-
Notifications
You must be signed in to change notification settings - Fork 6
longwosion/laps-mirror
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
 |  | |||
Repository files navigation
This LAPS README file is viewable on the WWW via the LAPS home
page at http://laps.noaa.gov/
TABLE OF CONTENTS
-----------------
1.0 ............ General LAPS info
1.1 .............. LAPS Software Disclaimer
2.0 ............ Installing and running the LAPS Ingest/Analysis code
2.1 .............. UNIX System Requirements
2.1.1 .................. NetCDF library
2.1.2 .................. Perl
2.1.3 .................. make
2.1.4 .................. C Compiler
2.1.5 .................. FORTRAN Compiler
2.1.6 .................. Disk Space
2.1.7 .................. Memory Etc. (ulimit)
2.1.8 .................. Plotting / NCAR graphics library (optional)
2.1.8.1 ................ Web Display
2.1.9 .................. GRIB2 external libraries
2.1.10 ................. GNUPLOT / ImageMagick for verification (optional)
2.2 .............. Installation Procedure Summary
2.2.1 .................. Untarring the Source Code
2.2.2 .................. Running Configure
2.2.2.1 ..................... Modifying Compiler Flags
2.2.3 .................. Ingest Software changes
2.2.4 .................. Running make
2.2.5 .................. Geography databases
2.2.5.1 ..................... High Resolution Terrain (sub-kilometer)
2.2.6 .................. Localizing for single or multiple data domains
2.2.6.1 ..................... Localization Method 1
2.2.6.2 ..................... Localization Method 2
2.2.6.3 ..................... Localization with LAPS GUI
2.2.7 .................. WRF Domain Wizard LAPS Support
2.2.8 .................. MPI support for LAPS wind analysis
2.3 .............. Raw data ingest
2.3.1 .................. Model Background (lga/lgb)
2.3.1.1 ..................... Acquiring Model Background Data
2.3.2 .................. Radar ingest
2.3.3 .................. Surface Data
2.3.4 .................. Wind Profiler / RASS
2.3.5 .................. PIREPS & ACARS from aircraft
2.3.6 .................. RAOB / Dropsonde / Radiometer
2.3.7 .................. Satellite
2.3.8 .................. GPS
2.3.9 .................. Other Data Sources
2.4 .............. Running LAPS Analyses
2.4.1 .................. Cron timing considerations
2.4.2 .................. Purging Output Files
2.4.3 .................. STMAS and other configurations
2.5 .............. Test data case
2.5.1 .................. Analysis Only Test
2.5.2 .................. Ingest + Analysis Test
2.6 .............. I/O of LAPS gridded files
2.7 .............. CHANGING THE HORIZONTAL DOMAIN
2.7.1 .................. Number of Grid Points
2.7.2 .................. Location of Analysis Domain (Map projections)
2.7.2.1 ..................... MAP PROJECTION FUNCTIONALITY/LIMITATIONS
2.7.3 .................. Domain Resolution
2.7.4 .................. Terrain Smoothing/Filtering
2.8 .............. CHANGING THE VERTICAL DOMAIN
2.8.1 .................. Sigma Height Grid
2.8.2 .................. Sigma Pressure Grid
2.9 .............. CHANGING THE CYCLE TIME
2.10 ............. LQ3 (HUMIDITY ANALYSIS) CHANGES
2.11 ............. OTHER RUNTIME PARAMETERS
2.12 ............. Detecting and Reporting Installation Errors
2.12.1 .................. Runtime Monitoring
3.0 ............ Description of LAPS Processes
3.1 .............. Localization Processes
3.1.1 .................. Gridgen_model (static.nest7grid generation)
3.1.2 .................. Surface Lookup Tables (gensfclut.exe)
3.1.3 .................. Satellite Lookup Tables (genlvdlut.exe)
3.2 .............. Ingest Processes
3.2.1 .................. LGA Model Background
3.2.2 .................. Surface Data Ingest
3.2.2.1 ..................... obs_driver.x
3.2.2.2 ..................... How to Blacklist stations
3.2.3 .................. Polar Radar Data (e.g. WSR 88D Level II, Level III)
3.2.4 .................. WSI / NOWRAD RADAR PREPROCESSING (VRC)
3.2.5 .................. Radar Mosaic
3.2.6 .................. PROFILER/VAD/SODAR (PRO) Ingest
3.2.7 .................. RASSs (LRS) Ingest
3.2.8 .................. PIREPS/ACARS Ingest
3.2.9 .................. Sounding (RAOB/Dropsonde/Sat/Radiometer) (SND) Ingest
3.2.10 .................. LVD (Satellite Image + Cloud Top Pressure)
3.2.11 ................. Cloud Drift Wind (CDW) Ingest
3.3 .............. ANALYSIS PROCESSES
3.3.1 .................. WIND
3.3.2 .................. SFC (LSX)
3.3.2.1 ..................... SURFACE ANALYSIS RUNTIME PARAMETERS
3.3.2.2 ..................... SURFACE ANALYSIS QUALITY CONTROL
3.3.2.3 ..................... SURFACE ANALYSIS VERIFICATION
3.3.2.4 ..................... STMAS-2D
3.3.3 .................. TEMP
3.3.4 .................. CLOUD
3.3.5 .................. WATER VAPOR (HUMIDITY PROCESSING)
3.3.6 .................. DERIV
3.3.7 .................. ACCUM
3.3.8 .................. SOIL MOISTURE
3.3.9 .................. BALANCE
3.3.10 ................. STMAS3D
3.4 .............. Model Initialization & Postprocessing
3.4.1 .................. LAPSPREP
3.4.2 .................. LAPS2GRIB
3.4.3 .................. WFOPREP
3.4.4 .................. LFMPOST
3.4.5 .................. FORECAST GRAPHICS
3.4.6 .................. VERIFICATION
4.0 ............ Porting code mods from LAPS users back to GSD
5.0 ............ LAPS Output Variables and netCDF File Organization
----------------- 1.0 General LAPS info --------------------------------------
Below is a description of the tar file containing the LAPS data ingest and
analysis code. The predictive component of LAPS (MM5, RAMS/SFM, ETA) is
set up separately (see Section 3.4).
Please note that GSD provides support for LAPS software only if a prior
agreement is made to that effect. Additionally, questions concerning LAPS
must be asked in reference to the latest released tar file; we cannot
support older versions of LAPS code. It is also recommended that LAPS
users try to take advantage of the latest LAPS updates by periodically
importing a fresh tar file every few months or so. Please check the
LAPS Software Page at 'http://laps.noaa.gov/cgi/LAPS_SOFTWARE.cgi'
for information about recent releases.
A flow chart for the LAPS software can be found at:
http://laps.noaa.gov/doc/Slide1.png
---------------- 1.1 LAPS Software Disclaimer ----------------
Open Source License/Disclaimer, Forecast Systems Laboratory
NOAA/OAR/GSD, 325 Broadway Boulder, CO 80305
This software is distributed under the Open Source Definition,
which may be found at http://www.opensource.org/.
In particular, redistribution and use in source and binary forms,
with or without modification, are permitted provided that the
following conditions are met:
- Redistributions of source code must retain this notice, this
list of conditions and the following disclaimer.
- Redistributions in binary form must provide access to this
notice, this list of conditions and the following disclaimer, and
the underlying source code.
- All modifications to this software must be clearly documented,
and are solely the responsibility of the agent making the
modifications.
- If significant modifications or enhancements are made to this
software, the GSD Software Policy Manager
([email protected]) should be notified.
THIS SOFTWARE AND ITS DOCUMENTATION ARE IN THE PUBLIC DOMAIN
AND ARE FURNISHED "AS IS." THE AUTHORS, THE UNITED STATES
GOVERNMENT, ITS INSTRUMENTALITIES, OFFICERS, EMPLOYEES, AND
AGENTS MAKE NO WARRANTY, EXPRESS OR IMPLIED, AS TO THE USEFULNESS
OF THE SOFTWARE AND DOCUMENTATION FOR ANY PURPOSE. THEY ASSUME
NO RESPONSIBILITY (1) FOR THE USE OF THE SOFTWARE AND
DOCUMENTATION; OR (2) TO PROVIDE TECHNICAL SUPPORT TO USERS.
---------- 2.0 Installing and running the LAPS Ingest/Analysis code -----------
---------------- 2.1 UNIX System Requirements ----------------
Supported UNIX platforms include...
<pre>
IBM rs6000 AIX4.3 NFS mounted disks should be mounted with
NFS version 2 instead of 3.
HP HPUX 10.20 Requires f90
SunOS (Solaris) 5.6 Requires f90
IRIX64 6.5 Requires f90
DEC (Alpha)
LINUX (i386,i686,x86_64) pgf90 is suggested
ifort (version 12.0.4 or later)
gfortran being tested
LINUX (MacOS-X) pgf90 is being tested
LINUX (Alpha) fort is suggested
</pre>
We are working on adding more supported platforms. We welcome suggestions
on how to modify LAPS for other platforms/versions. Note that we cannot
guarantee the portability of LAPS to all of these
other platforms (e.g. Windows NT).
--- 2.1.1 NetCDF library ---
The netCDF package is required for laps, we suggest using version 3.6.0
available here:
http://www.unidata.ucar.edu/downloads/netcdf/netcdf-3_6_0/index.jsp
Once netCDF is properly installed, check that the 'ncdump' and 'ncgen'
programs are in your path (e.g. 'which ncdump'), so that 'configure' will
find them and provide the laps package with the proper path. The path
setting will also help LAPS programs to run properly. Note that this path
specficiation will override anything supplied with the '--netcdf' command
line argument.
NetCDF is a general format structure. The detailed format of each data
file is self-describing (via 'ncdump'), and is mirrored in a separate
static file called the CDL. This CDL can be GSD's version or someone elses.
If possible, check that your netCDF library is built to be compatible
with the same FORTRAN compiler that you are using.
The netCDF library contains C routines that are to be linked with the
LAPS FORTRAN routines. Please see the discussion in section 2.2.2.1 for
details on troubleshooting this.
--- 2.1.2 Perl ---
The perl package is also required for laps, it is available via internet
at any perl site such as http://www.perl.com. Perl 5.003 or higher is
required. Check that 'perl' is in your path (e.g. 'which perl').
--- 2.1.3 make ---
Laps Makefiles work best by using gnu make (version 3.75 or higher).
This is downloadable from gnu sites such as the following URL:
'http://www.gnu.org/software/make/make.html'. You can check your version
of gnu make by typing 'make -v'. Some vendor provided make utilities may
also work, however if you find you are having problems in this area please
try obtaining and using gnu make. Check that 'make' is in your path.
If you are running AIX, 'make' should be used from the AIX toolbox:
http://www-03.ibm.com/systems/power/software/aix/linux/index.html
--- 2.1.4 C Compiler ---
In general, an ANSI compliant C compiler should be used. On some hardware
ANSI compliance requires a compiler flag, if you're not sure check the
documentation for your compiler. Some platforms such as Solaris and HPUX
do not come with an ANSI compliant C compiler by default. If you have
not purchased that additional product from the vendor, we recommend GNU C
(gcc) available at 'http://www.gnu.org/software/gcc/gcc.html'. Check that
the C compiler is in your path.
With the 'pgf90' FORTRAN compiler, 'pgcc' is recommended.
With the Intel 'ifort' FORTRAN compiler, 'icc' is recommended.
For AIX, Solaris, HP_UX platforms, 'cc' is recommended.
--- 2.1.5 FORTRAN Compiler ---
Please note that LAPS uses dynamic memory within the FORTRAN code in
the form of automatic and allocatable arrays, as well as other FORTRAN 90
constructs. This implies that you will need an 'f90' compiler or the
equivalent. LAPS will no longer work on most 'f77' compilers. Check that
the FORTRAN compiler is in your path.
For IBM/AIX platforms 'xlf' is recommended.
For Solaris & HP-UX platforms, 'f90' works well.
For Linux platforms (i386,i686,x86_64), 'pgf90' is suggested
'ifort' (version 12.0.4 or later) is being tested
'gfortran' is being tested
For Linux platforms (Alpha chip), 'fort' is suggested (normal serial use).
--- 2.1.6 Disk Space ---
The disk space requirements for LAPS vary depending on factors such as
domain size and purge parameters. As a general guide, 10MB would be needed
for source code. About 30MB are needed for executable binaries.
500MB to 1GB are typically needed for 12-24 hours worth of output data.
A similar amount of space is needed for the raw input data.
--- 2.1.7 Memory Etc. (ulimit) ---
'ulimit' settings should be placed at 'unlimited' if possible. Memory
requirements vary for LAPS. As a general guide, 128MB is needed and
256MB is preferred. More is needed for large domains. For very large
domains, a rough guide to the memory needed would be 100 x NX x NY x NZ
bytes.
--- 2.1.8 Plotting / NCAR graphics library (optional) ---
Lapsplot is an optional plotting program, thus NCAR graphics is optional.
If you wish to build the lapsplot process, access to NCAR graphics
libraries is needed so you will be able to run the 'ncargf77' command in
the LAPS Makefiles. You can download the free NCAR graphics (NCL) software
at the URL shown below. Note that NCAR graphics libraries should be built
against the same FORTRAN compiler being used in LAPS.
http://ngwww.ucar.edu
The 'lapsplot.exe' executable is an interactive program that reads in
the netCDF LAPS files and produces a 'gmeta' file as output. The 'gmeta'
file can be displayed using other NCAR graphics utilities like 'ctrans'
and 'idt'.
'Lapsplot' is designed to work with version 3.2 (or higher) of NCAR
graphics. The environment variable $NCARG_ROOT should be set when
configuring, compiling, or running 'lapsplot.exe'.
Before running 'configure', check that 'ncargf77' and/or 'ncargf90' is in
your path. If you are using a compiler other than 'f77', check after
running 'configure' to see that the right thing was done by inspecting
'NCARGFC' and 'FC' within 'src/include/makefile.inc'. 'NCARGFC' should
point to either the 'ncargf90' or 'ncargf77' script. If configure wants
to use 'ncargf90' and you don't yet have one, then consider making a soft
link called 'ncargf90' that points to the 'ncargf77' script, or copying
'ncargf77' to a new location and calling it 'ncargf90'.
If you only have an 'ncarg90' script (i.e. no 'ncargf77'), you may want
to also make a script called 'ncargf77' that lists the 'f77' compiler.
This can help 'configure' do its test for making the switch over from
'ncargf77' to 'ncargf90'.
Lapsplot is built as a special option to 'make', simply type
'make lapsplot' or 'make install_lapsplot'. It is not built with a plain
run of 'make'.
In order to get 'lapsplot' to compile and link properly it may be
necessary to edit your own version of 'ncargf90' or even the original
'ncargf77' script. Check that the proper FORTRAN compiler, load flags,
and load libraries are set in the script.
A possible alternative to fixing 'ncargf77/ncargf90' is to edit
'src/include/makefile.inc' with the full path for 'NCARGFC', and
appropriate compiler for 'FC' (and possibly compiler flags) for your
system (after running configure).
At times the linking of 'lapsplot' may show undefined references to
library routines. This often represets a mismatch between NCAR graphics
and various system libraries. Possible solutions for this include editing
the library list within the 'ncargf77/ncargf90' script or switching the
-Bstatic flag on or off.
'Lapsplot' can be modified to show political boundaries outside of the
U.S. The following data files are relevant from the 'static/ncarg'
directory: 'continent_minus_us.dat', 'state_from_counties.dat', and
'uscounty.dat'. These political boundary files are stored in big_endian
format. These would need to be converted manually prior to using
'lapsplot', if your machine is expecting little_endian. We will consider
automating this in the future.
To run lapsplot manually you can do the following...
1. setenv LAPS_DATA_ROOT to the correct path
2. run $LAPSINSTALLROOT/bin/lapsplot.exe (answer the questions it asks
interactively)
3. idt gmeta
Please note that 'lapsplot' is provided to help you check out how your
LAPS implementation is working. Aside from the pre-generated and
"on-the-fly" web products, we do not have any other plotting or
visualization packages available for distribution with LAPS at this time.
Many users have interfaced LAPS with their own display software (e.g.
IDV, VIS5D, AVS, IDL, NCL, NCVIEW, GEMPAK).
IDV is handy as it can read Grib-2 LAPS analysis output that can be
generated via the 'laps2grib' program. Feel free to post questions about
the various plotting packages to the online LAPS forum.
Another note of interest is that LAPS is visualized as an integral part of
the AWIPS & ALPS systems. If you have AWIPS (including either AWIPS-I or
AWIPS-II), then LAPS should be running on it and you can view its output
on the workstation.
---- 2.1.8.1 Web Display ---
The 'gmeta' file can be converted into a GIF/JPEG file for web display by
using 'ctrans' in conjunction with the 'netpbm' package of image
conversion programs that can be downloaded at the link just below. Within
this package our web display scripts use the 'rastopnm' and 'ppmtogif'
programs.
http://netpbm.sourceforge.net/
We have the option of making pre-generated GIF images that can be displayed
on the web by invoking the 'sched.pl -f dummy' command line argument.
Please see section 2.4 for more info on 'sched.pl'. The web images appear
as '.gif' files in the 'lapsprd/www/anal2d' directory. It will be necessary
to set the relevant NCARG environment variables (e.g. $NCARG_ROOT) in your
.cshrc file since several of the scripts called by 'sched.pl' are CSH
scripts.
The associated web related scripts (for analyses) such as
'etc/www/followup_ncarg.sh' are in the repository. These "wrapper" scripts
run 'lapsplot.exe' and output the GIF images suitable for web display.
The set of web image products are defined with configuration files in
'static/www/lapsplot.*'. Color tables are specified in 'static/www/*.lut'.
Other user definable plotting parameters are located in
'static/lapsplot.nl'.
A separate web related script is our "on-the-fly" page that is contained
in 'etc/www/nph-laps.cgi'. This CGI/PERL script can be run via a web
server. This also calls a set of scripts that wrap around 'lapsplot.exe'.
The file system(s) running LAPS should be made visible on your web server.
After running 'configure' the following steps will help in setting up
this web page.
1) edit 'etc/www/nph-laps.cgi' and set $web_root to be the root directory
of the web server (document root)
2) edit 'etc/www/laps.cgi' and set $web_root to be the root directory
of the web server
3) edit 'etc/www/nph-laps.cgi' and set $ncarg_root to be the root directory
of the NCL/NCAR Graphics installation
4) mkdir -p $web_root/request
5) cd $web_root/request; ln -s $LAPSINSTALLROOT/etc/www/nph-laps.cgi .
6) cd $web_root/request; ln -s $LAPSINSTALLROOT/etc/www/laps.cgi .
7) For each DOMAIN NAME (foo):
a) mkdir -p $web_root/domains/foo
b) cd $web_root/domains/foo; ln -s $LAPS_DATA_ROOT private_data
8) edit 'etc/www/nph-laps.cgi' and set $default_domain to be your
favorite domain 'foo' within the domain list established in
step (7)
At this point you should hopefully be able to use a web browser and
run the "on-the-fly" page with something like this URL:
http://yourdomain.something/request/nph-laps.cgi
--- 2.1.9 GRIB2 external libraries ---
The background models read by the model first guess ingest program (LGA)
include GRIB1 and GRIB2-formatted files. If you are reading model first
guess data in GRIB2 format, then you will want to install these libraries.
The external compression libraries required for processing GRIB2-formatted
files are libjasper.a, libpng.a, and libz.a. They are usually found in
/usr/lib or /usr/lib64. It is recommended to have a system administrator
install these external libraries if they are not already on your system.
(JPEG2000 and other image compression algorithms are built into GRIB2.
Library support for JPEG2000 is provided via the JasPer library. The
implementation of JPEG2000 compression reduces file sizes up to 80%.)
The 'configure' script will determine if these libraries are present.
If all are found, 'configure' prepares the file 'src/include/makefile.inc'
with DEGRIBLIBS, DEGRIBFLAGS and CDEGRIBFLAGS values allowing the lga
software to build to read both GRIB1 and GRIB2-formatted files. Without
these three specific compression libraries available, lga is built to
read only GRIB1-formatted files in addition to netCDF-formatted files.
There may be some occasions where the Jasper library isn't detected
automatically by 'configure'. For example, if the Jasper library is
placed in a location other than the system area (/usr/lib) then one can
set an environment variable CPP_INCLUDE_PATH for lga to build like this:
setenv CPP_INCLUDE_PATH /opt/jasper/1.900.1/include
After running 'configure', the DEGRIBLIBS value in 'makefile.inc' can
be manually edited to include the path information for the Jasper
library. Similarly the flags -DUSE_JPEG2000 and -DUSE_PNG can be added
to the value of DEGRIBFLAGS.
The unix/linux system command 'ldd' command prints the shared library
dependancies on an executable; running 'ldd lga.exe' is a helpful
command in the situation when you download the LAPS precompiled binaries
and need more information about shared libraries required by lga.exe.
See source directory: $LAPS_SRC_ROOT/src/lib/degrib/README_LIBS file
for additional information.
--- 2.1.10 GNUPLOT / ImageMagick for verification (optional) ---
LAPS has a built in verification package and this needs installation
of GNUPLOT and ImageMagick to run fully.
------------- 2.2 Installation Procedure Summary ------------------
To introduce this section, here is a hierarchical listing of some primary
directories and files in the laps tree. The default LAPS structure is
shown in the first tree below. These directories are created/addressed in
various portions of section 2.2 and beyond.
Various "root" directories are mentioned in the form of environment
variables. These can optionally be set to make it easier to follow the
instructions below more literally. The installation scripts can be run
without setting these variables if you'd like to enter the associated
paths directly as command line input.
$LAPS_SRC_ROOT - The full path that was created when the LAPS tar file
was untarred. This contains the source code and other supporting software.
$LAPS_SRC_ROOT is needed for building LAPS but is not needed at runtime.
$LAPSINSTALLROOT - The full path of installed binaries and scripts
(bin and etc). This is where you build the executables, configure the
scripts (converted the *.pl.in to *.pl), and configure
$LAPS_SRC_ROOT/src/include/makefile.inc. Note: $LAPS_SRC_ROOT and
$LAPSINSTALLROOT are in many cases the same but don't have to be.
$LAPSINSTALLROOT is needed at runtime.
$LAPS_DATA_ROOT - The full path to the output data and namelists. This
includes lapsprd subdirectories containing both LAPS output grids and
intermediate data files. $LAPS_DATA_ROOT is needed at runtime and it
contains all the files configured to run an analysis domain localized to
a location on earth. The $LAPSINSTALLROOT tree can drive several
$LAPS_DATA_ROOTs. Input data in its "raw" form is stored outside the
$LAPS_DATA_ROOT tree.
Note: $LAPS_DATA_ROOT is usually (and recommended to be) different than
$LAPS_SRC_ROOT/data and $LAPSINSTALLROOT/data but they don't need to be.
Also, $LAPS_SRC_ROOT/data/cdl and $LAPS_SRC_ROOT/data/static are
the repository versions and should be kept pristine.
Note: the namelists you get from the tar are configured for our Colorado
domain. More on localizing a domain for your own area later on.
To summarize, these three environment variables can either be part of one
directory tree or split out into separate trees as further discussed at
various times below.
