Help Using Duck Experiement Data Server

Organization of the Data

The data served by this system is organized by principal investigator (PI) and has an accociated PI code to identify the individual experiments. Two methods are available for obtaining the data: One method is through the DUCK Data Server that has a web based interface, which is based on the Joint Global Ocean Flux (JGOFS) Server. Data is delivered in columnar ASCII format and each data set has documentation describing what is known about the collection and analysis of that data. The server also provides simple plotting of the data of most data sets. The server provides easy access to the data and is useful for determining what is available and desired by a new user. The other method of obtaining data is through ftp. Statistics files are available as ASCII columnar format in each PI directory and has supporting documentation. Ftp will be useful for those people familiar with the experiment and desire quick downloading of data files. Zipped versions of these data files are also available in the "zipped" directory in each PI directory. The ftp directories can be reached from the links in each level of the Data Server.

The basic organization for the Data Server is through levels: PI->Date->Data. The Date may be on or more levels, such as 'month', 'month->day', or 'month->day->time'. In each level there are the following buttons:

   Directory - Return to data server main directory
   Documentation - Documentation provided by each PI for data collection, 
       analysis, quality, ...
   Plotting and Other Operations - Some simple web based plotting features.
   Level 0 - Return to the first level for the present data type.
   Next Level - Generally not used.  It delivers all data from the present 
       level.  For example, from the lowest level (0) with selectable levels of 'Aug', 
       'Sep', and 'Oct', using the 'Next Level' button will download all 3 months with 
       ASCII character headers separating the months.  It's best to just select 
       individual months. 
   Flat Listing - delivers columnar data at the present level with leading 
       character text (level information) and buttons removed.  Use this for saving 
       columnar data with the browser.  This format can be easily imported into 
       spreadsheets and database programs.  The server delivered data and the ftp files 
       have a leading character header that identifies the data type (e.g., Month, Day, 
       gage ID).

Note that missing or bad data is assigned the value -999 to zero or more decimal places. As a result columns with these values will not plot correctly. For MATLAB users there is a C MEX-file, called load2Mat.c that will read these ASCII columnar files with the character header label being assigned as the MATLAB variable name. In this program the -999 values are assigned NaN - if the user would rather have the -999 values simply change the code (search for -999) before compiling.

If you download data from this system, please take (and read) the documentation as well.

There is a good alternative for downloading large datasets using the 'Plotting and Other Operations...' button at the top of the data display. Upon first displaying the dataset, use the browser's 'Stop' button to abbreviate the lengthy download-to-display. Click on the 'Plotting and Other Operations...' button. Choose the menu item entitled, 'Other data listing formats'. Here you are presented with a few options for the format to download. The most popular is the 'CSV style listing'. Hold down the SHIFT key on your keyboard and click on the link. This will begin the download operation by presenting you with the familiar dialog box asking for a filename to which to save the data.

Return to Duck Experiment Data System

Or use the Back button on your browser to return to the previous screen.