UEMS Announcements & Other Stuff

7  September  2016
Whoomp there it is:  UEMS  V15.89  - The "In Purgamentum, Exit Purgamentum" release

This is what "Exit Purgamentum" means for you:

a. Fixed a problem with ems_post wherein the number of processes requested with EMSUPP was not being properly distributed across the specified nodes.

b. Change the layer for the most unstable CAPE from 0-300mb to 0-255mb to conform to accepted norms. As an expression of protest, the developer is wearing his underwear outside of his pants.

c. Modified (again) the calculation of total/high/middle/low-level cloud cover in the UPP. Previously the UPP followed the HRRR/RAP model by taking the mean value over a 30 km radius for each grid point, which resulted in a overly filtered (smoothed) appearance to the field. The new value is now calculated over a 12km radius and looks prettier, which is what really matters.

d. Turned OFF fields on PV surfaces in emsupp_cntrl.parm file by default. Previously, the -1.5 PV surface was turned on which is useless (almost) in the northern hemisphere and was causing a problem with the GRIB 2 complex packing routine. Enough said.

e. Fixed a bug (almost last one!) in that the mercator projection GRIB 2 files had an incorrect resolution & component flag value (8) indicating no DX, DY values were available when they really were. Change flag (56) in EMSUPP. That'll learn'm

f. Fixed (again) a bug when using the HRRR for initialization in that the specific humidity fields were not being properly initialized.

g. Turned on a feature in the EMSUPP that allows for all model hybrid surfaces to be written to GRIB instead of the maximum of 80. So if you have 201 model levels you can now write them all out to GRIB. Use of this option is described in the emsupp_cntrl.parm file.

h. Updated the ems_domain utility so now the "--refresh", "--restore", and "--import" options are properly working with the latest release. For those unfamiliar with the wonders of these features:

% ems_domain --refresh
Updates you domain configuration files with the latest versions while keeping your configuration intact.
% ems_domain --restore
Replaces your domain configuration files with the default versions, which are guaranteed to work. This is when you really screw something up and need a sense of calm.
% ems_domain --import [path to domain directories]
Import, refresh, and localize domain directories from another UEMS installation or from an area where they have been placed for safe keeping. This option does not harm the files in the original directory.
Also: % ems_domain --create
Creates a new domain
Note: "ems_domain --create" replaces "ems_domain --newdom" because I never liked "newdom"

i. Fixed bug in time where all nested domains were initialized from the parent domain data set rather than the initialized data set.

"Huh?" - you exclaim. This might be an issue if you were running a 21km -> 3km nest (yes 7:1 but don't do it) over Salt Lake City with both domains starting at the same time (T0), and you initialized your simulation with the 3km HRRR data set. Previously, the 21km domain would be used to initialize the 3km nest. Now, the 3km HRRR data are used for the 3km nest. Note that if you start integrating a nested domain after the primary simulation your meteorological fields will always come from the parent but the terrestrial fields (static) come from the localized dataset for that domain (higher resolution).

j. Some other stuff I've forgotten

8  February  2016
 UEMS  V15.58  now available - The "Apples to Xerotic" release

This is what "Apples to Xerotic" means for  you:

a. The archive of 0.25 degree GFS forecast GRIB files on the EMS3 data server been increased to 1 year. This means that future runs of the GFS will be accessible to personal tile users for much longer than the current 14 days.

b. The archive of 3km HRRR forecasts (full forecast GRIB) on EMS3 has been increased to 7 days AND the archive of (hourly) 3km HRRR 00-hour forecast GRIB files has been extended indefinitely.

Why archive only the 00-hour forecast data set? Because this is all you need to provide initial and boundary condition data for your simulations. By simply including the "--analysis" flag, the UEMS will use the 00-hour forecast files from consecutive hourly HRRR runs for the boundary conditions:

% ems_prep --dset hrrrpt --date 20160415 --cycle 15 --length 48 --analysis

Yes, you can run a 48 hour (or longer) simulation using just HRRR 00-hour forecasts as your initialization data set.

c. Changed the calculation of RADT when using the "Auto" option and fixed a small bug. Now there are three "Auto" choices:

1. Auto - Use a separate value for RADT based on the DX of each domain
2. Auto_1 - Use the value calculated for the primary domain for all sub-domains
3. Auto_2 - Use the value calculated for domain with the smallest DX for all domains

The default is "Auto" - Use a separate value for RADT based on the DX of each domain

d. Added a test to ensure that the number of grid points in each direction within a CPU tile is not less than the 8 for the primary domain and 10 for nested domains.

