Home > atmlab > retrieval > spareice > Contents.m

Contents

PURPOSE ^

SPAREICE Functions related to SPAREICE retrievals

SYNOPSIS ^

This is a script file.

DESCRIPTION ^

  SPAREICE   Functions related to SPAREICE retrievals

 Please note that, contrary to the rest of Atmlab, the functionality in
 this sub-toolbox relies on two toolboxes that require additional
 licenses to the Matlab core:

   - Neural Network Toolbox.  Cannot currently be evaded, although it
   should be possible to implement the resulting regression directly.

   - Parallel Computing Toolbox.  Can be evaded with a small adaptation,
   but that will make it much slower on multi-core machines.

 The functionality in this directory is not up to the same standards as
 the rest of Atmlab.  It is research code for Gerrits PhD project and it
 was not designed with multi-user interfaces in mind.  I (Gerrit) added
 it to Atmlab after my PhD, when I did not have the time for the cleanup
 needed.  Please use with care and I hope you can find what you need
 despite the messy situation.

 To perform SPARE-ICE retrievals, you will need:

   - The trained neural net, available in:

   /storage3/user_data/gerrit/neuralnets/avhrr_345_mhs_345_angles_tsurfcfsr_all_global_MEAN_AVHRR_Y,3,4,5,B_BT,3,4,5,B_LZA,B_LAA,B_SZA,B_SAA,CFSR_Skin_temperature,_to_MEAN_IWP_2C,_noise0.00,_v0.7_1.mat

   - Data for AVHRR and MHS, should be at /storage3/data/avhrr and
   /storage3/data/mhs, respectively

   - CFSR data, downloaded from http://rda.ucar.edu/.  Version  (ds093.1)
   is until March 2011, then it's (ds094.1).  Please see citation
   information at http://rda.ucar.edu/datasets/ds093.1/ and
   http://rda.ucar.edu/datasets/ds094.1/.  For the SPARE-ICE version, you
   will need skin temperature.  Note that skin temperature is actually
   included with the temperature profile!  So you should download the
   temperature field, then choose "ground level" on the next page.  You
   want "type of product" "all available" because the actual reanalysis
   is only every 6 hours and SPARE-ICE retrieval needs hourly data
   (reduces interpolation).  Choose the densest grid with all available
   data.  Do not forget to set "Output format" to "Converted to NetCDF".

   I downloaded the 'pure' reanalysis data to
   /storage3/user_data/gerrit/reanalysis

   - The Python script to convert this to daily netcdf files to be used
   with the atmlab retrieval.  Resulting daily files at
   /storage3/user_data/gerrit/reanalysis_per_day but change source code
   as needed, in particular 'indir', 'infile', 'outdir', 'outfile',
   defined near the top of the source code.  Note that this is Python3
   code.

   - The atmlab files in this directory.

 To process SPARE-ICE data, call process_ccniwp_data, for example:

   process_ccniwp_data([2013, 4, 2], [2013, 5, 30], 'noaa18');

 To read SPARE-ICE data:

   define_local_datasets();
   D = datasets();
   D.col_syn_iwp.read_...  (see SatDataset and SatDataset.read for details)

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:

DOWNLOAD ^

Contents.m


Generated on Mon 15-Sep-2014 13:31:28 by m2html © 2005