Hi Andrew -
I am not sure if I should have TEMPERATURE and SALINITY arrays with 4
dimensions
like TEMPERATURE(TIME,LATITUDE,LONGITUDE,PRESSURE) or just 1 dimension
like I have above i.e. TEMPERATURE(PRESSURE). ?
While we don't store profile data in OceanSITES, we are working on our
coordinates
to make them conform more to published standards, so I've been reading
the docs
on this lately. So, it's fresh in my mind - fresh, if not exactly clear.
The NetCDF Best Practices page
(unidata.ucar.edu/software/netcdf/docs/BestPractices.html)
recommends:
Make coordinate variables for every dimension possible (except for
string length dimensions).
And the CF manual has this recommendation about the order of the coordinates
and the use of scalar coordinates as dimensions, in section 2.4:
If any or all of the dimensions of a variable have the interpretations
of "date or time" (|T|), "height or depth" (|Z|), "latitude" (|Y|), or
"longitude" (|X|) then we recommend, but do not require (see
Section 1.4, "Relationship to the COARDS Conventions"
<http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.6/cf-conventions.html#coards-relationship>),
those dimensions to appear in the relative order |T|, then |Z|, then
|Y|, then |X| in the CDL definition corresponding to the file. All
other dimensions should, whenever possible, be placed to the left of
the spatiotemporal dimensions.
Dimensions may be of any size, including unity. When a single value of
some coordinate applies to all the values in a variable, the
recommended means of attaching this information to the variable is by
use of a dimension of size unity with a one-element coordinate variable.
These 2 recommendations would lead me to suggest that, *IF* your
pressure is
monotonic, you'd want to use:
TEMPERATURE(TIME,PRESSURE,LATITUDE,LONGITUDE)
If your pressure isn't monotonic, and so can't be a coordinate variable,
I don't know
whether it's preferable to demote ALL the coordinates to auxiliary
coordinate variable
status,
TEMPERATURE: coordinates = "TIME,PRESSURE,LATITUDE,LONGITUDE" ;
Cheers -
Nan
On 4/3/12 1:15 AM, andrew walsh wrote:
Hi Roy/All,
Agree totally it would be good to get this CTD netCDF done in an
interoperable way.
Regarding having pressure vs. depth. We liked to use pressure because:
1) It is the thing that is measured, let us store just that and
calculate depth if needed.
2) Depth can later be easily be calculated on the fly
using the 'Pressure to Depth' algorithm in "UNESCO (1983) Techinical
Papers in Marine Science 44, Algorithms for Computation of fundamental
properties
of Seawater. One can use the python seawater library (see
http://packages.python.org/seawater/ )
and seawater.csiro.depth(p, lat) to get depth from pressure and
seawater.csiro.pres(depth, lat)
to get pressure from depth.
3) I noted that the ARGO project (10000's CTD like profiles) and
others like
CSIRO Oceanography in Aust. make data available with just pressure.
4) It makes our processing and QC a whole lot simpler. We don't
have to worry about calculating and managing the extra 'depth' variable.
Is there any problem with having the "pressure" as a co-ordinate
which isn't really a "dimensional"
quantity like depth (z) in the 4-D sense i.e x,y,z,t ?
However I note that pressure (decibar) is allowed as a vertical axis
e.g see section
4.3. Vertical (Height or Depth) Coordinate of CF v1.6 conventions
document which says:
"A vertical coordinate will be identifiable by:
. units of pressure; or
. the presence of the positive attribute with a value of up or down
(case insensitive).
"
AND
section 4.3.1. "Dimensional Vertical Coordinate" which says:
"The units attribute for dimensional coordinates will be a string
formatted as per the udunits.dat3 file. The
acceptable units for vertical (depth or height) coordinate variables are:
. units of pressure as listed in the file udunits.dat. For vertical
axes the most commonly used of these include
include bar, millibar, decibar, atmosphere (atm), pascal (Pa), and hPa.
...
..."ΒΆ
Regards,
Andrew
----- Original Message ----- From: Lowry, Roy K.
To: andrew walsh
Cc: Jim Biard ; Upendra Dadi ; [email protected] ; Luke Callcut
; [email protected] ; [email protected]
Sent: Tuesday, April 03, 2012 2:16 PM
Subject: RE: [CF-metadata] Ocean CTD data following CF Conventions v1.6
Hi Andrew,
SeaDataNet will very soon be considering how it is going to encode
data, including single CTD casts, in CF-compliant NetCDF and so I
think the time is ripe for agreeing how the significant numbers of us
who indulge in this practice for whatever reason do it. That way
we'll end up with interoperable data.
