Seastep's WunderBlog
The Logical Temperature Record
Posted by: Seastep, 1:57 AM GMT on March 05, 2010 +0
Many thanks to atmoaggie for the great looking graphs. Beat mine by a long-shot. Just compare the first three to my fourth.

Sources are simply the Land+Ocean UAH satellite monthly temperature data set (averaged to annual) and the GISS annual data set, representing one trip around the sun.

Pre-reading note: The trend of the GISS temps up to 1978 and the UAH temps for 1979 forward is unchanged in the final result. i.e., all points remain the same distance apart. The idea is to create a data set that combines the spatial coverage of satellite obs with actual surface obs. The result should be a much more accurate data set. Certainly so over time, it is statistically so. Also, please ignore the "cold bias" reference in Figure 2. It might be accurate, but is irrelevant and should not be there. I had originally put it there only because I picked 1997, but had I picked 1998, it would be a "warm bias." Whether there is a bias or not is irrelevant, that is precisely the point of doing this.... objectivity.

First, I created an annual data set from the average monthly UAH data.

To calibrate the data sets to the same base year, select any year. I chose 1997, but it doesn't matter. The only difference would be where they have a common year. The Figure 2 still ends up the same, only did it to illustrate the calibration as a separate step. So, in the case of the UAH data set, all values are reduced by 0.05 (1997 value) and in the case of the GISS data set, all values are reduced by 0.40 (1997 value).

This is resulting graph:


Figure 1.


Next, I took the difference between the annual average of each data set and adjusted all the GISS values by that amount (0.094). The reason I started this project was to get a data set, objectively, as close as possible to the UAH data set for the years prior to 1979 to use in conjunction with the UAH data for longer term comparable data.

This is the result (much better):


Figure 2.


So now, all data points are calibrated and normalized resulting in the UAH data being a valid substitute for the GISS data for 1979-2009.

In order to substitute, the UAH needs to start at the same point the GISS leaves off, so the UAH data set is adjusted so 1979 is equal to the GISS 1979 (0.095).

Here is the full raw GISS record for comparison:


Figure 3.


The end result of the above is this graph with the UAH now substituting for the GISS data for 1979 forward. I believe this to be the most accurate temperature record. And it will only get more accurate over time.



Figure 4.


Please definitely let me know if there is some flaw in the logic.
Updated: 6:02 AM GMT on March 05, 2010   Permalink | A A A
View Full Blog Entry — View Comments (9)
Mea Culpa
Posted by: Seastep, 5:12 AM GMT on February 26, 2010 +1
Data is the data. I modified the original, rather than a new entry purposefully, so anyone clicking a link does not get false info.I was disorganized in my initial work. Upon doing my blog entry on the logic, I completely separated it from other work and realized the mistake.My apologies to all. The first two graphs are valid, and I post them here again. It was not fruitless, though. There is still a difference if you compare the raw GISS to the combined GISS+U...
Updated: 8:14 PM GMT on March 04, 2010   Permalink | A A A
View Full Blog Entry — View Comments (15)
Solar Cycles and Earth Temperatures
Posted by: Seastep, 3:37 AM GMT on February 11, 2010 +0
Correction: Sorry about the title I forgot to change on that first graph. It is relative to 1997.Rotating sunspots produce almost 300% more energy (Maria D. Kazachenko et al 2009 ApJ 704 1146-1158).The previous 11-year sub-cycle (cycle 23) that started in 1997 had 12% more rotating sunspots in the Northern hemisphere during the peak and the most rotating sunspots on record (X. L. Yan et al 2008 ApJ 682 L65-L68).Therefore, the location and, especially, rotation m...
Updated: 5:12 PM GMT on February 11, 2010   Permalink | A A A
View Full Blog Entry — View Comments (1)
Forecast validation: Actual Temps vs. IPCC AR4
Posted by: Seastep, 4:43 PM GMT on February 06, 2010 +0
Trend based on actual temps 2000-2009 relative to the year 2000.Dashed lines are the entire model range. Solid lines are IPCC most likely. Green line is actual.Note: 2005 appears to be missing, but that is a resolution issue. It is the IPCC most likely maximum line (slight uptick from 2004).Trend based on actual temps 1997-2009 relative to the year 2000.Sources:UAH Satellite temperaturesIPCC
Updated: 7:13 AM GMT on February 08, 2010   Permalink | A A A
View Full Blog Entry — View Comments (0)
The Cone
Posted by: Seastep, 5:52 PM GMT on November 27, 2009 +0
For us tropical buffs, I did an analysis of the IPCC predictions for first through fourth assessments (all available) in the same manner as the NHC cone.I used monthly mean sat temps only and used the UAH data.Actual temps are from Jan 1979 through October 2009. Left side of graphs is Jan 1979 and right side is Dec 2100.The dashed black lines are the entire model range, per IPCC, and the solid lines are their prediction (i.e., their "cone").Green line is actual UAH...
  Permalink | A A A
View Full Blog Entry — View Comments (5)

« View Older Entries

About Seastep
Calc 3 (thru calc 2 in HS) Physics Tons of prob and stats Actuarial science Heavy coding on computer end My citizen scientist creds in 150 charAL5

Tropical Blogs

Tropical Weather Stickers®

Contact This Author
Recent Photos
IPCC 2006 IPCC 2000
UAH vs. IPCC 4 2007 UAH vs. IPCC 4 2100
Seastep's Wunder Photos
Previous Entries for 2013
Personal Weather Stations
Galloways Farm in Country Acres
Parkland, FL
Elevation: 12 ft
Temperature: 78.5 °F
Dew Point: 73.3 °F
Humidity: 84%
Wind: 17.3 mph from the SW
Wind Gust: 17.3 mph
Updated: 1:26 AM EDT on June 20, 2013
About Personal Weather Stations