Skip to content

Latest commit

 

History

History
249 lines (206 loc) · 11.4 KB

README.md

File metadata and controls

249 lines (206 loc) · 11.4 KB

WyoSondes (Wyoming University Radiosondes Observations)

A GNU/Octave & MATLAB Automatic fetch and storage interface for Radiosonde data provided by the University of Wyoming

This is a simple GNU/Octave and MATLAB friendly function to fetch radiosonde data from the publich repository hosted by the University of Wyoming.

Table of Content

  1. Introduction

  2. Description

    2.1. Input variables

    2.2. Output variables

  3. Data storage

  4. Usage Example

  5. Plotting Example

  6. Apendix

    6.1. Other sources

    6.2. Octave requirements


Introduction

This repository has been mainly developed and intensily used with GNU/Octave v4.4.0 under Linux OpenSuse distribution. Some testing has been done for Matlab R2016a (the one I have at office), but some issues might appear with the latest Matlab versions, so any feedbacks are welcome.

Description

This GNU Octave/MATLAB function gets the DATA of the soundings (raob) from the Wyoming University internet site (http://weather.uwyo.edu)

The Function can be used by calling one or any combinations of the following set of parameters:

To download the latest radiosonde data given a specific station name:

>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station);

or to specify a range of dates given by years, months, days and hours:

>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station,year,month,day,hour);

The following optional arguments are possible:

>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station,year,month,day,hour,'outputpath','/whereto/storage/data/');
  
>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station,year,month,day,hour,'netcdf',true);
  
>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station,year,month,day,hour,'csvfile',true);
  
>> [data, metvar, {metadata}] = RASOBS_DOWNLOAD_DATA_RAW(station,year,month,day,hour,'matfile',true);

where,

INPUT VARIABLES:

% station-> (string) Code for the station to download;
% year  --> (numeric) Range for Year of the date we want to download, e.g. [2015:2018];
% month --> (numeric) Range for Month of the date we want to dowload, e.g. [10:12];
% day   --> (numeric) Range for Day of the date we want to download, e.g. [1:31];
% hour  --> (numeric) Range for Hour of the date we want to download [00,12];
%           hour will accept any range of hours, but Wyoming University only has two radiosonde per day at 00 and 12Z.

FOLLOWING OPTION INPUTS (keyword value pairs)

% 'netcdf', true/false --> whether to storage as NetCDF file (recomended for large datasets and portability);
% 'csvfile', true/false --> whether to storage as CSV files (creates one file per hour, feasable for short datasets);
% 'matfile', true/false --> whether to storage as v7 MATLAB files (recomended when use only with Matlab/Octave);
% 'outputpath','/path_to/storage/' --> Dir where files are storaged (see below for default storage directory);
% 'waiting', numeric seconds --> Number of seconds to wait between fetching profiles (default 3 secs.)

Default options:

% 'outputpath'->'../data/RASOBS/station/yyyy/',
% 'matfile'->true,
% 'csvfile'->false,
% 'netcdf'->false.

OUTPUT VARIABLES:

% data --> MATLAB Structure variable containing the Radiosonde data and related information. 
% metvar --> MATLAB Structure variable with Sounding Station Parameters and Indices for every profile.

In case of unsuccessful downloading (e.g. hour 10 doesn't exist) it returns an empty variable or skip that hour.

OPTIONAL OUTPUT:

% metadata --> structure with two fields containing the METADATA describing every profile and Indice variable (for a extended description of indices, see HTML http://weather.uwyo.edu/upperair/indices.html)
station_name --> variable containing a string with the Name of the Station, e.g. 'ENBJ' for the Bjornoya station with code 01028.

DATA STORAGE:

Downloaded data can be archived in one or any combination of following file formats -> CSV, NetCDF or MATLAB binary format. Depending on the input values a data archive will have a generic name like

RS_Y[year1]-[year2]_M[month1]-[month2]_D[day1]-[day2]_H[hour1]-[hour2].nc

or

RS_Y[year1]-[year2]_M[month1]-[month2]_D[day1]-[day2]_H[hour1]-[hour2].mat

For the case of CSV format one file per hour is created, therefore CSV format is suggested to use only with small datasets.

The new generated file is saved by default in the following folder: '../data/RASOBS/namestation/yyyy/' or in any other directory if specified by the 'outputpath' option.

The variable name, description and units of the profile variables (columnwise) are stored as a member variable named PROFILE_META in the optional structure output metvar, i.e.

>> metvar.PROFILE(1:3,:)
ans =
{
  [1,1] = PRES
  [2,1] = HGHT
  [3,1] = TEMP
  [1,2] = Atmospheric Pressure
  [2,2] = Geopotential Height
  [3,2] = Temperature
  [1,3] = hPa
  [2,3] = m
  [3,3] = °C
}
>> metadata.RSINDICES(1:3,:)
ans =
{
  [1,1] = OBST
  [2,1] = SLAT
  [3,1] = SLON
  [1,2] = Observation time
  [2,2] = Station latitude
  [3,2] = Station longitude
  [1,3] = [YYMMDD.HH]
  [2,3] = deg North
  [3,3] = deg East
}

The metadat can be readed independently with the following auxilliary function:

>> metadata = WyoRS_metadata;

In case of storage in NetCDF format, the units are displayed as argument for every available variable.

