[Wrf-users] WRF-NMM mis-handling of SEAICE for small lakes
Case, Jonathan (MSFC-VP61)[ENSCO INC]
jonathan.case-1 at nasa.gov
Fri Feb 18 12:01:46 MST 2011
Here is just a quick point of clarification on how the WRF-NMM incorporates sea ice into the initial condition for version 3.1.1. Upon looking at the real_nmm code, I realized that the WRF-NMM initialization routine does NOT allow the user to set any sea ice based on input skin temperature thresholds. The only way the WRF-NMM can initialize ice is by reading in a SEAICE grid from an external GRIB2 file (e.g. NAM or SPoRT's Great Lakes SEAICE via the icegl option in the EMS).
As a result, smaller water bodies in Canada and the Northeast U.S. are left as open water and assigned unrealistically cold water temperatures well below freezing in high-resolution NMM runs. I have not confirmed whether this has been fixed in WRF v3.2, but in case it hasn't been fixed, I have CC'd the wrfhelp and wrf-users forum email lists , as well as Bob Rozumalski to the reply.
(NASA SPoRT Center)
From: Dave Radell [mailto:David.Radell at noaa.gov]
Sent: Friday, February 18, 2011 12:13 PM
To: Michael Evans; Paul Sisson; Molthan, Andrew L. (MSFC-VP61); David Zaff; Robert LaPlante; Jeff Waldstreicher; Kenneth Johnson; Brian Miretzky; Joshua Watson; Fred Pierce; Christopher Mello; Chuck McGill; Ron Murphy; David Radell
Cc: Case, Jonathan (MSFC-VP61)[ENSCO INC]; Vasil Koleci
Subject: Update: Northeast WRF Ensemble
I figured I'd take this opportunity, while the winter weather is somewhat quieter, to give everyone an update on some recent WRF LES ensemble developments.
1. NMM Ice Cover Mask Issues
SSD discovered a problem with the SpORT ice mask used in the NMM core of the WRF-EMS. We noticed this last week during the Feb. 9-11 event in Buffalo's CWA, though the problem has likely persisted much longer. Essentially, the SpORT ice mask in the NMM-WRF was supposed to be set by the SpORT SST product (if SST<273K, set the point to ice) , and that was not happening correctly. So for each model run, the NMM would revert back to deriving ice cover from NAM skin temperatures which for last week's event covered most of Lake Ontario with ice---not very realistic. So the SpORT group quickly came up with a fix that involved creating a separate ice mask (independent of the SST product) grib file for use in creating the initial WRF conditions. We have tested this here with NMM runs, and it's working correctly, giving the proper lake ice cover. Note that this did not affect those running the ARW core. Here are the configuration options to check in your NMM configuration files to ensure that you're set up correctly for the SpORT SST and ice cover product:
To set up a run using both our SSTs (ftp://ftp.nsstc.org/outgoing/lafonta/sst/grib2/conus/) and the Great Lakes ice mask (ftp://ftp.nsstc.org/outgoing/lafonta/sst/grib2/great_lakes/) requires setting the following parameters in the conf/ems_auto/ems_autorun.conf file:
SFC = sstsport,icegl
BESTHR = sstsport (optional, for the diurnal matching of SSTs to model initialization hour)
If running the EMS manually, then the following options should be used in ems_prep (as a working example):
ems_prep --dslist namptile --sfc sstsport,icegl
Also, EMS users must be using at least EMS v184.108.40.206.1 for the "icegl" option to work.
A quick test of running 'ems_prep --dslist ' will reveal if the icegl option is available.
If you have questions on this, please let me know, and we'll coordinate with Andrew Molthan and Jon Case at SPoRT as appropriate.
2. BUFKIT Soundings
Ensemble soundings look to be generating regularly on the ER FTP server in \share\BTV\BUFKIT. A quick check this morning has sounding data for the following locations: CHD, CLE, COLT, EON, ERI, GOU, HZY, KBGM, KBTV, KITH, KMSS, KSLK, KSYR, RME, STAR. It appears CLE, BTV, and BGM are regularly providing their ensemble members for inclusion.
3. LES Ensemble Configurations
As promised and for those interested, I have posted the information I've received from you all on the individual ensemble members on the ER Local Model Wiki page at the address below. If would to like to make changes or updates to these configurations, please feel free to edit away directly on the Wiki page.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Wrf-users