<pre>
/home_disk/
raw_data/ (optional raw test data)
geog/
world_topo_30s
albedo_ncep
landuse_30s
soiltype_bot_30s
soiltype_top_30s
laps-m-n-o.tar
laps-m-n-o/ ($LAPS_SRC_ROOT=$LAPSINSTALLROOT)
Makefile
src/
ingest/
etc/ (laps scripts)
bin/ (executables)
data/ (original data tree that comes with tar file,
replicated and merged with templates during localization)
lapsprd/
product_list/ (laps output)
log/
static/
nest7grid.parms (namelist parameters)
*.nl (namelist parameters)
static.nest7grid (gridded topography)
time/
testdata/ (optional, can be relocated)
lapsprd/
product_list/
</pre>
In many UNIX environments, large data files are stored on a "data" disk
and the source code is stored on a smaller "home" disk. Below is a
typical laps directory structure for that setup. We recommend using
something like this setup for most LAPS users. This type of separation
makes it easier to update the LAPS source code while maintaining your
data intact.
<pre>
/home_disk/
builds/
laps-m-n-o.tar
laps-m-n-o/ ($LAPS_SRC_ROOT = $LAPSINSTALLROOT)
Makefile
src/ (source code)
ingest/
etc/ (laps scripts)
bin/ (executables)
template ($TEMPLATE parameters)
/data_disk/
geog/
world_topo_30s
albedo_ncep
landuse_30s
soiltype_bot_30s
soiltype_top_30s
raw_data/ (optional raw test data)
laps/
data*/ ($LAPS_DATA_ROOT, set up during localization)
lapsprd/
product_list/ (laps output)
log/
static/
nest7grid.parms (namelist parameters)
*.nl (namelist parameters)
static.nest7grid (gridded topography)
time/
testdata/ (optional, can be relocated)
lapsprd/
product_list/
</pre>
---- 2.2.1 Untarring the Source Code ---
Place the tar file in the directory '/home_disk' or '/home_disk/builds'.
Untar the laps source code using a command like...
prompt> gzcat laps-m-n-o.tgz | tar xf -
OR...
prompt> gunzip laps-m-n-o.tgz
prompt> tar -xf laps-m-n-o.tar
The $LAPS_SRC_ROOT directory will be set up one level below the tar file.
If you are having trouble running 'gunzip', the problem could be that
the 'laps-m-n-o.tgz' file was corrupted during the download. In that
case simply try downloading again.
---- 2.2.2 Running Configure ---
Go to the $LAPS_SRC_ROOT directory and run...
prompt> ./configure
'configure' supports many options, the most important is the
--prefix option which tells make where to install the laps system
(FORTRAN executables, Perl Scripts, etc.). The default (if you did
not use --prefix) is to install whereever the source is. The use of
the --prefix option is highly recommended to make it easier to update
your source code (e.g. importing a new LAPS tar file), without
disturbing the binaries, data, and runtime parameters that you are
working with on-site. This goes along with the second directory tree
diagram shown above in Section 2.0.
For example, to install laps in directory '/usr/local/laps' (i.e.
$LAPSINSTALLROOT) use...
prompt> ./configure --prefix=/usr/local/laps
One or more data directories for running laps can be specified
at runtime, if desired. A single set of binaries can thus support several
data directories as described below.
Another configure option is --arch. Configure tries to get the
architecture from a 'uname' command, but this can be overridden by having
an $ARCH environment variable or by using --arch. The allowed values for
'arch' include 'aix', 'hpux', etc.
For more information on passing in command line flags to 'configure' run...
prompt> ./configure --help
--- 2.2.2.1 Modifying Compiler Flags ---
The 'configure' script automatically modifies the compiler and compilation
flags by modifying 'src/include/makefile.inc' according to what type of
platform you are on. Hopefully the flags will work OK on your particular
platform. If you want to change the flags from the default set, you
can provide command line arguments to the 'configure' script.
Some examples based on our experience are as follows:
Solaris...
prompt> ./configure --cc=cc
For IBM/AIX platforms, you will want to override the default FORTRAN
compiler with 'xlf' using the command line option --fc=xlf as follows...
prompt>./configure --fc=xlf
For SGI platforms, certain flags may be needed. '-mips3' seems to help
on IRIX64 v6.2.
A second method of modifying the compiler flags is to edit
'src/include/makefile.inc', after running configure. If you find that the
default compiler flags don't work for your platform or that your platform
has no default, you'll need to experiment to find the right set of flags.
Changes in 'src/include/makefile.inc' will automatically modify the
flags used throughout laps. If you find flags that work for your platform
and would like us to add them to the defaults in 'configure' please let
us know via e-mail.
On Solaris for example, you may want to remove "-C" from the DBFLAGS with
an edit of 'src/include/makefile.inc' to allow compiling FORTRAN debug
versions of the software.
On some platforms (e.g. Linux) the linking of FORTRAN programs to netCDF
and other C library routines may need adjustment. This relates to the
existence and number of underscores in the C routine names when called by
FORTRAN routines. Fixes for this may include a combination of changing
the number of underscores in the C routines, changing the CPPFLAGS for
LAPS, or changing the FFLAGS for LAPS.
As an example, with errors linking to netCDF "nf" routines, you might
rebuild the netCDF C library with a different number of underscores and/or
adjust the FFLAGS according to the man page in your FORTRAN compiler.
On a Linux-Intel machine the netCDF library can be rebuilt with the
following flags...
<pre>
FC=pgf90
CC=gcc
CPPFLAGS=-DpgiFortran
FFLAGS=-O
</pre>
Errors linking to other LAPS C routines can be addressed with other
adjustments to the CPPFLAGS (among FORTRANUNDERSCORE and
FORTRANDOUBLEUNDERSCORE) or the FFLAGS.
--- 2.2.3 Ingest Software changes ---
In this file (mainly Sec 2.3), a number of potential manual changes to
ingest code are outlined prior to running 'make' and
'$LAPSINSTALLROOT/etc/localize_domain.pl', especially if one is using
ingest data formats other than "standard" ones used at GSD. After becoming
familiar with the changes needed for your implementation, it is
recommended that you develop a method to save the hand edited files in a
"safe" place outside of the laps directory structure, or by using a
revision control system such as CVS. This strategy would make it easier
to update your implementation of LAPS with the latest 'laps-m-n-o.tgz'
file from GSD, while minimizing the hassle involved with software
modifications for your local implementation.
---- 2.2.4 Running make ---
The next step is to build and install the executables, this can be done
by running the following (note the syntax might vary if the shell you
are using is different from Bourne shell)...
prompt> cd $LAPS_SRC_ROOT
prompt> make 1> make.out 2>&1
prompt> make install 1> make_install.out 2>&1
prompt> make install_lapsplot 1> make_install_lapsplot.out 2>&1
Check that the executables have been placed into the '$LAPSINSTALLROOT/bin'
directory. The total number should be the number of EXEDIRS in
'$LAPS_SRC_ROOT/Makefile' plus 2; this includes 'lapsplot.exe'.
Lapsplot can be installed only if you have NCAR graphics.
We recommend using Gnu Make Version 3.75 or later available via ftp from
any GNU site.
There are many other targets within the Makefile that can be used for
specialized purposes, such as cleaning things up to get a fresh start.
In particular, note that a 'make distclean' is recommended before running
'configure' a second time so that things will run smoothly.
---- 2.2.5 Geography databases ---
Currently there are three mandatory geography databases required to
localize a LAPS domain (with a fourth optional one). These are:
1) terrain elevation (required)
2) landuse category (required)
3) albedo climatology (required)
4) soil type [bottom/top] (optional)
The other geography data paths listed in 'static/nest7grid.parms' represent
data that can be processed by the localization though are unneeded by
the analyses. Hence it is unnecessary to download these and they aren't
available on our software download web page.
The 30" terrain elevation data is found in the tar files for 'topo_30s'.
The landuse data is global 30" data and required to compute a
land/water mask. The mask is used during localization to force consistency
between the other geography data at land-water boundaries. Land fraction
is derived from the landuse data using the water category, with valid
values ranging continuously between 0.0 and 1.0.
The global albedo climatology database has less resolution than either
the terrain or landuse data. The albedo is approximately 8.6 minutes
(0.144 degs) and was obtained from the National Center for Environmental
Prediction (NCEP). This data is used in the LAPS cloud analysis with
visible imagery data.
The geography data come in compressed tar files separate
from the rest of the LAPS distribution. The data are used in
process 'gridgen_model' which is the fortran code to process all
the geography data as specified by the user (see section 2.7.4 for more
information about gridgen_model). Only one copy of the geography
data is required no matter how many LAPS 'dataroot' installations you
are supporting. The paths to the geography data directories
(topo_30s, landuse_30s, and albedo_ncep) are defined as
runtime parameters within the 'nest7grid.parms' file (Sec 2.2.6).
The geography data is available on the LAPS Home Web page (software link).
You will find the following global data sets at this web/ftp site. Some of
the data have been subdivided into "quartershperes" for easier downloading.
Select the files needed for your application or get all of them if you
intend to generate localizations around the entire globe.
.
132446109 Aug 24 2001 topo_30s/topo_30s_NE.tar.gz
63435504 Aug 24 2001 topo_30s/topo_30s_NW.tar.gz
37194099 Aug 24 2001 topo_30s/topo_30s_SE.tar.gz
29204244 Aug 24 2001 topo_30s/topo_30s_SW.tar.gz
12324069 Aug 24 2001 landuse_30s/landuse_30s_NE.tar.gz
6118611 Aug 24 2001 landuse_30s/landuse_30s_NW.tar.gz
3355822 Aug 24 2001 landuse_30s/landuse_30s_SE.tar.gz
2808861 Aug 24 2001 landuse_30s/landuse_30s_SW.tar.gz
albedo_ncep/A90S000E
albedo_ncep/A90S000W
albedo_ncep/AHEADER
We are currently working on a procedure to access higher resolution
terrain and land use data from the USGS (at least to 1 arcsec).
Soil Type and Other Optional Databases:
The laps process 'gridgen_model' described below in section
3.0 can also process soil type, mean annual soil temperature,
and greeness fraction but these are not mandatory data required
in LAPS and therefore we do no describe them here. Soil Type
can however be used in the soil moisture analysis. You'll see
some reference to these data bases below and we have added
paths to this data in our namelist file (nest7grid.parms) but
you should enter dummy paths for these data in the event you
do not have them available. The gridgen_model
process will warn that these data are not available but you
should still see the localization run to completion (ie.,
static.nest7grid is generated).
--- 2.2.5.1 High Resolution Terrain (sub-kilometer) ---
Recently the localization was augmented to accept 30 meter terrain data.
The procedure involves setting up 1 degree tile data from the USGS in
an ArcGIS context, converted from 'adf' format to 'asc'. These tiles are
placed into a path containing the string 'topo_1s' and this path is given
to the 'path_to_topo_30s' variable in 'nest7grid.parms'. This allows the
30m (1s) data to supercede the 900m (30s) data.
High resolution terrain has also been imported (experimentally) into LAPS
via two methods, the WRF Wizard (see section 2.2.7), and Topograbber - see
http://laps.noaa.gov/topograbber/
The WRF Wizard can be used to generate a GEOGRID file on the same grid as
the LAPS analysis. The terrain from this file can be imported during the
LAPS localization by setting the 'nest7grid.parms' namelist path parameter
'path_to_topt30s' to contain the string 'wps' in the directory name.
Also, Topograbber is under development and this might be used with some
further work. The tiles produced may need some modifications to the
'gridgen_model.exe' program so they can be read in. A second way to use
Topograbber with LAPS is to generate a WPS GEOGRID terrain file, and
then read that in during the LAPS/STMAS localization process (see above
paragraph).
--- 2.2.6 Localizing for single or multiple data domains ---
Runtime parameter changes may be needed to tailor LAPS for your
domain(s); this includes ingest and geography data path names,
grid dimensions, grid location, and potentially other aspects of the
data processing. The parameter files are 'data/static/nest7grid.parms',
'data/static/*.nl', and 'data/static/*/*.parms'.
The localization involves several operations. The parameter files are
merged/updated with the repository versions if needed. The dimensions
in the 'cdl' files are also adjusted. Then several executable programs are
run including 'gridgen_model.exe' and 'gensfclut.exe' as per section 3.1.
Below are two mainly equivalent procedures for localizing LAPS to set
up one or more domains. The first is a newer, more efficient (and highly
recommended) method using domain "template" directories. The second is our
original method for localization. You'll want to use either Method 1 or
Method 2 but not both.
--- 2.2.6.1 Localization Method 1 ---
The first method is especially useful if you are using a separated
data tree and/or multiple domains. It is also recommended if you are
doing repeated software updates. Once you learn this method it can save
a lot of time and errors that may occur in the course of using Method 2.
SETTING RUNTIME PARAMETERS
If you are working in a separated data directory (e.g. using the
second tree shown above), you can set up a copy of the runtime
parameter files (for each window) in a new directory (called $TEMPLATE)
with a reduced parameter subset. The $TEMPLATE directory namelist files
should contain only those parameters that need to be changed for each of
the domain(s) from the settings in the repository,
$LAPS_SRC_ROOT/data/static. The remaining unchanged parameters should be
omitted from the $TEMPLATE versions. Otherwise the template namelist
looks exactly like the originally supplied namelist, except that the
comment section should be omitted. The modified $TEMPLATE parameters
generally include map projection settings, data paths, etc. The remaining
fixed parameters will later be automatically merged in from the
'$LAPS_SRC_ROOT/data/static' directory tree by the localization scripts
(next step).
Templates should be maintained in a location separate from the LAPS
distribution and LAPS_DATA_ROOT (e.g. see the template directory in the
tree diagrams above). This avoids them being erased during software
updates and relocalizations. Thus templates can be thought of as more
permanent, since they contain parameters dependent on the local
implementation and relatively independent of software updates. Once you
set up the template directory you'll be ready to run the
'window_domain_rt.pl' script. Here is an idealized example illustrating the
namelist merging process that is done during the localization...
<pre>
template repository tar file localized result
________ ___________________ ________________
$TEMPLATE/vad.nl $LAPS_SRC_ROOT/data/static/vad.nl $LAPS_DATA_ROOT/static/vad.nl
................ ................................. .............................
a=1 a=1
b=5 b=2 b=5
c=3 c=3
d=6 d=4 d=6
</pre>
And here is an example of an actual template for the 'nest7grid.parms'
file...
<pre>
&lapsparms_nl
C80_DESCRIPTION = 'NOAA/GSD LAPS running for Taiwan CWB'
C6_MAPROJ = 'lambrt',
STANDARD_LATITUDE = 10.0,
STANDARD_LATITUDE2 = 40.0,
STANDARD_LONGITUDE = +120.0,
GRID_CEN_LAT = +23.578,
GRID_CEN_LON = +120.91,
GRID_SPACING_M = 9000.,
NX_L = 153,
NY_L = 141,
path_to_topt30s='/home/data_disk/geog/world_topo_30s',
path_to_landuse30s='/home/data_disk/geog/landuse_30s',
path_to_albedo = '/home/data_disk/geog/albedo_ncep',
path_to_raw_profiler='/pj/fsldat/point/profiler/netcdf/',
path_to_qc_acars='/home/data_disk/dat/acars/',
c8_project_common='CWB',
/
</pre>
LOCALIZING with 'window_domain_rt.pl'
Generating new localizations, reconfiguring existing localizations, and
reconfiguring existing localizations without removing lapsprd or log
information is made easier with the perl script
'$LAPSINSTALLROOT/etc/window_domain_rt.pl' ("window" hereafter). The window
script makes use of namelist domain templates that specifically define a
user's localizations. The window script uses environment variables
$LAPS_SRC_ROOT, $LAPSINSTALLROOT, and $LAPS_DATA_ROOT, however, -s, -i,
and -d command-line inputs override those environment variables as
necessary depending on user needs. The -t command-line input specifies the
domain (input) template directory and the script saves the log/lapsprd
history if command line switch '-c' is not used; or, completely removes
$LAPS_DATA_ROOT, then does a 'mkdir $LAPS_DATAROOT' if '-c' is supplied.
The '-w laps' is always required. The window script can be run manually
when configuring or reconfiguring localizations. Window copies the domain
template namelists (partial nest7grid.parms or *.nl's) into a new "static"
subdirectory which, in turn, are merged with the full namelists by script
localize_domain.pl. Recall that $LAPSINSTALLROOT contains bin/ and etc/
while $LAPS_SRC_ROOT contains the untarred full namelists from the
repository. In summary the template parameters are an input of the
localization while the output appears in $LAPS_DATA_ROOT.
In the event that $LAPS_SRC_ROOT does not exist, a data/ subdirectory
containing static/ and cdl/ must be available for use by
'localize_domain.pl' (i.e. $LAPS_SRC_ROOT = $LAPSINSTALLROOT). Even
though it is possible to have $LAPS_SRC_ROOT/data = $LAPSINSTALLROOT/data
= $LAPS_DATA_ROOT, this is not recommended since it does not allow multiple
localizations. Templates will ensure that specific namelist modifications
are merged with the untarred full namelists. Templates also ensure that
specifics to a localization are merged into new software ports and new
namelist variables (available with new software) are merged into existing
localizations.
Examples:
<pre>
setenv LAPS_SRC_ROOT /usr/nfs/common/lapb/operational/laps
setenv LAPSINSTALLROOT /usr/nfs/lapb/operational/laps
setenv LAPS_DATA_ROOT "any new or existing LAPS_DATA_ROOT"
cd $LAPSINSTALLROOT/etc
a) perl window_domain_rt.pl -s $LAPS_SRC_ROOT -i $LAPSINSTALLROOT
-d $LAPS_DATA_ROOT -t "full path to template directory" -w laps -c
result: all required information is provided on the command line.
Window will use the command line info instead of getting
The paths from the environment.
b) perl window_domain_rt.pl -w laps:
result: lapsprd and log saved; operational namelists and cdl's are copied
into $LAPS_DATA_ROOT/static; $LAPSINSTALLROOT/bin/gridgen_model.exe
runs to regenerate static.nest7grid. "Saved" lapsprd and log are
restored into $LAPS_DATA_ROOT.
c) perl window_domain_rt.pl -c -w laps:
result: same as b) although lapsprd and log are removed and regenerated by
"etc/makedatadirs.pm"
d) perl window_domain_rt.pl -t "full path to template directory" -w laps
result: similar to b) but namelist specifics are copied to $LAPS_DATA_ROOT/
static and merged with full namelists in $LAPSINSTALLROOT.
setenv LAPS_SRC_ROOT /awips/laps
setenv LAPSINSTALLROOT /data/fxa/laps_data
setenv LAPS_DATA_ROOT /data/fxa/laps
e) perl window_domain_rt.pl -t /data/fxa/laps_template -s /awips/laps \
-i /awips/laps -c -w laps
result: Specific AWIPS relocalization command for lapstools GUI when run
within the AWIPS workstation. GUI writes user input to
laps_template/ (subset namelists; e.g., nest7grid.parms);
$LAPS_DATA_ROOT/static and cdl/ are moved to laps_data/;
$LAPS_DATA_ROOT is removed; new $LAPS_DATA_ROOT is generated and
subdirectory structure created by "etc/makedatadirs.pm";
laps_template namelist are copied to new $LAPS_DATA_ROOT;
localize_domain.pl merges $LAPSINSTALLROOT/ and regenerates
static.nest7grid.
</pre>
If you decide to manually change any parameters in $LAPS_DATA_ROOT/static
after running the localization, it is suggested to make the same change
in the $TEMPLATE directory as well. This will help preserve your local
changes in the future if you install an updated version of LAPS.
--- 2.2.6.2 Localization Method 2 ---
This method is included partly for historical reasons and can be useful
if you haven't yet learned how to use template directories and/or the
separated $LAPS_DATA_ROOT (see method 1). This procedure provides a result
equivalent to that from Localization Method #1 and provides an alternative
method (even if not recommended) of modifying the parameters.
For each domain you wish to create, run...
prompt> cd $LAPSINSTALLROOT/etc
prompt> perl makedatadirs.pl --srcroot=$LAPS_SRC_ROOT --installroot=$LAPSINSTALLROOT --dataroot=$LAPS_DATA_ROOT --system_type=laps
where the path name $LAPS_DATA_ROOT must be named differently for each
data domain if there is more than one. Recall that each domain can be set
up in a separate subdirectory under '/data_disk/laps'. Next, follow the
setup and localization steps below.
The order of the command line arguments is important, but only the
first one is required. If for example a $LAPS_DATA_ROOT is not supplied,
the dataroot tree location will default to where the LAPS binaries are
installed via configure. Thus, the default value of $LAPS_DATA_ROOT is
'$LAPSINSTALLROOT/data'.
The runtime parameters should be emplaced and/or modified within
each $LAPS_DATA_ROOT directory tree prior to running the localization.
More details on 'nest7grid.parms' and other parameter files are
discussed in subsequent parts of Section 2.
As one option you can edit the parameter files that are in
'$LAPS_SRC_ROOT/data/static' and tailor them for your domain. If you have
'$LAPS_DATA_ROOT' different from '$LAPS_SRC_ROOT/data/static', then a
good alternative may be to copy any parameter files you need to edit into
'$LAPS_DATA_ROOT/static' from '$LAPS_SRC_ROOT/data/static'.
Finally, you can create the static data files and look up tables specific
to the domain(s) you have defined in 'data/static/nest7grid.parms' and
other runtime parameter files. Shown below is an example of running the
localization for a particular laps domain. This should be repeated
(with a unique dataroot) for each domain if there is more than one.
prompt> cd $LAPSINSTALLROOT/etc
prompt> perl localize_domain.pl --srcroot=$LAPS_SRC_ROOT --install_root=$LAPSINSTALLROOT --dataroot=$LAPS_DATA_ROOT --which_type = 'laps'
--- 2.2.6.3 Localization with LAPS GUI ---
LAPS has a GUI interface under development that can be used to localize
the domain. This can be found in the '$LAPS_SRC_ROOT/gui' directory.
There it can installed using the 'install_gui.pl' PERL script as outlined
in the local README file.
--- 2.2.7 WRF Domain Wizard LAPS Support ---
The WRF Domain Wizard can be used to help specify correctly navigated
LAPS domain map parameters. When the Wizard is run it will write out
a 'nest7grid.parms' file for each nest that can be used as input
"templates" for LAPS localization.
http://www.wrfportal.org/DomainWizard.html
--- 2.2.8 MPI support for LAPS wind analysis ---
There is capability to compile and run the wind analysis (wind.exe) using
MPI. We do this by doing a separate software build with 'mpif90' and then
'sched.pl' runs the serial versions of most things while running the
parallelized version of the wind analysis. To build LAPS using 'mpif90'
edit the 'makefile.inc' file, between running 'configure' and 'make',
adding -DUSEMPI into the CPPFLAGS.
The 'sched' script presently submits multiple processor jobs using the
SGE queueing environment. We may consider adding an option to 'sched.pl' to
run/submit directly with 'mpirun' if that would be useful.
--- 2.3 Raw data ingest ---
There is a layer of "raw" data ingest code that may have to be modified
for the individual location depending on data formats. Its purpose is
to reformat and preprocess the various types of raw data into simple common
formats used by the subsequent analyses. It also helps to modularize the
software.
Working with the ingest code is usually the largest task within the porting
of LAPS. The supported component of the LAPS code is the analysis section.
