SSGM clouds
July 2003

SSGM will be used to make test scenes with moving clouds.

We used SSGM in 1999 to model some data for AGS - now there is another modeling task to import a time series of clouds into SSGM.

The SSGM cloud databases are stored in ssgm/modelin/cldsim. Two types of information are needed to set up the cloud - spectral information and spatial information. The spectral information is stored in files that end with 'db_v#', where # is the version number of the spectral database. The cloud spectral databases we currently have are:

Name version min - max
wavelength
(µm)		 cloud base
height
(km)		 cloud top
height
(km)		 file date
advfog.db_v220.25 - 16.80.1 0.14Mar 28 11:46
alto4.db_v2 20.25 - 12.54.0 4.50Dec 13 1997
alto6.db_v2 20.25 - 12.56.0 6.50Dec 13 1997
cirrus.db_v220.25 - 12.58.0 9.00Dec 13 1997
cumulo.db_v220.25 - 12.52.511.00Dec 13 1997
cumulo.db_v330.25 - 16.82.511.00Mar 15 2000

I believe we are missing several version 3 databases, which should extend the spectral information out past 12.5 um.

The spatial cloud distributions for any particular scene are stored in three files that give the cloud height, type index and type index definition. For example here is the VAASA_MID.DBD_v3 file:


*
* CLDDRV 3.00 DATA BASE DESCRIPTION FILE
*
HEADER/TITLE .......................... Aerospace subset of VAASA.
CLOUD ALTITUDE FILE (METERS, I*2) ..... VAASA_MID.ALT
CLOUD TYPE FILE (INDEX, BYTE) ......... VAASA_MID.TYP
SPATIAL RESOLUTION X AND Y (KM) ....... 0.2 0.2
NUMBER OF PIXELS (X, Y) ............... 2560 5850
MAXIMUM CLOUD ALTITUDE (KM) ........... 10.00
NUMBER OF CLOUD TYPES (MAX = 5) ....... 2
BRDF FILE NAME, CLOUD TYPE 1........... ./cldsim_input/alto6.db_v3
CLOUD BOTTOM ALTITIDE, CLOUD TYPE 1 ... .50
SPECULAR REFLECTANCE FLAG (T/F) ....... F
BRDF FILE NAME, CLOUD TYPE 2........... ./cldsim_input/cumulo.db_v3
CLOUD BOTTOM ALTITIDE, CLOUD TYPE 2 ... 1.50
SPECULAR REFLECTANCE FLAG (T/F) ....... F

It defines the files sizes and clouds types for the VAASA_MID cloud scene. The 'TYP' file gives a cloud type index for each pixel, the indices are defined in the example file as 1=alto6 and 2=cumulo. The cloud bottom height is constant for an particular scene, but the cloud top height is set on a per pixel basis in the 'ALT' file.

Looking back at old notes I made on running SSGM, I noticed that there are consistancy tests that should be run to verify the installation gives expected results. The next few steps for to start setting up a run will be:

  1. I will find and run the installation test
  2. Contact Shanks and/or Goldspiel and get version 3 cloud spectral databases? - until the installation test is run, we won't be able to tell if there are other model components that we are missing.
  3. Dennis needs to decide on the test spectral response filters (SRFs) for whatever bands we will be modeling -
  4. Decide on how to make background terrain - we'll need a classification image of the scene and a BRDF database of the classes - I'm not sure if we need to pick from a default set of classes or if we can make up our own.

July 24

Acceptance test is not finding some files it needs, before reinstalling from the '97 cds, I'm going to try to get a more recent set ('99) from the Navy - sent an email to user_support@vader.nrl.navy.mil

Aug 30

Dennis has received the SSGM '99 installation cd's and I installed them, the 99.3 update and the AGS files from Photon Research. The cloud databases are now at version 3 and do extend out to 16.8 um. I modified the AGS KIEV_N_STRETCH cloud database to point to the version 3 databases instead of version2.

Name version min - max
wavelength
(µm)		 cloud base
height
(km)		 cloud top
height
(km)		 file date
advfog.db_v220.25 - 16.80.1 0.14Mar 28 11:46
alto4.db_v3 30.25 - 16.83.5 4.00Dec 7 1998
alto6.db_v3 30.25 - 16.85.5 6.00Dec 7 1998
cirrus.db_v330.25 - 16.88.0 9.00Dec 7 1998
cumulo.db_v330.25 - 16.82.511.00Mar 15 2000

The installation test for '99 has successfully run and the '99.3 test is currently running.