This directory contains routines for comparing model data with observations. The
routines are:

argocmp          : Reads a argo profile netcdf file and interpolates hycom data
                   to argo positions.  Data is dumped to text files in directory
                   ArgoCmp.  Can also feed the hycom data into the argo profile file. 

argo_extract_profiles_new.sh  : Automates the processing of argo profiles with
                   argocmp. Input is a hycom file (DAILY or WEEKLY), and the script
                   will look for argo profiles for a matching date locally (on disk)
                   or remotely (ifremer ftp site). Upon successful download, the 
                   argocmp routine is run.

argo_extract_profiles.sh  :  Same as the above, but only for DAILY files.

argoreduce       : Uses output data from argocmp (or the two shell scripts
                   above) and calculates RMS and bias for temperature and salinity
                   for a given region.  Depth intervals and time interval(s) are 
                   given as arguments to the program, whereas the region definitions 
                   are specidief by the file "regiondefs.in". Output is files named 
                   "argoreduce.XX.locations" and "argoreduce.XX.timeseries" where 
                   "XX" are the names. see more info on regiondefs.in below

mldcmp           : Compares mixed layer climatology to model fields. The
                   climatology is the de Boyer Munteguet climatology. The path to 
                   this dataset must be specified in environment variable "MLD_PATH". 
                   This routine also  uses the regiondefs.in file to create
                   regional statistics

slacmp           : Compares observed sla fields to model fields. The sla fields
                   are those given by the Aviso/DUACS merged msla data sets. The
                   routine can use opendap access if enabled in netcdf
                   libraries, in this case only a hycom file is specified. it
                   can also use netcdf files from DUACS ftp, in this case 
                   the hycom file and duacs msla file is given as input.

slacmp_ftp.sh    : Automates the process of retrieving sla files from duacs.
                   Input is a hycom daily file. uses slacmp above

Other routines:
argo_mersea      : used to process a argocmp-processed argo profile (with
                   -ncdump). Sets some parameters needed by MERSEA file standards.

gmaps_update.sh  : Used to create plots/thumbnails of salinity/temp from argocmp
                   output. transfers them to NERSC webserver where they are
                   presented using google maps.

plotargocmp.py   : Used by the above script. Needs the scipy python package.
                   Converts files in ArgoCmp directory (created by argocmp) to
                   plots of salinity/temperature.


###############################################################################
Most of these routines require (in addition to hycom/obs files):
grid.info
regional.depth.[ab]
regional.grid.[ab]



FORMAT of regiondefs.in files - these files should contain the following: 
2 # Number of regions in file 
4 TEST1                       # Number of points and name of 1st region (name=TEST1) 
-30 40                        # lon/lat pair - 1st point 
30 40                         # lon/lat pair - 2nd point 30
60                            # lon/lat pair - 3rd point 
-30 60                        # lon/lat pair - 4th point 
4 TEST2                       # Number of points and name of 2nd region (name=TEST2) 
-30 20                        # etc etc ...  
30 20 
30 60 
-30 60 
... 
... 
...

Note that the regions must be traversed clockwise or anticlockwise. They can
consist of as many points as you want, but the region defined by the corners
must be Convex - see http://en.wikipedia.org/wiki/Convex_set . 

(Or, to be pedantic, it must be geodesic Convex on the sphere :
http://en.wikipedia.org/wiki/Geodesic_convexity)