Ingest code is supported only if your raw data looks has the same
configuration and format as GSD's raw data. It is the reponsibility of the
LAPS user to modify the LAPS ingest code if necessary to generate the
intermediate data files that are inputs to the analysis code.
A flow chart for the ingest processes may be found at this URL:
http://laps.noaa.gov/doc/slide1_v3.gif
The default LAPS ingest code obtains "raw" data, generally from the GSD
NIMBUS system. The raw data can either be in ASCII, netCDF (as point
data), or netCDF (as gridded data - generally not on the LAPS grid). Note
that the ingest code is also generally compatable with raw SBN/NOAAPORT
data as stored in netCDF files on the WFO-Advanced system. The ingest
code processes the raw data and outputs the LAPS "intermediate" data
files. The intermediate files are generally in ASCII for point data and
netCDF format for gridded data that have now been remapped onto the LAPS
grid. Most ingest code is located under the 'src/ingest directory'. When
netCDf format is used for the raw data, a cdl file for the raw data is
sometimes included in the source code directory.
Depending on the data source, you may generally prefer one of three
choices:
1) Convert your raw data to appropriate netCDF formats then run the
LAPS ingest code as is. The CDLs and sample raw/NIMBUS netCDF files
supplied with our test dataset can serve as a guide to writing the
software to do this. If the CDL is unavailable, doing 'ncdump -h'
on the actual data file will yield equivalent information. We
generally do not maintain or support any software for writing
"raw" netCDF files as this is done external to LAPS. Sometimes by
posting a message to 'laps-users' you can obtain information from
other LAPS users as to how they may have implemented this step.
2) Run a process independent of the LAPS ingest code that creates
the intermedate data file.
3) Modify LAPS ingest code to accept your own raw data format.
This often entails writing a subroutine that reads the data and
linking this routine into the existing ingest process. That
process then writes out the LAPS intermediate file. Note that
generating an GSD style raw data file is not here needed - all
that really counts is producing an intermediate data file.
Recommended only for advanced users or those who believe their
modifications have enough general interest for inclusion in the
baseline LAPS repository.
For the model background and in-situ observations generally (1) is the
best option. For gridded data (satellite or radar) options (1) or (2)
usually work best. Most external users should avoid option (3) unless it
is done in close consultation with LAPS staff. A key consideration
is how easy it will be to update your version of LAPS and have it work
with your local data.
You may note the following data sources used at GSD. These data sources
are what the GSD ingest code is tailored to for producing intermediate
data files. Note that LAPS will still run even if some of the data sources
are withheld, albeit in degraded fashion. A minimum dataset of model
background and surface observations is generally needed to get reasonable
results.
The pathnames for the ingest data sources are assigned within the
'./data/static/nest7grid.parms' and other '*.nl' files and can be set
accordingly at runtime. Doing a grep for 'path' in these files will
give you a quick listing of the relevant parameters.
Unless otherwise specified, the time window for data in the
intermediate data files should be '+/- laps_cycle_time'.
The time window for data in the raw data files is more variable and is
generally specified within the raw data (e.g. in the CDL).
Further information on specific LAPS ingest processes for the various
data sources is found in Section 3 of this README.
---- 2.3.1 Model Background (lga/lgb) ---
The model first guess (background) is generally on a larger-scale
grid than LAPS and is run independently. The model data is
interpolated to the LAPS grid by the LAPS ingest to produce
'lga/lgb' files. The interpolation is done in time, in space, and
can be from one map projection to another. This 'lga/lgb' output is
distinct from the 'fua/fsf' files that are first guess files of similar
format generated by the LAPS forecast model using an intermittent 4dda
mode.
The 'nest7grid.parms' namelist variable "fdda_model_source" controls
the background used in the analysis, including lga. A list of
"fdda" backgrounds that are available with this release are
specified in file etc/laps_tools.pm - module mkdatadirs. Even
though fdda subdirectories are populated with current backgrounds,
the analysis can be forced to override this by making the first
entry of "fdda_model_source" = 'lga'.
The acceptable models and formats for the background model are listed in
'data/static/background.nl'. Many models can be accepted in netCDF format.
A new capability in LAPS is to process GRIB input without first converting
to netCDF format. For Grib data to be decoded an associated Vtable.XXX
needs to be found in directory 'data/static/Variable_Tables'. The Vtable
can be configured for either GRIB-1 or GRIB-2. However we are unable to
guarentee that any model specified in 'background.nl' will work without
some software modification.
Rapid Refresh (RR) grids are ftp'ed from NCEP to GSD, then converted at
GSD from GRIB to netCDF. This netCDF file is the input for the LAPS ingest
process that writes "lga". For more information on RR check the
following URL for more info: 'http://rapidrefresh.noaa.gov/'
Note that we often read these into LAPS as RUC (Rapid Update Cycle) look
alike files.
RR is also available from UNIDATA and distributed to universities
through private companies like Alden.
The conversion from GRIB to netCDF is done outside of LAPS by GSD's
Information and Technology Services (ITS) group (in the NIMBUS system).
Having the CDL should mostly be sufficient along with general knowledge
of netCDF for writing out the data. Beyond that, you may wish to contact
the ITS group for more info (see the reference to them in section 3.2.1).
The Atlanta, Sterling, and Seattle WFOs have followed a more direct route,
going from the RR/Eta to the intermediate "lga" file, bypassing the
netCDF file on the model grid. This includes RR on isobaric surfaces.
---- 2.3.1.1 Acquiring Model Background Data ---
GRIB-formatted background model files are now supported and can be
directly read into lga.
--- Where Can Users Find GRIB Data? ---
At the NCEP ftp server for real time data sets located at
ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/. These products can be
downloaded from the web or via anonymous ftp. The following is a discussion
for locating and acquiring NAM, GFS, and RUC model backgrounds for use with
lga. The models are available in grib1 and grib2 formats as indicated.
NAM Model:
NAM 221 High Resolution North American grid, 32-km can be
found at ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/nam/prod/ with the
directory and filenames as follows nam.{YYYYMMDD}/nam.t{CC}.awip32{FF}.tm00{.grib2}
where YYYYMMDD is the current date, CC is the model cycle
time (00, 06, 12, or 18) and FF is the forecast hour (00-84).
awip32 indicates the 32 km North America (NCEP grid 221).
GFS Models:
GFS global longitude-latitude grid (360x181) 1.0 deg (fh 00-180) can be
found at ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/gfs/prod/
gfs.{YYYYMMDDHH}{CC}/gfs.t{CC}z.pgrbf{XXX}{.grib2}, and
GFS global longitude-latitude grid (720x361) 0.5 deg (fh 00-180) can be
found at ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/gfs/prod/
gfs.{YYYYMMDDHH}{CC}/gfs.t{CC}z.pgrb2f{XXX}
where CC is the model cycle time (i.e. 00, 06, 12, 18) and XXX is the
forecast hour of product from 00 - 180.
The 1.0 degree GFS uses file identifier 'pgrb' (pressure-based grib) and
is now available in grib2 as well when '.grib2' is present. The 0.5 degree
GFS uses 'pgrb2' (pressure-based) and is only available in grib2.
RUC Model:
RUC Rapid Update Cycle 40km and 20km pressure data sets can be found at
ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/ruc/prod/
ruc2a.{YYYYMMDDHH}/ruc2.t{CC}z.pgrb{XXX}{.grib2}
where CC is the model cycle time (i.e. 00, 06, 12, 18) and XXX is the
forecast hour of product from 00 - 12 (or more). File identifier 'pgrb'
is used for the 40km resolution and 'pgrb20' is used for the 20km.
Additional description of NCEP products can be found at
http://www.nco.ncep.noaa.gov/pmb/products/. A master list of NCEP GRIDS ID
numbers (e.g. 211) and other specifications can be found at
http://www.nco.ncep.noaa.gov/pmb/docs/on388/tableb.html
--- How Do Users Name The GRIB Data Files? ---
For LAPS ingest at NOAA/ESRL, we have a process that automatically
downloads GRIB files to a designated directory. For example,
'/data/grid/gfs/global_0p5deg/', '/data/grid/gfs/global_1p0deg/', and
'/data/grid/gfs/conus211/' are three directories for the GFS global
0.5 degree, global 1.0 degree and CONUS 211 domains. The files within
these directories are renamed from the complex patterns listed above to
filenames with the following pattern: 'YYJJJHHMMhhhh'. Here the 'hhhh'
part represents the number of hours into the forecast. Thus a file for GFS
CONUS 211 initialized on Jul 23 2008 at 1200 UTC, with a 6 hour forecast
would be named '/data/grid/gfs/conus211/grib/0820512000006'. The HRRR
model follows a slightly different convention of 'YYJJJHHMMhhmm', so
that forecasts of under one hour can be represented.
--- How Does lga.exe Know Where To Find The Data? ---
For lga.exe, the acceptable models, directory paths and file formats are
identified in 'data/static/background.nl'. In the example above if we wanted
to use the US-scaled data, we would set bgpath='/data/grid/gfs/conus211/grib/',
bgmodel=13 (for GRIB), and cmodel='GFS'.
---- 2.3.2 Radar ingest ---
The following are intermediate files for various forms of radar
data. These may have already been pre-processed (remapped) from "raw"
data, and at this stage are in Cartesian format on the LAPS grid.
<pre>
(vrc) - Low-level reflectivity from single or
multiple radars. For example, our ingest
at GSD processes WSI-NOWRAD, stored at GSD
in netCDF, to create the 'vrc' intermediate
file. Narrowband single-tilt data from AWIPS
is also stored in 'vrc' files.
(v01, v02, ... ) - WSR-88Ds or other radar data are stored as
a full volume.
- Each 'vxx' file has 3-d reflectivity,
velocity, and nyquist velocity for one
radar. Horizontal and vertical gaps are
filled in for reflectivity while sparse
arrays are used for velocity.
(vrz) - 3-D reflectivity mosaic from multiple radars
('vxx' files).
(ln3) - Layer reflectivity and echo tops, from a
single radar or mosaiced from multiple
radars. For example, WSI sends put a variety
of derived products from the WSR-88D's
which we call nexrad products. These
include 3 layer reflectivity products,
a composite reflectivity, echo tops, and
vil. FD also decodes these and writes
netCDF files. We have an experimental
process called 'ln3' that ingests these
data that we haven't yet approved for
release in our tar. This will probably
happen shortly though as the need
to further test/use this 3D reflectivity in
LAPS is increasing.
As of 9-22-98 we have committed the ln3 ingest
process to our repository and distribute this
source code with our tarfile. The reliability
of echo tops in conjunction with the layer
reflecitivity information still makes this a
problematical data set to use in the analyses.
Committing this source to the repository will
help to further investigate the utility of this
product.
The key fields from 'ln3' which are used
in the analyses are the Layer reflectivity
(0-4 km MSL, 4-8 km MSL, and >8 km MSL), as
well as echo tops (MSL).
</pre>
A description and flow chart showing polar radar data usage in LAPS is on
the Web at: 'http://laps.noaa.gov/albers/remapper_raw.html',
with some additional text details for various types of radar data in:
'http://laps.noaa.gov/albers/radar_decision_tree.txt'.
These include information on which types of radar data are processed
via the various intermediate data files.
Further information on using individual radar ingest processes is in
Section 3. Specifically we should establish whether your raw data is in
polar or Cartesian form. If polar, please take a look at "Polar Radar
Data" in section 3.2.3. NOWRAD / WSI (Cartesian) data is covered separately
within Section 3.2.4.
---- 2.3.3 Surface Data ---
Sfc Obs (lso): GSD uses surface observations as input with the default
being GSD's NIMBUS netCDF format. These are generally used when
running LAPS within ESRL/GSD using data from /public, and is only
available within GSD unless one is working with the supplied test
data case.
Surface observations of various types, covering much of the
world are available in realtime from GSD's MADIS system (with some
restrictions). This data, generally in 'WFO/AWIPS' netCDF format,
are distributed via the MADIS server at http://madis.noaa.gov.
Thus MADIS is available both inside and outside ESRL/GSD. The
MADIS netCDF has additional variables (such as QC flags) that go
beyond what is in the NIMBUS format.
The supported MADIS surface observation datasets include 'metar'
(METAR/SYNOP), 'maritime' (Buoy/Ship), 'mesonet', 'urbanet', and
others. This is an excellent source of surface observations for
most users outside of ESRL/GSD to start with. To request a real-time
data stream please go to the MADIS data application page at this
link: http://madis.noaa.gov/data_application.html
A few other METAR/SYNOP formats are now being supported in LAPS
software as listed in the 'static/obs_driver.nl' namelist. The GSD
code is in the '.../src/ingest/sao' directory, and includes routines
to read and reformat various surface data types (METAR/SYNOP, mesonet
or local/LDAD, buoy/ship or maritime, and GPS/profiler surface obs).
There is a subroutine tree outline in the 'src/ingest/sao/README'
file including information on the supported data formats for each
observation type. Paths to the datasets are specified in the
'obs_driver.nl' file.
In most cases users should be able to convert their surface
observations into the 'NIMBUS' or 'MADIS' NetCDF formats. Note that
the parameters and variable names in each NetCDF dataset or directory
will vary.
Only observations reasonably near the standard shelter height
(2 meters, except 10 meters for wind) should be included in the LSO
file. Tower mounted instruments should instead be placed in the SND
file using "TOWER" for the observation type.
---- 2.3.4 Wind Profiler / RASS ---
Profilers/RASS (pro/lrs) - The raw data are obtained from GSD's NIMBUS
database and/or AWIPS in netCDF format where they are stored in
four different directories. The data originally come from GSD's
Demonstration Branch (DB) from two main networks. The 30+ NPN
(National Profiler Network - NOAAnet) profiler network is located
mostly in the central U.S.
The second network supplies boundary layer profilers for both wind
and temperature, with formats including NIMBUS, MADIS Multi-Agency
Profilers (LDAD), and RSA (LDAD) format.
The profiler data for wind goes into the 'pro' intermediate file,
and RASS temperature profiles go into the 'lrs' intermediate file.
Note that the cdl's associated with each data source indicate the
time frequency of the data that our ingest code can process. The
path names for the profiler data are all set in 'nest7grid.parms'.
To summarize...
<pre>
Network Database(s) file frequency cdl(s)
------- ----------- ---- --------- -----------
NPN wind NIMBUS/AWIPS/MADIS/ PRO 404 MHz wpdn60.cdl
wpdn06.cdl
NPN RASS NIMBUS LRS rass60.cdl
rass06.cdl
BLP wind NIMBUS PRO 915 MHz wpdn60.cdl
wpdn30.cdl*
BLP RASS NIMBUS LRS rass60.cdl
rass30.cdl*
BLP wind MADIS-MAP/RSA PRO 915 MHz
50 MHz
SODAR
BLP RASS MADIS-MAP/RSA LRS 915 MHz
50 MHz
* Indicates that the data with this cdl is available, but our
ingest code would need modification to process it.
</pre>
The NPN wind profiler data is available via another route from
GSD with some restrictions. This data, in 'WFO/AWIPS' netCDF format,
is distributed via GSD's MADIS project at
http://madis.noaa.gov
---- 2.3.5 PIREPS & ACARS from aircraft ---
PIREPS (pin) - We are ingesting GSD NIMBUS and WFO/AWIPS (netCDF) pirep
files to translate the cloud layers from voice pilot reports into
intermediate "PIN" files.
ACARS (pin) - We are ingesting GSD NIMBUS, WFO/AWIPS (netCDF) and AFWA
databases for ACARS data to translate the automated aircraft
observations. The wind, temperature and humidity obs are appended
to our intermediate "PIN" file. A NIMBUS equivalent netCDF database
is available (with some restrictions) on the Web via MADIS at
http://madis.noaa.gov
Note the TAMDAR is presently being screened out from the NIMBUS
database while this data source is being validated.
---- 2.3.6 RAOB / Dropsonde / Radiometer ---
RAOBs (snd):
GSD NIMBUS, WFO/AWIPS, CWB, or AFWA databases. These are available
in real-time from GSD with some restrictions. RAOB data in
'WFO/AWIPS' netCDF format is distributed via GSD's MADIS project at
http://madis.noaa.gov
Dropsondes (snd):
A Dropsonde ingest module has been developed for the CWB database.
An ingest module has also been developed for AVAPS. We now allow the
SND format to be used as input (so far just for the "AIRDROP"
project). For the SND input option, the ingest program simply does a
time windowing of the raw data. We may include modules for other
(e.g. netCDF) databases in the future, such as NIMBUS or WFO/AWIPS.
Radiometers (snd):
A radiometer ingest module has been developed for the MADIS
database - http://madis.noaa.gov
---- 2.3.7 Satellite ---
Satellite Image Ingest (lvd): GOES data ingest. Data is acquired at
GSD's ground station and stored in netCDF. We also obtain
AWIPS/NOAAPORT/SBN data (stored in netCDF). Ingest of Air Force
Weather Agency (AFWA) satellite data is also possible.
Raw GVAR satellite data can be ingested and navigated using GIMLOC
routines. These files are in NetCDF. Further details can be found
in the file 'src/ingest/satellite/lvd/README'.
The ITS group at ESRL/GSD has put together a converter from McIDAS
AREA files to the GVAR netCDF format (lvd input). These files are
similar to the "raw" GVAR, except they have lat/lon arrays added
to make the files self navigating. The AREA files can be obtained
from sources such as the NESDIS ADDE server. The Java based
converter package can be found online at this URL:
http://laps.noaa.gov/software/sat/McArea2NetCDF.tar
Note that the NESDIS ADDE server can also supply worldwide
geosynchronous satellite data.
Some tweaking of satellite coordinate and image dimensions may be
needed when setting up the McIDAS package, as can be seen in the
sample illustration (link below). Programs like 'ncview' can be
helpful to check if the window is navigated properly in the GVAR
netCDF files prior to running the LAPS satellite ingest.
http://goes.gsfc.nasa.gov/pub/goes/GVAR.135W.gif
Other work has been done in Italy to ingest Meteosat Second
Generation data into LAPS, for example at ISAC.
http://satmet.isac.cnr.it/about.html
Another option under development is to use flat files (ascii files
generated by RAMSDIS or binary data) as input. The flat file ingest
was still under development as of 3-11-98.
Generally it is best to convert your data into either GVAR NetCDF
or remap it to create the intermediate LVD files.
Satellite Sounder Ingest (lsr): GOES satellite sounder data ingest.
Program lsr_driver.exe processes data from both satellites.
Product files are yyjjjhhmm.lsr and stored in subdirectories
lapsprd/lsr/'satid'. Nineteen channels. Output is Radiance. The
namelist 'data/static/sat_sounder.nl' defines the appropriate
parameters for this ingest process. Only the moisture analysis is
using this product. Currently GSD /public sounder files in netCDF
format are processed. This data is useful only when GOES Vapor
(GVAP) is unavailable.
Satellite derived soundings (snd):
We have interfaces to GOES binary and MADIS POES (Polar Orbiter)
formats. AFWA database format was previously used at GSD though not
currently. The output represents derived profiles of temperature and
moisture. For other formats you may wish to supply your own routine
to convert your raw data into the 'snd' format.
Cloud Drift Winds (cdw):
We are ingesting the ASCII satellite cloud-drift wind files
for use in the wind analysis. These come from NESDIS (via NIMBUS)
as well as from CWB and AFWA. We can also utilize netCDF files from
MADIS. Both NESDIS and MADIS files are included in our sample data
set.
---- 2.3.8 GPS ---
GPS:
LAPS uses GPS data from NIMBUS netCDF files. The precipitable water
is used in the humidity analysis. STMAS is being designed to use
the signal delay directly instead of the PW. The netCDF files
are available online at ftp://gpsftp.fsl.noaa.gov where they
are named according to 'GPSIPW_CDF_YYDDDHHMM0030o.nc'. The leading
'GPSIPW_CDF_' of the names would have to be stripped off to be
used in LAPS/STMAS.
There are plans to make files similar to the NIMBUS ones available
in AWIPS-II, though again filenaming conventions may need to be
addressed.
MADIS (LDAD) mesonet files also carry the GPS PW and related data,
including surface obs. However there may be some questions about
the latency of this data feed for GPS. Based on tests conducted in
2011, with a LAPS cycle that begins at about 20min past the top of
the hour, one can generally expect only 5-10% of the GPS data to
be avaliable via MADIS when the code is configured to seek the data.
---- 2.3.9 Other Data Sources ---
Radar VAD Algorithm winds (pro)
GSD NIMBUS netCDF database, from WSR-88D algorithm output. GSD
obtains this from NCEP and does not presently redistribute it.
SODAR data (pro) - This is treated in a similar manner to wind profilers
and can be processed by LAPS ingest to appear in the PRO file.
This is available as part of the RSA project at Kennedy and
Vandenberg Space Centers and comes into netCDF format via
AWIPS/LDAD.
Met Tower data (snd) - This is treated in a similar manner to RAOBs
and can be processed by LAPS ingest to appear in the SND file.
This is available as part of the RSA project at Kennedy and
Vandenberg Space Centers and comes into netCDF format via
AWIPS/LDAD.
Radiometric Profiler (snd):
We have an interface to radiometric profilers (in netCDF via NIMBUS)
that can be used for the temperature and humidity analyses.
Lightning Data:
Although the LAPS repository doesn't yet have any lightning data
ingest it is being considered to do this in terms of a simulated
2-D reflectivity that is one of the components of the VRC
intermediate file.
--- 2.4 Running LAPS Analyses ---
LAPS runs in real-time under cron; there is a sample cron script in
'$LAPSINSTALLROOT/util/cronfile'. Referring to this cron, you can
see that once each hour (or other cycle time), the main './etc/sched.pl'
runs. As an example at ESRL, we run the 'sched.pl' hourly at :20 after the
top of the hour. By inspecting the 'sched.pl' file you can see the various
executables that are run in a certain order. Various command line arguments
are documented within 'sched.pl' (such as '-d 0.25' that is useful for a
15-min cycle when the latency is ~20 minutes). You might want to modify
the 'sched.pl' file for your needs.
In the sample cron script several ingest processes are run separately
from the 'sched.pl'. For example the satellite ingest (lvd) is run
several times per hour and utilizes './etc/laps_driver.pl'. NOWRAD
Radar ingest (vrc) is also run at more frequent intervals. You might also
choose to run 'remap_polar_netcdf.exe' for radar ingest in this manner.
On many unix systems jobs that run in cron do not have access to the
environment defined by the user. They instead use a system default
environment defined in '/etc/profile'; thus 'perl' may not be in the $PATH..
The cron file uses the full path to 'perl' to ensure that this will
not be a problem. If the path to 'ncgen' is not in '/etc/profile', then
you may want to add this to your own '.profile' file.
Each script in the cron requires the path to laps as a command line
argument. A second optional argument specifies the path to the laps data
directory structure; this path defaults to '/fullpathto/laps/data' if not
provided.
The 'util/cronfile' is created by the configure step. Much of the needed
editing has already been done in the creation of this file. You might see
some remaining '@....@' constructs though that can be edited either
manually or by running the 'cronfile.pl' (next paragraph). The
@laps_data_root@ can be replaced with your path to $LAPS_DATA_ROOT and the
optional @followup@ can be replaced with anything you wish to run after
the 'sched.pl' has completed (using a semicolon to separate the two
commands).
