On 30/09/15 15:50, Jonathan Gregory wrote:
Dear Maarten
Thanks for your new standard name proposals. I have various comments.
Name: uv_aerosol_index
Description: UV means ultraviolet radiation. The UV-Aerosol Index is
a UV colour index that represents the deviation of the outgoing TOA
radiation in the UV from that of a standard atmosphere, featuring
Rayleigh scattering and gas absorption, in particular by ozone. No
cloud droplets or suspended liquid or solid particles (aerosols) are
present in the standard atmosphere. It is bounded below by a
Lambertian surface, featuring isotropic reflection, assumed
independent of wavelength. The UV-AI is computed from the Earth
reflectances at two UV wavelengths. A positive deviation from the
standard atmosphere is often, but not exclusively, attributed to the
absorption of radiation by aerosols, while negative values represent
increased scattering, not necessarily by aerosols.
From the definition, I don't properly understand what this quantity is, in
physical terms. Is it possible to describe it more precisely and, if so, would
it be possible to give it a standard_name which conveyed that meaning? This
is not essential. As you know, we have added some other quantities to the
table whose names are opaque, such as atmosphere_stability_total_totals_index.
That quantity is used in practice with that name and is hard to describe as a
general-purpose geophysical quantity. But if a physical description is
possible, it seems preferable to me.
It isn't a physical quantity (unit = "1"). It indicates a deviation of the
top-of-atmosphere reflectance from the reflectance of a pure Rayleigh atmosphere
including absorbing trace gases (notably Ozone). Such a deviation can be caused by
absorbing aerosols, but the value has many complicating contributions, including the
altitude of the aerosols, the single scattering albedo, presence of clouds, and
probably more. This parameter is however one of a few that can identify some
important classes of aerosols in the presence of clouds, for instance biomass burning
aerosols and volcanic ash.
The parameter has been derived from TOMS up to the present on various instruments, at
first to identify ozone retrievals that were affected by aerosols. I've asked several
specialists in the field to come up with a description of this parameter, and they
warned: "Here is my attempt at defining a quantity that is far from clear". The
description does cover what is in the parameter, and the name is easily found in
literature.
A smaller note: we haven't used "uv" before. It is well-known, of course,
but spelling it out as ultraviolet might be better,
That is fine with me, so: ultraviolet_aerosol_index
Name: cloud_optical_centroid_pressure
Description: The cloud optical centriod pressure is a pressure level
inside the cloud, near the optical thickness center of the cloud.
I think this should be air_pressure, to be precise, and I would suggest
air_pressure_at_cloud_optical_centroid, by analogy with air_pressure_at_
cloud_base and various other air_pressure_at_X names.
Ok: air_pressure_at_cloud_optical_centroid.
Name: cloud_area_fraction_assuming_fixed_cloud_albedo
Description: "X_area_fraction" means the fraction of horizontal area
occupied by X. "X_area" means the horizontal area occupied by X
within the grid cell or the satellite footprint. Cloud area fraction
is also called "cloud amount" and "cloud cover". The cloud area
fraction is for the whole atmosphere column, as seen from the
surface or the top of the atmosphere. Cloud area fraction assuming
fixed cloud albedo is the cloud area fraction by assuming the clouds
in the X_area having the same fixed cloud albedo value (Y). The
clouds having cloud_area_fraction_assuming_fixed_cloud_albedo and
cloud albedo Y yield the same reflectance at TOA as the real clouds
in the X_area. A phrase assuming_condition indicates that the named
quantity is the value which would obtain if all aspects of the
system were unaltered except for the assumption of the circumstances
specified by the condition. The assumed albedo shall be attached to
the variable either by an attribute 'assumed_cloud_albedo' (if the
value is independent of geolocation) or through the
ancillary_variables to a variable with standard_name 'cloud_albedo'.
To me, this name doesn't appear to say what you mean. The name appears to
imply that if you change the albedo of clouds, you change their area. You
mean the cloud fraction of specified albedo which would give the area-average
albedo as the actual clouds do, if I understand correctly. That description
would be a cumbersome standard name. Is this a commonly used quantity which
needs a standard name, and if so, is this what it's normally called?
Yes, you do change the resulting cloud fraction if you assume a different cloud
albedo. :-)
It is frequently called an unhelpful 'effective cloud fraction'.
Background: UV-VIS-NIR space-borne spectrometers (GOME, GOME2, Sciamachy, OMI,
TROPOMI) have observation areas that are large, especially when compared to typical
dedicated cloud instruments (MODIS, VIIRS, Seviri). Think 300x40 km for GOME down to
7x7km for TROPOMI. These instruments are geared towards trace gas retrievals, but
need a matching cloud parametrization to correct the trace gas retrievals for the
presence of clouds. We can't assume a homogeneous scene (like many of the cloud
specific instruments implicitly do). While it is possible to retrieve three cloud
parameters (cloud albedo, cloud fraction and cloud pressure) using a very simplified
cloud model, there are instrumental limitations that make this unstable in practice
(mainly the combination of NIR spectra with a UV signal is tricky, as col-location is
a limiting factor). Instead many algorithms have taken a different approach, fix the
cloud albedo to a fairly high value (0.8) and fit just two parameters:
cloud_area_fraction_assuming_fixed_cloud_albedo and
air_pressure_at_cloud_optical_centroid. Obviously if you assume a lower cloud albedo
you will end up with a higher cloud fraction, most likely at a different pressure level.
The albedo we assume is (much) higher than 'typical' clouds, and the resulting cloud
fraction is much lower than a geometric cloud fraction. But the top of atmosphere
radiance of the _whole_ pixel is accounted for. So it is pixel averaged, not cloud
area averaged. That leads I think to the name I proposed.
Name: cloud_albedo_assuming_cloud_area_fraction_of_1
Description: cloud_albedo_assuming_cloud_area_fraction_of_1 means a
derived (retrieved) cloud albedo in an area assuming cloud area
fraction is 1 in this area (the grid cell or satellite footprint). A
phrase assuming_condition indicates that the named quantity is the
value which would obtain if all aspects of the system were unaltered
except for the assumption of the circumstances specified by the
condition.
Is 1 the only cloud area fraction you're interested in? If so, I think it would
be better to use a phrase in words, meaning the opposite of assuming_clear_sky,
which is used in many standard names to mean cloud area fraction of 0. If you
might want to use arbitrary area fractions, you could use a coordinate variable
of cloud_area_fraction.
No, 1 is the only one we're interested in. To refer back to the cloud model, if we
have a scene with a high surface albedo (i.e. snow or ice) we cannot distinguish
between cloud and surface. In this case we use a different cloud model: assume the
whole scene is cloud-covered, and retrieve both cloud albedo and cloud pressure. For
scenes where we could potentially use the other cloud model, this model is more
sensitive to very light cloud cover, which becomes visible as a scene pressure (see
below) that deviates from the true surface air pressure.
There is a clear_sky label that is frequently used, but I haven't come across the
opposite. What do you suggest? cloud_albedo_assuming_full_cloud_cover
Name: cloud_optical_centroid_pressure_assuming_cloud_area_fraction_of_1
Earlier comments apply.
Which combined leads me to:
air_pressure_at_cloud_optical_centroid_assuming_full_cloud_cover
Best,
Maarten Sneep
--
KNMI
T: 030 2206747
E: maarten.sn...@knmi.nl
R: A2.14
_______________________________________________
CF-metadata mailing list
CF-metadata@cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
