Friday, August 21, 2009

Pielke and Matsui (2005) revisited


So, a whole lot of hot air has passed, and maybe the dust is starting to settle. Mt noted that the Klotzbach et al work is "Pielkes all the way down", and several references have been made to the Pielke and Matsui 2005 paper which underpins a lot of this work. Mention of this prompted a distant memory, so I checked my mailbox (amazing the power of computers to recover old memories these days). I also re-read the PM05 paper more critically, in light of its current relevance and the previous emailed comment.

Well, well. That was...enlightening.

PM05 examines the near-surface lapse rate that arises on calm and windy nights, with a simple one-dimensional analytical model. They further calculate the changes in lapse rate as the cooling rate varies.

In all this work, they apply the radiative cooling at the surface, even though they explicitly portray this forcing as being representative of the effect that arises from a change in GHG concentrations. Standard climate theory holds that the radiative forcing is applied the top of the atmosphere - indeed this is the level at which the forcing is defined. It is simply wrong to claim that a doubling of CO2 will generate a forcing of 3.7Wm-2 at the surface, for example.

The startling impact of this odd application of "bottom of the atmosphere" forcing is apparent from their Table 1. A change in this "forcing" of a mere 1Wm-2 leads to a temperature difference of a whopping 1.5C (at the 2m level) over a single calm night! This is the simple result of applying 1Wm-2 of cooling to the fairly shallow layer at the bottom of the atmosphere, which has relatively low heat capacity due to its shallowness.

Consider a simple thought experiment. Start with a pre-industrial atmosphere, wait for a calm night, and instantaneously double the CO2 level. According to PM05's interpretation of GHG forcing, this will result in a temperature rise of about 6C (compared to the unforced case) in the space of 12 hours. This simply doesn't pass the sniff test, not by a million miles. It's not just wrong, it's several orders of magnitude wrong!

Remember, PM05 is the celebrated theoretical underpinning for this entire night-time warming edifice.

For the standard approach, check out this article which compares various radiation models. Here's their plot of depth-varying warming rate due to an increase in GHGs (from 1860 to 2000 values) holding everything else constant:
Note the warming rate is basically uniform in the bottom 200hPa of the atmosphere, in stark contrast to the PM05 assumption that the forcing acts at ground level. Thus, a large increase in GHGs generates a warming rate of about 0.04K per day across the boundary layer, as compared to the Pielkian ~1K over a single night (depending on wind speed).

This paper (pdf here), co-authored by all the big names in the field, ("The radiative effects from increased concentrations of well-mixed greenhouse gases (WMGHGs) represent the most significant and best understood anthropogenic forcing of the climate system.") has essentially the same plot in their Figure 12. There does not appear to be the slightest bit of debate over this subject.

So I don't think that the PM05 calculations can have any relevance to the response of the lower part of the atmosphere to changes in GHG forcing.

12 comments:

James Annan said...

Roger Pielke Sr sent the following comment by email (actually immediately prior to me publishing this post, but it seemed the most appropriate place to put it):

"James - You clearly do not understand the fundamental boudary layer physics issue in our paper. As I have written to you in an e-mail

"Our paper is as solidly mainstream in boundary layer physics as one can have.

That the Parker and Peterson papers are brought up does not make them necessarily the final word on this subject. Moreover, they deal with only the rural/urban issue [Do you really conclude that there is no difference in long term trends between urban and rural sites?]. Regardless, our paper involves all sites, pristine, rural and urban."

If you want to discuss the actual science of our work, I would be glad to do that."

Roger Sr.

Deep Climate said...

So let's see if I understand all the implications of PM05 Table 1.

Apart from rapid changes in warming trends with every meter (at least on calm nights), the table also appears to imply that relatively calm areas will warm much faster near the surface than windy areas, all other things being equal.

That sounds, well, unphysical. Or am I missing something?

Hank Roberts said...

"the finding by Parker [2004,
2006] that there is no difference in trends between windy and
calm nights is curious since the fundamental understanding
of stable boundary layer physics requires that trends near the
surface, for the same boundary layer warming, will be
greater on light wind nights ..."