There is also a script called 'etc/cronfile.pl' that creates a modified
version of 'util/cronfile' tailored to a given domain. This script can be
run manually and the output location of the cronfile is located in
'$LAPS_DATA_ROOT/cronfile'.
---- 2.4.1 Cron timing considerations ---
The frequency of the cron entries for running 'sched.pl' is defined to be
the LAPS cycle time. This should correspond to the value of the
'laps_cycle_time' parameter within the 'nest7grid.parms' file.
The best timing of the cron is often related to the arrival time
of the raw surface observations. For example, if most of the surface
data arrives within 20 minutes of the observation time, then running
the cron 20 minutes after the 'systime' would be optimum. The time window
for acceptance of surface stations in the LSO file can be controlled
by runtime parameters in 'obs_driver.nl'.
In most cases, the data cutoff time window for 3D observations is
+/-laps_cycle_time/2 or +/-laps_cycle_time. For example an hourly
LAPS cycle accepts RAOB data from a +/-60 minute time window and
ACARS from a +/-30 minute window.
---- 2.4.2 Purging Output Files ---
The script '/etc/purger.pl' purges the 'lapsprd' output files and
is in turn run by the 'sched.pl'. There are default settings in place for
the number of files and age of files to be kept. These can be overridden
in three ways.
1) The 'sched.pl' command line options '-r -m N', where "N" is the
(default) maximum number of files to be kept in each product directory
by the purger
2) Overrides can be read in from '/data/static/purger.dat'. This file
can be modified by the user to optimize the purging in various
domains. One can review the 'purger.pl' script to see how the
'purger.dat' information is used.
---- 2.4.3 STMAS and other configurations ---
Within the LAPS cron the call to 'sched.pl' can have some optional command
line arguments that adjust the runtime options. The default is to run
both surface and 3-D analyses from the "traditional" version of LAPS.
Here are examples of some other alternatives:
1) STMAS-2D surface analysis only
prompt> perl sched.pl -M stmas_mg.x [other regular options]
Note that when running STMAS-2D analyses, the 'lgb_only' parameter
in the 'background.nl' namelist can be set to .true. for improved
runtime efficiency.
2) "traditional" LAPS surface analysis only
prompt> perl sched.pl -M laps_sfc.x [other regular options]
3) STMAS-3D
prompt> perl sched.pl -V STMAS3D [other regular options]
--- 2.5 Test data case ---
Tar files containing test data (called 'lapsdata*') are available that
contain a snapshot of several hour's worth of laps data from the Colorado
domain using namelist settings taken from the repository. The tar files
include intermediate files from the 'ingest' code plus outputs from
the 'analysis' code. Several consecutive analysis cycles are posted with
one file per cycle. Included are the contents of the 'lapsprd', 'time',
'static', and 'log' subdirectories under 'data' or $LAPS_DATA_ROOT. The
log files are useful for diagnosing any differences in output you may
observe. The contents of the various directories are outlined elsewhere
in this README file. The data was created using the latest software
release. Our users can download this data at this URL:
'http://laps.noaa.gov/cgi/LAPS_SOFTWARE.cgi'.
It is suggested here to test the localization procedure to ensure that all
the static files needed to run LAPS are present. To do this, check that
the paths to the geography data are correct in '$TEMPLATE/nest7grid.parms'
and/or '$LAPS_DATA_ROOT/static/nest7grid.parms'.
When running LAPS as a whole for the archived data, the 'etc/sched.pl'
script will accept a '-A' command line argument. This forces the script
to run for the time you are inputting instead of the current time.
An example call is shown as follows...
prompt> perl sched.pl -A dd-mmm-yyyy-hhmm $LAPSINSTALLROOT $LAPS_DATA_ROOT
...where the inputted 'dd-mmm-yyyy-hhmm' value is the date (for example
28-Aug-2007-1500). This date can be inferred from the contents of
'$LAPS_DATA_ROOT/time/systime.dat'. Best results are obtained when using
a time just prior to the latest raw data tarfile time.
One can also initiate individual executables (bin directory) listed in the
'sched.pl' to run on the test data. This often helps in getting a better
match between your output and ours. Note that $LAPS_DATA_ROOT needs to
be set as an environment variable when executables are run individually.
The time of the run is specified in '$LAPS_DATA_ROOT/time/systime.dat'.
This can be modified if needed if you want to try a slightly different
time from the one supplied. To do this, interactively run the script
'$LAPSINSTALLROOT/etc/systime.pl' and write the standard output to
'$LAPS_DATA_ROOT/time/systime.dat'.
Note that for any given process or set of processes, deviations from the
GSD output may be caused by differences in the inputs as well as machine
roundoff error. Most, but perhaps not all of the input data is supplied.
One main area to check would be differences in available "raw" background
data files. Having all of the data history from 'lapsprd' may also be an
issue; this may be less of a problem if you run laps for the latest hour
of data that is supplied. The history is then supplied from earlier
'lapsdata*/lapsprd' output. Output differences can be tracked down by
recompiling specific analyses with the '-g' option. This can be done by
typing 'make debug' in the appropriate 'src' directories. Various
debuggers can then be used such as 'dbx'. Examination of the log files
again is helpful.
We have a new script (in 2004) called 'casererun.pl' that can be used
for archive data runs. We have yet to try it on the supplied test data
case though it could prove to be useful.
---- 2.5.1 Analysis Only Test ---
You may want to check that any analysis outputs from this time are not
present, leaving only the 'ingest' outputs in place. This may improve
the results of comparisons of your own output with GSD analysis output,
though this step is not always necessary. You might consider adding the
'-T' command line option when you run 'sched.pl' so that we run the
analysis executables only thus skipping the ingest processes. This can
be done if the ingest outputs (i.e. analysis inputs) are already present
in the various 'lapsprd' subdirectories.
One way to supply the analysis inputs is as follows for each input (taken
from a list of ingest outputs, see section 3.2):
prompt> cp testdata/lapsprd/[inputlist]/*
$LAPS_DATA_ROOT/lapsprd/[inputlist]
OR
prompt> cd $LAPS_DATA_ROOT/lapsprd
prompt> ln -s testdata/lapsprd/[inputlist] .
---- 2.5.2 Ingest + Analysis Test ---
For this type of test, you will want to download the 'rawdata*' tar files
into your 'raw_data' directory to start the processing of LAPS. Recall
that the 'raw_data' directory is on a separate tree than $LAPS_DATA_ROOT.
Raw data formats and filename conventions are consistent with the default
namelist settings taken from the repository. This is generally in NIMBUS
(self describing netCDF) format with associated file naming conventions.
A typical filename on NIMBUS looks like this: '0606701000100o' meaning
'yydddhhmmHHMM' where 'ddd' is the day of year, 'hhmm' is the time of day
and 'HHMM' is the file recurrence interval. The 'o' at the end means that
observations are binned into files according to observation time (instead
of 'r' for receipt time) More about NIMBUS is detailed in publications on
the web at this URL:
http://www.fsl.noaa.gov/its/papers/jb_ams94.html
Note that with the RUC grib data there are two directories. The one with
soft links (and without the ".grib" at the end of the filenames) is the
one to use.
Time information will be needed in the form of 'data/time/systime.dat';
this can be extracted from the 'lapsdata*' tar file.
The 'raw_data' directory is a convenient place to store test data. User
supplied raw data for operational runs can be stored anywhere on your
system, often outside of the LAPS trees.
Note that the 'lapsdata*' tar files contain intermediate plus analysis
output files only. The 'rawdata*' tar files supply much of the "raw" data
that are inputted to the ingest processes. The times for the raw data
match the 'lapsdata*' output approximately but not always exactly
(one example being the raw background data files). As a hint with the
background data check that the available raw files bracket the systime
of interest. If needed one can change the 'use_analysis' flag in
'background.nl' to get 'lga.exe' to work better.
In many cases, a user could independently generate the intermediate data
files (ingest output) and then compare them with ours. If other "raw" files
are needed as they appear on GSD's NIMBUS & MADIS systems, please let us
know and we can try to add them to our test data case or send them
separately.
---------------- 2.6 ------ I/O of LAPS gridded files -------------------------
Once the laps library is compiled (as outlined above), laps grids can
be read. There are three levels of software that can access the data.
<pre>
1. Lowest Level - netCDF c routine calls (not recommended unless you're
a netCDF hacker)
2. Medium Level - READ_LAPS_DATA - look at the source code in
lib/readlapsdata.f for the arguments.
3. Highest (and easiest) level - get_laps_3d or get_laps_2d. The source
is contained in src/lib/laps_io.f. The various grids available are
listed later in this README file under the heading "NetCDF organization"
</pre>
To link to the reading routines, you will want to link to:
laps/src/lib/liblaps.a
libnetcdf.a
-------------- 2.7 ---- CHANGING THE HORIZONTAL DOMAIN-----------------------
Laps will allow you to change the horizontal domain after compilation
and before the running of the localization scripts. Below is a list of
the relevant changes.
The dimensions and location of the horizontal domain can be changed at
run time. Prior to running 'window_domain_rt.pl', set the following
parameters in 'data/static/nest7grid.parms' or in the corresponding
template directory (needed only if you are outside the default Colorado
domain). This script in turn runs 'gridgen_model.exe' and other programs.
---- 2.7.1 Number of Grid Points ---
Adjust the horizontal dimensions in terms of the number of grid points
(NX_L, NY_L) in './data/static/nest7grid.parms'.
NOTE: Various files in the ./data/cdl directory are automatically edited
by ./etc/localize_domain.pl using the values found in
'./data/static/nest7grid.parms'.
---- 2.7.2 Location of Analysis Domain (Map projections) ---
<pre>
1) Modify the 'grid_spacing_m' parameter (only if you want to
change from the default 10000m for the grid spacing).
Grid spacing in meters on the projection plane. Used for all
projections. The grid spacing is valid at the standard
latitude and may vary elsewhere.
2) Modify the 'grid_cen_lat' and 'grid_cen_lon' parameters.
These are the latitude and longitude of the center of the
domain, expressed in degrees. The parameters are needed for
all projections.
3) c6_maproj:
Polar stereographic:
Set to 'plrstr'.
Lambert Conformal:
Set to 'lambrt'.
Mercator:
Set to 'merctr'.
Latitude / Longitude
Set to 'latlon'
In most cases, the Lambert projection is recommended.
Mercator is recommended if the domain includes the
equator, or for domains centered in the tropics where
sin(latitude) varies by more than a factor of two over the
domain.
If the domain includes one of the geographic poles, then
Polar Stereographic should be used instead.
Latitude / Longitude is recommended only for a global LAPS
run. This is because certain fields aren't yet generated with
this projection, such as divergence and balanced parameters.
See the note below regarding current map projection limitations.
4) standard_longitude:
Polar Stereographic:
This defines the longitude which is straight up and down
(parallel to the "y" axis) in the map projection.
Lambert Conformal:
This defines the longitude which is straight up and down
(parallel to the "y" axis) in the map projection.
Mercator:
N/A
5) standard_latitude:
Polar Stereographic:
This is the latitude at which the grid spacing is exactly
the nominal value ('grid_spacing_m' e.g. 10km).
This parameter is usually set to +/-90 degrees to match
the latitude of the projection pole ('standard_latitude2'),
given that the projection pole is at one of earth's
geographic poles. The actual grid spacing (measured on
the earth's surface) matches the 'grid_spacing_m'
parameter at the projection pole, which may or may not be
located within your domain. For domains distant from the
projection pole, the actual grid spacing inside the domain
becomes noticeably less. The value of 'grid_spacing_m'
can be increased to compensate. The projection plane is
tangent to the earth's surface.
When the projection pole is at a geographic pole,
'standard_latitude' can be set to values other than +/-90.
The 'grid_spacing_m' parameter then represents the
true grid spacing (measured on the earth's surface) at a
latitude of 'standard_latitude'. The projection plane
is secant to the earth's surface.
Consider the angle 'psi' which is the angular distance from
the pole of the projection. 'phi' = 90 - 'psi'. The map
factor 'sigma' is (1+sin(phi0))/(1+sin(phi)) and becomes
unity when 'phi' for a particular grid point is equal to
'phi0'. This occurs when you are located at the
'standard_latitude' for the case of a "secant" projection.
Note that the grid spacing for a particular location in
the domain is equal to 'grid_spacing_m'/'sigma'.
Example 1: grid_spacing_m = 10000.
standard_latitude = +90.
standard_latitude2 = +90.
grid_cen_lat = +40.
grid_spacing at projection (north) pole = 10km
grid_spacing at domain center (+40) ~ 8km
Example 2: grid_spacing_m = 10000.
standard_latitude = +40.
standard_latitude2 = +90.
grid_cen_lat = +40.
grid_spacing at projection (north) pole ~ 12km
grid_spacing at domain center (+40) = 10km
Example 3: grid_spacing_m = 10000.
standard_latitude = -90.
standard_latitude2 = -90.
grid_cen_lat = +40.
grid_spacing at projection (south) pole = 10km
grid_spacing at domain center (-40) ~ 8km
Example 4: grid_spacing_m = 10000.
standard_latitude = -40.
standard_latitude2 = -90.
grid_cen_lat = -40.
grid_spacing at projection (south) pole ~ 12km
grid_spacing at domain center (-40) = 10km
Note that the 'Dx' and 'Dy' values that appear in the
'static.nest7grid' should equal the value of
'grid_spacing_m'.
Lambert:
This is the latitude at which the grid spacing is exactly
the nominal value (e.g. 10km). The projection cone will
intersect the Earth's surface at this latitude.
Mercator:
This is the latitude at which the grid spacing is exactly
the nominal value (e.g. 10km). Equivalently this is the
latitude at which the projection cylinder intersects the
Earth.
6) standard_latitude2:
Polar Stereographic:
This must be set to +90. or -90. and defines the pole
latitude of the polar stereographic projection
(Earth's North or South Pole).
Lambert:
For a tangent lambert (e.g. CONUS), set this equal to
the 'standard_latitude' parameter. For a secant
(two-latitude) lambert, set this to the second true
latitude where the projection cone intersects the surface.
Mercator:
N/A
</pre>
When you run ./etc/localize_domain.pl, the netCDF static
file 'static.nest7grid' will be automatically generated by process
'gridgen_model.exe'. This contains grids of latitude, longitude,
elevation, and land (vs. water) fraction.
The following output message, "topo_30s file /U50N119W does not exist",
does not necessarily mean there is a problem. It may signify that
your domain runs outside the available 30" data, and should still be
covered by the 10' worldwide data, if you are using the 'topo_30s'
dataset. Other WARNINGs or ERRORs may be more significant.
---- 2.7.2.1 MAP PROJECTION FUNCTIONALITY/LIMITATIONS ---
LAPS runs with the polar stereographic, lambert, and mercator projections.
Please let us know if you encounter any problems.
The polar stereographic projection has a pole that may be set to either
earth's north or south geographic poles.
Setting the pole to an arbitrary lat/lon (local stereographic) is a
possible future enhancement. A test local stereographic domain gave an
error of 2km in the grid points; the test code works in cases where the
projection pole coincides with the center of the domain. Further
improvement of this may include more fully converting library subroutines
'GETOPS' and (possibly) 'PSTOGE' to double precision.
The projection rotation routine 'projrot_laps' also has some
approximations when local stereographic is used. These need to be checked
for their validity and refined if needed. Cases of interest include a
projection pole point at the domain center, as well as offset from the
center.
The local stereographic projection also ignores 'standard_latitude' from
the namelist so this is internally assumed to be +90. This means that the
grid spacing is valid at the projection pole location, regardless of
both where on the earth the pole is and the pole's latitude.
The map projection calculations are performed with a spherical earth
assumption.
---- 2.7.3 Domain Resolution ---
The default value of the 'grid_spacing_m' parameter is 10000m. This is
one of the parameters used in constructing the static file (as mentioned
above). To date, we have run LAPS with resolutions ranging from 1000m to
48000m.
---- 2.7.4 Terrain Smoothing/Filtering ---
Edit the file 'data/static/nest7grid.parms'...
<pre>
1) silavwt_parm:
Default value of 0. This parameter allows the potential use of
silhouette terrain which is the maximum elevation in the local area.
Useful range is anywhere between 0-1. A value of zero uses the
average terrain instead of the maximum. Note that a value of 1
may reduce the apparent effect of filtering with 'toptwvl_parm'.
2) toptwvl_parm:
For example, a value of 4 represents 4 delta-x filtering of the
terrain. You can change this to alter the smoothness of the
terrain. Higher numbers mean smoother terrain.
</pre>
--------- 2.8 ----- CHANGING THE VERTICAL DOMAIN------------------------------
PRESSURE OF THE LEVELS (and vertical resolution):
To do this, perform the following between untarring the tar file and
localizing LAPS
1. Copy 'data/static/pressures.nl' to your $TEMPLATE directory, then edit
it with to have the new set of levels.
Update the list of pressures that go in sequence from higher to lower
pressures (bottom to top)
Note that the default vertical grid uses constant pressure coordinates and
that the vertical pressure interval can vary between levels. For example
one might want to use higher density in the boundary layer (~100Pa interval)
and make it coarser higher up (~250Pa interval).
Of course the top pressure should be greater than zero mb. The bottom
level should extend below the terrain and below the observations. The
pressure values must be in multiples of 100 pascals, corresponding to an
integer number of millibars.
NUMBER OF LEVELS:
1. The default value of 'nk_laps' is set to 21 levels in
'data/static/nest7grid.parms' and will automatically be reset
during the localization (based on the contents of 'pressures.nl').
2. Note that compatibility with model background data will depend of the
vertical extent of that data source.
Note: If you are feeding LAPS output into an AWIPS workstation, then
additional workstation related changes may be needed.
------ 2.8.1 ------ Sigma Height Grid ---------------
UNDER DEVELOPMENT - mainly for STMAS-3D
Similar to 2.8 except that one changes the 'vertical_grid' parameter in
'nest7grid.parms'. Also the 'heights.nl' namelist is used instead of
'pressures.nl'. Note the heights in this namelist are scaled sigma values
where the namelist (idealized) height = sigma * (height_top - height_bottom)
Presently the height_top and height_bottom values are hard wired to 20000.
and 0. meters, respectively.
------ 2.8.2 ------ Sigma Pressure Grid ---------------
UNDER DEVELOPMENT - mainly for STMAS-3D
Similar to 2.8 except that one changes the 'vertical_grid' parameter in
'nest7grid.parms'. Also the 'sigmas.nl' namelist is used instead of
'pressures.nl'.
-------- 2.9 ----- CHANGING THE CYCLE TIME -----------------------------------
The default cycle time is 60 minutes. To change this, do as follows...
1. Edit runtime parameter file 'data/static/nest7grid.parms'
to change the value of 'laps_cycle_time'.
----- 2.10 ----------LQ3 (HUMIDITY ANALYSIS) CHANGES---------------------
Recent changes as of February 26, 2006
NAMELIST
The namelist file ./laps/static/moisture_switch.nl controls the data
assimilation within the moisture analysis. This file is self-documented, refer
to it for details. This file has not changed in this latest update however, one
of it's controlling aspects is GVAP or GOES vapor (total precipitable water,
product data) and the application of this data has changed since a major
implementation change March 2005.
OPTRAN
The NESDIS Community Radiative Transmittance Model (CRTM) and forward radiance
model called OPTRAN is incorporated into the current release of LAPS. Details
of OPTRAN are available from:
Tom Kleespies
NOAA/NESDIS
[email protected]
Also OPTRAN can be used by any U.S. Government or U.S. Military entity without
problem. ALL other users need to contact NESDIS (Tom Kleespies) and receive
authorization to use this software. Generally a simple acknowledgement to give
full credit to the program author is all that is required. GSD assumes no
obligation or responsibility in integrating this software as part of LAPS. To
disable the use of OPTRAN in LAPS, simply assign the GOES option in the
moisture_switch.nl namelist file to zero.
The version of OPTRAN in LAPS is configured to work with GOES-8 and -10 sounder
or imager at this time. Note also that GOES imager channel 5 (water vapor split
window) is currently not available on GOES 11, 12 and future satellites since it
was replaced with a different band. Therefore, the GOES imager data should not
be used in the moisture algorithm for any GOES satellite 11 and beyond. There
are simply not enough moisture channels available to offer any useful
information about moisture depth due to this change. Furthermore sounder
radiances for GOES-10 are deemed about 98% reliable, they are regarded to be
100% reliable for GOES-8. NaN values have been observed being generated from
the GOES-10 sounder coefficients that currently accompany this software. At
this time there are only basic provisions to handle the NaN state conditions.
They have not been observed to crash the moisture analysis and seem to be
handled gracefully to date. Any observation otherwise needs to be communicated
to:
Dan Birkenheuer
NOAA/GSD
[email protected]
To model the atmosphere with OPTRAN, an atmosphere is formulated that extends to
0.1 hPa. This is a composite of the normal LAPS analyzed vertical domain
(nominally extending to 100 hPa), spliced together with a climatological
atmosphere of 20 levels that extends to 0.1 hPa. The joining of the two
vertical coordinate systems is computed automatically and is continuous. This
will automatically take place even if the nominal LAPS levels are extended
beyond 100 hPa. In this upper region, temperature, and mixing ratio are
functions of latitude and Julian day. Ozone is based on the U.S. Standard
Atmosphere.
ADDENDUM: routine RAOB_STEP.F
It should be noted that some users have had to modify the parameter that defines
dimensions in routine raob_step.f due to the fact that this can overflow array
limits on some machines. The current parameter snd_tot is set to 1000. The
primary reason for this is to accommodate satellite soundings of which there can
be many in even a small area. This parameter ties in to the dimensions of the
weight matrix (ii,jj,snd_tot). If a large horizontal domain is defined, and you
don't have a lot of RAOB data and are not using satellite processed soundings,
you may have better success at compiling this routine by reducing the value of
snd_tot to a smaller value.
GVAP
GVAP data are GOES sounder total precipitable water data acquired from the
sounding retrieval process. These data were added to LAPS under a grant from
NOAA NESDIS.
The analysis for GVAP data has recently changed from the prior application. Up
until the March 2005 release, GVAP data were used as a direct total moisture
data source in that the integrated state variable in the moisture routine (q)
was compared to GVAP totals and part of the minimization procedure was to
improve this match through variational techniques.
It was learned during the IHOP 2002 experiment that the GVAP data were badly
biased, especially at asynoptic times. (see
http://laps.fsl.noaa.gov/cgi/birk.pubs.cgi for all publications, and
specifically http://laps.fsl.noaa.gov/birk/papers/Birkenheuer2005a/j.pdf for the
article about IHOP, or it can be located in the literature at:
Birkenheuer, D., and S. Gutman, 2005: A comparison of the GOES moisture-derived
product and GPS-IPW during IHOP. J. Atmos. Oceanic Tech. 22, 1840-1847.)
As a result, the algorithm was modified to use GVAP gradients and to compute
gradients in the solution field and match these gradients to those from GVAP.
The advantage to using gradients in this procedure was that it eliminates bias
and improves data structure. There is not a Tech Memo that has been published
and is also available on line that describes this new technique. (refer to:
http://laps.fsl.noaa.gov/birk/papers/tech_memos/GSD_Tech_Memo_32.pdf or a copy
can be gotten directly from GSD)
------ 2.11 ----------------OTHER RUNTIME PARAMETERS---------------------------
It is worthwhile to check the 'nest7grid.parms' and other namelist
files in 'data/static' to make sure all the runtime parameters are
correct. Some parameters worth noting are:
<pre>
nest7grid.parms
_______________
c8_project_common - Depends on which "realization" of LAPS you are running.