Violating this requirement results in a corrupted BC input file, which can compromised a simulation and lead to an unexplained crash.

e. Added verbiage to the run_ncpus.conf file explaining how to prevent this silent killer of simulations (and the developer's patience) world-wide.

f. Fixed a bug in the navigation of nested lat-lon domains.

g. Fixed a "designed flaw" WRF wherein the SEAICE field over lakes was being set to zero when the point was initialized as ice covered.

h. Fixed Registration file entry for Morrison MP scheme as per "Known Problems" page on NCAR WRF site.

i. Recompiled WRF code, again, and again, and again.

j. Changed the dominant precipitation type category fields back to EMS V3.4 output where values are either 1 (that type) or 0 (no precip or not type)

k. Fixed problem when using the adaptive timestep with a regional lat-lon domain. Now fft_filter_lat is assigned a value of 89 degrees to avoid using of the default registry value (91 degrees).

l. Corrected an error in the NNRP gribinfo configuration files wherein the wrong METGRID tables were being used leading to an incorrect skin temperature initialization.

m. Users can no longer begin a nested simulation after the start of the parent domain when initializing with multiple data sets (--dset dataset1%dataset2). This is due to changes in the WRF registry that determine which fields from the parent are used to initialize a child domain.

n. Fixed the Perl code to allow the QNSE PBL scheme to run with CU schemes that include shallow cumulus effects by turning OFF the EDMF scheme (mfshconv=1). Also provided a better explanation in the updated run_physics.conf file.

o. Switched the order of the pressure levels for PBL layer fields in the GRIB 2 files. Now they are consistent with the operational fields from NCEP.

p. Added "lightning Potential Index" LPI as an option in the run_lightning.conf file. To view the details, be sure to copy the new configuration file from $EMS_CONF/ems_run/arw/run_lightning.conf into the conf/ems_run directory within the domain you are using.

Additionally, look for the "LPI_ON_SURFACE" field in the updated emsupp_cntrl.parm file (uems/data/tables/post/grib2/emsupp_cntrl.MASTER).

q. Modified HRRR Vtable (Vtable.HRRR) to fix a problem when initializing a simulation with 3km HRRR pressure surface GRIB 2 files. The RH on pressure surfaces were not being used and thus specific humidity values at T0 were too low. Now initialization is done directly with specific humidity from GRIB files.

r. Added some new options to the NOAH MP configuration file that include:

OPT_SNF = 4 - Surface snowfall partitioning based on output from MP scheme OPT_STC = 3 - Semi-implicit Snow/Soil temperature time scheme where Ts use snow cover fraction

s. Fixed a bug in the WRF NOAH MP LSM source code. The snow accumulation field (ACSNOW) was using "mm/s" instead of "mm".

t. Fixed (another) ACSNOW accounting bug in the NOAH LSM. When using the mosiac option, the accumulated snow was off by a factor of "mosaic_cat" or the number of sub-tiles being included in the simulation.

u. Added a few radiation fields to the WRF output because I was asked very nicely. More to come as time allows.

v. Commented out a fatal error message in the NOAH MP scheme that was causing a simulation to fail when scheme was turned ON. The error causing termination appears benign and only occurs during first time step of a nested simulation along a land-water interface. Recompiled WRF code again.

w. Fixed a small bug by setting the NO_STOP_MESSAGE environment variable in strc/EMSutils/Eenv.pm

x. Fixed bug that was causing ems_prep to point to wrong CFSR data set when date was before 1 April 2011.

Some configuration files were updated for this release. You will need to copy the following files into your local conf/ems_run directory to see what's new:

(Located in uems/conf/ems_run/arw)

* run_lightning.conf
* run_physics.conf

Additionally (again), to get the benefits of the new LPI field, you must replace your current static/emsupp_cntrl.parm with the updated emsupp_cntrl.MASTER file located in uems/data/tables/post/grib2.
Be sure to rename it to emsupp_cntrl.parm.

18 December 2015
UEMS Update: Version 15.51  now available

Announcing the first UEMS  update in over three and a half  weeks!

So, "how can this update possibly make me a better person than I already am?"

Well, allow me to explain what this update means to  you:

a. Added period 10m mean U & V-component wind fields. Default is OFF Turning this field ON requires installing the updated emsupp_cntrl.parm file:

% cp $EMS/data/tables/post/grib2/emsupp_cntrl.MASTER static/emsupp_cntrl.parm

And then turn ON:


b. Fixed bug where topo_wind was missing from namelist.wrfm file when using YSU PBL scheme but set to 0 (topo_wind = 0).

No big issue as default is topo_wind = 0 anyway.

c. Added convective cloud cover to GRIB files. Turning this field ON will require installing the updated emsupp_cntrl.parm file:

% cp $EMS/data/tables/post/grib2/emsupp_cntrl.MASTER static/emsupp_cntrl.parm


d. Added the much anticipated "Temperature at inland lake level depth (K)" field (WTMPIL_ON_DEPTH_BELOW_WATER_SFC) to the GRIB output. This field will only be available when the simple lake surface physics scheme is turned ON in the run_sealakes.conf file (SF_LAKE_PHYSICS=1). This field should be of interest to those planning day trips to the Edmund Fitzgerald or are "sleeping with the fishes" and want a bedtime temperature forecast.

e. Made slight correction to calculation of personal tile request for global data sets. That's all you need to know about that.

f. Fixed bug that was causing the GrADS & GEMPAK image processing scripts not to be called.

g. Fixed small bug when running "sysinfo" on a system where the Perl "Time::Hires" module is not installed. You should install it anyway.

h. Fixed bug in WRF REAL (again) that was causing the routine to fail when initializing with data sets on different vertical coordinate systems, specifically RAP (hybrid) and anything on pressure surfaces (GFS,NAM,HRRR).

i. Now users can define the model (process) ID and subcenter ID to be written to the GRIB2 files. Previously, users were limited to the default values and liked it. Also updated post_grib.conf to explain the new functionality.

j. Modified ungrib so that RAP surface pressure was output to the WRF intermediate files. Was previously only working for RUC.

