Seasonal forecasts by Liqiang Sun


file package for the class:

Log in to one of Scripps machines and type bash

Create a directory for your work and copy rsm-fcst-course.tar and untar.

To view observed climatology of precip anomaly MAM
go to map
open obs.pcp71-00.mam.ano.ctl
set lat -20 0
set lon -50 -30
set mpdset ../grads/mres
set gxout grfill
d obs

To create contingency table
go to prog directory
pgf90 table.f
run the executable
output will be in output directory
to view output, go to map and open rsm.pcp71-00.mam.contable.ctl




A: above-normal, N: normal, B: below-normal
e.g., AA means RSM forecasts above-normal and observation is above-normal too

March 2004 forecast run
go to prog
pgf90 fcst-tercile.f
run the executable - rsm.2004.mam.02asst.tercile is created in output
pgf90 obs-tercile.f
run the executable - obs.pcp2004.mam.tercile is created in putput
go to map and open grads in portrait mode (grads -p)
and run to plot a figure

To correct forecast
go to prog
pgf90 c-tercile.f
run the executable - crsm.2004.mam.02asst.tercile is created in output
go to map and open grads in portrait and run

go to prog
pgf90 rpss.f
run the executable - rpss.2004.mam.02asst is created in output
go to map and run

Demonstration/experiments - Hideki Kanamaru and Haiqin Li

  • G-RSM on Linux cluster - demonstration by John Benoit
  • Testing different modes of G-RSM - experiments led by Hideki Kanamaru and Haiqin Li
  • Comparison of cloud parameterization schemes - lecture and experiments led by Song-you Hong
  • G-RSM system update (new system) - demonstration on COMPAS by Masao Kanamitsu
  • RSM-ROMS coupling - demonstration by Masao Kanamitsu

------ screen -----
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20

workstation 1 to 4 connect to
5 to 8 connect to
9 to 12 connect to
13 and 14 to
15 to 18 to
19 and 20 to

Log in your machine.
Open Xmanager on your Desktop and double click Xstart and choose a correct session profile (bagamoyo, gewex, etc)
username: maui2009
password: will be given in class

We set up CVS path etc already. Type bash in the terminal.

Create your own directory for your experiment.

50x50 grids on 60km resolution takes about 10-15 mins on single processor

Testing different modes of G-RSM

cvs co Install
./inst rsm_latest (replace with your choice) lo

  • rsm_latest: RSM latest (trunk) version
  • crsm_frozen: c2r version of RSM (CRSM) frozen version
  • gsm_latest: GSM latest (trunk) version
  • scm_latest: SCM (single column) latest (trunk) version
  • gdas: global data assimilation
  • and others

---- scm version
you need to change scm/configure-model
you have to add "albaer", "cnvaer", and "cnvalb" after "elif [ $MODEL = scm ] ; then" line.

--- to use Grads
$ source .bashrc
use $ grads on bagamoyo and ceop
use $ gradsc on other machines

and use libs/etc/grmap to create .ctl file from GRIB files

---- for advanced users; if you want to try scm, crsm and gdas use this version
new version of ECPC G-RSM can be installed by
cvs co -r kana_roms_branch_20090504 INSTALL
Use ./install script instead of ./inst for this new version

--- cloud water prediction
edit rsm/def/physics.h
CLD1 default
for CLD2, 3, 5, 6, T93, turn on DIFUC
for CLD3, 5, 6, turn on VCI

after editing def/physics.h
make clean

and run the model

See also