Description of the profile variables

A detailed description of the variables in the profile is provided by the Wyoming University web site: http://weather.uwyo.edu/upperair/columns.html

Description of the Sounding indices

Similarly a detailed description of the Radiosonde indices is provided by the Wyoming University web site: http://weather.uwyo.edu/upperair/indices.html

USAGE EXAMPLE

This is an example to download data from the Station Norderney with station number '10113' for the year 2015, the months of Mai, June and July ([5:7]) and all available hours i.e. 00 and 12Z ([0,12]) and storaging the data as NetCDF file.

>> [data,metvar] = RASOBS_DOWNLOAD_DATA_RAW('10113',2015,[5:7],[1:31],[00,12],'netcdf',true);
   
warning: Data from 06.31.2015_00UTC does not exist!
warning: Data from 06.31.2015_12UTC does not exist!
warning: Data from 07.27.2015_00UTC does not exist!
warning: called from RASOBS_DOWNLOAD_DATA_RAW at line 175 column 21

for the example above, the process to download and storage took approximatelly 2.7 minutes with GNU Octave.

>> whos data
Variables in the current scope:
Attr Name        Size                     Bytes  Class
==== ====        ====                     =====  =====
     data        1x183                  1705968  struct

>> whos metvar
Variables in the current scope:
 Attr Name         Size                     Bytes  Class
 ==== ====         ====                     =====  =====
     metvar      1x1                      41013  struct

The data will the then storaged in a NetCDF file in the default directory ../data/norderney/2015/ with a file name RS_Y2015-2015_M05-07_D01-31_H00-12.nc.

SIMPLE VISUALIZATION EXAMPLE

To help visualize the downloaded Radiosonde data, a simple GNU Octave/Matlab script helps to navigate along the profiles. This script is more useful when a large number of profiles are present, for less than 5 or single profiles it isn't worth it.

The visualization script has been tested and used under GNU/Octave v4.4.0 and Matlab R2016a.

For MAT binary files

Three options to start the GUI visualization script:

  1. From workspace as follow:
>> RASOBS_GRAPH_DATA_RAW;

and a file browser will pop-up to select a .mat file with the downloaded Radiosonde data.

  1. Directly read the file RS_Y2015-2015_M05-07_D01-31_H00-12.mat locate at /tmp/data/ do:
>> RASOBS_GRAPH_DATA_RAW('/tmp/data/RS_Y2015-2015_M05-07_D01-31_H00-12.mat');
  1. When the data are alredy loaded as workspace variable, use
>> RASOBS_GRAPH_DATA_RAW(data,metvar);

where data is the structure variable with the profiles, and metvar is the structure variable with the Radiosonde indexed (see above).

The script creates a Figure as it is shown in the following screenshot:

>> ha = RASOBS_GRAPH_DATA_RAW(data,metvar);

where ha is the structure of handles for the graphical elements, i.e. plot lines, axes, GUI's controls, etc.

the top panel shows the vertical profile or the selected variable (default TEMP), the middle panel shows the time-series of vertical profiles in a 2D surface plot (same variable as the top panel but only 10km height is considered) and in the bottom panel the time-series of the selected Radiosonde Indexes (default CAPE).

By using the pop-up menus it is possible to select other variables to show, and by moving the horizontal slice on top it is possible to change the observation time, alternatively the pick-up T0 buttom allows to select a time point with the cursor from any of the two time-series plots.

The profile is shown in color black for the selected time and for reference it also includes 2 profiles before and 2 after which are indicated on the legend by T0-2, T0-1, etc. where T0 is the date and time indicated at the top, e.g. 16.5.2015 T0=00Z in the figure above.

For NetCDF files

When the Radiosonde data has been stored as netCDF file, it can easily be visualized by any third-party software, for a quick peak ncview can be used

/home/user:~/wyorasobs/data/2015> ncview RS_Y2015-2015_M05-07_D01-31_H00-12.nc &

and by selecting any profile variable a 2D time series can be easily plotted, for instance for the temperature variable TEMP

For CSV Files

The visualization for the .csv files is not supported.

APPENDIX

Other Sources of Radiosonde data

* NOAA ESRL Radiosonde Database

NOAA's Radiosonde data archive including a good web based GUI to download a series of profiles with four differnet file formats, namely FSL ascii, original FSL ascii, FAA604 WMO/GTS ascii and netCDF format. General information and link to data archive is here

* IGRA Weather Balloon

Alternative Radiosonde data can be found at the Integrated Global Radiosonde Archive (IDRA), with most of the stations also covered by Wyoming University there are others which are additionally available at IDRA data bank. The IDRA data files, however, can be accessed directly from their FTP server as compressed files for the entire period of availability. Therefore a tool like WyoSonde is not needed.

Octave requirements

GUI script ROAS_GRAPH_DATA_RAW under GNU/Octave requires QT to be enable. In Octave command line write:

  > available_graphics_toolkits
      ans =      {
        [1,1] = fltk
        [1,2] = gnuplot
        [1,3] = qt
      }
    > graphics_toolkit("qt")

to verify whether or not your GNU/Octave installation support QT.


(c) 2018 Pablo Saavedra G.,
Geophysical Institute,
UNIVERSITY OF BERGEN

See LICENSE