In fact what is measured is the property of the *stuff at the point* - i.e. this is a sampling issue.
The stuff is the atmosphere, its property is its thermodynamic potential, and it is sampled at the
point. This is what O&M Part 2 Sampling Features is about, and is why every sampling feature (of
which SamplingPoint is a special case) has a "sampledFeature" such as "the
atmosphere".
So, I suggest a good look at O&M Part 2 is in order. I know that Andrew W has
already done this, since he was on the RWG and is credited as one of the
contributors since he provided a lot of feedback.
Simon
-----Original Message-----
From: galeon-bounces@xxxxxxxxxxxxxxxx
[mailto:galeon-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Gerry Creager
Sent: Friday, 14 March 2008 2:03 AM
To: Unidata GALEON
Subject: Re: [galeon] Fwd: CDM feature and point types docs
Ron, et al,
Ah, but then it becomes a science-discipline semantics issue, too.
I do think in terms of making "point" observations of in-situ weather data.
The observation is made at a fixed location, at a particular finite time, and its
geometric property is not a bounded region or polygon, but a point. Realizing there are
gaps in this (wind is measured at 10m above ground, temperature, humidity, pressure at 2
meters, precipitation at 1 meter, direct and diffuse solar radiation at nominally 2m but
may vary, etc) the data are represented to end-users as being at a single spatial point.
Think of it as semantic collision rather than assimilation.
And, while I don't think I'm completely clueless, I've spoken at the TCs, and mentioned
to you in the past, about "point features" in my use of WFS to represent
observations, without discussion.
gerry