I think there are a number of people on this list who have already
encoded single CTDs into NetCDF and so it would be useful to start by
asking for descriptions (like Andrew's examples) of how this has been
done and what tools are dependent upon that encoding.
The z co-ordinate parameter (pressure/depth) is also an issue worth
resolving. Whilst interconversions are relatively straightforward,
agreement would make life much easier. My preference leans dowards
having depth as the dimension with pressure as an optional variable.
That way we interoperate with other kinds of oceanographic profile
data such as bottle data.
If we can get that far, we can then look at how to aggregate multiple
CTDs into a file according to the CF point data conventions.
Cheers, Roy.
From: andrew walsh [[email protected]]
Sent: 03 April 2012 04:39
To: Lowry, Roy K.
Cc: Jim Biard; Upendra Dadi; [email protected]; Luke Callcut;
[email protected]
Subject: Re: [CF-metadata] Ocean CTD data following CF Conventions v1.6
Hello Roy, Upendra, Jim and CF list,
Thanks for all your feedbacks.
My proposal relates to single CTD profile data (a vertical profile of
pressure, Temp. Salinity)
not a trajectory in x,y,z,t and I have put :featureType = "profile" as
recommended in the global
attributes section. As Roy has mentioned to Jim CTD data is usually
processed and QC'ed
as a single profile per netCDF file so that's why I am doing it this way.
Aggregations using multiple CTD profiles per netCDF file
may be constructed later at say national/international data centres
and these aggregrations
would follow the CF conventions v1.6, Chapter 9 - Discrete Sampling
geometries
and also the new netCDF templates provided by NODC, thanks NODC -:)
Roy,
The recent NODC netCDF templates don't have an aggregation example for
CTD
however the "Profile/World Ocean Database Observed Level" example
comes close
(see http://www.nodc.noaa.gov/data/formats/netcdf/) and click on
Profile/World Ocean Database Observed Level link. This example appears
to be for ocean station/bottle samples with vertical dimension of
depth (z) (m) rather
than case of CTD which would use pressure (dbar) as the vertical (z)
dimension.
It would be useful I think to have a NODC netCDF template for an
aggregation of
CTD casts.
Upendra,
Based on your responses and what I have seen at NODC and other places
it seems there are 2 methods to do this:
1) You can have dimensions for lat, long time i.e an array
with a single value. That is:
dimensions:
TIME=1
PRESSURE=729
LATITUDE=1
LONGITUDE=1
variables:
double TIME(TIME) ;
+ std. attrs.
double LATITUDE(LATITUDE) ;
+ std. attrs.
double LONGITUDE(LONGITUDE) ;
+ std. attrs.
double PRESSURE(PRESSURE) ;
+ std. attrs.
double TEMPERATURE(PRESSURE) ;
+ std. attrs.
TEMPERATURE:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
double SALINITY(PRESSURE) ;
+ std. attrs.
SALINITY:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
(2) Other way is to have Lat, long and time are single valued scalars:
dimensions:
PRESSURE=729
variables:
double LATITUDE ;
+ std attrs
double LONGITUDE ;
+ std attrs
double TIME ;
+std attrs
double PRESSURE(PRESSURE) ;
+ std. attrs.
double TEMPERATURE(PRESSURE) ;
+ std. attrs.
TEMPERATURE:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
double SALINITY(PRESSURE) ;
+ std. attrs.
SALINITY:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
I can see Method 2 above is much simpler and I would prefer to use
this.
However I have another small concern about all of this. Given there is
more than
1 way to treat dimensions does the netCDF plotting and visualising
sofware out
there understand both approaches. That is can we expect something like
ncBrowse, IDV or ODV? give me a plot of say salinity vs. pressure
and work OK with BOTH approaches to coding the netCDF?
Thanks and Regards All,
Andrew Walsh
----- Original Message ----- From: Lowry, Roy K.