Allowed values are listed within 'nest7grid.parms'.
cloud.nl
________
l_use_vis - Boolean set to indicate whether we are confident in the
calibration of the visible satellite data and albedo
fields for use in the cloud analysis. This is normally
set to .true. at GSD and .false. for WFO and other
ports unless we are confident in the vis data
normalization.
</pre>
------ 2.12 -----Detecting and Reporting Installation Errors-------------------
To determine how well LAPS was installed, verify that all (31 at last check)
executables were built OK ('bin' directory) with no errors in the output
of 'make'.
Similarly, check the output of the localization script.
If you have any problems during the configure, install and localization
process, there are several things to check. For certain platforms, you can
compare your build output with ours by clicking on "Results of Latest LAPS
Builds" on the LAPS Software page. Also double check that you've followed
all the installation steps in this section of the README. There is also a
FAQ available at http://laps.noaa.gov/birk/LAPS_FACTS.htm
Finally, check the release notes at the
http://laps.noaa.gov/software/release_notes.html URL.
If you don't find the answer in these documents, send mail to
[email protected] Include in your mail:
<pre>
LAPS version number (hopefully you're using the latest version?)
The type of system (often, uname -a)
The system limits (ulimit -a)
The applicable compiler versions (often a -v or -V option to the compiler)
The entire output of configure
The entire output of make (standard output + error output)
The entire output of localize_domain.pl (found in $LAPS_DATA_ROOT/log/localize_domain.log*)
</pre>
---- 2.12.1 Runtime Monitoring ---
To see how well LAPS is running, check if output files are being placed in
the various 'lapsprd' subdirectories. A graphical product monitor that can
help with this is available in 'etc/laps_monitor.pl'. This script may need
some simple editing to suit your needs (e.g. to specify the
$LAPS_DATA_ROOT[s]). The monitor script writes HTML output to 'stdout'. This
HTML output, if routed to a file or hooked up to a Web server, can be viewed
with a browser. You can click on
http://laps.noaa.gov/monitors/Laps_Monitor.cgi to see an example of the
monitor output. Green means optimum product continuity, red means the
product is failing to generate, yellow means it is generating OK now but has
failed in the past. The numbers in the columns indicate the number of files
in each directory, as well as the youngest and oldest file ages in hours.
The data flow is generally from top to bottom on the product list, starting
with analyses and ending with forecast model (fua/fsf) output being listed
at the very bottom. In general the root cause of missing products would be
the first one that is missing along the data flow.
To check what model background and observational data were used in the
analyses as well as some QC and error (verification) statistics, you can
view the log file summaries in the files
'$LAPS_DATA_ROOT/log/*.wgi.yydddhhmm'. Generally each named 'wgi' file
corresponds to the name of an analysis process, except that 'sfc.wgi.*'
is generated by one of several executables than can provide the LAPS
surface analysis depending on the runtime configuration.
For more details, check the various log files in the '$LAPS_DATA_ROOT/log'
directory for occurrences of the string 'error' and 'warning'. The errors
are generally more significant. If any core dumps occur they can usually be
flagged by searching for the 'sh:' string in 'sched.log.*'.
If you are reporting runtime errors it can be useful to tar up your entire
$LAPS_DATA_ROOT and make it available on your web or FTP server as a
compressed "tgz" file. If the data set is very large you might consider
mailing us a CD or DVD. Alternatively if you have the untarred
$LAPS_DATA_ROOT files on your web server we can browse through the
directories for the log and product files as needed to help diagnose the run.
If the $LAPS_DATA_ROOT is large it can be pared down as there is a script
called 'etc/tarlapstime.sh' that works by just tarring up the current hour's
worth of files. If you need to narrow this down further just the inputs to
the particular analysis would be needed as shown in section 3 of the README.
Also, things like the 'cdl', 'time' and 'static' subdirectories should be
included.
---------- 3.0 ---- DESCRIPTION OF LAPS PROCESSES ---------------------------
The following section contains information on which LAPS processes
generate which LAPS output products. Static data (like lat and lon
grids) are included in section 3.1. These are the processes contained
within the LAPS tar file and built with the localization script.
"Inter data" is an ascii file containing non-gridded data
(intermediate data files). Examples of this are surface obs, profiler obs,
etc.
This list contains all outputs generated by LAPS processes.
The products listed under each process are the outputs produced by
that process. Inputs are listed here for some analyses. If the cron
including 'sched.pl' (see section 2.4) is run according to the flow therein,
the necessary inputs will be available.
------- 3.1 --------------- Localization Processes-----------------------
------ 3.1.1 ------ Gridgen_model (static.nest7grid generation) ---------------
Package: gridgen_model.exe - Writes static file, run by localization script.
Contact: Steve Albers - [email protected]
Inputs:
Geography databases (topography, land fraction, landuse, soiltype top/bot)
greenness fraction, mean annual soil temperature, and
albedo. Files are typically in 10, 30, or 180 deg tiles.
See section 2.2.5 for details on the geography data.
static/nest7grid.parms
<pre>
Outputs:
static/static.nest7grid netCDF grid geography data mapped to LAPS
grid
'LAT' latitude in degrees
'LON' longitude in degrees
'AVG' mean elevation MSL
'LDF' land fraction (1.0=land; 0.0=water)
'LND' land-water mask (1=land; 0=water)
'USE' usgs 30s (24 vegetation categories)
landuse data (currently dominant category
for each grid point).
1: Urban and Built-Up Land
2: Dryland Cropland and Pasture
3: Irrigated Cropland and Pasture
4: Mixed Dryland/Irrigated Cropland and Pasture
5: Cropland/Grassland Mosaic
6: Cropland/Woodland Mosaic
7: Grassland
8: Shrubland
9: Mixed Shrubland/Grassland
10: Savanna
11: Deciduous Broadleaf Forest
12: Deciduous Needleleaf Forest
13: Evergreen Broadleaf Forest
14: Evergreen Needleleaf Forest
15: Mixed Forest
16: Water Bodies
17: Herbaceous Wetland
18: Wooded Wetland
19: Barren or Sparsely Vegetated
20: Herbaceous Tundra
21: Wooded Tundra
22: Mixed Tundra
23: Bare Ground Tundra
24: Snow or Ice
'U01-U24' Fractional distribution of landuse category (not active).
'STL' Soil type - top layer (0-30cm)
'SBL' Soil type - bottom layer (30-90cm)
(currently dominant category for each grid pt)
FAO/WMO 16-category soil texture:
1 SAND
2 LOAMY SAND
3 SANDY LOAM
4 SILT LOAM
5 SILT
6 LOAM
7 SANDY CLAY LOAM
8 SILTY CLAY LOAM
9 CLAY LOAM
10 SANDY CLAY
11 SILTY CLAY
12 CLAY
13 ORGANIC MATERIALS
14 WATER
15 BEDROCK
16 OTHER (land-ice)
'T01-T16' Fractional distribution of top layer soil texture class (not active)
'B01-B16' Fractional distribution of bottom layer soil texture class (not active)
'TMP' Mean annual soil temperature
'G01-G12' Monthly (center of month) greenness fraction
'A01-A12' Monthly (center of month) albedo climatology
'ALB' Not used
static/latlon.dat Binary grid latitude
longitude
static/topo.dat Binary grid mean elevation
static/corners.dat ASCII lat/lon of 4 corner points
static/latlon2d.dat ASCII latitude/longitude
</pre>
Source directory: laps/src/grid
Sample Output: Should be available in the test data case. The grids start with
gridpoint (1,1) in southwest corner of the domain and end with
gridpoint (ni,nj) in the northeast corner. The bottom
(southernmost) row of the domain is written first (I increases
with consecutive grid points, then J increases). I increases as
you're moving east on the grid, J increases as you're moving
north.
------ 3.1.2 ------ Surface Lookup Tables (gensfclut.exe) ---------------
Package: gensfclut.exe - Writes surface lookup tables, run by localization
script. (contact: John McGinley / Steve Albers)
Source directory: laps/src/sfc/table
<pre>
Output:
static/drag_coef.dat Binary grid Drag Coefficients
</pre>
In 'gensfclut.exe', the friction parameter has been configured by automatically
producing a scaling factor based on the range of elevations across the domain.
This factor can be changed in the 'drag_coef' section of 'build_sfc_static.f',
if so desired.
------ 3.1.3 ------ Satellite Lookup Tables (genlvdlut.exe) ---------------
Package: genlvdlut.exe - Writes satellite lookup tables, run by localization
script. (contact: John Smart)
Source directory: laps/src/ingest/satellite/lvd/table
<pre>
Output:
static/lvd/*.lut Satellite Lookup tables
</pre>
Additional information on the lookup tables can be found in the file
'laps/src/ingest/satellite/README'.
------- 3.2 ------------------ Ingest Processes ---------------------------
As mentioned above, a flow chart for the ingest processes may be found at
'http://laps.noaa.gov/doc/slide1_v3.gif'.
------ 3.2.1 ----------------- LGA Model Background -------------------------
Package: lga.exe - ingest background model data
(contact: Steve Albers - [email protected]).
LGA LAPS analysis grids from RUC or other analysis/forecast grids.
Inputs:
Raw model data on the model's native grid. The acceptable models and
formats for the background model are listed in
'data/static/background.nl'. We have recently added support for the
background models to include GRIB-formatted files. See source directory:
$LAPS_SRC_ROOT/src/lib/degrib/README_LIBS file for detailed information.
For some models you might want to do a
separate conversion of GRIB to netCDF prior to running LGA. One software
option for this is available from the GSD/ITS group as described in
section 3 of this web document at the following URL:
http://www-its.fsl.noaa.gov/dsg/FSL-WhitePaper.html
Tropical cyclone bogusing information is also an input in the form
of the 'tcbogus.nl' namelists. These are generated independently of LAPS
as "raw" data, yet are placed in the '$LAPS_DATA_ROOT/lapsprd/tcbogus'
subdirectory. The filename convention should be 'yydddhhmm_tcbogus.nl'.
To see the format check the sample file located in
'data/static/tcbogus.nl'.
Outputs: (Feeds various analyses)
<pre>
lga grid background model 3-D data analysis/forecast
lgb grid background model Sfc data analysis/forecast
</pre>
Source directory: The source code for this is in 'src/background'.
Library directory: Associated library modules are in 'src/lib/bgdata'.
Parameter namelist file: 'static/background.nl'
Sample Input/Output: May be available in the test data case.
This software currently supports nearly 10 different models. If additional
models are required, then software mods may be needed, potentially a new
source file added to 'src/lib/bgdata/read*.f'. A key variable that
relates to which model you're using is 'bgmodel'.
Note that time interpolation is used if the required LAPS analysis time(s) are
between the valid forecast times for two of the set of input files. In this
context output files are produced for the LAPS analysis time as well as +/- one
LAPS analysis cycle time. Input data for LGA should thus be available over an
appropriate time span.
------ 3.2.2 -------------- Surface Data Ingest ------------------------------
------ 3.2.2.1 obs_driver.x ---
LSO process - obs_driver.x - Ingest surface data
(author: Pete Stamus/Steve Albers)
<pre>
Input:
METAR/SYNOP data
Buoy/ship (maritime) data
MADIS (or LDAD) mesonet/urbanet data
GPS (sfc obs - without precipitable water)
Profiler surface data (via LDAD)
Output:
LSO ascii LAPS surface obs intermediate data file: METAR, Mesonet,
and Buoy/ship obs. The format of this file may be
determined from looking at library access routines such
as 'read_surface_data', the obs_driver code, or by
looking at sample LSO files in the test data case.
Sample Input/Output: May be available in the test data case.
</pre>
Source directory: $LAPS_SRC_ROOT/src/ingest/sao (contains a README file)
Parameter file (specifies input data paths and formats): 'static/obs_driver.nl'
The LSO file is fairly self explanatory. The easiest way to see what
goes where is to look at the routine 'read_surface_data' in the file
'src/lib/read_surface_obs.f', and the corresponding format statements in the
file 'src/include/lso_formats.inc'.
The routines are pretty well commented, and should be enough to tell you what
you need to know if you want to make a decoder that outputs an LSO-type
formatted file directly. This direct route would allow you to bypass the
step of producing "raw" netCDF surface observation data.
Here are a few recommended settings for the observation type
variables (reportType and autoStationType)
if you are constructing your own LSO file:
<pre>
raw data reportType autoStationType
________ __________ _______________
metar METAR UNK (unless an automated A01 or A02 station)
synop SYNOP UNK (unless an automated A01 or A02 station)
buoy MARTIM FIX
ship MARTIM MVG
</pre>
The expected accuracies are based on "offical" NWS numbers where possible.
For LDAD observations, they're just a best guess, since no one really knows
how good the obs are. These expected accuracies will be used in the quality
control routines sometime in the future. The lat/lons are in decimal degrees.
Gross "climatological" QC error checks are applied to several variables
including temperature, wind, and pressure. MADIS QC flags are checked as
can be controlled via namelist.
------ 3.2.2.2 How to Blacklist stations ---
(author: Steve Albers/Pete Stamus)
As part of the 'obs_driver' code, a Blacklisting function has been added. This
allows users to tell LAPS to skip stations with known bad variables (one or
several), or to skip a station completely. As of this writing, the user will
have to edit a "Blacklist.dat" file...in the future we hope to include this
function in the LAPS GUI.
An example file, called "Blacklist.example" has been included in the same
directory as this README file. It shows the format that *must* be followed for
the Blacklist to work properly. An error in the format will either allow the
bad station(s) through, or crash the program completely. Let's decode the
"Blacklist.example" file:
The first line is the number of obs to blacklist...in this case, 5.
Each station goes on a new line. The number of variables to blacklist
for that station is next, then the codes for the variable(s) follow. For
the first station (KFCS), we are blacklisting the 3 pressure variables.
To blacklist the entire station (KDTW) use 1 for the number of variables,
and "ALL" as the variable. All the stations from a particular provider can
be blacklisted by adding 100 to the number of variables (third example).
The last two examples show 1 and 2 individual variables, respectively.
These are the valid codes for variables to blacklist:
<pre>
"ALL" - Set all variables at this station to bad
"TMP" - Set temperature bad
"DEW" - Set dew point temperature bad
"HUM" - Set relative humidity bad
"WND" - Set wind bad (this does both speed and direction, and gusts)
"ALT" - Set altimeter bad
"STP" - Set station pressure bad
"MSL" - Set MSL pressure bad
"VIS" - Set visibility bad
"CLD" - Set clouds to bad (this does all cloud layers reported)
"PCP" - Set precipitation amount to bad (all reported, 1-12 hrs)
"SNW" - Set snow cover to bad
"SOL" - Set solar to bad (if reported)
"SWT" - Set soil/water temperature to bad (if reported)
"SWM" - Set soil moisture to bad (if reported)
</pre>
You might keep in mind that some variables act as a group. For example
both "HUM" and "DEW" variables feed into the LAPS dewpoint analysis so
consideration should be given as to whether to blacklist one or both of
these variables. "ALT", "STP" and "MSL" are a similar group of pressure
variables. An incorrect variable code generates a warning message, and
the code should hopefully continue without acting on the station in question.
Note that when a station is blacklisted, its name, latitude, longitude,
elevation, and time, will still be stored in the LSO file. However, the
selected variables (up to "ALL" of them) will be set to the 'badflag' value
and skipped in the analyses.
To actually get this stuff working, edit the file called "Blacklist.dat" in the
'data/static' (or template) directory. The "Blacklist.dat" being used at GSD is
supplied in this directory as a default, and this provides additional examples.
Format the file *exactly* as the 'Blacklist.example' file (using your station
information, of course). Save the file, and the next time 'obs_driver' runs,
it will use the blacklist information. This will be noted in the
'obs_driver.log' file.
You may eliminate element specific or ALL data from a particular provider by
replacing the leading 0 with a 1 in the second column. See the WXforYou example
in the Blacklist.example file. To ensure the elimination of the data by
provider, care must be taken to make certain the correct provider is listed in
the Blacklist.dat file.
Primarily, offending datasets are from stations received through LDAD. To find
the provider for a given station you can look in the "log/sfc.wgi.*" files or
in the input LDAD netCDF files.
For AWIPS users a list of these stations are kept in ~ldad/data on your AWIPS
system. Each .txt file in ~ldad/data will have a .desc file associated with
it which describes the data being ingested etc., by that provider. Look in a
particular .desc file of interest. Go to the 3rd word of the 1st line which
is not commented out (e.g. aprswxnet : -9999.00 | APRSWXNET). For this
example, the APRSWXNET (3rd word) is the provider name and should be the entry
used if utilizing the elimination by provider feature of the Blacklist.
------ 3.2.3 --- Polar Radar Data (e.g. WSR 88D Level II, Level III) ---
Process: remap_polar_netcdf.exe
Author: Steve Albers (Steve.Albers AT noaa.gov)
Every volume scan
Initiation: Completion of volume scan
Inputs:
Wideband Radar Data (reflectivity and velocity in polar coordinates,
in netCDF format) These have one tilt per file and at least 4 tilts
per volume scan (all with the same volume timestamp in the filenames).
This data can be obtained from a WSR 88-D Level-II data feed or the
equivalent. A description of how we obtain these Polar netCDF files
for Level-II is at 'http://laps.noaa.gov/albers/remapper_raw.html'.
Note that narrowband reflectivity data (e.g. WSR 88D Level-III RPG)
can also be used as long as it is converted to the required polar
coordinate, netCDF format. This is in fact being done for the AWIPS
implementation of LAPS for a low-level tilt from a single radar, via
the 'etc/LapsRadar.pl' script running in the AWIPS environment. The
comment section at the top of this script explains how this 4 bit
processing of reflectivity data works. 'etc/LapsRadar.pl' runs two
executables. The first executable 'tfrNarrowband2netCDF' from AWIPS,
writes out the polar netCDF files in the directory
'$LAPS_DATA_ROOT/lapsprd/rdr/???/raw' where ??? is the radar number.
The second executable 'remap_polar_netcdf.exe' is run as part of LAPS.
We haven't been using Level-III velocity data since it is of limited
4-bit resolution and we're running only with the lowest tilt at present.
For both Level-II and Level-III the polar netCDF files are named
according to 'yydddhhmm_elevxx' where 'xx' is the tilt number.
Sample polar netCDF files including a CDL may be found at:
http://laps.noaa.gov/software/radar/wideband/
The transformations between counts and dbz can be found in the
'netcdfio.f' source code file in subroutines 'counts_to_dbz' and
'counts_to_vel'.
In Nimbus NetCDF the range for dBZ and RadialVelocity is -128 to +128.
Z_nimbus_netcdf = (2 X Z_measure) -62
V_nimbus_netcdf = (2 X V_measure)
Outputs (LAPS intermediate files - depending on input parameters):
<pre>
v01 grids 3-D Radar reflectivity, velocity, and Nyquist vel
v02 " "
rdr/???/vrc " 2-D Radar reflectivity (??? = radar number)
vrc " "
etc. (for each radar)
</pre>
The outputs from this process, on the Cartesian LAPS grid, are used by the
LAPS wind analysis, and also potentially by cloud and precip accumulation
analyses. One output file is written per volume scan.
When running the remapper, files such as v01, v02, vrc, etc. are produced
depending on which radar is being used and on the input parameters.
A further description of how the remapper software functions may be found on
the World Wide Web at 'http://laps.noaa.gov/albers/remapper_laps.html'.
Also recall the flow chart showing the inputs and outputs for
'remap_polar_netcdf.exe' at
'http://laps.noaa.gov/albers/laps/radar/laps_radar_ingest.html'.
Source directory: The source code for this is in 'src/ingest/radar/remap'.
Compile time parameters: 'src/include/remap_dims.inc'
Runtime parameter namelist file: 'static/remap.nl'
Sample Input/Output: May be available in the test data case.
------ 3.2.4 -------- WSI / NOWRAD RADAR PREPROCESSING (VRC) ------------------
Process: VRC (vrc_driver.x)
Author Steve Albers ([email protected])
Parameter namelist file: 'static/vrc.nl'
<pre>
Inputs:
Raw WSI NOWRAD radar reflectivity data
Outputs (Intermediate data file):
vrc grid 2-D reflectivity
</pre>
The WSI data are decoded externally to LAPS and written as
netCDF files in NIMBUS format. The vrc_driver.x process reads these netCDF
files. WSI sends out many types of radar data. We use the files
that are labeled "_hd" (15 min freq). They also send out an "_hf"
(5 min freq) file. We use hd because WSI hand edits these for
ground clutter. The hf files are not edited. The hd and hf files
are composites of "low-level" elevation scans from the 88D's around
the country. The vrc_driver.x also maps from conus to laps domain
for the wfo data set. The map transformation software is found in
lib/gridconv, lib/nav, and lib/radar/wsi_ingest. The switch to use
wsi versus wfo in variable c_raddat_type in the 'nest7grid.parms' namelist.
Pathway to data is variable path_to_wsi_2d_radar in 'vrc.nl'.
The output reflectivity is used by the cloud and precip accumulation analyses.
------ 3.2.5 --------------- Radar Mosaic ------------------------------------
Process: (mosaic_radar.exe)
Author Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs:
v01, v02, etc. 3-D reflectivity
rdr/001/vrc, rdr/002/vrc, etc. 2-D reflectivity
Outputs (Radar Mosaic - intermediate data file):
vrz grid 3-D reflectivity
vrc grid 2-D reflectivity
</pre>
This program runs once per LAPS cycle in the 'sched.pl'. The default is to
write just one mosaic file for the cycle valid at 'systime'. A namelist option
allows this program to produce multiple mosaic outputs within a given LAPS
cycle. The multiple mosaics are all run at the same wall clock time, while the
valid mosaic times are spaced throughout the previous LAPS cycle.
The nearest radar with valid data is the one chosen to contribute at each
grid-point.
The output reflectivity mosaic is used by the cloud and precip accumulation
analyses. Further QC is done within these analyses.
Parameter namelist file: 'static/radar_mosaic.nl'
------ 3.2.6 ---------------- PROFILER/VAD/SODAR (PRO) Ingest -----------------
Process: PRO (ingest_pro.exe) LAPS Wind Profile Ingest
Author: Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs: (located in separate netCDF directories)
NPN 404-MHz profiler wind data in netCDF format
NetCDF CDLs from both GSD-NIMBUS and AWIPS/MADIS are accepted.
Boundary layer 915-MHz profiler wind data in netCDF format
(GSD-NIMBUS/DD, AWIPS-LDAD & MADIS-MAP CDLs).
50-MHz profiler in netCDF (AWIPS-LDAD & MADIS-MAP CDLs).
Doppler Radar VADs in netCDF (GSD-NIMBUS CDL) format
SODAR data in netCDF format (AWIPS-LDAD & MADIS-MAP CDLs).
Output: (feeds wind)
pro inter data wind profile direction and speed (ASCII-true North)
Parameter namelist file (for data paths): 'static/nest7grid.parms'
</pre>
Source directory: laps/src/ingest/profiler
Parameter namelist files: static/nest7grid.parms, static/vad.nl
Sample Input/Output: Should be available in the test data case.
