I can find AOD time series for brief periods and regions, e.g. 1984-2002 over China. I can’t find any long-term time series for it. So, I hate to echo Victor, but how do we know the 1940-1970 pause was due to aerosols? We don’t seem to have consistent time series data for aerosols.
Mean dependent var −0.132763 S.D. dependent var 0.389642
Sum squared resid 2.850456 S.E. of regression 0.137394
R-squared 0.878870 Adjusted R-squared 0.875662
F(4, 151) 273.8998 P-value(F) 4.11e-68
Log-likelihood 90.83099 Akaike criterion −171.6620
Schwarz criterion −156.4127 Hannan-Quinn −165.4684
rho 0.123729 Durbin-Watson 1.739247
Excluding the constant, p-value was highest for variable 4 (TSI)
I’m a little surprised that sunlight seems to have no effect over the whole period, or no significant effect. How can that be? But that’s what the numbers say. Everything else–carbon dioxide (of course), ocean temperature sloshing, aerosols, all matter.
BTW, does anyone have AOD data past 2011, or AMO before 1856? I’m unable to find either on the web (probably just my poor search skills).
Okay, this is interesting: I did a test for non-linearity on the variables’ relation to dT, and found a significant result for TSI. With square of TSI included in the regression, sunlight suddenly becomes significant at the 95% level:
Model 2: OLS, using observations 1856-2011 (T = 156)
Dependent variable: dT
Between ice core data and sampling from 1983 on, I was able to construct an 1850-2019 series for methane. Unfortunately, due to very high multicolinearity with CO2 (r = 0.98), the coefficient in a multiple regression comes out perversely signed. I’ll have to look up how to compensate for multicolinearity. I used to know this stuff… *sigh*
Oops. Apparently there are two Ls in the middle of “multicollinearity.” Boy, I really am out of it… Gotta refresh my statistics…
jgnfldsays
BPL…Agree you very likely have an MCL problem as you suspect. You don’t exactly compensate for MCL, generally. Basically, either you construct combinations using such techniques as PCA (or in this case simple addition of normalized scores since the cor is so high). Or you do separate analyses.
Reporting of Notz et al (2020) ‘Arctic Sea Ice in CMIP6’ will likely re-invigorate the shouty denialists about predictions of an ice-free Arctic Ocean. (So far, the Wattsupian bog-house wall has only had “2050? What happened to 2014?” scratched into the paintwork.)
Notz was interviewed by USToday about the research.
“”If we reduce global emissions rapidly and substantially, and thus keep global warming below 2 degrees Celsius relative to pre-industrial levels, Arctic sea ice will nevertheless likely disappear occasionally in summer even before 2050,” said study lead author Dirk Notz, who heads the sea ice research group at the University of Hamburg in Germany. “This really surprised us.”‘
nigeljsays
Video by Kate Marvel on climate models, feedbacks and tipping points, aimed at the general public.
Really clearly presented, but starts with saying “all models are wrong, some models are useful”. This is apparently the standard scientific understanding, but it really doesn’t seem like a great opening line for the general public, because its like this massively negative statement, will confuse the general public, and is quite naieve, and is spoon feeding the denialists who will just repeat scientists admit climate models are wrong (omitting the bit about being useful). Just dont say things like that. Surely there are more subtle ways of talking about strengths and weaknesses of models when talking to the general public?
Having updated my TSI plot with last data to the end of Feb (graphed here – usually two clicks to ‘downoad your attachment), it is showing no sign of any up-tick into Sunspot Cycle No25, TSI still bobbling along at the same low level seen through 2019.
Re 59 Where Prof Notz says:
“Arctic sea ice will nevertheless likely disappear occasionally in summer even before 2050,”
Does anyone agree with me that if we have one summer where Arctic sea ice disappears, then it will also disappear during every following summers?
This is because the sea ice will not be able to regrow during the following winters to the extent that it will produce enough ice to last the following summer. Once it is ice free during one summer it will have passed a tipping point.
I can find AOD time series for brief periods and regions, e.g. 1984-2002 over China. I can’t find any long-term time series for it. So, I hate to echo Victor, but how do we know the 1940-1970 pause was due to aerosols? We don’t seem to have consistent time series data for aerosols.
MAR 51,
Thanks! That’s exactly what I needed!
Okay, here’s my latest regression (using the GRETL package):
Model 1: OLS, using observations 1856-2011 (T = 156)
Dependent variable: dT
coefficient std. error t-ratio p-value
———————————————————–
const 10.3734 42.5121 0.2440 0.8076
CO2 0.0126877 0.000405670 31.28 7.44e-068 ***
AMO 0.437422 0.0637090 6.866 1.61e-010 ***
TSI −0.0106398 0.0312496 −0.3405 0.7340
AOD −1.98960 0.549712 −3.619 0.0004 ***
Mean dependent var −0.132763 S.D. dependent var 0.389642
Sum squared resid 2.850456 S.E. of regression 0.137394
R-squared 0.878870 Adjusted R-squared 0.875662
F(4, 151) 273.8998 P-value(F) 4.11e-68
Log-likelihood 90.83099 Akaike criterion −171.6620
Schwarz criterion −156.4127 Hannan-Quinn −165.4684
rho 0.123729 Durbin-Watson 1.739247
Excluding the constant, p-value was highest for variable 4 (TSI)
I’m a little surprised that sunlight seems to have no effect over the whole period, or no significant effect. How can that be? But that’s what the numbers say. Everything else–carbon dioxide (of course), ocean temperature sloshing, aerosols, all matter.
