Sunday, March 13, 2016

Huge rise in GISS in February - now 1.35°C.

GISS in Jan was 1.14°C, revised up from 1.13. We had been expecting a big rise - maybe to 1.3°C. But this rise of 0.21°C way exceeded expectations. And of course it is the hottest month ever, etc, and puts 2016 well on the way to being hottest year. As well as the surface TempLS, the reanalysis and satellite measures showed similar rises (satellites even higher).

Here is a plot of the comparison with 1998. In fact the jump was slightly less than the big Jan-Feb jump in that year, and just lately, the two sequences move in parallel, with 2016 0.4-0.5°C warmer.

I'll show the world GISS map below the fold. It shows the same general pattern as TempLS; a band of warmth from Europe through Russia to the East, and another through Alaska to central Canada, plus Arctic. I think part of the reason for GISS rising more than TempLS is that it had the Arctic warmer. As usual, the WebGL GHCN/ERSST plot has more detail.

Here is the GISS map:



And here is the corresponding TempLS spherical harmonics plot:







14 comments:

  1. I was going to say that this is the new normal, but of course, it's not. Temps will regress to the mean sometime after the El Nino runs its course later on this year. However... they will settle at a higher 'normal' this time, probably around 0.85 C above the 1951 - 1980 mean. That's a prospect more than a bit scary for those who understand what's actually happening to the world around us, but meaningless to the authoritarian types that largely inhabit the denialosphere.

    ReplyDelete
  2. They will regress to vary around the trendline, not the mean, of course, but that's what you meant, I'm sure. Assuming white noise, the annual residual puts 1998 at around a 2sigma event depending on series. It took 18 years for the trend to catch up to that cherrypickable value. It is looking like 2016 could turn out to be on the order of a 3 sigma event. Boy will that make the denier crowd happy on a go forward basis!

    ReplyDelete
  3. jgnfld, I think Matthew England's 2014 paper (doi: 10.1038/nclimate2106 ) is interesting. It implies that trade wind strengthening contributed something like -0.05 C/decade until recently and that's linked to things like Pacific temperature patterns (e.g. Kosaka & Xie, 2013) and the Pacific Decadal Oscillation (some of Trenberth's papers)

    The England et al. results suggest that instead of the data falling along a +0.17 C/decade trend we should expect more like +0.22 C/decade if the winds stay at the same strength, or +0.27 C/decade if the winds return towards normal strength. Assuming nothing else major changes, it looks like the odds favour accelerated warming. So long as you calculate that by forcing the trend lines to join up and don't allow magical temperature jumps between them.

    ReplyDelete
    Replies
    1. I agree with him. Unless England's wind returns and stays for awhile, the trend by 2020 could exceed .20 ℃ per decade.

      Delete
  4. There is a large difference between GISS and TempLS in the Arctic. GISS is much warmer there especially around Svalbard. GISS 250 km smoothing covers not the Arctic and has 'only' 1.25°C compared the 1200 km smoothing with 1.35°C. So which Arctic extrapolation is more realistic? GISS 1200 km or TempLS mesh? If I look at NCEP/NCAR reanalysis data for the Arctic it looks more like GISS. There is also very warm air (compared to climatic mean) above the Arctic.

    ReplyDelete
  5. Uli, TempLSmesh extrapolates SST over Sea ice (and land) in places where meterological stations are few. There is a lot of SST up in the Arctic this Winter since the ice extent is very low. Watch the green SST crosses on this map
    Giss and Hadcrut kriging use ice (and land) masks, which only allows air temps to be infilled over those areas. Months like the recent ones, when air over ice anomalies are very high, whereas SST-anomalies are close to zero, there can be significant differences between the TempLSmesh and the Giss method.

    ReplyDelete
  6. @Olof: Yes, I think so too. An additional difference seems to be that GISS does not use SST near (100km) land stations even over open ocean.

