"Martin Rowley" wrote in
message ...
... following on from this (and other, earlier) discussion about
roof-top sites, with London being the most notorious, would there
be having any merit in having the tag "roof" or "rooftop"
'hard-wired' into the name of the reporting station?

Thus, Cardiff (WMO= 03717) becomes "Cardiff MO Roof" or perhaps
"Cardiff rooftop (MO) or somesuch.

Sensible idea.

These sites have such peculiar behaviour, never mind the 'local'
influence (e.g. air-conditioning stacks, hard-standing surfaces,
significant elevation above above street-level etc.) and are
creeping into the network like a cancer, that some routine
distinction could usefully be made.

There seems to have a serious change in attitude inside the MO in
the early-1980s. Temperature observations have been taken on
the roof of the LWC (and before that "The Air Ministry Roof")
at various sites in inner London since 1929, but the figures were
never used or published formally, except for the daily list of
health resorts and other sites compiled for the daily newspapers.
Sunshine, visibility and wind were reported in the MWR, but
not temperature or rainfall which were clearly regarded as being
irrelevant in-house. There is no mention of temperatures on the
Air Ministry/LWC roof in W.A.L.Marshall's "Century of London
Weather" (1952) or in J.H.Brazell's "London Weather" (1968),
or indeed in T.J.Chandler's "The Climate of London" (1965),
and none of three gentlemen deemed it necessary to explain this
absence. Marshall and Brazell presumably worked there, and
Brazell formally acknowledges help from his colleagues in
extracting the climatological data, so the understanding
described above clearly extended to the LWC staff.

In 1983, temperature and rainfall data from a variety of
Weather Centres first appeared in the MWR, appropriately
flagged as non-standard sites. I can understand and appreciate
that any data, *suitably annotated*, may be worth publishing
(it's a philosophical argument, really, which we needn't go into
here). But the crucial thing is that all potential users MUST
know what they are looking at. Nowadays I get the feeling
that half (or more) of the MO staff think that rooftop sites
provide legitimate climatological records. One knowledgeable
meteorologist (not sure about his climatological credentials,
though) on a message board recently said that he believed
LWC readings were adequately representative of central
London. Sigh He's clearly never compared LWC's obs
with those from St James's Park, for instance.

The climate information on the MO's website refers to all
sorts of rooftop sites without the slightest acknowledgment of
the site characteristics. In the various networks there are now
dozens of rooftop sites, and I have to admit I'm not sure that
even I know which ones are which. Some have the same name
as discontinued surface sites (some of long standing), to add to
the general confusion.

My preference would be to follow Brazell and co, and simply
not even acknowlege these non-standard sites. I have to admit
being inconsistent over the years in my "Weather Log" copy, but
that has now changed [sic :-) ]. The trouble is, with data from
non-standard sites proliferating in MO publications, people think
you have missed something if those data are absent from your
own work.

Talking about Weather Centres reminds me of another source
of confusion -- the naming of stations. Broadly speaking, until
the WCs appeared in the MWR, all names of climatological
stations were geographical ... they could be found on an
OS map. Thus, even in urban areas it was possible for
anyone to locate the station ... eg Birmingham (Sparkhill),
Cambridge (Botanic Garden), Huddersfield (Oakes),
Worcester (Perdiswell), etc, etc. There were a few
anomalies, usually where the station was on MOD property
which had a specific MOD name (eg "Bedford" instead of
"Thurleigh"). But it was not and still is not possible to locate
the WCs unless you actually know where they are. The
Air Ministry Roof was properly called "Kingsway" in the MWR,
but the LWC should have been known as "Holborn" or
"High Holborn", and latterly as "Clerkenwell" or
"Clerkenwell Road". I could go on. Watnall, anyone?

The point is, what is a name for? It's to help a user of
the information locate the site. How do you do that?
You use names they are most likely to know, or, failing
that, that they ought to be able to find on a map.

Philip Eden