To do an update:

% uems_install.pl --update

Finally, If you are registered but don't have the current UEMS install utility then feel free to contact me for the update; otherwise, you have to register.

13 November 2015
Major UEMS Announcement: UEMS version 15.46  - The "Magnum Opus" release

Announcing the first major UEMS* release in over three and a half years! And only three years after the first anticipated release date!

I present to you the almost nearly more perfect than before, UEMS V15.46, the "Magnum Opus" release!

* Formerly know as WS ETA, WS WRF, UEMS, EMS, and ξ


It's taken so long, just what the %^&$ have you been doing?!

Allow me to explain:

a. The "Unified" has been prematurely added to the Environmental Modeling System to create the “UEMS.” The “unified” refers to the inclusion of the NEMS NMM-B package, which is still undergoing development, and thus not functional. But at least something is ahead of schedule!

b.The package now features WRF version 3.7.1. It’s finally your chance to be up-to-date with the latest NWP technology. This release includes all the physics schemes that are in the current WRF, which might be too much of a good thing.

c.The UEMS has slimmed down, as it no longer carries any NWP “junk in the trunk” in the form of support for the WRF NMM core, i686 (64-bit only) binaries, and GRIB 1 post processing output (GRIB 2 only).

The NMM was eliminated since neither NCEP nor the NCAR DTC no longer supports it. This decision was made following a lengthy consultation with myself in a mirror.

d.The STRC Perl modules were completely re-written to facilitate future UEMS releases, which means that you won’t appreciate all the time and effort that went into developing something that just works, but you will be happy that it just works!

e. There is an all-new UEMS post processor (EMSUPP). The EMSUPP was completely rewritten to support GRIB 2 output and to make my life easier by not having to work on code that bares the scars of too many developers. The code is better organized and “field tested” to ensure you get the simulation output that you deserve. It even has that new post processor smell, which has been described as an “acquired scent.”

In addition, the EMSUPP control files are skillfully documented and packed with new and exciting fields, such as synthetic satellite output for multiple observing systems, hail and graupel simulation and period accumulation as well as maximum and instantaneous precipitation rates. Way too many fields to list, so I’m not going to even bother listing any, except those I’ve already listed.

f. There is an all-new UEMS BUFR file processor (EMSBUFR), for all the same reasons as sited above.

g. The UEMS now includes source code! That’s correct, all my blood and sweat encrusted source code files are provided for those looking to fill weeks of time attempting to compile your UEMS executables locally; however:

1. I don't provide the WRF source code. you will need to get the official release from NCAR and replace some files with the UEMS modified versions. It’s trivial though.

2. The instructions and build scripts are provided as well as all the libraries and utilities, but since the package was build with PGI you will need to do some work if planning to use the GNU compilers.

h. The UEMS now includes personal tile support for the 3km HRRR , 0.25 degree GFS, 5km NAM CONUS nest, and 6km Alaska nest. Here’s a complete list of the personal tile data set options (note that “ptile” has been replaced with “pt” in the moniker): cfsrpt, gfsp25pt, gfsp50pt, hrrrpt, lispt, nam4kmpt, namak6kmpt, namakpt, namcapt, nampt, narrpt, rappt, rtgsstpt, rucpt, sportsstpt

i. There is much more that has changed in the UEMS package since the previous release, but after three and a half years I can’t begin to recall everything. Besides, you would not want this announcement to get any longer.


If you're stuck in the age of EMS V3.4.X, well then gramps, here is your opportunity to cool down from chasing kids off your lawn and inject some "better than bingo" excitement into your life. All you will need is the current UEMS install utility and a willingness to let go of the past, as in your current release.

To do a new install:

% uems_install.pl --install

Warning, the new WRF terrestrial data sets are very large. I've broken them down into 70 smaller files but it still takes a long time to download.

Should you have an interruption during the download process, I've added a "--continue" flag so you don't have to start from the beginning. If it fails, then just re-run uems_install.pl with:

% uems_install.pl --install  --continue

And if you simply want to simply freshen up your old "analogue"  EMS, well, you can't, because the new release requires a fresh install as well as NO horse-trading of files with the old installation; otherwise,  your system will likely break and you will wake up in a bathtub full of ice with your kidney missing. So just don’t do it (NIKE copywrite?).

Finally, If you are registered but don't have the current UEMS install utility then feel free to contact me for the update; otherwise, you have to register.