that spirit.
I don’t have any strong criticism of your data acquisition methods; the missing data might have been handled a little better, but I don't think it affected the outcome materially. I won't quibble with that.
What I do criticize, however is your process of fitting curves to the data set and the implications that you then draw from those curves.
@ TheRedneck--I can't promise I'll ever git-r-done, but I'd love to give your templates a shot. How do I get my hands on them?
Redneck, a very interesting analysis, thank you for taking the time to do it. I have to say that I disagree with 1947
boomer's use of forces as a comparison for your analysis. Forces a governed by set laws and mathematical theorms.
originally posted by: DBCowboy
a reply to: TheRedneck
*applause*
This has spurred me to conduct a study of my own. We often hear about water levels rising and since I live on a coastal area, I can look at water levels here.
Thanks for making me think!