For the 'pro' output, each profile starts with an ASCII header and the
formatted entries are defined in sequence...
<pre>
1) WMO ID or other ID number. The use of this is optional and zero can
be used if you don't know the number.
2) Total number of levels for which data is provided. This can include
the surface data as the first level.
3) Latitude (degrees)
4) Longitude (degrees)
5) Station Elevation (meters MSL)
6) Station Name
7) Time of observation (UTC). This is the middle of the observation period
if time averaging is used.
8) Data type. Can be either "PROFILER" or "VAD"
</pre>
After the header, the data entered for each level is as follows...
<pre>
1) Elevation (meters MSL)
2) Wind Direction (degrees)
3) Wind Speed (meters/second)
4) Estimated Root Mean Square (RMS) error of measurement
</pre>
------ 3.2.7 ----------------- RASSs (LRS) Ingest -----------------------------
Process: (ingest_lrs.exe) LAPS local data RASS ingest
Author: Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs:
NPN RASS temperature data in NIMBUS netCDF format
Boundary layer RASS data in netCDF formats
(NIMBUS, RSA-LDAD, MADIS-MAP)
Outputs: (feeds LSX and temp.exe)
lrs inter data RASS Virtual Temperatures (ASCII)
Parameter namelist file (for data paths): 'static/nest7grid.parms'
</pre>
Source directory: laps/src/ingest/rass
Sample Input/Output: Should be available in the test data case.
------ 3.2.8 ----------------- PIREPS/ACARS Ingest ----------------------------
Process: (ingest_aircraft.exe) LAPS Pireps / ACARS
Author: Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs:
Aircraft voice pireps (cloud layer reports)
NetCDF files using GSD-NIMBUS or WFO/AWIPS CDLs
ACARS data
NetCDF files using GSD-NIMBUS CDLs
Uses pressure altitude
Hourly netCDF filename convention is 'yydddhh00q.cdf'
NetCDF files using MADIS CDLs
Same format as GSD-NIMBUS
Software can automatically switch to the MADIS filename
convention of 'yyyymmdd_hhmm' if needed.
NetCDF files using WFO/AWIPS CDLs
Uses pressure altitude
AFWA ASCII format also allowed for ACARS
WISDOM Balloon Data
NetCDF files using MADIS CDLs
Outputs: (Intermediate output written to the 'pin' file. Feeds cloud.exe,
wind.exe, lq3driver.x)
pin inter data voice pireps/clouds
ACARS/(wind, temp, mixing ratio - using
pressure altitude)
</pre>
Source directory: The source code for this is in 'src/ingest/acars'.
Parameter namelist file (for data paths): 'static/nest7grid.parms'
Sample Input/Output: Should be available in the test data case
------ 3.2.9 ----- Sounding (RAOB/Dropsonde/Sat/Radiometer) (SND) Ingest -------
Process: (ingest_sounding.exe) LAPS Soundings
Author: Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs:
RAOB in various formats:
(GSD-NIMBUS CDL - netCDF)
(WFO/AWIPS CDL - netCDF)
(AFWA and CWB ASCII formats also allowed)
Satellite Soundings (GOES binary, MADIS POES, or AFWA format)
Dropsonde (AVAPS, CWB, or SND format)
Radiometer (GSD-MADIS CDL - netCDF)
Met Tower (LDAD netCDF)
Outputs: (Feeds temp.exe, humid.exe, wind.exe)
snd inter data (ASCII) sounding temp, dewpoint, wind (true N)
</pre>
Source directory: laps/src/ingest/raob (contains a README file)
Parameter namelist file (for data paths): 'static/snd.nl'
Sample Input/Output: May be available in the test data case. If not, the README
in the source directory contains a description of the
'snd' file.
Note: Sounding data is used if the observations lie in the time window
of +/- 'laps_cycle_time' centered on the analysis time. There
are flags to toggle usage of the sounding (i.e. snd) data in
'wind.nl', 'temp.nl' and 'moisture_switch.nl'.
------ 3.2.10 --------- LVD (Satellite Image + Cloud Top Pressure) ------------
LVD process - lvd_sat_ingest.exe - takes raw sat. data and puts it on LAPS
grid.
(author: John Smart, contact Kirk Holub - Kirk.L.Holub AT noaa.gov)
Input:
GOES or other satellite data
<pre>
Output:
LVD/'SATID' grid LAPS satellite data file
SATID (e.g. goes08 or goes09)
CTP grid Cloud-top pressure information
</pre>
Parameter namelist file: 'static/satellite_lvd.nl'
Source directory: laps/src/ingest/satellite/lvd (contains a README file)
------ 3.2.11 --------- Cloud Drift Wind (CDW) Ingest -------------------------
Process: (ingest_cloud_drift.exe) LAPS Cloud Drift Winds
Author: Steve Albers (Steve.Albers AT noaa.gov)
<pre>
Inputs (supported cloud drift wind formats):
GOES cloud drift winds in NESDIS (ASCII) format
MADIS netCDF format
AFWA & CWB formats are also allowed.
Outputs: (Feeds wind.exe)
cdw intermediate data (ASCII) Satellte cloud drift winds
</pre>
Parameter namelist file: 'static/cloud_drift.nl'
Source directory: laps/src/ingest/satellite/cloud_drift
Note: Sounding data is used if the observations lie in the time window
of +/- 'laps_cycle_time' centered on the analysis time.
------- 3.3 ------------- ANALYSIS PROCESSES --------------------------------
A flow chart for the analysis processes may be found at this URL:
http://laps.noaa.gov/doc/LAPS_flow_v_02.png
Listed below is a summary of each analysis process in the order it
is typically run by the 'sched.pl' script.
------ 3.3.1 -------------------- WIND ----------------------------------------
Process: wind.exe - WIND analysis and related fields
Author: Steve Albers (Steve.Albers AT noaa.gov)
Generate a wind analysis using surface observations, profiler, cloud
drift wind, and aircraft reports. VAD and SODAR can also be read in.
Background model grids are used as a first guess and to do quality
control on new observations. Time tendencies from the background model
are applied to the aircraft/cloud-drift wind reports when they are
taken before or after the nominal analysis time. The quality control
rejects any observations deviating from the background by more than a
threshold depending on observation type as in the following table.
<pre>
ACARS 10 m/s
Cloud-Drift winds 10 m/s
Profiler 22 m/s
Doppler Radar 12 m/s
Other 30 m/s
</pre>
The wind analysis is done in three steps. The first step analyzes the
non-radar data with the background wind field using a multiple
iteration successive correction technique.
For the second step, the first step results are used as the background.
The data used includes non-radar data; any grid-points with multiple-
Doppler radial velocities are also mixed in. Radial velocities are
taken from the Doppler radars after dealiasing and other quality
control steps are done. If two or more radars illuminate a given
grid-point, a full wind-vector is constructed from a combination of
the radial velocities and the preliminary non-radar analysis. This is
done via a "successive insertion" process, beginning with the
background (non-radar analysis), then followed with the radial velocity
from each radar in sequence.
For the final step the background field comes from the result of
the second step. All point data is now used, including grid-points
illuminated by only a single radar. The tangential component for each
radar observation is estimated by using the background from the
previous step (i.e. non-radar data and/or multi-radar data).
The omega field is calculated by kinematically integrating the
horizontal wind divergence. The lower boundary condition is specified
by the surface wind and terrain gradient.
<pre>
Auxiliary functions:
write out graphical products
Inputs: '*' = essential input
* lga/fua grid model data analysis/forecast
needed for current and previous cycle times
pro inter data profiler, VAD, SODAR winds (all true north)
snd inter data RAOB/Dropsonde data including winds
pin inter data ACARS Winds (using pressure altitude)
cdw inter data cloud drift wind
* lso inter data LAPS surface obs file (e.g. mesonet & METAR)
(remapped from raw radar data)
v01-vxx grid 3-D radar reflectivity/radial velocity
Outputs: (LW3 is main output)
pig inter data acars, cloud drift winds (prior to QC,
m/s, true north)
prg inter data profiler, sounding winds (prior to QC,
m/s, true north,
corrected for time)
sag inter data surface winds (prior to QC, true north)
d01-dxx inter data derived radar vector obs (grid north)
lw3 3d grids 3-D winds (U and V are wrt GRID NORTH), omega
lwm 2d grid surface winds
</pre>
Source directory: laps/src/wind (contains a README file)
Parameter namelist files: 'static/wind.nl', 'static/nest7grid.parms'
Further description and reference is at:
http://laps.noaa.gov/albers/laps/talks/wind/sld001.htm
------ 3.3.2 ----------------- SFC (LSX) -------------------------------------
Surface processing - laps_sfc.x (LSX)
(authors: John McGinley / Pete Stamus / Steve Albers)
The surface package collects surface data from the LSO intermediate data file
(METARs, local mesonets via LDAD, buoy/ship obs), IR brightness temperatures,
and fields from selected background models. Places surface data on LAPS grid
and performs a simple quality control of the obs (climo + standard deviation
checks). The quality control is described in the section below at (3.3.2.2).
A flow chart can be seen at this URL:
http://laps.noaa.gov/albers/laps/talks/sfc/Sfc_anal.gif
The background fields come from the locally-run LAPS model (FSF file), other
large-scale models (RUC, ETA, AVN - via the LGB file), or a previous analysis
(if all else fails). If the background model terrain is on a coarser grid than
LAPS, this is accounted for so that the LGB fields have the fine-scale terrain
related structure. For wind fields, the background comes from the 3-D wind
interpolated to the surface or LWM file.
Prior to analysis of each field, another quality control step is done that
rejects observations that deviate from the background by more than a threshold.
This threshold is proportional to the standard deviation of the observation
increments. The proportionality constant is set depending on the field.
The next step in the analyses is done with a successive correction technique
similar to the 3-D wind and temperature analyses (see those sections and their
web references). Observation increments are used for T, Td, U, V, MSL, P
and straight observations are used for visibility. The temperature and dewpoint
observations are also corrected for deviations of the station elevation from
the LAPS terrain. Standard lapse rates are applied to this elevation difference.
The analysis innovation is constrained to vary from the background by no more
than the magnitude of the observation rejection threshold discussed above. This
helps prevent overshooting (ballooning) of gradients into data sparse areas.
For relative humidity, the RH observations are converted into dew point using
the station temperature (if the dew point isn't directly reported). The
analyzed variable for moisture is dew point. After the analysis is performed
the gridded dew point field is converted back into relative humidity using the
analyzed temperature.
A land fraction term is factored into the weighting whenever the observation
and grid point are on either sides of a 0.01 land fraction threshold. This
helps prevent situations such as heating over the land having undue effects
over the water areas. This weight is applied mainly to the T, Td, U, and V
fields.
For pressure analysis, three fields are computed including reduced pressure (P)
at reference height 'redp_lvl', surface pressure (PS), and mean sea level
pressure (MSL). Background pressure fields come from the LGB or FSF files.
The MSL background is used as read in upon input. The (PS) background is
converted from the background model terrain to the LAPS terrain within the
LGB/FSF file. The (P) background is generated by reducing the (PS) background
to the reference analysis height 'redp_lvl' using Poisson's equation.
This reference height should be approximately equal to the mean elevation of
stations reporting surface pressure or station pressure.
Continuing the pressure analysis the altimeter setting observations are
converted and added to the set of station pressures using the standard
atmosphere. Station pressure observations are in turn converted to reduced
pressure using Poisson's equation. The (P) analysis uses the (P) background
plus the reduced pressure observation increments. The (P) analysis then
uses variational techniques to constrain the surface winds and reduced
pressures (P) to the full equations on motion. In contrast, mean sea level
pressure (MSL) is a direct analysis of the MSLP observation increments
together with the model background 'MSL' field. The station pressure analysis
(PS) is calculated using the model background gridded 'PS' field, multiplied
by the ratio of the (P) analysis to the (P) background.
Visibility is arrived at by first analyzing the surface visibility
observations. A second step is applied to decrease the visibility in areas
that have high RH and are near the cloud base that is given by the cloud
analysis (in the previous time cycle).
Several derived variables are calculated before the LSX file is written. Also,
a dependent data validation is done by interpolating several variables back to
the observation locations and comparing the analysis to the obs. Output from
this check is written to files located in
'$LAPS_DATA_ROOT/log/qc/laps_sfc.ver.hhmm', where 'hhmm' is the analysis
'systime'.
<pre>
Inputs:
LSO surface observations
- or -
LSO_QC QC'd surface observations
LGA Background model (3-D fields) on LAPS grid
(State variables from 700mb and 500mb)
LGB Background model (surface fields) on LAPS grid
(TSF, PSF, SLP, DSF, P, VIS) fields
- or -
FSF Background local model (surface fields)
(T, PS, MSL, TD, P, VIS) fields
- or -
LWM (background sfc wind from 3-D analysis - used for wind only)
SND Soundings (possibly containing surface obs)
LC3 Cloud cover (for visibility)
LM1 Soil moisture (for fire wx calc)
LM2 Snow cover (for fire wx calc)
Output:
LSX LAPS surface data grids (24 2-d fields packed in one file)
Includes various fields such as...
T, Td, Wind, MSLP, Reduced P (reference height sfc), Surface P
Fire Danger:
LAPS fire weather index is driven mainly by the surface fields
of current humidity, wind, and temperature. RH and wind have
the most weight with temperature having a lesser weight for
this index that ranges from 0 to 20. Snow cover, elevation, and
land fraction are given secondary consideration. High
elevations, assumed to be above the treeline, are given a lower
maximum value of 10. This index was developed primarily by
Matt Kelsch of GSD.
Colorado Severe Storms Index (CSSI):
Severe storm potential mostly geared to the Colorado area. This
uses a decision tree and various empirical functions. For more
info please check the documentation in subroutine 'make_cssi'
in file 'src/sfc/lapsvanl_sub.f'.
Reference: Rodgers, D.M., and R.A. Maddox, 1981: Surface
Meteorological Features Associated with Eastern Colorado Severe
Convective Storms. NOAA Technical Memorandum ERL OWRM-13. DOC,
NOAA, NWS, Boulder, Colorado, 21p.
Link: http://www.crh.noaa.gov/crh/?n=tm-114
Heat Index:
A function of temperature and humidity for discomfort due to heat.
This is based on a formula from Lans Rothfusz, NOAA/NWS. It is
calculated only when the surface air temperature exceeds 75 deg F.
The idea is to give a "feels like" temperature. For example, if
the temperature is 85 F but the heat index is 100 F, most people
would respond physically like it was 100 F actual temperature.
It's generally used to warn people that the temperature and
humidity will combine to make it seem hotter than it actually is,
and that they should take precautions like drinking more water,
stay out of the direct sun, take frequent breaks if working
outside, etc.
Ground Temperature:
A single level analysis using land/sea surface (skin) temperature
observations combined with background model information (if
available).
</pre>
Source directory: laps/src/sfc
Parameter namelist file: 'static/surface_analysis.nl'
--------- 3.3.2.1 ------SURFACE ANALYSIS RUNTIME PARAMETERS -------------------
PRESSURE REDUCTION
You will need to select an elevation for the reduced pressure analysis.
The reduced pressure is the only one really used in the variational
portion of LAPS, and the idea is to select an elevation that is
representative of the domain (or portion of the domain) you are
interested in. For example, the Colorado LAPS domain includes 4000m
high mountains over the western 1/3, and plains that slope below
1000m at the eastern boundary. We use 1500m as the Colorado LAPS
reduced pressure. This is close to the elevations over the eastern
2/3's of the domain, and requires a smaller reduction over the
mountains compared to MSL, for example. Change the namelist variable
in '/data/static/nest7grid.parms' when you localize LAPS.
--------- 3.3.2.2 ------SURFACE ANALYSIS QUALITY CONTROL -------------------
LAPS has a layered QC approach that gives us several opportunities
to flag erroneous observations. To start with, a variety of gross
"climo" checks are applied to the observations in the 'obs_driver.x'
ingest program.
The next steps in quality control are encountered in 'laps_sfc.x'.
This first checks the observations against climatologically
reasonable ranges. Next, the observations (most fields except wind)
are checked to see which ones are outliers (at 5 sigma) relative to
the average observation value in the domain. As a further check, the
Temperatures, Dewpoints and MSL pressures are checked to see if they
deviate from the background field by more than a threshold absolute
amount. The output from these checks is in both 'laps_sfc.log' and
'sfcqc.log'. The 'sfcqc.log' file contains the 'rely' (positive=retain,
negative=reject) values designated as follows:
<pre>
| STANDARD DEVIATION CHECK (against other obs)
|
CLIMO | PASS N/A FAIL
|____________________________
|
PASS | +35 10 -15
|
FAIL | -99 -99 -99
-25 failed model background comparison
-99 observation was missing
</pre>
If you wish to skip over these steps, you can change the
'surface_analysis.nl' namelist file.
A new check looks at the temporal history of the obs where a 24 hour
bias check flags temperature observations. Winds that aren't changing
in speed or direction over the 24 period are also flagged.
There is an additional check for all analyzed fields (except visibility)
within the 'spline' routine that rejects stations deviating from the
background by more than a threshold number of standard deviations of
the observation increments. This threshold can be independently
adjusted (i.e. tightened or loosened) for each field via the
'surface_analysis.nl' namelist. If you see any bulls-eyes in the
surface analysis that you don't believe, try contacting Steve Albers
at GSD for more information on making these quality control namelist
adjustments.
--------- 3.3.2.3 ------SURFACE ANALYSIS VERIFICATION-------------------
Verification statistics for the surface analyses are written to the
'log/qc/laps_sfc.ver.hhmm' files. These contain information of
obs differences relative to the background and the analysis. The
obs listed have had most of the QC checks already applied, though
an ob may have been rejected in the analysis by the final "spline
standard deviation" check yet still have a non-missing value listed
in the QC files.
--------- 3.3.2.4 ---------------STMAS-2D--------------------------------
Source directory: laps/src/mesowave/stmas_mg
Author: Yuanfu Xie (Yuanfu.Xie AT noaa.gov)
Parameter namelist files: static/stmas3d.nl, static/nest7grid.parms
------ 3.3.3 ---------------------TEMP----------------------------------------
Process: temp.exe - Temperature-Height analysis
Generate a temperature analysis using model background, sfc temp
analysis, and RASS data.
Quality control is applied to the temperature soundings. If any
level in a sounding differs from the model background by more than
a threshold (~10 deg), the entire sounding is rejected.
<pre>
Inputs: (from LGA, LSX, FUA [if available], LRS) '*' = essential input
* lga/fua grid model data analysis / forecast
lrs inter data RASS vertical temp profile
snd inter data sounding temperatures (RAOB/Dropsonde/
Satellite Sounding)
pin inter data ACARS Temperatures (using pressure altitude)
* lsx grid LAPS surface data grids
Outputs:
lt1 3d grid 3-D temperature (K), 3-D Heights (M-MSL)
pbl 2d grid 2-D Boundary Layer Depth (m), and BL top (Pa)
tmg inter data temperature obs used for lapsplot plotting
</pre>
Source directory: laps/src/temp
Further description and reference is at:
http://laps.noaa.gov/albers/laps/talks/temp/sld001.htm
------ 3.3.4 ---------------------CLOUD---------------------------------------
Process: cloud.exe - Cloud analysis package
Author: Steve Albers (Steve.Albers AT noaa.gov)
Several input analyses are combined with METARs of cloud layers. These
input analyses are the 3D temperature analysis, a three-dimensional
LAPS radar reflectivity analysis derived from full volumetric radar
data, and a cloud top analysis derived from GOES IR band eight data.
Vertical cloud soundings from METARs and pilot reports are analyzed
horizontally to generate a preliminary three-dimensional analysis.
This step provides information on the vertical location and
approximate horizontal distribution of cloud layers.
The satellite cloud-top temperature field is converted to a cloud-top
height field using the three-dimensional temperature analysis. The
cloud-top height field is then inserted into the preliminary cloud
analysis to better define the cloud-top heights as well as to increase
the horizontal spatial information content of the cloud analysis.
A set of rules is employed to resolve conflicts between METAR and
satellite data. Finally, the three-dimensional radar reflectivity
field is inserted to provide additional detail in the analysis.
<pre>
Inputs: '*' = essential input
* lsx grid LAPS surface data grids
* lt1 grid LAPS 3-d temperature/height grid
* vrc/v01/vrz grid 2-D or 3-D radar reflectivity
* lvd grid Infra-red and Visible satellite data
(requirements are adjustable)
pin inter data pireps/clouds
lm2 grid composite snow cover (prev hour normally)
* lga/fua grid model data analysis / forecast
* lso inter data surface (e.g. METAR) obs
Outputs:
lc3 3d grid (ht) 3d clouds (fractional cover)
lps 3d grid 3D Radar Reflectivity (filled in)
lcb 2d grid cloud base/top (LCB,LCT) - all clouds are
considered (> .1 cover). Heights are MSL.
cloud ceiling (CCE) - only areas analyzed
with a cloud fraction > 0.65 are considered.
Units are meters AGL.
lcv 2d grid column max cloud cover / snow cover
satellite/radar diagnostic fields
downward short wave radiation
</pre>
Source directory: laps/src/cloud
Parameter namelist files: static/cloud.nl, static/nest7grid.parms
Further description and reference is at:
http://laps.noaa.gov/albers/laps/talks/cloud/sld001.htm
------ 3.3.5 -----------WATER VAPOR (HUMIDITY PROCESSING)----------------
Last updated: 2/24/2006 by Daniel Birkenheuer
Code organization:
The moisture code is coordinated by the LQ3 modules all of which (with the
exception of libraries) exist under ./src/humid/. The main driver, lq3driver.x
contains only one subroutine call at this time.
./src/humid/lq3_driver1a.f (Module) <active>
is the primary moisture processing module that sequences the various
subroutines.
There is a second routine that formerly was used for HSM satellite
processing; it is currently deactivated:
./src/humid/lq3_driver1b.f (Module) <deactivated>.
Now, using the CRTM forward radiance model and more advanced techniques, the
satellite inclusion takes place in the above "1a" module. Treat the "1b" module
as orphan code. Furthermore, a FORTRAN 90 compiler is required to fully compile
the forward model along with the rest of the moisture analysis system.
Control file:
./data/static/moisture_switch.nl
An ASCII file intended for easy editing and control of the moisture modules
activities. The first record controls usage of RAOB data (0=off, 1=on). The
second record controls usage of satellite data (LVD files) and again (0=off;
8=on, use GOES-8, 9=on, use GOES-9). This module is exported with the RAOB
feature OFF and the satellite feature ON and SET FOR GOES-8. The third switch
enables (1) or disables (0) saturating air in cloudy areas. The fourth now
enables using sounder data in lieu of imager data (GOES only). This should be
set to (0) for the current time.
Data Particulars:
CLOUD DATA
A "switch" has been added to enable cloud data use to saturate air in cloudy
areas. This is included as the last item in the moisture_switch.nl file that is
maintained under the static area. To enable cloud data for saturating the air
this is (1) to disable the feature, set the character to (0).
You might wonder why we need such a switch. During October (1996), we
experienced problems with the cloud analysis. This was inadvertently causing
problems in the moisture analysis through the cloud saturation adjustment. The
incorrect moisture was in turn causing the models runs to fail. Hence we added
this switch so that we could easily reactivate the feature once the cloud
analysis was repaired without having to worry about recompiling any code.