    I don't know which topic it fits but looking to the year zonal GISS temps the 64S-90S was -0.32, the coldest anomaly since 1976. So the last year record was despite this Antarctic cold. Also the previous 64S-90S anomalies in the first year of en El Nino were negative, f.e. -7 in 1997. Is this systematic?

    ReplyDelete
    Replies
    1. @Uli: Antarctica was extremely cold during the SH winter season (Apr-Sep), dragging down global temperatures by approx 0.2K for 6 months in a row. Check this plot (difference between black and grey line is caused by the Antarctic region): Timeseries GFS anomalies 2015

      Makes sense as (tropospheric) SH winter circulation is much more bound to be stable due to the very strong stratospheric vortex which only breaks up on spring. While ozone and GHGs are the main players in establishing the trend (in terms of changes in the circumpolar jet position and strength and hence temperature over Antarctica itself), ENSO does have an impact on the jet as well. There is a clear teleconnection pattern which tends to weaken the circumpolar zonal wind (not shown) and to increase the mean sea level pressure (MSLP) in response to El Nino. This plot shows the PDV/PDO response in NCEP Reanalysis and our HadAM3P model ensemble (200 member average): NCEP and HadAM3P MSLP teleconnection

      So yes, there is a systematic El Nino signal down south. MSLP increase should cause cooling. Whether the record cold 2015 Antarctic winter season is only due to El Nino? I strongly doubt it. Disentangling all three factors isn't straight forward, which is particularly true for the temperature response, even though we have some good grasp of the expected forced circulation changes.

      Delete
    2. Just noticed. I did fool myself a bit. Higher MSLP usually correlates with warmer conditions (simple feedback especially over sea where warmer water means more buoyant conditions and hence reduced MSLP). It's less straightforward over land, in winter in particular. This is what led me to the (perhaps) hastily conclusion that higher MSLP over Antarctica might mean lower temperatures. In fact, the same plot for 2m temperatures doesn't quite support my idea: NCEP and HadAM3P 2m Temperature teleconnection

      HadAM3P does cool Antarctica. NCEP Reanalysis doesn't. However, as reanalysis over the SH is essentially just another model result, my feeling is that NCEP isn't doing particularly great over Antarctica. In any case, what we mighgt see instead of an ENSO response is an anomalous annual cycle caused by forcing which only happens to be stronger recently ...

      Delete
    3. Thanks. I just asked me why the difference between a very hot NH and a cooler Antarctica becomes so large in the last time.

      Delete
  7. Hadcrut4 kriging (Cowtan&Way) has reported the February anomaly now. There was only a modest rise, up 0.07 from January to 1.12 C.
    I had expected more, a rise of maybe 0.15-0.20 C, slightly less than Gistemp since HadSST3 decreased more than ERSSTv4 in February. So I'm a little surprised with the moderate rise in Hadcrut kriging, and can't really find an explanation when I break down the global components, ie Land, HadSST, air over Arctic Ice..

    However, Feb 2016 was 1.53 C over the preindustrial temp (ie the 1850-1899 february average)

    ReplyDelete
    Replies
    1. Maybe the surprise is that January was high for C&W; it went up by 0.07 when some others went down from Dec.

      Delete
  8. Yes, in the long run such differences will likely even out..
    Looking at the details for Feb there might some significant differences in station coverage between Gistemp and Hadcrut, eg in Antarctica but also in the Arctic. There is no Feb data for Franz Josefs land in Hadcrut4 , a northern outpost that will be kriged over large portions of the Arctic ocean. According to other sources there is a very large anomaly there, about +15 C from 1961-1990.

    ReplyDelete
    Replies
    1. Btw, I made a zonal comparison between Gistemp loti and Hadcrut4 kriging

      The Feb temps are 1.23 and 1.12 C respectively on the same 1961-1990 base. I guess that half the difference is due to a cooler HadSST, and half the difference due to cooler temps north of 70 degrees. Can it be the missing Franz Josef Land only? Normally the Arctic should be warmer in Hadcrut, since it doesn't suffer from the GHCN v3 Arctic cooling bias

      Delete