I probably should’ve looked at this before placing my bet, or giving my update. However, here’s a graph showing the average daily anomalies for Aqua CH5 from the AMSU website for a month vs. the reported UAH anomaly for that month.
This starts in August 2002, since that is the earliest data available on the website for the daily anomalies. For anyone curious, the r-squared value is 0.80, with the linear trend equation of UAH Anomaly = 0.9283 * AQUA + 0.1854
So we’re looking at about 0.18 C warmer anomaly report from UAH than the average daily anomalies, if you’re using those anomalies to guess the upcoming temperature report.
Update (12/20)
The daily anomaly averages were calculated relative to the “Average” line available at the AMSU website. Given the difference we generally see between these daily anomalies and the reported UAH anomaly, it’s easy to think that they are just being reported relative to a different baseline.
However, according to Roy Spencer here , the average line is indeed the 1979-1998 average. The UAH dataset here also reports relative to 1979-1998, as can be confirmed by summing those anomalies (and getting 0).
To dispell the notion that the “Average” line on the AMSU website is simply the 2002-2009 average, here is the difference:
Hmmm. I haven’t thought deeply enough about this. But why shouldn’t the Channel 5 (blue) and UAH (red) lines superimpose?
Comment by AMac — December 21, 2010 @ 2:37 pm
In theory, my thought would be that they should (which is why I originally suspected a different baseline). According to the link to Spencer’s blog above, the daily AMSU uses NOAA-15 rather than NASA’s AQUA satellite, although this is confusing because the AMSU website graph says (AQUA ch5v2). But I suppose the NOAA-15 satellite could be suffering from drift, thereby skewing the results.
Still, it seems odd that the NOAA-15 mid-tropospheric data would consistently show such a smaller difference from the 1979-1998 average than the overall average tropospheric temperature. I still can’t find any of that daily data from 1979-2002 either that forms the “Average” line on the graph.
Comment by troyca — December 21, 2010 @ 10:53 pm