date: Mon, 08 Jan 2007 15:31:18 -0700
from: Kevin Trenberth <trenbertatXYZxyzr.edu>
subject: Re: Science presentation for Paris
to: Susan Solomon <Susan.SolomonatXYZxyza.gov>
One too many 0's. 0.005.
Susan Solomon wrote:
Thanks. This comes up both in the presentation and in SPM language.
A suggested merge of Phil's text below with the SPM language we have implies replacing the
sentence on page SPM-5, 6-7 with the following proposal:
Sites affected by the urban heat island effect are identified and excluded from these
averages, so that remaining uncertainties due to this effect are negligible (less than
0.0005�C per decade).
This would address several comments asking us to explain what is done with UHI.
At 3:52 PM +0000 1/8/07, Phil Jones wrote:
On the UHI (slide 9) we should probably change the middle bullet. The first
and third are not in dispute. May be better to spell out SSTs though, or say
marine air temperatures. SSTs are used as anomalies though to approximate
Middle bullet currently says
o Major influences are identified and excluded from the records used to create the
continental and global values
Perhaps we should refer directly to David Parker's paper on UHIs, where he
couldn't detect any difference in trends (averaged for 200+ cities) in temperatures
on calm nights (when you'd expect the biggest effect) compared to
windy nights (when you'd expect the least).
There are two aspects to the major influences.
1. Some sites are removed. This isn't many as a % of the total (about 1%).
2. We include in Brohan et al (2006) an estimate of urbanization in
the calculation of the errors. This is 0.0055 deg C/decade since 1900.
It is a one-sided 'error'. If you look very closely the error range in
this paper and in some of the Ch 3 figures is slightly one-sided.
This figure comes from Jones et al. (2001) , which came from
Jones et al. (1990).
Difficulty with all UHI work is that there are countless papers looking
at individual sites - which generally use a site in the city centre. This
site is rarely one used in the dataset - generally an airport is instead.
It is made worse by then looking at individual days and not monthly
averages. Only Jones et al. (1990), Parker (2005,2006) and Peterson
have looked at large scales.
Affected site are identified and excluded from the records used to create the
continental and global values (as not all sites are tested, part of the error range
assumes an urban component of 0.0055 deg C/decade)
At 22:47 07/01/2007, Kevin E Trenberth wrote:
Many thanks for the feedback. My comments and explanations follow. I'll
expressly ask Phil to respond to us on the UHI issues and what we should
say succinctly. I am keen to get further feedback on what to exclude. I
had decided to exclude the full slide on all the regional precip trends
becuase it is too detailed and would take too long to go through and so
the zonal mean latitude-time series captures a lot of the changes.
Personally I would like to have both but the issue will be time and
simplicity of message, and hence my decision to drop the series:
implicitly those are included of course because they are in the chapter.
> Many thanks for the preview. I agree that the
> presentation has improved, thanks for that. I
> would like to offer the following suggestions:
> 1) Ramaswamy will cover radiative forcings, and
> will do so comprehensively including aerosols,
> ozone, etc. Calling out CO2 and N2O on your
> title slide will likely raise queries about why
> you cite those and not others. I suggest that
> you drop that bullet from your first slide.
Yes slide 1 at present is more comprehensive and perhaps more appropriate
for you to use. In general with these slides that context will be
desirable but perhaps not for Paris.
> 2) The chapter relates changes in DTR to clouds,
> and possibly aerosols and land use. The chapter
> doesn't explicitly say DTR changes are linked to
> dimming. While I personally would agree this is
> scientifically quite reasonable, your slide 8
> would be easier for people to understand and will
> avoid confusion if its language followed the
> chapter so replacing the word dimming on the
> slide with clouds, possibly linked to aerosols
> and land use, would be helpful.
I understand: indeed we did not expressly say "dimming" but in the
discussion of dimming it clearly relates to clouds and aerosol. My
thinking here is that some may well be aware of dimming but not of changes
in clouds, so I thought that terminology might be helpful rather than add
confusion. Other views appreciated.
> 3) Slide 9 says major influences of UHI are
> identified and excluded. Can this slide please
> be clearer as to what is meant by this and what
> exactly is done? I think it will benefit all if
> we avoid spending a lot of time explaining what
> 'major influences' are and what 'minor
> influences' aren't covered, how big those are,
Let me ask Phil to suggest a couple of bullets.