RAOB DATA
The capability to ingest RAOB data into the moisture module has been available
since 1996.
There are two important items to know about:
1) The RAOB data are contained in lapsprd/snd/*.snd files. The moisture module
will automatically use .snd data if present. If you do not wish to use sounding
data there are 2 ways to exclude these data, the most obvious is to not provide
.snd files.
2) In the event that you wish to exclude the use of sounding data and want them
to be present in the data directory (possibly for some other application) you
can avoid using them in the moisture code by modifying the file:
./data/static/moisture_switch.nl
The first record of this ASCII file is used for the RAOB data inclusion. The
file itself is documented internally following the second record. If the first
record is "1" (nominal case), the use of sounding data will be on, and .snd
files will be processed if present. If this character is "0", the moisture code
will not process sounding data.
Input files:
Inputs (status as of August 1996) ("grid" designates LAPS netCDF grid file
unless otherwise stated):
<pre>
*LGA/FUA grid MAPS/RUC background analysis or forecast (FUA)
LSX grid LAPS surface analysis
LC3 grid LAPS 3-D clouds
*LT1 grid LAPS 3-D temperatures
SND ASCII RAOB observation file
LVD grid Satellite data from AWIPS NOAAPORT/SBN
<off> LH1 grid LAPS grid of VAS total precipitable water.
<off> LH2 grid 3 LAPS grids of VAS/radiometer modified precipitable
water.
Outputs (note LH3 contains 2 fields):
LQ3 grid 3D Specific Humidity (floating point number)
LH3 grid1 3D (RH3 field) Relative Humidity units of percent
0-100 (floating point number) with respect to
liquid water if ambient temperature is warmer
than 0 C, with respect to ice if ambient
temperature is equal to or less than 0 C
LH3 grid2 3D (RHL field) Relative Humidity units of percent
0-100 (floating point number) with respect to
liquid water at all temperatures
LH4 grid 2D Total precipitable water (meters) (floating point
number)
</pre>
PRIMARY ALGORITHM SUMMARIES
RAOB ENHANCEMENT:
The RAOB data are added to the analysis via a second pass Barnes analysis.
Normally, a Barnes analysis consists of two parts. The first fills the entire
domain with values weighted by the distance to the neighboring points. In the
second pass, a difference field (derived from the difference of the first pass
and the observations) is added to the result from the first pass with adjusted
weights to better tune to the scale of interest.
In this application we skip the first pass using instead the "background"
analysis in place of the result of the first pass Barnes' result. The
difference field is then generated and applied using a set of weights
appropriate for the LAPS domain resolution and density of observations.
SATELLITE ALGORITHM:
An essential ingredient of the variational method is a satellite forward
radiance model. The forward model produces a simulated radiance based on
temperature, moisture, and ozone profiles along with the temperature of the
surface or cloud top, and the pressure of that radiating surface (i.e.,
surface pressure or cloud top pressure whichever applies). Also needed are
the zenith angle, used to determine the air mass path and optical depth
between the radiator and the satellite. The forward model used for this work
was obtained from NESDIS. The forward model coefficients used for this study
were vintage late 1995.
In order to apply the forward model appropriately, a determination of clear
and cloudy fields-of-view (FOV) need to be determined. The LAPS cloud
analysis is used to identify clear and cloudy LAPS grid points. The analysis
as presented here is only working from FOVs classified as clear. Cloudy FOVs
probably can be used, but this is an early attempt at this technique, so a
conservative approach was chosen. Later research may focus on using a
combination of both clear and cloudy FOVs in the algorithm.
The first step in the algorithm is to assure all the data needed for proper
execution are present. These include channel radiances derived from AWIPS
imagery, the LAPS cloud analysis output, the LAPS surface temperature output,
and LAPS 3-D temperatures. The forward model also requires an ozone profile
along with moisture and temperature profiles above 100 hPa. These are gotten
from climatology since LAPS extends only to 100 hPa. The entire ozone
profile is provided by the forward model since LAPS does not analyze this
parameter.
Next, the forward model is run to verify "clear" LAPS gridpoints, where clear
is defined as those points in which both the modeled and measured GOES image
radiances in channel 4 (11 micron) agree to with 2K. This step uses the LAPS
thermal and as yet unmodified moisture profiles. Disparity in the channel 4
brightness temperature comparison indicates that the LAPS thermal profile is
too far off or perhaps it is really cloudy where the LAPS cloud analysis is
indicating it is clear. (It doesn't have to be totally cloudy for a
disparity to exist, it can be partially cloudy and this will still be
detectable in this difference test.) This is a conservative test; it really
goes beyond simple cloud detection though that is a likely cause of
differences, the forward model check is very sensitive and in many ways
eliminates any thermal profiles that subsequent variational technique will
find difficult to deal with. We are basically saying that we will not worry
about moisture adjustment unless the thermal profiles are reasonable. The
current LAPS system uses an older forward radiance model named OPTRAN and
this is now being switched to CRTM. However, this test is not that satellite
specific and the older OPTRAN model can be used by stating that newer
satellite data is of the vintage that OPTRAN uses. For this reason, the user
should not be that concerned with the exact satellite specified for this
process. At this point, all gridpoints offering promise of moisture
adjustment have been identified. If the domain is totally cloudy, the GOES
adjustment is discontinued and returns unmodified moisture values which are
passed to the final QC step. Assuming some gridpoints have been classified
as clear, the next step is a variational adjustment at those locations. The
functional evaluated at each gridpoint has and is best described in the
literature (see articles under http://laps.fsl.noaa.gov/cgi/birk.pubs.cgi.
Basically a funcational is minimized that differences the perturbed solution
against observation. The best perturbation is accepted as the "answer". The
first term in the functional maximizes agreement between the forward model
and observed radiance at the expense of only modifying the water vapor
profile. The second term adds stability and gives more weight to solutions
in which the coefficients departure from unity (no change to the initial
profile) is minimized. The stability term was discovered to be necessary
since without it some very good radiance matches were solved but with
unreasonable coefficients.
Note that differences in all three channels are minimized in this technique,
not only the moisture channel. Thus, any improvement in the "dirty window,"
channel 5, will also contribute to the solution. A variational technique is
used to minimize this function and typically requires three to 10 iterations
to converge. A limit of 50 iterations was set as the maximum number to
attempt. If limit was reached, that particular gridpoint was excluded and
treated as cloudy. Once the coefficients are determined, Laplace's equation
is solved for interior points for which coefficients have not been determined.
Then the entire domain is averaged using a spatial invariant filter; simply
averaging the values in a 3x3 gridpoint window, assigning that average to the
window's central grid location.
When the coefficients have been determined, they are applied to the specific
humidity field at each pressure level for which they are designated. The
modified specific humidity field is then advanced to the final analysis step.
As a final note it should be mentioned that owing to the unknown bias in
radiance data. If it is available, it is far better to use water vapor
gradient fields derived from satellite, than to assimilate satellite
radiances directly. If the GVAP option is turned on, it is recommended that
the direct assimilation is disabled in moisture_switch.nl. This is one
reason why direct radiance assimilation has been slow in development with
CRTM. Its value still can be reasonably questioned given the unknown bias.
It is far more straight forward to rid the system of bias by using the first
derivative structure of the radiance or PW field that to try to accept/deal
with the bias.
Reference: http://laps.fsl.noaa.gov/cgi/birk.pubs.cgi
GPS ASSIMILATION ALGORITHM:
One of many terms in the humidity variational minimization step, the
GPS total water is used to constrain the integrated water computed
every iteration. Like the other terms in the functional used in the
variational minimization, this term will reach a relative minimum when
the state variables and specifically Q, best match this and other
observations in a simultaneous manner (simultaneous here is respect to
heterogeneous observational fit and not the more traditional state
variable multi-variate solution sense).
The GPS algorithm traditionally read internal GSD netCDF files for
input. It now has the capability (12/2010) to read MADIS surface data
files for GPS data. The MADIS data are typically built every 5 minutes,
so to read GPS data from these files, one should look back to the prior
hour's MADIS file for the most recent GPS data. This is due to the
fact that typically the GPS data are not ready for use until about an
hour after acquisition time. So for a typical "20-min after the hour"
LAPS run, the current hour's MADIS file will not contain any GPS data.
The software is currently tuned to open the prior hour's MADIS file and
seek the "latest" GPS data that can be found in that file. This step
is required since the MADIS file will be adding GPS data to it as it
arrives and by the end of a given hour, MADIS files will contain 2
different GPS ingests.
Therefore the user, should be aware that if a LAPS start time other
than 20-min past the hour is chosen, the software may have to be
changed to acquire the most recent data. Right now things should be
pretty stable in this regard. If one starts LAPS at the top of the
hour, say 16 UT, this module will read the 15UT MADIS file for GPS data.
It will likely find the "latest" data in that file to have been written
about 15:20 UT. This is what the traditional read of internal GPS
files would have returned.
Furthermore, if the LAPS system starts at 20-min after the hour, the
same 15UT MADIS file will be opened and the GPS data read will be from
about 15:45UT. Again, the routine will find the same that the
traditional GPS file read would acquire. On the other hand, if one
were to run at 15:50UT, there is a chance to miss the latest GPS data.
The code as now written will open the 14UT MADIS file for data, when in
fact the 15UT MADIS file may at this time contain data from 15:20UT.
On the other hand, if this mistake is made, the data obtained will
likely be nearly within the nearest hour of analysis time (14:50UT) and
depending on the cycle time, may or may not be a critical issue. The
user will have to determine whether this can be tolerated.
Cautions for STMAS: When reading GPS data in a 4DVAR context, GPS data
reading will have to spend more time concerned with the actual data
time associated with the GPS data. In this regard, multiple MADIS
files will likely need to be opened, their contents matched with their
respective observation times, and then the data will need to be
temporarily stored, sorted, and processed according to the needs of
4DVAR.
------ 3.3.6 ---------------------DERIV---------------------------------------
Process: deriv.exe - Derived products
Author: Steve Albers (Steve.Albers AT noaa.gov)
These derived products are cloud, wind, stability, and fireweather
related.
<pre>
Inputs: '*' = essential input
* lc3 3d grid (ht) 3d clouds (fractional cover)
* lt1 grid LAPS 3-d temperature grid
lps 3d grid 3D Radar Reflectivity (filled in)
lsx 2d grid sfc pressure, temperature
lcv 2d grid column max cloud cover / snow cover
lh3 grid LAPS 3-d Relative Humidity (normally previous
file)
lso inter data Sfc (METAR) obs - for precip type verification
lw3 3d grid 3-D winds (U and V are wrt GRID NORTH)
vrc/v01/vrz 2-D or 3-D radar reflectivity
Outputs:
lcp 3d grid (pres) 3d clouds (fractional cover) (pressure grid)
cty 3d grid 3D cloud type (CTY)
threshold for cloud cover in CTY is 0.65.
pty 3d grid 3D precip type (PTY)
lwc 3d grids Cloud liquid water content (LWC)
Cloud ice content (ICE)
Hydrometeor Concentration (PCN)
Rain+Snow+Precipitation Ice Concentration
Rain Concentration (RAI)
Snow Concentration (SNO)
Precipitating Ice Conc. (PIC)
The last four are specific contents in
kilograms/meter**3. These can be converted to
mixing ratio if desired by dividing through
by the air density.
lil 2d grids Vertically integrated cloud fields:
Cloud Liquid (LIL)
Cloud Ice (LIC)
Cloud Optical Depth (COD)
Cloud Albedo (CLA)
Visibility (VIS)
lct 2d grid SFC precip type (SPT,PTT)
Types are:
0 - No Precip
1 - Rain "R"
2 - Snow "S"
3 - Freezing Rain "Z"
4 - Ice Pellets "I"
5 - Hail "A"
6 - Drizzle "L"
7 - Freezing Drizzle "F"
SPT uses simple 0 dbz reflectivity
threshold to define areas of precip.
PTT uses a 13 dbz threshold for non-snow
precip (~.01"/hr), 0 dbz is still used
for snow though a surface dewpoint
depression threshold is used to filter out
areas of snow virga not reaching the ground.
The latter may be more useful for display
purposes by end users. PTT also utilizes
METAR data (in concert with analyzed cloud
fraction) to delineate areas of drizzle,
freezing drizzle, rain, freezing rain,
and snow - in areas that radar does not
detect echoes.
SFC cloud type (SCT)
This is the type of the lowest cloud layer
in the CTY (3-D cloud type) file. The
cover threshold is 0.65. The presence of
a CB higher up has priority. There are
10 possible cloud types.
Types are:
0 - No Cloud
1 - Stratus "St"
2 - Stratocumulus "Sc"
3 - Cumulus "Cu"
4 - Nimbostratus "Ns"
5 - Altocumulus "Ac"
6 - Altostratus "As"
7 - Cirrostratus "Cs"
8 - Cirrus "Ci"
9 - Cirrocumulus "Cc"
10 - Cumulonimbus "Cb"
lmd 3d grid mean cloud drop diameter
lmt 2d grid max echo tops (LMT),
Low level reflectivity (LLR)
lco 3d grid Cloud omega - computed where cloud cover > .65
lrp 3d grid 3D icing index (integers 0-6)
0 is no icing
1 is light continuous
2 is mod continuous
3 is heavy continuous
4 is light intermittent
5 is mod intermittent
6 is heavy intermittent
lst 2d grids CAPE, CIN, and LI are calculated by lifting
a surface parcel taken from the LAPS surface
T and Td fields, as well as LAPS surface
(terrain following) pressure. LAPS 3-D
temperatures are also used.
CAPE (Convective Available Potential Energy)
This is a net positive energy, so any
negative area is subtracted from the
positive area.
CIN (Convective Inhibition)
Negative area in the sounding.
LI (Lifted Index)
Environmental minus parcel temperature at
500 mb.
SI (Showalter Index)
TT (Total Totals Index)
K (K Index)
LCL (Lifted Condensation Level)
WB0 (Wet Bulb Zero)
lhe 2d grid Helicity (Storm Relative Environmental). This
is integrated from the sfc to 3km AGL. It
is numerically equal to -2. times the hodograph
area.
A calculated storm motion vector is used
according to the Bunkers method used by the
National Weather Service. First a layer from
the sfc to 6km AGL is used to calculate the
mean wind. A shear vector through the sfc-6km
layer is also calculated. The storm motion
vector is assumed to equal the mean wind vector
plus a vector with a magnitude of 7.5 m/s and
a direction 90 degrees to the right of the
shear vector.
Mean Winds (sfc - 6km layer)
liw 2d grid log(LI*Omega) (partially derived from 3-D winds)
lmr 2d grid 2-D column max radar ref
lfr 2d grids Fire weather indices as follows:
HAH (High Level Haines Index)
HAM (Mid Level Haines Index)
FWI (Fosberg Fireweather Index)
VNT (Ventilation Index)
UPB (PBL Mean U Wind Component [grid-east])
VPB (PBL Mean V Wind Component [grid-north])
CWI (Critical Fire Weather Index)
</pre>
Source directory: laps/src/deriv
Further description and references are at:
http://laps.noaa.gov/albers/laps/talks/cloud/sld005.htm
and
http://laps.noaa.gov/albers/laps/talks/wind/sld007.htm
------ 3.3.7 ---------------------ACCUM---------------------------------------
Process: accum.exe - Snowfall/Liquid Equivalent Precipitation
Author: Steve Albers (Steve.Albers AT noaa.gov)
LAPS incremental/storm total snowfall/liquid equivalent accumulation.
<pre>
Inputs:
* lsx grid LAPS surface data grids
* lt1 grid LAPS 3-d temperature grid
lh3 grid LAPS 3-d Relative Humidity (normally previous)
lso inter data Rain gauge measurements of 1-hr precip
* vrc/v01/vrz grid 2-D or 3-D radar reflectivity
Outputs:
L1S 2d grid Snowfall over LAPS cycle time (S01 field)
Storm total snow accumulation (STO field)
Time interval is listed in the comment field.
Rain/Liquid Equivalent Precip (R01 field)
over LAPS cycle time.
Storm Total Rain/Liquid Precip (RTO field)
Time interval is listed in the comment field.
</pre>
Source directory: laps/src/accum
Parameter namelist file: static/nest7grid.parms
The precipitation analysis uses radar estimated precip rates as the primary
dataset. The radar reflectiivty can be obtained from any combination of NOWRAD
2-D (Section 3.2.4) or low-level reflectivity from mosaiced 2-D or 3-D radar
reflectivity data. The source can be narrowband or wideband radar
(section 3.2.3). The mosaics can be performed with either 2-D or 3-D inputs
(section 3.2.5).
We presently use a Marshall Palmer Z-R relationship to obtain liquid equivalent
precipitation. Snow is also estimated using a snow/rain ratio derived as a
function of column maximum temperature. More on the basic accumulation
processing is in Albers et. al., 1996.
In the present LAPS version, areas without radar coverage switch over to a
gauge only analysis of 1-hr precipitation - using a background or model first
guess field (if available) or zero field as a first guess.
Areas having both radar and rain gauges present can be bias adjusted. An
algorithm is presently be tested that determines this bias as a function of
reflectivity.
Reference: Albers S., J. Mcginley, D. Birkenheuer, and J. Smart 1996: The Local
Analysis and Prediction System (LAPS): Analyses of clouds,
precipitation, and temperature. Weather and Forecasting, 11,
273-287.
Available at http://laps.noaa.gov/frd-bin/LAPB.pubs_96.cgi
------ 3.3.8 ---------------------SOIL MOISTURE------------------------------
Process: lsm5.exe - Soil Moisture
Author: John Smart (John.R.Smart AT noaa.gov)
LAPS soil moisture and snow cover
<pre>
Inputs:
* lsx grid LAPS surface data grids
l1s grid LAPS surface precipitation
lcv grid LAPS satellite derived snow cover
lm2 grid Previous soil moisture information
Outputs:
LM1 3d grid Soil moisture (3-layer)
LM2 2d grids Snow Cover and additional Soil Moisture info
</pre>
This program is in the early stages of development and provides a three
layer analysis of soil conditions. The three layers are as follows:
<pre>
layer 1 (0-6in [0-0.152m])
layer 2 (6-12in [.152-.305m])
layer 3 (12-36in [.305-0.914m])
</pre>
A snow cover analysis is included.
The fractional snow cover is a composite over time of information from the
cloud analysis (visible and IR satellite), and snow accumulation (derived
mainly from radar). More documentation can be found within the source code
(e.g. soilmoisture5.f, calc_evap.f).
Note that a soil temperature analysis is not included at this time. The closest
thing we have to this is a single layer "ground temperature" analysis in the
LSX surface output file.
Source directory: laps/src/soil
------ 3.3.9 ---------------------BALANCE------------------------------------
Process: qbalpe.exe - "Quasi-geostrophic balance of height, wind and clouds.
authors: John McGinley/John Smart/John Snook/Ed Tollerud
contact: Edward.Tollerud AT noaa.gov
LAPS quasi-geostrophic balance of height and wind with temp adjustment.
Cloud fields are now balanced with the other fields.
<pre>
Inputs:
* lw3 3d grid LAPS wind analysis (grid north)
* lt1 3d grid LAPS height analysis
* lsx 2d grid LAPS sfc station pressure (PS field)
* lwc 3d grid LAPS Cloud Liquid/Ice/Precip
lh3 3d grid LAPS humidity
lco 3d grid LAPS Cloud Omega
* lga 3d grid Model first guess grids (including omega)
Outputs:
lt1 3d grid in lapsprd/balance/lt1 (ht and t field)
lw3 3d grid in lapsprd/balance/lw3 (u3, v3 and om field)
(grid north)
lh3 3d grid in lapsprd/balance/lh3 (rh field)
lq3 3d grid in lapsprd/balance/lq3 (sh field)
lsx 2d grid in lapsprd/balance/lsx (surface fields)
</pre>
Source directory: laps/src/balance
Parameter namelist file: 'static/balance.nl'
The balance package starts by inputting the results from a simple, offline
cloud model which retrieves liquid and ice partitioning and an estimate of
vertical motion from the observed clouds (lwc/lco). The variational scheme
is designed to accept cloud vertical motion estimates and ice and water
content as observations. The cloud observations are fully coupled to the
three dimensional mass and momentum field using dynamical constraints
which minimize the local tendency in the velocities and ensure continuity
is satisfied everywhere.
The scheme performs the analysis on the difference from an input model
background with the benefit that existing background model balances need not
be recreated each model cycle and that background model error daily compiled
is input explicitly on a gridpoint by gridpoint basis.
Reference: McGinley, J.A. and J.R. Smart, 2001: On providing a cloud-balanced
initial condition for diabatic initialization. Preprints, 18th Conf. on Weather
Analysis and Forecasting, Ft. Lauderdale, FL, Amer. Meteor. Soc.
------ 3.3.10 ---------------------STMAS3D------------------------------------
Process: STMAS3D.exe - Space-Time Mesoscale Analysis System in 3D
This analysis can be run with the other appropriate executables by
using the -V STMAS3D option in 'sched.pl'
------- 3.4 ---- Model Initialization & Postprocessing ---------------------
LAPS analyses are used to initialize various mesoscale models (e.g. WRF, MM5,
HIRLAM, BOLAM) to accomplish the prediction component. The forecast models
themselves are obtained separately from the LAPS analysis tar file. There is
some documentation for the model interfacing (for MM5) at this URL:
http://laps.noaa.gov/mm5doc/README_lapsmm5.htm
For the WRF model we have a flow chart that illustrates an example of the
initialization process: http://laps.noaa.gov/doc/HMT-m1.png
------ 3.4.1 --------------LAPSPREP------------------------------------------
Process: lapsprep.exe - Post-processes LAPS analysis files into formats
that can be used to initialize a local forecast
model (e.g., MM5, RAMS, WRF)
This process reformats LAPS data into files suitable for initializing
a mesoscale forecast model. The output format is controlled by the
"output_format" entry in lapsprep.nl and can be set to one of the
following:
output_format = 'wps'
This causes the program to output a file in the WPS format (as needed
for WRF version 3). Note that WPS has a constraint that the vertical
levels of the initial condition (LAPS) be matched with those from the
lateral boundary condition. This matching can be done either when
running LAPS or in the WPS processing steps by three methods.
1) In an example with the GFS as a lateral boundary condition one can
reduce the levels in the LAPS 'static/pressures.nl' namelist as in
this example: http://laps.noaa.gov/wps/pressures.nl.gfs
2) In an example with the NAM as a lateral boundary condition and if
LAPS has more analysis levels than the NAM one can use the WPS
utility 'mod_levs.exe'. It uses the 'namelist.wps' to rip out the
NAM levels not in the namelist. The NAM data has levels from 1000
to 100 at a 25 mb interval plus a surface level.
3) The third method is the most desirable option that we recommend.
We first run 'metgrid.exe' for the boundary conditions (e.g. NAM),
starting at the initial time and proceeding through the forecast
times. We then run 'real.exe' for the boundary conditions over
the entire period. This will produce 'wrfbdy_d01' and 'wrfinput_d01'
files. We next run 'metgrid.exe' for LAPS, followed by running
'real.exe' for LAPS only at the initial time. In this way the
'wrfinput_d01' (initial time file) will be overwritten by the LAPS
initial condition.
output_format = 'cdf'
Writes a NetCDF file of the output
output_format = 'wrf'
This causes the program to output a file in the WRF
Standard Initialization "grib_prep" format. These
files can be read by the WRF SI "hinterp" process.
output_format = 'mm5'
This causes the program to output a file in the MM5v3
pregrid (v4) format that can be read in by MM5 the "regridder"
pre-processor. See the NCAR MM5 REGRID documentation for
the format specification of this output file.
output_format = 'rams'
This causes the program to output a file in the RAMS 4.x
"RALPH2" format. These files can be read in by the
RAMS ISAN pressure stage process. Note that RALPH2 files
are in ASCII, so these files are actually human-readable.