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, D24S08, doi:10.1029/2006JD008229, 2007

So is this a disjunction between those who measure and those who know what should be found?

EliRabett said...

Hmm, appears you have mastered the zen, getting those who you disagree with to respond before you speak.

Of course, it is not necessarily in their interest to have done so, Obi wan

Deep Climate said...

So is this a disjunction between those who measure and those who know what should be found?

Or: "...those who claim to know what should be found?"

Hank Roberts said...

Well ....

http://maeresearch.ucsd.edu/kleissl/255/

MAE 207/255: BOUNDARY LAYER AND RENEWABLE ENERGY METEOROLOGY
--- briefest of excerpts ----
...
Homework 4: Surface Energy Balance / Convective Fluxes, Stability...
1. ... Material on practical issues and some theory Conduct measurements ... during daytime (ideally midday) and after sunset....
5. ... # Compute the mean and turbulent parts of the vertical kinematic heat flux [K m/s] .... Compute the sensible heat flux [W/m2] and comment on its relation to net radiation.
# Assuming neutral conditions compute and plot the profile \bar{M}(z) of the mean horizontal velocity M = (sqrt(u.^2+v.^2).
# Compute the flux Richardson number using d\bar{M}/dz from the previous question. Also compute the Obukhov length L. ...

---- much else there ----

I swear it's just amazing what these kids these days are doing.

Dirk said...

James:

Thanks for dismantling this paper and for showing how flawed it is. May I suggest publishing a comment to PM05 (and the associated Lin, Pielke et al 2007 paper) to GRL?

"Our paper is as solidly mainstream in boundary layer physics as one can have.

For someone who fails to understand how urban surfaces alter temperature and turbulence profiles near the surface - or at least less than 10 m - that's a stunning admission of his ignorance.

If you want to discuss the actual science of our work, I would be glad to do that."

Hard to objectively do that when his blog has been closed to comments.

Hank Roberts said...

Nice maps are included here, from 2005 description of the current Hadley:

http://www.cru.uea.ac.uk/cru/data/temperature/HadCRUT3_accepted.pdf

"... A new version of this dataset, HadCRUT3, has been produced; benefiting from recent improvements to the sea-surface temperature dataset which forms its marine component, and from improvements to the station records which provide the land data. A comprehensive set of uncertainty estimates has been derived to accompany the data: estimates of measurement and sampling error, temperature bias effects, and the effect of limited observational coverage on large-scale averages have all been made...."

James Annan said...

a comment to PM05

I should emphasise that the original observation was not by me (and I doubt the author would really want to be dragged into this catfight, though I may email him to check I haven't said anything silly). And although I think I can follow the arguments, it's not a topic that I can claim to speak with any real authority on.

I think I may have got to the bottom of the misunderstanding though...as post may shortly explain.

DC,

Yes, it seems that the claim is that nights should be 0.3-1.5C warmer at ground level, depending on wind. That is relative to the daytime warming, ie a reduction in daily range. Here in Japan the daily range is only about 5C at the moment.

Hank Roberts said...

The word "bias" again:

http://www.agu.org/pubs/crossref/2009/2009GL039152.shtml



O'Connor, F. M., C. E. Johnson, O. Morgenstern, and W. J. Collins (2009),

Interactions between tropospheric chemistry and climate model temperature and humidity biases

Geophys. Res. Lett., 36, L16801, doi:10.1029/2009GL039152

18 August 2009

All I see is the abstract; have you time to look at the paper?

James Annan said...

All I see is the abstract; have you time to look at the paper?

Yes, but it's not hugely interesting to me. They are of course using "bias" in the standard way (in this context) to mean a persistent discrepancy between reality and a model. They checked the effect of removing these biases for the atmospheric chemistry model. It didn't have a very large effect.

EliRabett said...

The PM05 model fails for low wind speeds because it leads to an infinite temperature difference between the surface and the scale height (which goes to zero). Something is missing, either a limit on the scale length at low wind speeds to account for turbulence or some other process to move heat to and from the surface from the top of the inversion layer.