NEXRAD to HDF or netCDF

[Thomas Pfaff]

Hi guys,

this is not really a wradlib question, but I does someone on this group
have experience converting NEXRAD level III data, mainly the digital
precipitation array data, to netCDF or HDF5?

I tried the official converters found e.g. here
http://www.ncdc.noaa.gov/data-access/radar-data/radar-decoding as well
as the weather climate toolkit's exporter.
Both fail with error messages on which no information is to be found on
the web.

I wanted to use the data to demonstrate the use of HDF and netCDF on
large amounts of data, so the only thing I need would be gridded, easily
displayable radar images.

Thanks in advance.


Thomas

[Jonathan Helmus]

Thomas,

    What type/product code NEXRAD Level III files are using trying to convert?  Despite all being in a same base format, Level III data encompassed a number of different products which use different message types to encode the data.  Not all products are supported by the various software products.

    I've had good results using Unidata NetCDF Java Library [1] to convert NEXRAD Level III (and II) data to NetCDF files which is one of the recommendation on the NCDC page.  Despite being a NetCDF library the software actually implements Unidata's Common Data Model (CDM) which supports a number of additional formats including NEXRAD Level III, also known as NIDS, and NEXRAD Level II.  

    The two tools I've used from Unidata are the toolsUI.jar file which can be run as a GUI to convert files using " java -Xmx1g -jar toolsUI.jar"and the netcdfAll-4.6.jar library which can be used to convert files from the command line using:
    java -Xmx1g -classpath netcdfAll-4.6.jar ucar.nc2.dataset.NetcdfDataset -in NEXRADFILE -out out.nc
   

I should also mention that I recently added support in Py-ART for reading some NEXRAD Level III files, specially those which define radial data.  This includes a few precipitation products (see the table starting at line 490 in nexrad_level3.py [2]).  Py-ART does not support the gridded formats which sounds like what you are after.  You could use Py-ART gridding routines to map the radial data to a Cartesian grid or use the the structures and code in the nexrad_level3.py file as a base to implemented a class to read gridded NEXRAD Level III data.

Cheers,

    - Jonathan Helmus

[1] http://www.unidata.ucar.edu/software/thredds/current/netcdf-java/documentation.htm
[2] https://github.com/ARM-DOE/pyart/blob/master/pyart/io/nexrad_level3.py#L490