To: Upendra Dadi ; andrew walsh
Cc: Luke Callcut ; [email protected] ; [email protected]
Sent: Tuesday, April 03, 2012 9:31 AM
Subject: RE: [CF-metadata] Ocean CTD data following CF Conventions v1.6
Hi Upendra,
I like the idea of a station dimension. It goes a long way to
resolving the issue raised in my response to Jim which was based on
the tunnel vision of having pressure/depth as a dimension. I have yet
to look at the recently published NODC NetCDF templates. Is this CTD
encoding included in them? If so, I'll bump up looking at them on my
'todo' list. I'd also recommend that Andrew and my colleagues in
SeaDataNet take a look.
Cheers, Roy.
From: [email protected]
[[email protected]] On Behalf Of Upendra Dadi
[[email protected]]
Sent: 02 April 2012 17:21
To: andrew walsh
Cc: Luke Callcut; [email protected]
Subject: Re: [CF-metadata] Ocean CTD data following CF Conventions v1.6
Hi Andrew,
Either way it should be okay as far as CF compliance is concerned.
But the dimensions - latitude, longitude and time are not really
required. If it is required to indicate that there is only one
station(profile) in the file, there could be a dimension for number of
stations instead, with a value of 1. Also, using a station dimension
is the way to go if storing a collection of profiles in a single file.
Here at NODC, we took the approach that we would use the same
consistent representation whether there is a single instance or a
collection in a file.
Upendra
On Mon, Apr 2, 2012 at 2:51 AM, andrew walsh <[email protected]> wrote:
Hi CF lis
We are working on coding up some 1000's netCDF files off CTD
instruments and want to make usre we are following the
latest netCDF conventions (v1.6) OK. As background the CTD records
a profile pressure, temperature and salinity.
Here is a summarised CDL version (not all attributes+variables+qc
flags there, just majors for now)
of what we propose:
dimensions:
TIME=1
PRESSURE=729
LATITUDE=1
LONGITUDE=1
variables:
double TIME(TIME) ;
TIME:standard_name = "time" ;
TIME:units = "days since 1950-01-01 00:00:00Z" ;
TIME:axis = "T" ;
TIME:valid_min = 0. ;
TIME:valid_max = 999999. ;
double LATITUDE(LATITUDE) ;
LATITUDE:standard_name = "latitude" ;
LATITUDE:units = "degrees_north" ;
LATITUDE:axis = "Y" ;
LATITUDE:valid_min = -90. ;
LATITUDE:valid_max = 90. ;
double LONGITUDE(LONGITUDE) ;
LONGITUDE:standard_name = "longitude" ;
LONGITUDE:units = "degrees_east" ;
LONGITUDE:axis = "X" ;
LONGITUDE:valid_min = -180. ;
LONGITUDE:valid_max = 180. ;
double PRESSURE(PRESSURE) ;
PRESSURE:standard_name = "sea_water_pressure" ;
PRESSURE:units = "decibars" ;
PRESSURE:axis = "Z" ;
PRESSURE:valid_min = 0. ;
PRESSURE:valid_max = 12000. ;
PRESSURE:positive = "down" ;
double TEMPERATURE(PRESSURE) ;
TEMPERATURE:standard_name = "sea_water_temperature" ;
TEMPERATURE:units = "degrees_C" ;
TEMPERATURE:_FillValue = -99.99 ;
TEMPERATURE:valid_min = -2. ;
TEMPERATURE:valid_max = 40. ;
TEMPERATURE:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
double SALINITY(PRESSURE) ;
SALINITY:standard_name = "sea_water_salinity" ;
SALINITY:units = "psu" ;
SALINITY:_FillValue = -99.99 ;
SALINITY:valid_min = 0. ;
SALINITY:valid_max = 40. ;
SALINITY:coordinates="TIME LATITUDE LONGITUDE PRESSURE"
// global attributes:
:conventions = "CF-1.6" ;
:featureType = "profile"
:cdm_data_type = "profile"
+ several other attributes later for ISO19115 metadata generation
I am not sure if I should have TEMPERATURE and SALINITY arrays with 4
dimensions
like TEMPERATURE(TIME,LATITUDE,LONGITUDE,PRESSURE) or just 1 dimension
like I have above i.e. TEMPERATURE(PRESSURE). ?
Any feedback on the above is greatly appreciated.
Andrew Walsh
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
--
*******************************************************
* Nan Galbraith Information Systems Specailist *
* Upper Ocean Processes Group Mail Stop 29 *
* Woods Hole Oceanographic Institution *
* Woods Hole, MA 02543 (508) 289-2444 *
*******************************************************
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
