Chapter 1
Introduction
 

       This documentation describes the 1997 unix version of the NCEP (National Centers for Environmental Prediction) RSM (Regional Spectral Model). The documentation describes several how-to?s. How the model is distributed in the Unix file system? How to set up the initial configuration to start up the RSM? How to design and run an RSM experiment? How to display RSM output? What the relevant input and output files are, and where are the model major routine descriptions?

      The RSM is a limited area atmospheric numerical model system, which used primarily for daily weather forecasts, and climate simulation or forecasts. The initial motivation is from the authors' experiments with a limited-area grid-point model. Since spectral computation has higher accuracy in term of gradients and spectral interpolation. We switched from grid-point model to spectral model. The difficulty to do limited area model with spectral method is overcome by the time-dependent perturbation method. The extension of this perturbation method is used to step into the non-hydrostatic model from the existed hydrostatic model. Thus current version has hydrostatic and non-hydrostatic version of RSM. For simplicity and convenience, the hydrostatic version we call it RSM as usual and the non-hydrostatic version is called MSM (meso-scale spectral model).

      Current version distributed has two major directories; one is machine independent part another one is machine dependent. The machine independent part is under a common directory and the users should not change it, the model developers or any advanced users may upgrade it. We suggest that users should contact the model developer or authors to make any suggestion or modification inside this part. The machine dependent part has to be configured by the user once after installation and before any experimental try. All the distributed directories provide the possibilities of the model configuration for either run global only, run regional using global files, run regional using regional files, run hydrostatic regional mode, or nonhydrostatic mode. The other distributed machine-dependent directories are utilities for creating machine-dependent constant files, small scripts to read or check files, packing or unpacking GRIB files, and the display system called GrADS.

      Besides the descriptions of the file system and how to run the model in sections 2 to 4, we also provide some details of the model preprocessor, forecast and postprocessor related their source codes for advanced users. We will describe more in details how to change the source code and put it into the system without disturbing the common place.

      Any institution can get this system with free, except a request letter describes the purpose of the use is required to send to the Director of the NCEP. And all the material can be used as it is, we are not responsible for any installation, debugging and further maintenance. However, the collaboration between institutions with NCEP can be arranged through existed project. Any suggestions should be sent to either one of the developers. Hann-Ming Juang (henry.juang@noaa.gov), or Masao Kanamitsu (masao.kanamitsu@noaa.gov) or Song-You Hong (songyou.hong@noaa.gov).

      After you get permission from our institution, the system can be obtained either by an ftp download or by an untar from a CDROM or a 4 mm or 8 mm DAT tape. Please refer to the ftp manual and tar manual for details of their uses. For simplicity, we give you an example for ftp and untar from tape.

(1) From ftp in your Unix system which should be in network as

ftp nic.fb4.noaa.gov

user> anonymous

password> yourid@machine.orgonization.domain

ftp> cd   /pub/rsm/port

ftp> binary

ftp> get common97.tar.Z

ftp> get   [machine]97.tar.Z

ftp> get example.tar.Z

ftp> quit

Then you can uncompress and untar them in your unix system by

unix % uncompresses   common97.tar

unix % tar xvf   common97.tar

unix % uncompress   [machine]97.tar

unix % tar xvf   [machine]97.tar

where italic type characters indicate what you should type. Depending upon which machine you will be using, the [machine] can be sgi, hp, ibm, sun, or linux etc.

(2) From tape in your unix system unix % tar   xvf   /dev/rmt0l

Once you have unloaded all of the files from the tape, uncompress and untar the example.tar.Z, common97.tar.Z and [machine]97.tar.Z. If it is distributed on an SGI tar tape, you can do following to untar from tape on a non-SGI system, such as a Dec alpha system, by the using the command

unix % dd  if=/dev/rmt0l   bs=20b   conv=swab    tar   xvf   -

where /dev/rmt0l is an example of the tape point in Dec alpha system. This device could be /dev/0mn in the HP system, and /dev/rmt/tps0d6 in the SGI system.

After you have unloaded the entire system, you can read through the next section to find out what they are inside.


webmaster: Hann-Ming Henry Juang
henry.juang@noaa.gov