No, but the sum total of every period other than your cherry-picked 8 year period is more meaningful.
You completely missed the point. The 9 years he cited shows a much faster rise than the long-term trend. The 9 years you cite shows a much slower rise. They’re both equally meaningless.
Beechnut: it’s not just that the 9-year period is more meaningful because of some arbitrary decision. It’s because EVERY period you can choose EXCEPT THE ONE THAT YOU HAVE CHOSEN shows clear evidence of a warming trend. The 9-year period from 1997 to now, the 7-year period, the 40-year period, the 4-year period, the 66-year period, the 21-year period: all of them show a clear warming trend.
You have repeatedly demonstrated your very clear lack of understanding for how statistical models work. Have you noticed that not a single person has jumped in to speak up for your point of view? I can guarantee you it’s not because everyone else on this board is foolish and naive.
A follow up on the 1998 / 2005 global temperature anomalies. To say that 2005 was higher than 1998 has no basis in fact, not even close to a correct statement. Based of the t-2 means they are NO different! Another fact that the there has been no increase in global temperature from 1998 through today:
Yr Mo LT O-S Met
1998 1 0.529 0.520 0.570
1998 2 0.673 0.790 1.010
1998 3 0.475 0.560 0.700
1998 4 0.773 0.560 0.710
1998 5 0.648 0.610 0.780
1998 6 0.574 0.670 0.830
1998 7 0.511 0.700 0.890
1998 8 0.509 0.620 0.750
1998 9 0.462 0.420 0.470
1998 10 0.412 0.390 0.570
1998 11 0.166 0.430 0.540
1998 12 0.271 0.510 0.700
Avg 0.500 0.565 0.710
2005 1 0.440 0.680 0.850
2005 2 0.320 0.550 0.760
2005 3 0.301 0.710 0.860
2005 4 0.414 0.650 0.790
2005 5 0.223 0.560 0.660
2005 6 0.247 0.600 0.730
2005 7 0.325 0.540 0.660
2005 8 0.156 0.580 0.670
2005 9 0.338 0.700 0.880
2005 10 0.366 0.720 0.850
2005 11 0.338 0.630 0.750
2005 12 0.240 0.580 0.690
Avg 0.309 0.625 0.763
t - two means:
0.430 -0.135 -0.118
sigma 1998 - 2006 = 0.1814
.625 isn’t bigger than .565? .763 isn’t bigger than .71? Moreover, they are so much lower as to be “not even close”? :rolleyes:
No, they are both important, however, obviously older data is less relevant than more recent data, wouldn’t you agree? i.e., the exclusion today of the cooling in the 70s.
Ooops … correction, t-2 means … 1998 versus 2005 …
LT 1998 was significantly higher than 2005, L-O and Met were not …
Sorry … 
Yr Mo LT L-O Met
1998 1 0.529 0.520 0.570
1998 2 0.673 0.790 1.010
1998 3 0.475 0.560 0.700
1998 4 0.773 0.560 0.710
1998 5 0.648 0.610 0.780
1998 6 0.574 0.670 0.830
1998 7 0.511 0.700 0.890
1998 8 0.509 0.620 0.750
1998 9 0.462 0.420 0.470
1998 10 0.412 0.390 0.570
1998 11 0.166 0.430 0.540
1998 12 0.271 0.510 0.700
Avg 0.500 0.565 0.710
2005 1 0.440 0.680 0.850
2005 2 0.320 0.550 0.760
2005 3 0.301 0.710 0.860
2005 4 0.414 0.650 0.790
2005 5 0.223 0.560 0.660
2005 6 0.247 0.600 0.730
2005 7 0.325 0.540 0.660
2005 8 0.156 0.580 0.670
2005 9 0.338 0.700 0.880
2005 10 0.366 0.720 0.850
2005 11 0.338 0.630 0.750
2005 12 0.240 0.580 0.690
Avg 0.309 0.625 0.763
t- 2 means: 2.582 -0.810 -0.709
sigma 1998 - 2006 = 0.18143
No significant statistical difference, t’s less than 1 let alone 2, thus not “bigger”, much more a random event, right? :rolleyes:
WTF? You are the one trying to claim that even though two of the three temperatures are larger, it’s “not even close” to being warmer than 1998 because it’s not a statistically significant change??? And yet, you’re the one who wants to use the aberrant year of 1998 to compare the recent data to?
