What integrity is there in global SST trends ?
August 11th, 2006 by Warwick HughesI made this map of sea surface temperature (SST) trends at the GISS website module, and it shows the 27 year trends 1979-2005 for an SST dataset GISS names HadReyn_v2 which I take in to mean the Hadley Centre version 2 SST’s with Reynolds data grafted on post 1982.
There may be another explanation of exactly what the dataset is but for sure it is the SST data Jim Hansen’s GISS group chooses to use.
I have annotated several warm anomalies that caught my eye and have compared UAH satellite trends over those warm patches with the the combined land sea HadCRUT2 data calculated at www.co2science.org.
See the Table for stunning differences in HadCRUT2 minus MSU, up to 1 degree C in 27 years !! The table also sets out the Lat & Long for various five degree grid cells that include these warm SST patches. For some history of the development of the SST datsets go to my page showing graphics of huge corrections required in raw SST data.
Posted in Climate indicators, IPCC, Jones et al, Surface Record | 9 Comments »
August 13th, 2006 at 7:14 pm
It looks like the 1200 km radius smoothing used by GISS picks up information from both SSTs and land air temperatures and averages them together. The result is an overestimation of the trends over the oceans near the coasts of every continent. You also see it about isolated islands such Tahiti in the south Pacific.
In effect, the reported trends in SST are no longer purely arising from oceanic properties. Rather, UHI effects in coastal cities can influence trend values up to 1200 km away from the coast. That would be my interpretion of what is going on.
August 14th, 2006 at 11:16 am
Three things I’ve often wondered about:
1) impacts caused by use of seawater for cooling systems
2) impacts caused by sewage outfalls in non tropical seas
3) impacts of urban run off in non tropical seas
August 14th, 2006 at 12:52 pm
Steve, I’m sure such things can and do have local effects in restricted shallow coastal waters, but considering that the effect of 50 or so cubic miles of annual glacier melt isn’t all that apparent off East Greenland, I can’t imagine they could be significant in terms of a global data set. The ocean is very, very large.
Re Doug’s comment: I think there are pure SST data sets, so why not give it a try with one of those? I didn’t follow the literature all the way through before the lovely John A. shut down the Climate Audit thread, but as far as I can tell a) an awful lot of work has been done on every aspect of this (unsurprisingly given the length of time the UAH controversy has been going on and b) the anomalies should flatten out given enough time, although that statement was in a single abstract so I couldn’t tell what it was based on. The abstract I refer to has Hegerl as one of the authors and cited the second linked paper (if I recall right, but it definitely cites one of the three and so should be easy to find if anyone really cares).
Warwick, it seems pretty clear you didn’t do a literature check before doing your data comparison. I think that would be appropriate due diligence in future.
Anyway, here is the comment I was prevented from posting at CA:
“Re #36: Doug, considering the 2LT/TLT is taken at a 3.5 kilometer height, why exactly would we expect anomalies between that height and the sea surface to match up, even over a long period of time? Certainly some degree of correlation would be expected, but complete? An obvious way to resolve the issue is to look at the different atmospheric layers measured by the MSU instruments and see if the globally averaged anomalies are even faintly comparable. It turns out they aren’t. If these charts look that different, one would think there would have to be substantial geographic anomalies.
“A quick resort to Google Scholar (what Warwick should have done before going to all that effort) shows that the existence of substantial regional anomalies beyween the surface and tropospheric data sets was known about and analyzed more than a decade ago; see e.g. here. The cites to that paper, in particular this one, show the vast amount of scientific attention given over to such comparisons, including by Christy and Spencer, but somehow nobody has noticed up until WH just now that there’s am unresolved scientific issue? Far more likely, and as the literature demonstrates, they’ve all been very aware of it but have found it not especially remarkable.
“I should note that in going over all of this I haven’t directly double-checked any of WH’s work.”
August 14th, 2006 at 1:57 pm
Somewhat related, in that it reveals an effect of overreliance on flawed models. This is the 3 month lead temperature forecast from NOAA. It was issued in July. When you get to a 3 month lead, this is where weather forecasting starts to overlap into climate forecasting and where the same structural flaws in the GCMs that have buggered the world of “Climate Science” reveal themselves. Here is the forecast:
www.weather.gov/climate/l3mto.php
This forecasted higher than normal temperatures for much of California during the period. In fact, August to date has been either normal or below normal, depending on the specific location. The model failed to forecast a series of long wave troughs and short waves brought on by a prominent dip in the jet stream. It is obvious that the models would have had the jet stream actually in a complete opposite pattern.
August 15th, 2006 at 7:01 am
Warwick (and others),
Perhaps you can help correct the simple yet very significant misrepresentation of the data that’s present in almost every SST map I’ve seen: significant magnification of polar area, which leads to serious visual interpretation errors.
It would not be difficult to correct this. Equal Area projection algorithms are widely available. I recommend taking a look at this (and specifically this or perhaps this)
A very interesting projection from Snyder is discussed here, including source code etc.
August 15th, 2006 at 11:27 am
Mr. Pete – good point. Flying from the US to Europe it is apparent just how short the distances between landmasses at the North end of the Atlantic truly are. You fly over Greenland then next thing you know you are over Iceland. Then Ireland comes up pretty fast after that. It’s no wonder the Vikings made it to North America when they did.
August 17th, 2006 at 6:48 am
Just in case you missed it Warwick, here’s what Steve Bloom says about you on other weblogs:
and
You should know that Steve Bloom attacks your character and scientific integrity on other websites and then acts all prissy when its pointed out. Try getting him to apologize or retract his statements before hosting any of his posts.
August 25th, 2006 at 6:14 pm
Somewhat related Pielke Sr. has a couple of new posts at his blog about a recently accepted paper regarding claimed 2003 – 2005 apparent SST / upper water column cooling. Recommend checking them out.
climatesci.atmos.colostate.edu/2006/08/14/more-information-on-the-geophysical-research-letters-article/
climatesci.atmos.colostate.edu/2006/08/10/new-geophyical-research-letters-paper-accepted-recent-cooling-of-the-upper-ocean-by-jm-lyman-jk-willis-and-g-c-johnson/
October 19th, 2007 at 6:17 pm
Re #3:
Steven Bloom exemplifies one of the qualities the RC Team share in common: a failure to grasp the defining concepts of the scientific endeavor. One must show one’s work; mere riff-raff are obligated to reproduce your results; your work must result in testable predictions; you are graded on the outcome.
Even if the SST results above are legitimate, they bode ill for AGW, we expect random variation between SST and MSU measurements, not these uniform gradients.
No matter, if one compares SEC daily data with that of the last SS minimum we can see AGWer’s are all lost and adrift.