Examining in closer detail Jones et al 2008 eastern China land/SST grid box where they conclude urban warming at ~0.1 degrees per decade
March 23rd, 2009 by Warwick HughesRemember from the previous post, that Jones et al 2008 in their section 3.3. Comparisons with SST Data, say.
“As it is difficult to develop a network of specifically rural sites in China, we now compare the CHINA-LI and CRUTEM3v series with one based on HadSST2 [Rayner et al., 2006] for the region (20–45 deg N by 110–125 deg E) to the east of China.”
A quick look at the map in an article below, will show you where this grid area fits in east China.
Their results are given in Jones et al 2008 Table 5 which shows temperature trends in their land based CHINA-LI and CRUTEM3v series compared with sea surface trends of HadSST2. Table 5 quotes comparisons over 3 periods, 1951-2004, 1954-1983 and 1981-2004.
The two longest periods measured in their Table 5 do show the land warming more than the sea, at 0.11/decade 1951-2004 and 0.13/decade 1954-1983. These results back their urban warming conclusion. Only the shortest period they checked, 1981-2004 shows a contrary result.
Jones et al quote the HadSST2 sea surface temperature data but there are at least two other global SST datasets worth comparing to attempt to get a fuller picture. We also compare with the Reynolds V2 and HadISST1 SST data, all downloaded from the KNMI Climate Explorer website. Now Reynolds data is only available from Nov 1981, 1982 on a full year basis so I do two comparisons to get as close as I can to the Jones et al 1981-2004 period.
My results on KNMI data are as follows, all numbers are temperature trends per decade
.
1982-2004
Trend per decade CRUT3=0.49, HadSST=0.53, Reynolds=0.42, HadISST1=0.42
So differencing CRUT3 and Reynolds & HadISST1 over this 23 year period indicates urban warming of circa 0.07 degree per decade.
1982-2008
Trend per decade CRUT3=0.45, HadSST=0.40, Reynolds=0.29, HadISST1=0.32
So differencing CRUT3 with Reynolds & HadISST1 over this 27 year period indicates urban warming of circa 0.16 degree per decade (Reynolds) and 0.13 degree per decade (HadISST1).
To conclude, these extra SST comparisons strengthen the Jones et al 2008 finding of urban warming over China.
Posted in IPCC, Jones et al, Surface Record, Urban Heat Islands | 2 Comments »
March 28th, 2009 at 4:36 pm
And as you know, Warwick, there’s an upward step change in the Hadley Centre SST data in 1997 that is caused by their merger of two incompatable datasets at that time. So that shift in the Hadley Centre’s SST data would reduce the apparent urban warming. For those who aren’t aware of the step change, I discussed it in this post:
bobtisdale.blogspot.com/2008/12/step-change-in-hadsst-data-after-199798.html
Regards
March 30th, 2009 at 2:15 am
Thanks for reminding us of that Bob, HadSST trends closer to land rates above. HadISST1 compares more closely with Reynolds. I have not examined the data in detail to see if your step change exists here.