XPost: alt.global-warming   
      
   (This is me here: kaggle.com/kymhorsell1).   
      
   The history of what charged particles do to the earth's climate is   
   somewhat fraught and amusing at the same time.   
      
   50y ago climate deniers on WUWT & elsewhere used cosmic rays and solar   
   radiation (some but not total overlap between these 2 things ;) to   
   "explain" why the earth's global warming -- responsible for keeping   
   the planet 30C warmer than it would be for its distance from the sun   
   -- was caused by "sunspots" and particle radiation from the sun.   
   (There is no scientific argument from the point of view that the sun's   
   light radiation warms the earth).   
      
   While cloud chambers obviously are a thing -- they are the traditional   
   instrument used to find charged-particle radiation because protons and   
   electrons whizzing through a chamber with air set to condense actually   
   DOES condense small water particles around them, leaving a visible   
   trail in the chamber -- it was speculated that they should also have   
   an observable effect on the earth's climate via cloud formation.  It   
   was then a matter whether the clouds formed would be "high" -- which   
   tend to bounce more heat radiation coming up from the earth back down   
   to the earth thereby creating more GW -- or "low", which tends to   
   bounce light from the sun back out to space thereby cooling the   
   earth. So which was it?   
      
   CERN set up a big experiment to verify what was what one way or the   
   other.  The results I saw published argued along the lines for being a   
   very small and not significant effect whichever way it was. And the   
   subject seemed to be closed.   
      
   But we have the data. So we should look at it ourselves.   
      
   I worked at one point for some Russian Academy on a project setting up   
   a global cosmic ray telescope. My end of the deal was doing some of   
   the data processing, not designing cosmic ray telescopes and setting   
   them up in Greenland or Uzbekistan, or wherever. But as usual part of   
   the fee was to get their database to play with at some later point.   
   That point has never really come up in the past 20+ years. But maybe   
   now is the time. :)   
      
   We can ask an AI-boosted stats package to check whether there is a   
   robust relationship between cosmic rays and changes in the earth's   
   climate, and if so what they are. It should be pointed out first up   
   that cosmic ray telescopes don't actually detect cosmic rays as a rule.   
   They detect "secondary radiation" that comes down from the upper atm   
   after extremely high cosmic ray particles -- usually bits of small   
   atoms -- come charging into the upper atm and mightily smack one or   
   other of the air molecules up there. So there's some ambiguity usually   
   whether what is being see on the ground is from the sun or the wider   
   universe.  It's generally accepted that really high energy cosmic   
   particles come from way out there, possibly left over from some super   
   event, maybe even been travelling toward us since the start of the   
   universe itself.  (There are some super high energy particles that   
   move so fast time has virtually stood still for them since just after   
   the Big Bang).   
      
   So I have divided the Russian data -- 100s of different stations   
   around the world since the 1970s -- into some time series based on   
   monthly averages, and compare that with data for global temperature   
   and e.g. cloud formation (for which I have some sat data left over   
   from another job).   
      
   And the net result is this:   
      
   Cosmic ray 	Temp 	Params  R2   
   cosmic-wtlat	glb	12o2	0.33778916   
   cosmic-LARC	glb	1o2	0.13584963   
   cosmic-CLMX	glb	1o2-x	0.13457051   
   cosmicseg-60	glb	6o2	0.13331150   
   cosmic70lng120	glb	6o2-x	0.13193978   
   cosmic-THAI	glb	12o2	0.13128931   
   cosmicseg120	glb	6o2-x	0.13096583   
   cosmicseg150	glb	6o2	0.12558309   
   cosmic30lng-110	glb	1o2-x	0.12406043   
   cosmic-IRKT	glb	6o2-x	0.12270517   
      
   The first column is the "X" used in a robust time series regr   
   (ARMA(1,1) with bells and whistles). The 2nd is the global temp ("Y")   
   data-file -- in this case NASA's surface temp going back to the 1850s   
   in 1/100th deg C. The "params" column tells me what mucking around the   
   AI did to get the result its posting. And the "R2" shows what fraction   
   of the monthly ups and downs of the Y are "explained" by matching ups   
   and downs of the "X".   
      
   All results are tested 2 different ways and check out 95% or better to   
   be not just lucky data.   
      
   We see the best model found is the "cosmic-wtlat". This is an   
   invention of my own. Using the value of cosmic radiation at a given   
   station we calculate a "weighted average latitude" over all the   
   stations.  This very very very roughly gives the direction relative to   
   earth where the big source of cosmic rays was for each given month.   
      
   Sounds spurious, but turns out this is the best measure to predict 1/3   
   of all those little ups and downs in global temps.  While the trend   
   has been steadily up for the past 150+ years according to the NASA   
   data-set (or 2000 years according to 100s of data-sets maintained by the   
   PAGES2k group; and 100s of 1000s of years according to groups that use   
   tree rings, shellfish, ice cores and lake sediments as temperature   
   proxies) maybe 1/3 of all the little jags up and down we see in the   
   "glb" plots seem to relate to cosmic rays.   
      
   But more importantly. What is the "sign" of the change?  Given this is   
   an "average latitude" is the earth slightly warmer if the big source   
   of cosmic rays that month is in the NH sky or is it bigger if the   
   source is in the SH sky?   
      
   Here's the model:   
      
   y = -2.038454e+00*x + 9.096947e+01   
   beta in -2.03845 +- 0.236692  (90% CI)   
   alpha in 90.9695 +- 1.86585   
   T-test: P(beta<0) = 1.000000   
   Rank test: calculated Spearman corr = -0.860809   
   	Crit val = 0.432 2-sided at 1%; reject H0:not_connected   
   r2 = 0.33778916   
      
   I.e. we are 99% sure in 2 different ways (we might like to think of   
   that as 1 chance in 1000 or even maybe 1 chance in 10,000) this is not   
   just down to luck. There is a link.   
      
   So it seems the sign is -ve. The further NORTH the big source of   
   cosmic rays is this month the COOLER will be the earth to the tune of   
   .02 deg of temp for each deg of latitude further north.  The data-set   
   for wtlat shows it varies erratically N and S of the equator from   
   month to month and has no statistically robust trend.   
      
   If we look at the station that seems to predict global temps best we   
   find this:   
      
   Cosmic data     Temp    Params  R2              Beta          stderr(Beta)   
   cosmic-LARC	glb	1o2	0.13584963	-0.0117559 +- 0.0024414   
      
   I.e. the station "LARC" (in Antarctica) seems to find about 14% of   
   variations in global temp are down to cosmic rays detected down there.   
   For a month with e.g. 100 more events per sec average for the month   
   global temps go down .012 deg. (Remember, the "glb" temps are in   
   1/100th deg C).   
      
   So there is some evidence we can find that more "cosmic rays" (i.e.   
   charged ions that come down from the upper atm for some reason -- many   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)