Weather Banter - View Single Post - WARMEST SOUTHERN HEMISPHERE SEPTEMBER IN 130 YEARS OF NASA DATA!

Tom P[_3_]

Peter Muehlbauer wrote:
Tom P wrote:

Peter Muehlbauer wrote:
Zorro wrote:

On Oct 28, 7:21 pm, Roger Coppock wrote:
WARMEST SOUTHERN HEMISPHERE SEPTEMBER IN 130 YEARS OF NASA DATA!
It's 2.4 SIGMA above the mean Southern Hemisphere September
and 1.5 SIGMA above the 130-year linear trend.

In the real world,
outside the fossil fuel industry's spin and lies,
global mean surface temperatures continue to rise.
Please see:

http://www.metoffice.gov.uk/corporat...20080923c.html

These hemispherically averaged temperature data come from NASA,http://data.giss.nasa.gov/gistemp/tabledata/SH.Ts.txt
They represent the results of millions of readings
taken at stations covering the lands of the Southern
Hemisphere over the last 130 years. Yes, the data
are corrected for the urban heat island effect.

The Mean September temperature over the last 130 years is 14.023 C.
The Variance is 0.1214.
The Standard Deviation, or SIGMA, is 0.3484.

Rxy 0.5073 Rxy^2 0.2573
TEMP = 13.71464 + (0.00471 * (YEAR-1879))
Degrees of Freedom = 128 F = 44.355302
Confidence of nonzero correlation = approximately
0.999999999 (9 nines)
The month of September in the year 2009,
is linearly projected to be 14.327,
yet it was 14.85. - 1.5 SIGMA above the trend,
therefore, the warming accelerated.
The sum of the absolute errors is 30.7605

Equal weight exponential least squares fit:
TEMP = 13.721718 * e^(.0003341 * (YEAR-1879))
The sum of the absolute errors is 30.7380

Rank of the months of September
Year Temp C Anomaly Z score
2009 14.85 0.827 2.37 --
1882 14.82 0.797 2.29
2008 14.80 0.777 2.23
1996 14.65 0.627 1.80
2007 14.64 0.617 1.77
2003 14.63 0.607 1.74
2001 14.63 0.607 1.74
2005 14.62 0.597 1.71
1989 14.61 0.587 1.68
1983 14.57 0.547 1.57
2002 14.55 0.527 1.51
1988 14.55 0.527 1.51
2006 14.49 0.467 1.34
1991 14.48 0.457 1.31
MEAN 14.023 0.000 0.00
1917 13.58 -0.443 -1.27
1968 13.56 -0.463 -1.33
1954 13.56 -0.463 -1.33
1964 13.54 -0.483 -1.39
1931 13.53 -0.493 -1.42
1923 13.52 -0.503 -1.44
1906 13.52 -0.503 -1.44
1903 13.51 -0.513 -1.47
1902 13.48 -0.543 -1.56
1892 13.44 -0.583 -1.67
1887 13.38 -0.643 -1.85
1925 13.29 -0.733 -2.10
1935 13.27 -0.753 -2.16
1894 13.26 -0.763 -2.19
1891 13.13 -0.893 -2.56

The most recent 40 continuous months, or 3 years and 4 months,
on this SH.Ts.txt data set are all above the 1951-1980
data set norm of 14 C.
There are 1557 months of data on this data set:
-- 791 of them are at or above the norm.
-- 766 of them are below the norm.
Remember, this is Roger's data, I didn't make it up.

Rank of the months of September
Year Temp C Anomaly Z score
2009 14.85 0.827 2.37 --
1882 14.82 0.797 2.29

OH NO, September 2009 in the Southern Hemisphere, is 0.03 C warmer
than September 1882 in the Southern Hemisphere.

In 127 years, we have warmed by 0.03 C.

Quick, we have to do something fast, before the temperature drops
below what it was in 1882. Quickly everybody, generate some more CO2.
See below for an explanation of your foolishness...

I'm looking and looking, but can't see any explanation...

More CO2 won't affect temperature.
There is no link between temperature and CO2.

And Roger et al is still not able to explain, where all that doooooming
warming is coming from but from a ****ty peace of computer software.

I looked at the GISS data using Open Office spreadsheet and sure enough,
the results are the same as Roger's. Peter M is welcome to repeat the
analysis himself using any other commercial package like Excel.

If you plot the data out, several things become apparent.
Firstly, the trend is monotonous, there is a continuous linear rise in
temperature of around 0.04 per decade.
Second, the scattering of the data points is very large from 1880-1890
then becomes steadily more stable.
Thirdly, data prior to 1880 is missing - there is none available to give
the DJF and D-N fields.

These last two observations suggest that the 1880-1890 data are based on
a very small number of recording stations. Bear in mind that in the
19th century, this would be someone looking at the thermometer and
writing down a value to the nearest whole degree. In other words, what
we see in the data for the first decade is simply digital noise caused
by averaging a small number of discrete values.

That means that singling out and comparing a single 1880 data point with
the 2009 data point is simply comparing a white noise signal peak with a
highly reliable value - totally useless and misleading.

... but only 91% defective data, you used to draw your conclusions.

How much significance would you give your pleadings?

Peter, you still haven't told us the result of your analysis.