|
How to Write EPIC Data Files
EPIC is a system for management, display and
analysis of oceanographic in-situ data. If your data
is written into EPIC format, then EPIC programs can be used
to diaplay and analyze your data, or presesnt it on the Web.
Please read the EPIC System
Overview and other EPIC documentations
for a basic understanding of the architecture of the EPIC system
before proceeding. The EPIC system supports multiple
data file formats, but the recommended format is netCDF. Many
characteristics of EPIC netCDF files are general and apply to
all the types of data stored in EPIC, whereas others are specific
to the type of data stored in the data file. Supported data
types include:
Profile data (i.e.,
CTD, XBT, Bottle data)
Profile data is stored with a single profile in a single
datafile. This is a 1-dimensional data, indexed by (generally)
pressure or depth. Since profile data is taken at a single
point, the time, latitude, and longitude axes each consist
of a single value, whereas the depth/pressure axis will contain
many points. The profile data file includes all variables
measured or calculated for the profile. An example is a CTD
data file with measured variables of pressure, temperature,
conductivity and oxygen and calculated variables of salinity,
density, potential temperature, and potential density.
Time series data (e.g.,
from an instrument on a moored buoy)
Time series data is stored with a single 1-dimensional time
series in a single datafile. This 1-dimensional data is indexed
by time. Since time series is taken at a single point, the
depth, latitude, and longitude axes each consist of a single
value, whereas the time axis will contain many points. The
time seies data file includes all variables measured or calculated
by the instrument. An example is a current meter data file
with measured variables of zonal and meridianal velocity and
calculated variables of speed and direction.
Track data (i.e., shipboard
ADCP data, or other data collected along the track of a moving
platform)
Shipboard ADCP data (in the scope of this project) consists
of processed acoustic data collected along the shiptrack. The
shipboard instrumentation records average measurements, derived
from individual acoustic ping "events". The time interval
over which an "ensemble" of individual pings are processed
is fixed. Current velocities (and some other measurements) are
calculated for some number of depth bins during each ensemble
interval, representative of the time between the start and end
of the interval, and therefore of the section that spans the
ship's initial and final positions. Thus, the resulting data
is a mix of 1-D fields spanned by time (e.g. position of one
ensemble) and 2-D fields spanned by time and depth (e.g. velocities).
This structure needs to be reflected by the netCDF data files.
Please refer to this commented frame of a shipboard ADCP file
for detailed information. The original netCDF file can be found
here. For further information look on PMEL's anonymous ftp server:
ftp://ftp.epic.noaa.gov/epic/adcp/.
Steps to write your data file
- Assemble your data, and decide which data type it is (profile,
time series or track).
- Decide which variable codes match the variables or parameters
in your data file. See Variables
and the EPIC key file.
- Look at the Attributes
to see which must be used to include required metadata in
your data file, and which you may choose to use, both general,
and those specific to your data type.
- Look at the example
programs for writing EPIC data files. Select one to use
as a template, and customize it for your own use.
- Obtain the EPIC I/O
system library (epslib) and install it. Compile, link
and run your program.
- Refer to the manual
for the EPIC System I/O Library (epslib), as needed.
- If you have further questions, please contact the EPIC team
at epic@noaa.gov.
|