Dear List,
There are two variables from the CMIP5 and 6 data request with standard names of the form "histogram_of_X_over_Z". Following some discussion on the thread "Usage of histogram_of_X_over_Z <
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/2016/059038.html> " which has clarified the definition of these terms and the intended use of "histogram_of_..." in CF, the conclusion was reached that the variables in question are actually frequency distributions rather than histograms. So, for CMIP6, we would like to propose new standard names to be used instead of the existing names:
histogram_of_backscattering_ratio_over_height_above_reference_ellipsoid
histogram_of_equivalent_reflectivity_factor_over_height_above_reference_ellipsoid
Proposed new names:
frequency_distribution_of_backscattering_ratio
frequency_distribution_of_equivalent_reflectivity_factor
I've shortened the names by omitting the "over_...." component, as this is not essential to the definition of the physical quantity. Jonathan has suggested that the dimension(s) over which the data is aggregated to create a frequency distribution or histogram should be represented in a cell_methods string.
Further: there are two variables in the CMIP6 request which represent fractional occurrence of clouds
with a given range of cloud-top-heights and optical depths. Summing over cloud-top-heights and optical depths gives the total cloud fraction in a grid cell. The CMIP variables are clmisr (new in CMIP6) and clisccp (from CMIP5). In CMIP5, clisccp carried the standard name of "isccp_cloud_area_fraction". I'm concerned that this standard name does not capture the fact that the variables is a distributions, and so would like to propose a new standard name:
fractional_cloud_area_fraction_over_cloud_top_altitude_and_optical_depth
This differs from the histogram and frequency distribution in that it is neither counts nor relative frequency. It should carry coordinates cloud_top_altitude and optical_depth with bounds to give the bin sizes. The CMIP variables will be functions of latitude, longitude, time, cloud-top-height and optical-depth. The proposed standard name will clearly indicate the special role of the last two dimensions. The generic definition would be "fractional_X_over_Y[_and_Z] indicates the amount of X in bins defined by axes Y and, if present, Z. The axes must be present as coordinates of the variable and must have bounds specified. The variable has the same canonical units as X. Summing over Y [and Z] gives the quantity X".
The CMIP5 isccp_cloud_area_fraction variable is specified as a function of altitude, which could be interpreted as relating to area fraction of cloud in an altitude layer, rather than a fractional distribution with respect to cloud top height, and I'd like to remove this ambiguity.
regards,
Martin
Received on Thu Nov 03 2016 - 10:08:33 GMT