> 4) A number of governments have asked for more
> clarity on where heavy precip has increased. You
> show it nicely in slide 16 but language on the
> slide will help us when the discussion of
> language comes up. In the extremes table we say
> that heavy precip has increased 'over most land
> areas' and if the title of this slide were
> 'Proportion of heavy rainfalls have increased
> over most land areas' that would be very helpful
> in laying ground for that.
Heavy precip is confusing, because some analysis are in absolute terms:
and others are in terms f the percentage of precip that is heavy. The
latter change is much more universal, and the main exceptions are where
precip amounts have decreased, implying a drier regional climate. Since
our report there is anew report in Science on extremes in India in the
monsoon increasing and there they talk about real extremes. In the slide
we already say "proportion of heavy rainfalls are increasing" so the
suggestion is to add "most land areas"? OK.
> 5) What is the reference for slide 20? it's a
> nice image but if it's not in the report then
> we'll need to discuss that. Slide 19 covers
> similar content very well, I think so the second
> one on pdsi could be dropped.
Slide 20 is from Dai et al 2004. It is extensively discussed in the full
report in section 3.3.4 and was featured in some email discussions for
the TS related to the trend in the previous slide, resulting in some
refinement of the FAQ 3.2. Whereas slide 19 is for all of PDSI, slide 20
separates out PDSI above and below a threshold of 3 and -3 and takes it
apart to examine the precip and temperature contributions. It is quite
complementary in that regard and shows more explicitlt that it is the dry
spells that increased first from precip decreases and second from
> 6) The Emanuel (2007) slide is nice but that
> paper has not been assessed in our report. If
> you are seen by governments to be making your
> argument for the hurricane statement based on the
> Emanuel (2007) paper, we will almost certainly
> have challenges to the hurricane statement on
> procedural grounds -- which is not what we want
> to invite. Even though it is an update, it is
> substantially different from the published one
> that is assessed.
No that is not true. In our discussion in section 3.8.3 we note that the
original Emanuel (2005a)set of curves was revised and discussed in Emanuel
(2005b) in response to the comment by Landsea. But that response did not
publish the revised curve; instead it appeared on Emanuel's website. It
was that curve we discuss in the report (and the main reason we did not
show it was because it had changed) and we say "the PDI increasing by
about 75% (versus about 100%) since the 1970s (Emanuel 2005b)." The 100%
was the original finding. Now there is a further minor refinement in the
2007 paper (in response to further complaints by Landsea, the corrections
to the record to make the surface p and wind estimates compatible was not
done at the highest wind speeds: very small changes) but an advantage is
that it is updated to include more years: through 2005. It is standard
practice for obs time series to be updated and that is mainly what the new
curve does. It is not at all at odds with what we discuss already.
>You can make a similar basic
> point using assessed material by putting one of
> the two Webster et al panels next to the SST
> trend in slide 27, highlighting the recent trends
> in both SST and intense storms with your nice
> animated ovals (and replacing the ACE figure,
> which uses non-satellite data). While the
> Webster figure itself wasn't explicitly in the
> chapter, the paper was referenced so I think that
> can be defended.
The SST curve though is for N Atlantic only and the Webster stuff is global.
We could replace the ACE curve with the numbers curve from slide 28?
With these explanations, I look forward to further suggestions.
> To respond to some of your other queries: I
> think slide 5 is better than slide 6 - showing
> all the data is nice. I agree with the idea of
> removing the Sahelian series.
Agree with both.
I suggest putting
> back the large-regions rainfall trends slide for
> several reasons ( replacing the zonal mean time
> series figure with the trends figure). It is
> the trends figure that maps to the language in
> the SPM which is what we are trying to explain
> here - the zonal means are not what we explicitly
> talk about in the SPM. If you don't explicitly
> defend our SPM paragraph, then we certainly risk
> losing it or at best wasting a lot of time on it.
See comments above. I'll see if I can do something else.
> I also think the trends image is clearer for the
> non-expert than slide 15 showing the zonal means
> (although as you know I am a big fan of slide 15
> personally on a scientific level).
> There probably still are too many slides and it
> will be helpful if we all think hard about which
> of these is most needed. In cases where queries
> are from just one or two governments, or are more
> technical than they are likely to raise in the
> plenary, etc., it will be better to be shorter.
I look forward to comments from others as to which, if any, should be
excluded. Of course I love them all.
> The comments make clear that we are going to be
> queried on the increases in heat waves statement
> as being too weak and only backed up in the FAQ.
