[CF-metadata] Standard_name for cloud-cover by phenomenon

From: Wright, Bruce <bruce.wright>
Date: Tue, 15 May 2012 10:07:59 +0100

All,

Sorry to wade into this discussion late, but I believe part of the
difficulty experienced in the discussions here are a consequence of
mixing two distinct 'concepts':

1. Cloud Height Classification Based on Cloud Types

There is a well-recognised allocations of cloud types to height-bands.
These types and bands are nicely illustrated both in tabular form and
visually on the Cloud Appreciation Society website at:
http://cloudappreciationsociety.org/collecting/about-cloud-classificatio
ns/
http://cloudappreciationsociety.org/collecting/

I believe that this allocation to height bands is sufficiently
well-known to be characterized without attributing an owner (e.g. WMO)
or an observation process (e.g. SYNOP), as Heiko argued. Thus, (if
required) these should probably be given the standard names:

low_type_cloud_area_fraction
medium_type_cloud_area_fraction
high_type_cloud_area_fraction

*However*, at present I would argue that these can only be accurately
determined by a human inspection of the sky, which leads us to the
second concept...

2. Cloud Height Classification Based on Height Ranges

Most automated systems, be they cloud base recorders, numerical models
or other forecasting processes, will assign a cloud height class based
on a height range. In this case, I would argue that the following set of
standard names are more appropriate:

low_cloud_area_fraction
medium_cloud_area_fraction
high_cloud_area_fraction

I acknowledge that different height ranges will be adopted by different
users, but, as Heiko states, this approach will at least allow
Intercomparison, and the exact details of the height ranges used could
be included as additional (non-CF Standard) metadata.


Having presented these two 'concepts', I would suggest that the second
is likely to be the most useful, in an age where the human observers are
significantly outnumbered by automated observing and forecasting
systems. However, there is no reason why both sets of standard names
could not to adopted.

My contribution to the debate - I hope it's helpful.

Regards,
Bruce

-- 
Bruce Wright  Strategic Advisor on Data Management
Met Office FitzRoy Road Exeter EX1 3PB United Kingdom 
Tel: +44 (0)1392 886481 Fax: +44 (0)1392 885681
E-mail: bruce.wright at metoffice.gov.uk http://www.metoffice.gov.uk 
-----Original Message-----
From: cf-metadata-bounces at cgd.ucar.edu
[mailto:cf-metadata-bounces at cgd.ucar.edu] On Behalf Of Heiko Klein
Sent: 15 May 2012 09:10
To: Cameron-smith, Philip
Cc: cf-metadata at cgd.ucar.edu; Jonathan Gregory
Subject: Re: [CF-metadata] Standard_name for cloud-cover by phenomenon
Philip,
I agree that referring to versioned std_name tables is generally a good
idea. The WMO members have (and still have) problems with versioning, in
particularly when I think about grib-tables.
I still have a bad feeling about using a name like
wmosynop_high_cloud_area_fraction or isccp_high_cloud_area_fraction in
particularly since they give a over-specification of a generally well
understood phenomenons. Both wmosynop and isccp are
measurements/measurement networks, and CF generally does not have enough
metadata to cover all measurement-details. E.g. for air-temperature, it
is from a measurement point of view often important what type of
instrument has been measuring it (automatic, human, scale,...). We don't
have a 'human_quicksilver_air_temperature' in CF (or a
wmosynop_air_temperature). For comparing 'air_temperature' between
measurements and possibly different models, the generally understood
'air_temperature' is best.
With high-clouds, this is a similar problem. high, medium and low clouds
is generally well understood and well documented in literature (a simple
search on the net gives more than enough hits).
 From a model point of view, I cannot estimate if a cloud is exactly the
one type or the other. There are often estimations like:
sigma < 0.4 -> high cloud
0.4 > sigma > 0.7 -> medium cloud
sigma > 0.7 -> low cloud
It will be difficult for the modeller to say: This is a cloud according
to the wmosynop definition.
Therefore, I think we need at first a general CF-name for
high/medium/low cloud to be able to compare in between measurement
networks and models. ISCCP, WMO synop and models have already the
concept of these clouds, they are not a 100% match, but close enough.
If then one of these networks needs a more exact definition of
high/medium/low clouds, they should ask for a std_name of their own.
Best regards,
Heiko

Received on Tue May 15 2012 - 03:07:59 BST