Friday, February 19, 2016

NOAA down by 0.08°C in January

As expected, NOAA global anomaly reduced in January, from 1.12°C to 1.04°C (report here). It was expected because grid methods like theirs give lower weighting to the Arctic, which was very warm, and so relatively more weight to Eurasia, much of which was cold. But it has been a month which emphasises these divergences. I'll show the recent months plot from the results page:



The NCEP/NCAR index pretty much follows GISS, upward. The satellite indices rise in unison. NOAA follows TempLS mesh down, with TempLS grid dropping further. Normally I would expect TempLS mesh to go with GISS, and grid with NOAA. The difference is there, but with TempLS generally dropping more. I'm not sure why.

13 comments:

  1. JMA dropped 0.14 C from Dec, which falls in between NOAA lo (-0.08) and TempLS Grid (-0.18). I guess that NOAA has more infill over the northern lands and pick up more of the Arctic heat.
    However, it is puzzling that the TempLS indices are about 0.08 C lower than their comparisons. I suspect that it is SST "contamination" of land and sea ice air temps. The map over reporting stations and SST, show several green crosses (SST) in the Canadian arctic archipelago. There may be some SST reports from occasionally open straits there, but it is not right if such anomalies (likely -1,8 C = anomaly 0 C) are allowed to dominate over the few land stations that correctly report anomalies of at least +5 C.

    ReplyDelete
    Replies
    1. Olof,
      I think SST diffusion is a problem with TempLS mesh, and could have cooled the Arctic. TempLS grid is less susceptible, since it does really no infilling.

      Of my other integration methods, Spherical Harmonics went from 0.997 to 0.897, and Infilled Grid went from 0.959 to 0.821. It's not so surprising that SH agreed with mesh - I think these are the best methods - but they are very different. Infilled grid was like grid, but not quite so down.

      Delete
    2. Nick,

      The Thing to GISS is, that always be some differences caused by different climatologie. In speaking of 1981-2010, GISS went down by 0.02 from Dec 2015(0.7K) to Jan 2016(0.68K)

      Use their Tool: http://data.giss.nasa.gov/gistemp/maps/

      Delete
    3. Nick, try to make a zonal mean comparison of Giss and TempLSmesh. Or more simple, cut out the segment north of 70 degrees. If I do that with Giss, January is actually down 0.13 C from December. And the piece north of 70 degrees is +6, hotter than ever before..

      Delete
    4. Olof,
      Yes, I'm planning to make comparison maps, in Karsten's style. But not till next week - busy weekend ahead.

      Delete
  2. It seems TempLS is not following GISS because of Arctic warmth. Really impressive GISS with +7°C for 80 to 90°N. February still very hot in the Arctic.
    Forecasts for 2016 ? NCEP CFSv2 looking very, very warm in march... Could be a surprise with GISS announcing +1,2/1,3°C in march... I would say such a start in 2016 could bring the annual anomaly as high as +1°C.

    ReplyDelete
    Replies
    1. Anomaly numbers that big would change my mind on 2016, and we'll soon know.

      Delete
  3. Nick, just noticed the latest spike ncep/near. A new record?

    ReplyDelete
    Replies
    1. Bill,
      Yes, it is. Beats a 0.975 in December. And seems to be continuing.

      Delete
    2. Normally, we should see a small drop for the next days but it's still impressive. Your NCEP GISS equivalent is showing +1,379. COP21 wanted to limit global warming to +1,5°C above preindustrial. We're way above in february (something like +1,6°C maybe) but it seems not getting so much attention, curiously...

      Delete
    3. Beats a 0.975 in December. - Nick

      These records are so long lasting. It just not realistic to expect that one to fall during my lifetime.

      Delete
  4. Cowtan &Way for January is here, 1.04 C, up 0.05 from Dec.

    The prospects for February are rising, the most far reaching by Karsten Haustein and Weatherbell suggest + 0.15 and 0.16 respectively, compared to January.
    Right now there is an impressive spike in Karsten Haustein Haustein GFS, a 24 hour average of 1.03 above 1981-2010.

    ReplyDelete
  5. Up to Feb 28, February is averaging 0.832 °C in NCEP/NCAR, 0.17 °C over January and 0.21 °C above December. If this has any reflection in GISTEMP and other global surface data the month will be another remarkable first.

    ReplyDelete