BTW, does anyone have AOD data past 2011, or AMO before 1856? I’m unable to find either on the web (probably just my poor search skills).
Okay, this is interesting: I did a test for non-linearity on the variables’ relation to dT, and found a significant result for TSI. With square of TSI included in the regression, sunlight suddenly becomes significant at the 95% level:
Model 2: OLS, using observations 1856-2011 (T = 156)
Dependent variable: dT
coefficient std. error t-ratio p-value
————————————————————–
const 250714 126332 1.985 0.0490 **
CO2 0.0127623 0.000403535 31.63 3.10e-068 ***
AMO 0.437046 0.0630983 6.926 1.18e-010 ***
TSI −368.338 185.604 −1.985 0.0490 **
AOD −1.86337 0.548144 −3.399 0.0009 ***
sq_TSI 0.135284 0.0681713 1.984 0.0490 **
Warning: data matrix close to singularity!
Mean dependent var −0.132763 S.D. dependent var 0.389642
Sum squared resid 2.777534 S.E. of regression 0.136077
R-squared 0.881969 Adjusted R-squared 0.878035
F(5, 150) 224.1711 P-value(F) 1.04e-67
Log-likelihood 92.85240 Akaike criterion −173.7048
Schwarz criterion −155.4057 Hannan-Quinn −166.2725
rho 0.093784 Durbin-Watson 1.798654
Between ice core data and sampling from 1983 on, I was able to construct an 1850-2019 series for methane. Unfortunately, due to very high multicolinearity with CO2 (r = 0.98), the coefficient in a multiple regression comes out perversely signed. I’ll have to look up how to compensate for multicolinearity. I used to know this stuff… *sigh*
Oops. Apparently there are two Ls in the middle of “multicollinearity.” Boy, I really am out of it… Gotta refresh my statistics…
BPL…Agree you very likely have an MCL problem as you suspect. You don’t exactly compensate for MCL, generally. Basically, either you construct combinations using such techniques as PCA (or in this case simple addition of normalized scores since the cor is so high). Or you do separate analyses.
Time to back to your old notes deep in the files!
Reporting of Notz et al (2020) ‘Arctic Sea Ice in CMIP6’ will likely re-invigorate the shouty denialists about predictions of an ice-free Arctic Ocean. (So far, the Wattsupian bog-house wall has only had “2050? What happened to 2014?” scratched into the paintwork.)
Notz was interviewed by USToday about the research.
Video by Kate Marvel on climate models, feedbacks and tipping points, aimed at the general public.
https://www.youtube.com/watch?v=kZNAlEPMrOY
Really clearly presented, but starts with saying “all models are wrong, some models are useful”. This is apparently the standard scientific understanding, but it really doesn’t seem like a great opening line for the general public, because its like this massively negative statement, will confuse the general public, and is quite naieve, and is spoon feeding the denialists who will just repeat scientists admit climate models are wrong (omitting the bit about being useful). Just dont say things like that. Surely there are more subtle ways of talking about strengths and weaknesses of models when talking to the general public?
Having updated my TSI plot with last data to the end of Feb (graphed here – usually two clicks to ‘downoad your attachment), it is showing no sign of any up-tick into Sunspot Cycle No25, TSI still bobbling along at the same low level seen through 2019.
While the denialist numpties seem happy to declare No25 as arrived (and doing their best to wave the possibility of a coming Grand Minimum to stop AGW in its tracks), those monitoring the sun are in no rush to declare such an event with the latest being a December NOAA ‘prediction’ of April 2020 +/- 6 months.
The sunspot count does feature a minimum-so-far in SSN late last year according to this daily data with a long-enough pause in sunspots (14Nov to 23Dec) for some to declare the beginning of No25. Mind, while reverse polarity sunspots were reported in December, the old-polarity sunspots are still appearing at the end of April so No24 is not over just yet.
Here’s a circa 2017 discussion between Neil deGrasse Tyson and Alan Alda about science communication. Good stuff.
https://www.youtube.com/watch?v=syIb73RQqVU
Re 59 Where Prof Notz says:
“Arctic sea ice will nevertheless likely disappear occasionally in summer even before 2050,”
Does anyone agree with me that if we have one summer where Arctic sea ice disappears, then it will also disappear during every following summers?
This is because the sea ice will not be able to regrow during the following winters to the extent that it will produce enough ice to last the following summer. Once it is ice free during one summer it will have passed a tipping point.