> I personally like the European example but if you
> could also possibly put some text on that slide
> to help back it up more broadly, that will help
> to avoid challenges (please see the comments).
I included slide 22 which shows the shift in distribution of hot days and
cold nights, and I thought this might be better than the Alexander et al
maps. Again we run into too any slides. The change in hot days of course
relates to heat waves, because the change in extremes relates to the whole
pdf. The term heat waves is very subjective and the time scale is not
always clear. There was a heat wave on east coast (New York 71F yesterday)
although part of a month long warm period. The other main discussion of
heat waves in our text is for Australia and I took out the slide of
Australia temperatures vs precipitation in the first version (that Brian
and Matilde have not seen). There is not much we can do here. The
preponderance of evidence from all the statistics and studies demonstrates
a clear increase in heat waves, even if there is not a definitive study
just on heat waves. That is what we have to say.
> I'll probably have more comments when we talk but I hope this is helpful.
> At 2:17 PM -0700 1/5/07, Kevin Trenberth wrote:
>>I received some very helpful comments from
>>Jurgen and I have revamped the slides in the
>>light of the comments. I am cc'ing Matilde and
>>Brian as they are part of telecon. Please see
>>the attached. In all cases I have simplified the
>>presentation by placing the take home message at
>>the top. There are 30 slides here. At present
>>3 are hidden as possible alternates. Also some
>>should be dropped: your choice. The slides are
>>designed to address what was seen as the biggest
>>sources of misunderstanding in the comments on
>>The telecon will presumably discuss whether my
>>perceptions on that are the same as others.
>>Slide 4 may now be somewhat redundant with the
>>added years on slide 2. Turns out the cleanest
>>separation is for top 8 years graphically, but
>>they do not include 1999 or 2000. Suggestions?
>>I made a new graphic of the land T vs SST
>>differences, and that is slide 6 but it could be
>>replaced by slide 5. Your choice.
>>I simplified slide 14 (on precip) and removed
>>the slide with all the time series.
>>I have cleaned up many others somewhat.
>>I would be inclined not to show the slide on the Sahel drought (21).
>>I added an extra new slide on hurricanes using
>>Kerry Emanuel's updated and corrected series.
>>So at present there are 5 slides on hurricanes
>>and at least 2 of those should be removed. The
>>Emanuel one has the advantage over the Webster
>>one of including SST. Of these only slide 27
>>includes figures from the chapter, yet I would
>>be inclined to drop that one. Your views on this?
>>Slides 2 thru 12 are on aspects of temperature
>>13-16 and maybe 17 are on precipitation
>>17 to 21 are on drought
>>22 and 23 are on extremes and heat waves
>>24 and 25 deal with circulation and relations between T and precip
>>26 to 30 deal with tropical cyclones.
>>To wrap up I repeated the first slide: and I
>>added a little piece to the first slide (I know
>>this will not make Susan happy, and I would not
>>include in Paris, but I thought it was funny).
>>Please view as slide show.
>>That would leave about 24 slides. Some could
>>count as 1, e.g. 9 and 10 go together and would
>>take less than a minute. But I would guess a
>>minute average: order 25 minutes here.
>>Please do not use these slides at least until after the report is
>>Kevin E. Trenberth e-mail: firstname.lastname@example.org
>>Climate Analysis Section, www.cgd.ucar.edu/cas/trenbert.html
>>P. O. Box 3000, (303) 497 1318
>>Boulder, CO 80307 (303) 497 1333 (fax)
>>Street address: 1850 Table Mesa Drive, Boulder, CO 80305
>>Attachment converted: Discovery:C3IPCCParis.ppt (SLD3/�IC�) (00377B45)
Dr. Kevin. E. Trenberth
Climate Analysis Section
PO Box 3000
Boulder CO 80307
ph: (303) 497 1318
Prof. Phil Jones
Climatic Research Unit Telephone +44 (0) 1603 592090
School of Environmental Sciences Fax +44 (0) 1603 507784
University of East Anglia
Norwich Email email@example.com
Kevin E. Trenberth e-mail: firstname.lastname@example.org
Climate Analysis Section, www.cgd.ucar.edu/cas/trenbert.html
P. O. Box 3000, (303) 497 1318
Boulder, CO 80307 (303) 497 1333 (fax)
Street address: 1850 Table Mesa Drive, Boulder, CO 80305