Help

The AB/CPC Data Page allows you to (1) plot, (2) retrieve (ftp), and obtain documentation (doc) about various NCEP datasets.

What is WEASD or ...

Most of the variable names are standard NCEP names which have been documented in NCEP's Office Note 388 (GRIB). Some of the names are pretty strange such as UGRD=zonal wind, and WEASD=snow. Some of the variable are also very strange such as ABSD=absolute divergence and RELD=relative divergence. Even NCEPers have to refer to a table to "decode" a variable name and its units.

In the web plot, you can "include variable definitions and units" to get more information. This button is on the page with the .ctl selection. (By the way the "-v" option in wgrib will give you the same information.)

The last case part of the name refers to the level. Some times the suffix is dropped for "obvious" variables and for pressure level data. 

    sfc     =  surface
    toa     =  top of atmosphere
    mwl     =  maximum wind level
    trp     =  tropopause
    tht     =  isentropic data, constant theta levels
    prs     =  pressure level data
    sig995  =  0.995 sigma level (pressure = 0.995/Psurface)
    lcb     =  low cloud base
    lcl     =  low cloud layer
    lct     =  low cloud top
    mcb     =  middle cloud base
    mcl     =  middle cloud layer
    mct     =  middle cloud top
    hcb     =  high cloud base
    hcl     =  high cloud layer
    hct     =  high cloud top
    2m      =  2 meters above ground
    10m     =  10 meters above ground
    clm     =  atmospheric column average
    10cm    =  0-10cm below ground average
    200cm   =  10-200cm below ground average
    1829m   =  1829 meters above ground (6000 feet)
    2743m   =  2743 meters above ground (9000 feet)
    3658m   =  3658 meters above ground (12000 feet)

Of course not all the datasets follow this convention. This is the US government, what else would you expect?

Plotting Problems

Network Problems, or Services Refused:

  • The ftp/plot server may be down. Try another server or try later.
  • You may have a "firewall from hell®".

Network/Configuration Problems:

  • The workstation with the database may be down.
  • The network may be down.
  • Stale NFS handles. Try another server.
  • The database may be unavailable or may have been moved.
  • I made a mistake.

Blank Plot

  • The plot could not generated because the data was not available.
  • Some datasets claim to start at Jan 19xx but the dataset really starts several months later.
  • Some variables are not saved on all levels such as humidity above 300 mb and omega above 100 mb.
  • Computer or software problems can make the data unavailable.

Only old forecasts/analyses

  • Some datasets are not updated in real-time.
  • CDAS is updated around 3-5 days after the end of the month.
  • If the data are really missing. Please send email to person responsible for the dataset.

Problems with Downloading the Data

To retrieve the data, click on "ftp" or "FTP" on the previous page. "FTP" indicates the datasets can be downloaded using an anonymous FTP server, while "ftp" indicates that you have to use the FTP2U facility. If the data is only available through FTP2U, then you will need to fill out a form so that our computers can FTP the data to your machine. We realize that this is impractical for cron jobs and some secure computer facilities but this is the best we can do given our limited hardware.

I never had any problems filling out the FTP2U screen for obvious reasons (I wrote it). The help screen from a previous version may help you.

Common mistakes include:

  1. wrong IP address or using e-mail address in the "computer" section
  2. using an e-mail address in the "user-id" section
  3. sending the data to a protected or non-existent directory
  4. not having an FTP server running on the computer (common on Windows PCs)
  5. trying to send data through a firewall

FTP transfers can fail for unexplained reasons. It is more common for large files, long distances, and busy times of the day. If you have problems with the FTP2U transfers, reduce the size of the file transfers (GRIB FILTER), move closer to NCEP :-) or try again when the internet traffic is lighter.

If the file transfer is too slow (big files, slow connection, slow computers), Netscape may timeout and say "no data". Before you give up, check to see if the file is still downloading. (The file size should increase until finished.) You can tell when the download is complete by comparing the file size to that on the FTP2U screen.

FTP2U may fail when you try to obtains small amounts of data from a large number of files. The problem is caused by FTP2U using a separate FTP session for each file. Unfortunately a rapid sucession of FTP sessions looks like a denial of service attack. (On linux boxes, you need to change the wait parameter in inetd.conf.)

Finally FTP2U transfers probably will not work if you want to transfer data through a firewall. If you have a "firewall from hell®". nothing will work. Other firewalls allow you to initiate FTP transfers from within the firewall. In this case, FTP2U the data to a third machine set up to receive data by FTP. Normally you would send the data to one of your machines outside of the firewall. If that is not possible, you can use my PC as a last resort.

Problems using the Data

Common mistakes include:
  1. If you only retrieve part of the data file, you must remake the GrADS control and index files. See grib2ctl.
  2. Some of the "plot" variables are computed on the fly and obviously cannot be found in the data file.
  3. non-grib binary files are for SUN/SGI/HP workstations and are in big endian order. We like making life tough for x86 and Alpha users.
  4. Not reading the (missing) documentation. If the documentation is missing, ask the responsible person for the dataset information.

Documentation

Several of the datasets do not have any on-line documentation. Contact the person responsible for the dataset. He/she may have references or copies of paper documentation. Given sufficient demand, people tend to place the documentation on-line.

Control File Modification

Point of Contacts can enhance their control files to make better plots.

back

comments: Wesley.Ebisuzaki@noaa.gov

Privacy Policy