Newsgroups: sci.geo.meteorology,sci.environment
Subject: Re: Ozone Layer over Australia
From: rparson@spot.Colorado.EDU (Robert Parson)
Date: 23 Jun 1997 22:37:03 GMT

In article <jscanlon-2206971530370001@sp119.linex.com>,
Jim Scanlon <jscanlon@linex.com> wrote:

>You might want to read an article, in the same issue of JGR, on page 8737
>"Analysis of long-term behavior of ultraviolet radiation measured by
>Robertson-Berger meters at 14 sites in the United States" a completely
>baffling paper which shows the sorry state of ultraviolet studies.

 I don't see why you call it "baffling". It's a very clear and
 thorough analysis of a very difficult problem. In fact I strongly
 recommend it for anyone who is interested in this subject - it
 explains a lot of the details that most specialized papers take
 for granted. I certainly learned a great deal from it.

>Nine years later we get a new study on this "burning question"  which
>extends the pioneering data for the 8 stations, and includs 6 more,
>extending it to 1991. The measurements obtained confirm the previously
>reported downward trend in surface UVB.(-6% per decade)
>
>The authors select 4 of the 14 stations and analyse the data differently
>to  produce an increase of 2% in 3 of the sub set of 4. Although the 2%
>increase is not statistically significant, and the conclusion of the study
>is that "...the data, by themselves, do not allow for definitive trend
>analysis results", the lead author is quoted in the Washington Post (April
>28, 1997, page A2) saying that UV has increased over the continental US by
>a few percent in the 1980s and has continued to increase in the 1990s"

 You left out one of the major findings of the paper: that the earlier
 trends were contaminated by calibration shifts. When allowances are
 made for these, the trends become positive; however, they depend upon
 just how the error correction is made, and the information required
 to do a proper reanalysis is not available. The overall
 conclusion of the paper is that the RB network data is messed up, and
 that there isn't much that can be done to repair it. This isn't
 a reflection on the experimenters - as the paper points out, the
 network wasn't designed for this purpose in the first place:

 "The RB meter network was originally established to determine the
 relative amounts of UV at different locations around the earth,
 with most sites in the United States. The data have been useful for
 their intended purpose, that is, to help explain differences in skin
 cancer at different locations. There was no original plan to use
 the network to determine trends, and therefore the network was not
 maintained using the high level of standards necessary for accurate
 trend determination. The network management, calibration techniques,
 and in some cases instrument location, underwent changes over the
 20 years of operation. Unfortunately, most of the records documenting
 the maintenance and calibration of the network were misplaced during
 transfer of the network among different managers."

 I'm not completely familiar with the history here, but I think it
 goes something like this: after the discovery of the antarctic ozone
 hole in the 1980's, people started looking for evidence of long-term
 trends in UV and ozone at other latitudes. For ozone, there were
 a number of sources to draw on: the worldwide Dobson spectrophotometer
 network which had been operating since 1957 (with individual station
 records going back as far as 1928), the TOMS and SBUV satellite
 measurements, and ozonesonde records from a number of places. After
 a couple of years of work reconciling these sources with each other,
 it was possible to report long-term trends in total ozone with well-
 defined error limits. For surface UV, however, there was no such
 database to draw on. It would have been convenient if someone had
 set up a worldwide network of spectroradiometers back in 1957,
 but nobody had. So they had to make use of data that had been
 collected for different purposes. It has turned out not to be
 very useful.

 This points up an important difference between laboratory science and
 field measurements. In the lab, when you find that an experiment was
 not properly controlled, you toss out the results and do it over again.
 In the field you don't have that option - we can't go back to 1976
 and repeat the measurements using fancier equipment and more carefully
 designed calibration procedures. Sometimes people just have to publish
 what is available, with appropriate caveats, in the hope that in the
 future someone might find a way to use it. As a result, the earth
 sciences are fruitful territory for "lawyer's scientists" of all
 persuasions.

 On a related subject, many people have an oversimplified notion of
 what these "ozone measurements" and "UV measurements" really are.
 In fact, they are both really different kinds of UV measurements.
 Most of the techniques used to measure stratospheric ozone actually
 measure UV. The Dobson spectrophotometer measures the ratio of
 radiant intensity at two specified wavelengths in the UV-B region
 to intensity at two other wavelengths; from this one can calculate
 the total ozone column overhead using well-understood radiation
 physics. (It's a little more complicated than a Beer's law calculation,
 but not much so.) The satellite-borne SBUV instrument works on a
 similar principle: instead of measuring UV transmitted through the
 ozone layer it measures UV that scatters back from the ozone layer.
 Of course all of these techniques have to be constantly calibrated
 with respect to each other and to measurements based on independent
 physical principles, such as balloon-borne electrochemical ozonesondes
 that actually do measure stratospheric ozone by chemical methods.
 One reason why people who are familiar with stratospheric ozone science
 act puzzled when asked why they think a reduction in stratospheric
 ozone will mean an increase in surface UV-B is the simple fact that
 the measured ozone reductions are actually indicators of UV-B
 increases! Nevertheless, they are _indicators_, not complete
 measurements. The Dobson spectrophotometer measures relative UV
 intensities, not absolute intensity, and in most applications it
 measures direct solar UV, not the UV that is scattered as it
 passes through the lower atmosphere (an appreciable portion of the
 total.) You can't turn Dobson or TOMS/SBUV data into total surface
 UV without further assumptions. But that is true of the RB data
 as well - it also doesn't directly measure what you are really interested in,
 UV at biologically damaging wavelengths below 300 nm. It measures
 a broad spectrum of UV radiation, weighted towards 315 nm. So it
 also is an indicator of biologically effective UV, rather than
 a direct measurement of it.

 Basically, the situation is that there is no "pre-ozone-depletion"
 baseline with which to compare present-day UV measurements. It's
 not that the trends have been measured and found not to be positive,
 it's that the trends were never adequately measured in the first
 place.

 -------
 Robert