I agree completely.
You’ve convinced me that eight years is exactly, precisely, the only meaningful measurement of global warming; and that we must only examine the most recent eight years of data, throwing out anything older — because anything older is irrelevant, and anything less than eight years is insufficient.
That means global warming doesn’t exist this year. Yay!
Next year, of course, we’ll throw out that old, irrelevant 1998 data and global warming will be back.
- 1998 aberrant? … please explain, basis for that statement. Statistical treatment?
- 1998 lower troposphere temperature was significantly greater than 2005.
- 2005 land-ocean and met data were not significantly higher than 1998 with t’s less than 1.
Are you saying if you toss 12 coins in your left hand and 12 coins in your right hand and the results show 6 heads left, 7 heads right, you can draw a conclusion that the 12 coins in your right hand when tossed will result in more heads than the 12 coins in you left hand?
You’re trying to enforce rigorous statistical methods in certain areas, and not in others. It’s as simple as that.
All I have said here is that there has been no global warming in 8 years based on the global temperature anomalies. I have asked how this could be given that CO2 concentration and fossil fuel usage has continued to go up over these past 8 years? What is the explanation for this and how is it any different than when the temperature went up for 9 years?
What areas didn’t I apply statistical treatment to on the LT, O-L and met data for the 8 years from Jan 1998 through Mar 2006?
Let’s say you are comparing theoretical* murder rates over the past years:
1997 535
1998 540
1999 545
2000 550
2001 1000
2002 555
2003 560
2004 565
2005 570
Can I then claim that because the last year’s murder rates are far lower than the mean over 2001-2005, that there is no evidence of a long-term murder rate increase, despite the fact that 2001 is an anomalous year? Indeed, if I were shown that 2001 was an anomalous year, would it not be disingenous of me to insist that all analyses of long-term murder rates use 2001 as a base year?
*Not intended to imply an actual trend in murder rates nor categorization of various acts as murders for statistical purposes.
Data from 9 years ago is less relevant — you said so yourself. Who cares? Global warming doesn’t exist this year; you’ve proved it.
And as I’m not a statistician, I’m not allowed to disagree.
If you applied it correctly, you would know that this data set is insufficient to rule out the hypothesis (i.e. ~0.6 deg/yr rise in global temperature) with any statistical significance. If you disagree, please tell us what upper bound is placed on global temperature rise by this particular data set.
No. As others have pointed out, what this shows is that your method of looking at the data sucks. If I had been around in 1998, I could have used it to show that temperatures were going through the roof. (And, by the way, if I had looked instead over a 5 year period instead of a 9 year period ending in April 1998, I could have made the rate of increase something like 25 C over 110 years.)
The overarching point is that you can prove anything if you take noisy data, cherry-pick either or both of your beginning and endpoints, and make no attempt to do things like look at running averages over a long enough period to get rid of some of the noise.
Here on the SDMB, we don’t keep repeating arguments without responding to the counterarguments of others. We have explained to you that, not only is it not surprising that there is internal variability in the climate system, but in fact such variability is seen in the very models that are used to predict the amount of temperature increase one will get with increasing CO2 concentration.
This means that the apparent contradiction that you see between theory and data is simply not there. You are simply mischaracterizing the theory in order to pretend that it is there.
I didn’t know the SDMB was exempt from using data that wasn’t significantly relevant? Let’s do it this way, how about we inculde ALL years? Fair enough?
What conclusions are you going to draw from this graph that are significantly relevant about today’s “global warming?”
Note: CO2 concentrations varied proportionally to the temperatures shown. Will show both temperature and CO2 shortly.
The increase in temperature and CO2 today are completely insignificant statistically to the historic variations in both. Historically we are talking variations on the order of several degrees C, making today’s temperature variation random. See above graph.