By Steve Goddard
Satellite data vs. normalised surface data during the 1998 El Niño
A few days ago I discussed the amplification of satellite temperatures relative to surface data. And Dr. Roy Spencer posted a very helpful description of the root cause.
(1) It is well known, and widely published, that year-to-year surface temperature anomalies are magnified with height, This is due to moist convective transport of heat from the surface to the atmosphere by evaporation and precipitation; It’s been seen in radiosonde data before there were ever satellites.
(2) It is similarly well known that this heat transport takes some time to occur, especially on large space scales (e.g. the tropics) with a 2 month or so average time lag between peak surface temperatures and peak tropospheric temperatures in the case of El Nino/La Nina.
The magnified warming with height is the same effect as the so-called “hot spot” that is expected with *long-term* warming, but which the satellite data do not seem to support so far. Some think this is a big deal, others not so much, and still others think is an artifact of errors in one or more of the measurement systems.
Had-Crut shows this same amplification in the current El Niño (seen below) with positive ENSO months consistently amplified in RSS relative to normalised Had-Crut.
GISS also uses surface temperatures, so it should so the same pattern as Had-Crut. And during the 1998 El Niño it did, but since then has been tracking RSS quite closely with no evidence ENSO effects. During the current El Niño, normalised GISS is running almost identical to RSS. Both have increased by 0.7 degrees since the bottom of the last La Niña.
Perhaps GISS has moved their thermometers up to 14,000 feet, per Dr. Spencer’s description? The graphs below show a steady divergence between GISS and Had-Crut since 1997.
GISS explains the difference with Had-Crut this way: “A likely explanation for discrepancy in identification of the warmest year is the fact that the HadCRUT analysis excludes much of the Arctic ….. (whereas GISS) estimates temperature anomalies throughout most of the Arctic.” In other words, they “estimated” the bright red region across the top of their 2005 map below, which heavily skewed their data towards warmer temperatures.
Below is the GISS 1200 km anomaly map for April, 2010 :
GISS estimated that the entire region above 80N was 4-8.6 C above normal in April, and using their “estimated” dark brown areas – GISS forecasts 2010 to be the warmest year ever.
Not surprisingly, GISS shows 2010 (so far) as being the warmest year ever, whereas Had-Crut (table below) has 2010 in the number five position for January through March.
GISS Arctic estimates appear high relative to other data sources. DMI showed most of April less than 4C above normal.
Similarly, NSIDC showed the area above 80N generally 3-5 C above normal in April.
Conclusion : GISS data vs. satellites does not show any El Niño amplification at 14,000 feet, and in fact appears to have found a mid-troposphere “non-hot spot.” This would seem to nullify the idea that CO2 is causing the observed warming. I personally don’t believe that and consider it more likely that the GISS Arctic data temperature estimates are too high, heavily skewing their global temperature estimates towards the higher temperatures. The fact that GISS El Niño temperatures are not muted relative to satellite temperatures is strong evidence they are incorrect.
Disclaimer: All normalisation uses only shifting, not scaling – so please don’t take the discussion there!
As for the calendar year, it is likely that the 2010 global surface temperature in the GISS analysis also will be a record for the period of instrumental data.