See the RAMS RALPH2 format specification for
documentation.
There are three other namelist entries in the lapsprep.nl
file:
hotstart: Set to '.true.' if you wish to include the
cloud species from the cloud analysis in the
output files. This currently only applies
when output_format is equal to 'mm5' or
'wrf'.
balance: Set to '.true.' if you wish to use the
wind and temperature, height, and surface
analysis files from the balance package.
This will only work if LAPS is running the
balance package.
adjust_rh: Set to '.true.' if you wish to use the
adjusted RH analysis from the balance
directory.
This program essentially replaces part of the "dprep.exe"
functionality, in that it produces initial conditions files
for your local forecast model. If running a forecast model
in real time, then this program should be executed immediately
following the LAPS analysis during the hours in which the
model will be initialized. It can simply be run as the
last entry in sched.pl, which means you will always have
an initial condition file avaialble immediately following
your LAPS analysis.
To actually initialize a forecast model, you will still need
to run the appropriate program to build the lateral boundary
condition files, as LAPSPREP does not provide this function.
<pre>
Inputs: '*' = essential input
* lw3 grid LAPS 3-d wind analysis (grid relative)
U, V, Omega are written out
* lt1 grid LAPS 3-d temperature & height analyses
* lh3 grid LAPS 3-d relative humidity analysis
lq3 grid LAPS 3-d specific humidity analysis
lwc grid LAPS 3-d cloud/precip hydrometeor analyses
* lsx grid LAPS 2-d surface analyses
lm2 grid LAPS 2-d snow cover analysis
Outputs:
mm5_init:YYYY-MM-DD_HH grid MM5 init. file (pregrid v3 format)
ram_init:YYYY-MM-DD_HHMM grid RAMS init. file (RALPH2 format)
wrf_init:YYYY-MM-DD_HH grid WRF init. file (gribprep format)
</pre>
Parameter namelist file: 'static/lapsprep.nl'
Source directory: 'laps/src/lapsprep'
------ 3.4.2 --------------LAPS2GRIB------------------------------------------
Process: laps2grib.exe - Converts LAPS analysis output into a single GRIB2
file located in the 'lapsprd/gr2' subdirectory.
The parameters to convert are entered into a configuration
file; the choice of parameters and the scaling of the
parameters is controllable.
(author: Brent Shaw)
Parameter namelist file: 'static/laps2grib.nl'
lrun_laps2grib = .false. (default) or .true. (to create grib2 lapsprd/gr2 files)
Data file: 'static/laps2grib.vtab'
Source directory: 'laps/src/laps2grib'
E.g. parameters in data file: 'static/laps2grib.vtab'
3d 0,'lt1','t3 ', 1000.,110000., 1., 0,0, 0, 0
3d 0,'lt1','ht ', 1000.,110000., 1., 0,0, 3, 5
2d 'l1s','r01',1000.,4, 1, 0, 0,255,255,255,0, 1, 8
There are two numbers in the laps2grib.vtab file that immediately follow the
file name extension and variable name: the conversion factor and the scale
factor. The conversion factor will be multiplied by the LAPS variable coming
out of the file in order to make the units conform to WMO specs (e.g. like
cloud cover, which WMO defines as a percent from 0-100 whereas LAPS uses a
fraction of 0-1., so we set the conversion factor to 100). In the case of
precipitation, the units need to be in mm for GRIB, so if LAPS has precip
specified in meters, then your conversion factor needs to be 1000 so you can
get the data into mm. The scale factor specifies how many digits of precision
to preserve after the decimal. It can be negative (for example, -1 would have
precision to the nearest 10, 0 would give you to the nearest, and 1 gives you
1/10th, and so forth). So, if you have something that is typically very small
(say, mixing ratio which in kg/kg ranges from 0.0001 to about 0.01, you might
use a scale factor of 4 to preserve 4 digits after the decimal. On the other
hand, with cloud cover you may only need the nearest integer value from 1-100,
so you could use 0.
See laps/src/laps2grib/laps2grib.doc for more detailed information.
------ 3.4.3 --------------WFOPREP------------------------------------------
Process: wfoprep.exe - Processes AWIPS/WFO large-scale model forecast files
into formats that can be used as lateral boundary
conditions to initialize a local forecast model
(e.g., MM5, RAMS [SMF], WRF). The input files come from
the SBN and are in netCDF format.
(author: Brent Shaw)
This an optional program that can be used in the AWIPS/WFO environment.
In other environments you'll want to use a different program to generate
lateral boundary conditions.
Parameter namelist file: 'static/wfoprep.nl'
Source directory: 'laps/src/wfoprep'
------ 3.4.4 -------------- LFMPOST ------------------------------------------
Process: lfmpost.exe - Post-processes WRF/MM5 model forecast files into formats
that can be used to feed back into LAPS analysis or
plotting software.
authors: Linda Wharton, Brent Shaw, John Snook, Steve Albers, Isidora Jankov
contacts: Linda Wharton / Steve Albers
Parameter namelist file: 'static/lfmpost.nl'
Source directory: 'laps/src/newlfmp' (new default version)
'laps/src/lfmpost' (old version)
The default (newer) version of the lfmpost program consists of a Fortran
executable: $LAPSINSTALLROOT/bin/lfmpost.exe
This has been tested so far with WRF version 3.
There is also an "old" version of lfmpost. To build this version run
'make' and 'make install' in the 'src/lfmpost' directory. Then
in $LAPS_DATA_ROOT/static (or the template) copy 'lfmpost_old.nl' to
'lfmpost.nl'.
LFMPOST is used to post-process raw model output files from the
following models:
1. MM5 (Version 3 binary output format)
2. WRF (NCAR EM core, Version 1.3 netCDF output format) [old lfmpost.exe]
3. WRF (NCAR EM core, Version 2 netCDF output format) [old lfmpost.exe]
3. WRF (Version 3) [new lfmpost.exe]
It performs the following functions:
1. Read in model output for each time
2. Destagger variables to LAPS grid points
3. Vertically interpolation to isobaric levels
4. Output various formats, including LAPS fua/fsf netcdf
format, Vis5D format, GRIB-1, and tabular text
point forecast files.
It is controlled by the namelist file "lfmpost.nl". If processing
point forecasts, you also need to set up "lfmpost_points.txt". Samples
of these two files can be found in your LAPS_SRC_ROOT/data/static
directory. To use lfmpost, you will need to copy these two files into
MM5_DATA_ROOT/static or MOAD_DATAROOT/static (for MM5 or WRF, respectively)
and edit them to your liking. If you are going to output LAPS fua/fsf
files with lfmpost, you will need a valid LAPS_DATA_ROOT for the same
model domain, and your pressure levels selected in lfmpost.nl must be
the same levels selected in LAPS_DATA_ROOT/static/pressures.nl. Note that
for this option, lfmpost expects the horizontal domain (dimensions,
projection, etc.) to identically match for LAPS and the model being used.
To execute lfmpost, you should set the following environment variables
as necessary:
MM5_DATA_ROOT (if running MM5)
MOAD_DATAROOT (if running WRF)
LAPS_DATA_ROOT (if fua/fsf output is desired)
LFMPOST expects to find the raw output files in:
MM5_DATA_ROOT/mm5prd/raw (MM5)
MOAD_DATAROOT/wrfprd (WRF v1 and v2)
Output from lfmpost goes into:
MM5_DATAROOT/mm5prd/d##/ (for MM5)
MOAD_DATAROOT/wrfprd/d##/ (for WRF)
Within the output directories, the following subdirectories need
to exist to contain the specific output formats:
fsf -> For LAPS fsf files (2d and surface fields)
fua -> For LAPS fua files (3d isobaric output)
grib -> GRIB data [old lfmpost.exe]
points -> Tabular text point forecasts
v5d -> Vis5D files
After setting the appropriate environment variables and ensuring your
namelists are configured properly, the syntax (old lfmpost.exe) is:
lfmpost.exe NAME DOMNUM
where NAME is one of "mm5", "wrf", or "wrf2" for MM5, WRFv1.3,
or WRFv2, respectively. DOMNUM is the nest to process.
Additional arguments are needed for the new (default) version of
'lfmpost.exe'.
lfmpost.exe NAME WRFOUT NEST RCTIME FCSTTIME LAPS_DATA_ROOT
NAME - one of "mm5", "wrf", or "wrf2" for MM5, WRFv1.3,
or WRFv2, respectively.
WRFOUT - full filename of WRFout file
NEST - nest number (1 is outer)
RCTIME - CTIME in seconds of model initialization
FCSTTIME - number of seconds into the forecast
LAPS_DATA_ROOT - LAPS_DATA_ROOT where static files are set up
(or the equivalent in the WRF directory)
LFMPOST is designed to operate on "incremental" raw model output data, so
when you run WRF, be sure to output each time period to a separate file.
When running lfmpost in real-time for WRF output there is a Perl script that
can be used:
$LAPSINSTALLROOT/etc/lfmpost.pl (use wrfpost.pl for older lfmpost.exe)
There is also a driver script located in 'etc/models/lfmpost_test.csh', often
used for non-realtime case runs, that can be executed as in this example:
./lfmpost_test.csh $LAPSINSTALLROOT $LAPS_DATA_ROOT $RUNTIME mvoutput
RUNTIME is model initialization time with format 'yyyymmddhh'
------ 3.4.5 -------------- FORECAST GRAPHICS -------------------------------
A script can be run in cron (after the FUA/FSF files are created) to make
GIF images of various forecast fields. This is located in
'etc/followup_fcst.pl'. Output images will appear in 'lapsprd/www/fcst2d'.
------ 3.4.6 -------------- VERIFICATION ------------------------------------
LAPS has a built-in verification package. This can be run after a model
is run and the verifying observations and analyses are available. The
driver script is in 'etc/verif/verif_fcst_driver.csh'. For real time runs
it can be run via cron once for each model cycle. The script has
several command line arguments that are described in comments at the top.
The script will produce stats files and PNG/GIF image output in the
'lapsprd/verif' directory tree. To help in setting this up or troubleshooting
the results please note the input data that are being used:
1) FUA/FSF forecast files should be located in
$LAPS_DATA_ROOT/lapsprd/f??/[model]/*.f??'
2) Observation and analysis files should be located in various other
'lapsprd' subdirectories. Some examples are as follows:
<pre>
LSO - surface obs
LT1 - 3-D Temperature
LMR - Composite Reflectivity
LPS - 3-D Reflectivity
LCV - Solar Radiation (GHI)
</pre>
3) Several parameters are relevant in 'static/nest7grid.parms' including:
model_cycle_time, model_fcst_intvl, model_fcst_len, fdda_model_src
4) Log files are in $LAPS_DATA_ROOT/log/[*fcst*][*verif*]
5) Animated montages are an option that will work if 'followup_fcst.pl'
is run prior to the verification (see previous sub-section 3.4.5).
-------- 4.0 -- Porting code mods from LAPS users back to GSD --------------
We would like to encourage suggestions from LAPS users on how to improve
LAPS, both scientifically and in the software itself. The changes should
be made by downloading the most recent source code tree. Edit your changes
in the source files, and then retar part or all of the source tree to send
back to us. Please state the LAPS version number you had used. Any
documentation pertaining to the reasoning behind the changes would be
appreciated.
In some cases, a less formal process may be easier to go by. Here, the
user can provide documentation of suggested mods either in descriptive form,
or in terms of before and after code. The code author can then implement
the changes in the GSD version. This can be useful in the event the mods are
simple, or if the user has been working with a relatively old version of
the software and/or there have been significant recent GSD mods to the
software. This can also be useful if the user has an idea of a desired
functionality within LAPS, but has not actually looked at the software
details associated with implementing the functionality.
------- 5.0 -- LAPS Output Variables and netCDF File Organization -----------
LAPS Variables and netCDF File Organization
LAPS output is written in netCDF format as summarized below. Each file
extension contains a set of variables that goes into a separate directory under
'$LAPS_DATA_ROOT/lapsprd/'. This directory includes so-called pre-balanced
files, while the final balanced output is in the 'lapsprd/balance' subdirectory.
For example the LT1 temperature grid is written with the pre-balanced version
in 'lapsprd/lt1' and the balanced version in 'lapsprd/balance/lt1'.
Map projection attributes are specified in the NetCDF files. Here are some
of their definitions:
Lat1: latitude of lower left corner grid point
Lon1: longitude of lower left corner grid point
Lov: longitude on map projection where grid is oriented along true north-south
Note that netCDF information on the units of the fields, etc. is contained in
the '$LAPS_DATA_ROOT/cdl/*.cdl' files. At the bottom of the list is a section
on the "intermediate" files that are computed while the ingest is running.
File LAPS CDF Num
Ext Var Var Lvl Field
Process surface:LSX U su 1 Surface (10m) wind u (grid north)
V sv 1 Surface (10m) wind v (grid east)
P fp 1 Reduced Pressure (constant height sfc)
PP pp 1 Perturbation Pressure (if available)
T st 1 Temp (2m)
TD std 1 Dewpt Temp
TGD tgd 1 Ground Temp (land surface/SST)
VV vv 1 Vertical Velocity
RH srh 1 Relative Humidity
MSL mp 1 MSL Pressure
TAD ta 1 Temp Advection
TH pot 1 Potential Temp
THE ept 1 Equivalent Potential Temp
PS sp 1 Station Pressure (terrain following)
VOR vor 1 Vorticity
MR mr 1 Mixing Ratio
MRC mc 1 Moisture Flux Convergence
DIV d 1 Divergence
THA pta 1 Potential Temp Advection
MRA ma 1 Moisture Advection
SPD spd 1 Surface Wind Speed
CSS cssi 1 CSSI
VIS vis 1 Surface Visibility
FWX fwx 1 Fire Danger (LAPS / Kelsch)
HI hi 1 Heat Index
Process temp: LT1 T3 t 21 Temperature
HT z 21 Height (geopotential meters)
PBL PTP ptp 1 Boundary Layer Top (pressure)
PDM pdm 1 Boundary Layer Depth (in meters)
Process accum: L1S S01 s1hr 1 Snow Accum Cycle
STO stot 1 Snow Accum Storm Tot
R01 pc 1 Liq Accum Cycle
RTO pt 1 Liq Accum Storm Tot
Process humid: LQ3 SH sh 21 Specific Humidity
LH3 RH3 rh 21 Relative Humidity
RHL rhl 21 Relative Humidity with resp to liquid
LH4 TPW tpw 1 Integrated Total Precipitable Water Vapor
Process wind: LW3 U3 u 21 Wind u (wrt GRID NORTH)
V3 v 21 Wind v (wrt GRID EAST)
OM om 21 Wind omega
LWM SU u 1 Surface wind u (wrt GRID NORTH)
SV v 1 Surface wind v (wrt GRID EAST)
Process cloud: LC3 LC3 camt 42 Fractional Cloud Cover (levels 1-42)
LCB LCB cbas 1 Cloud base
LCT ctop 1 Cloud Top
CCE cce 1 Cloud Ceiling
LCV LCV ccov 1 Cloud Cover
CSC csc 1 Cloud Analysis Implied Snow Cover
ALB 1 LAPS derived albedo
S3A 1 3.9u satellite data
S8A 1 11u satellite data
RQC 1 Radar QC information (2D vs 3D)
SWI 1 Downward Shortwave Radiation
LPS REF ref 21 LAPS Radar Reflectivity
Process deriv: LCP LCP ccpc 21 Fractional Cloud Cover Pressure Coord
LWC LWC lwc 21 Cloud Liquid Water
ICE ice 21 Cloud Ice
PCN pcn 21 Hydrometeor Concentration
RAI rai 21 Rain Concentration
SNO sno 21 Snow Concentration
PIC pic 21 Precipitating Ice Concentration
LIL LIL lil 1 Integrated Liquid Water
lic 1 Cloud Ice
cod 1 Cloud Optical Depth
cla 1 Cloud Albedo
vis 1 Visibility
LCT PTY spt 1 Sfc Precip Type
PTT ptt 1 LAPS Sfc Precip Type
SCT sct 1 Sfc Cloud Type
LMD LMD mcd 21 Mean Cloud Drop Diameter
LCO COM cw 21 Cloud omega
LRP LRP icg 21 Icing Index
CTY CTY ctyp 21 Cloud Type
PTY PTY ptyp 21 Precip Type
LMT LMT etop 1 Max Echo Tops
LLR llr 1 Low Level Reflectivity
LST LI li 1 Lifted Index
PBE pbe 1 Positive Bouyant Energy
NBE nbe 1 Negative Bouyant Energy
SI si 1 Showalter Index
TT tt 1 Total Totals Index
K k 1 K Index
LCL lcl 1 Lifted Condensation Level
WB0 wb0 1 Wet-Bulb Zero
LWM SU u 1 Surface wind u (grid north)
SV v 1 Surface wind v (grid east)
LHE LHE hel 1 Helicity
MU mu 1 Mean wind u (grid north)
MV mv 1 Mean wind v (grid east)
LIW LIW liw 1 log(LI*omega)
UMF umf 1 Upslope Component of Moisture Flux
LMR R mxrf 1 Column Max (Composite) Radar Reflectivity
LFR HAH hah 1 High Level Haines Index
HAM ham 1 Mid Level Haines Index
FWI fwi 1 Fosberg Fireweather Index
VNT vnt 1 Ventilation Index
UPB upb 1 PBL Mean Wind U-component (grid north)
VPB vpb 1 PBL Mean Wind V-component (grid east)
CWI cwi 1 Critical Fire Weather Index
Process soil: LM1 LSM lsm 3 Soil Moisture
LM2 CIV civ 1 Cumulative Infiltration Volume
DWF dwf 1 Depth to wetting front
WX wx 1 Wet/Dry grid point
EVP evp 1 Evaporation Data
SC sc 1 Snow cover
SM sm 1 Snow melt
MWF mwf 1 Soil Moisture content Wetting Front
LAPS Fcst Model:
FUA U3 ru 21 Fcst Model Wind u (grid north)
V3 rv 21 Fcst Model Wind v (grid east)
HT rz 21 Fcst Model Height (geopotential meters)
T3 rt 21 Fcst Model Temperature
SH rsh 21 Fcst Model Specific Humidity
FSF USF usf 1 Fcst Model Surface wind u (grid north)
VSF vsf 1 Fcst Model Surface wind v (grid east)
TSF tsf 1 Fcst Model Surface Temperature
DSF dsf 1 Fcst Model Dewpoint
RH rh 1 Fcst Model Relative humidity
LCB lcb 1 Fcst Model Cloud base
LCT lct 1 Fcst Model Cloud top
P p 1 Fcst Model 1500m pressure
SLP slp 1 Fcst Model MSL pressure
PSF psf 1 Fcst Model Surface pressure
LIL lil 1 Fcst Model Integrated cloud liquid water
TPW tpw 1 Fcst Model Total precipitable water vapor
R01 r01 1 Fcst Model Liquid accum cycle
RTO rto 1 Fcst Model Liquid accum storm total
S01 s01 1 Fcst Model Snow accum cycle
STO sto 1 Fcst Model Snow accum storm total
TH th 1 Fcst Model Potential temperature
THE the 1 Fcst Model Equivalent potential temp
PBE pbe 1 Fcst Model Positive buoyant energy
NBE nbe 1 Fcst Model Negative buoyant energy
PS ps 1 Fcst Model Surface pressure
CCE cce 1 Fcst Model Cloud ceiling
VIS vis 1 Fcst Model Visibility
LCV lcv 1 Fcst Model Cloud cover
LMT lmt 1 Fcst Model Max echo tops
SPT spt 1 Fcst Model Sfc precip type
LHE lhe 1 Fcst Model Helicity
LI li 1 Fcst Model Lifted index
HI hi 1 Fcst Model Heat index
SWI swi 1 Downward Shortwave Radiation
SWO swo 1 Fcst Model Outgoing Shortwave Radiation
LWO lwo 1 Fcst Model Outgoing Longwave Radiation
FWI fwi 1 Fcst Model Fosberg fire weather index
FWX fwx 1 Fcst Model Kelsch fire weather index
RSM LSM lsm 11 Fcst Model Soil Moisture
Intermediate LAPS files:
Process vrc_driver:
VRC REF ref 1 NOWRAD 2D radar reflectivity
Process mosaic_radar:
VRZ 21 (3D reflectivity mosaic)
Process remap: V01 REF refd 21 Radar reflectivity
VEL veld 21 Radial Velocity
NYQ nyqd 21 Nyquist velocity
files V02, V03, V04, V05, V06, V07, V08, V09, V10, V11,
V12, V13, V14, V15, V16, V17, V18, V19, V20 same format
Process lga.exe (background model):
LGA HT ht 21 Model isentrop height interp to LAPS
isobaric (geopotential meters)
T3 t 21 Model isentrop temp interp to LAPS isobaric
SH sh 21 Model specific humidity
U3 u 21 Model u wind component (grid north)
V3 v 21 Model v wind component (grid east)
OM om 21 Model vertical velocity (Pascals/second)
LGB USF usf 1 Model Surface wind u (grid north)
VSF vsf 1 Model Surface wind v (grid east)
TSF tsf 1 Model Surface Temperature
TGD tgd 1 Model Ground Temperature
DSF dsf 1 Model Dewpoint
SLP slp 1 Model MSL pressure
PSF psf 1 Model Surface pressure
RSF rsf 1 Model Specific Humidity
P p 1 Model reduced pressure
PCP pcp 1 Model Precipitation
Process lvd_sat_ingest:
LVD S8W s8w 1 GOES IR band-8 bright temp warmest pixel
S8C s8c 1 GOES IR band-8 bright temp coldest pixel
SVS svs 1 GOES visible satellite - raw
SVN svn 1 GOES visible satellite - normalized
ALB alb 1 albedo
S3A s3a 1 GOES IR band-3 bright temp averaged
S3C s3c 1 GOES IR band-3 bright temp filtered
S4A s4a 1 GOES IR band-4 bright temp averaged
S4C s4c 1 GOES IR band-4 bright temp filtered
S5A s5a 1 GOES IR band-5 bright temp averaged
S5C s5c 1 GOES IR band-5 bright temp filtered
S8A s8a 1 GOES IR band-8 bright temp averaged
SCA sca 1 GOES IR band-12 bright temp averaged
SCC scc 1 GOES IR band-12 bright temp averaged
Note: band-8 is approx 11.2 microns.
Static LAPS file - run by localization:
gridgen_model.exe: creates file 'static.nest7grid'
LAT 1 Latitude (degrees)
LON 1 Longitude (degrees)
AVG 1 Mean elevation MSL (m)
STD 1 Unused
ENV 1 Unused
ZIN 1 Z coordinate - used for plotting in AVS
LDF 1 Land Fraction (0=water,1=land)
USE 1 Landuse
About
mirror source for LAPS (version 0-57-17)
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published