25-06-2012, 04:07 PM
Ocean Data View 4.0
Ocean Data View 4.0.pdf (Size: 20.51 KB / Downloads: 53)
Abstract
Ocean Data View, a multiple-variable graphical analysis and display package for
oceanographic station data (station metadata, temperature, salinity, nutrients, others),
compatible with several commonly used international marine data formats, is reviewed.
Standard derived variables or those based on user-provided formulas can be displayed.
Station charts, station data profiles, multi-variable scatter plats, section profiles, and
surface plots (one variable plotted on a specific value of any other variable) can be
created easily within a user-specified physical layout of Cartesian frames. Plotted data is
easily gridded and/or contoured in nearly all views.
The Holy Grail of oceanware, a really good, Windows data browser, has finally appeared: Ocean Data View!
When word spread this past winter that Reiner Schlitzer at the Alfred Wegener Institute-Bremerhaven had
released a sensational new graphic program for marine data (http://www.awi-bremerhaven.de/GPH/ODV/), this
reviewer wasted no time in downloading and testing it. My intention was just to see if it really worked, and -if
so- could it handle some of the analysis and quality control lessons we offer in Intergovernmental Oceanographic
Commission (IOC) training classes. Ocean Data View (ODV), now on version 4.0.4, turned out to be a powerful’
versatile, yet easy-to-use package that instantly transforms this klutziest of users into a savvy hydrographer
(well, almost). ODV is now my program of choice -its overall design, simplicity, capabilities and compatibilities
set a high water mark in oceanware that will be hard to match.
I’ve heard about folks who are absolutely addicted to another browser, OceanAtlas (Swift, J. et al;
http://odf.ucsd.edu/OceanAtlas/), and have seen the outstanding examples of its work in Mattias Tomczak’s webpublished
lecture notes (http://gaea.es.flinders.edu.au/~mattom/E...tents.html). But the OceanAtlas is a
Macintosh application, vaguely descended from Peter Rhines’ old PC ATLAST
(http://podaac.jpl.nasa.gov:2031/dataset_...TLAST.html) which sprang from the Living Ocean Atlas of
Stommel. Reading the literature on OceanAtlas certainly convinces me that it might be a terrific mousetrap, but
alas I am a PC-head, son of a PC-head, and will never know.
Although, as I said, you never really see what’s going on "inside" ODV’s obviously very robust data model, the
capabilities for producing graphical analyses are vast. The "data collection" (the author’s term to describe all data
currently combined from various sources into an internally consistent, memory-resident, integrated database)
available to ODV includes the usual cruise, station, sample, and measurement fields. These can be sliced and
diced in an almost limitless way, by mapping selected fields against the X and Y axes in a Cartesian coordinate
system. The Z axis (if needed) is represented by the color-coding schema for the X- Y data points, resulting in
plots similar to those from ATLAST. Distance along cruise tracks (or latitude, or longitude) can be an axis,
resulting in section plots; time can be an axis, resulting in time-series analyses. [The ease of cutting the data into
useful X-Y planes (or X-Y-Z spaces) from the inherent n-dimensional dataset reminds this reviewer of the everpopular
Grid Analysis and Display System (GrADS) meteorology data program (http://grads.igesgrads/),
which by the way is out now in new Windows versions (!). In fact, I’d probably be recommending various kluges
between ocean data and GrADS, if its data format requirements weren’t so...well...intellectual.]
There are three very special capabilities in ODV that extend the above analytical model in powerful ways:
· The user can select from a wide suite of "canned" derived variables (including the depth derivative) to
supplement the measured variables. These new fields are added instantly to the data collection. In the most
recent version of the program, geostrophic velocities can be calculated from appropriate survey transects;
the new values are used to populate a new "virtual" data collection tied to the original data and displayed
separately.
· When the X- Y plane is selected such that X=longitude and Y=latitude (i.e. the Cartesian graph is also a
map), then the user can plot a third variable (the Z variable) on any selected "surface". Surfaces are specific
values of another variable. For instance, you can plot oxygen on the 18-degree temperature surface.
· There is a macro capability, yet unexplored by this reviewer, based on a higher-order syntax defined in the
on-line HELP facility.
Additional bells and whistles are also provided to enhance the visual impact of the charts. The default, lowresolution
global coastline (ca. 1:20,000,000) can be replaced by "high resolution" versions (ca. 1:1,000,000) for
the globe or for the Mediterranean. The old ETOPO5 bathymetry
(http://www.ngdc.noaa.gov/mgg/global/global.html) is present for bottom depiction in sections. In the graphing
department, all kinds of nifty things can be done to the gridded data depictions, such as contour lines, grid
coarseness control, error limits, and data-point notation. The overall menu is far too rich to cover here.
So, ODV can make all kinds of graphs, but with all this power are we locked into specific variables
combinations or prescribed layouts? Are flexibility and control sacrificed in order to bring us this rich menu of
possibilities? Author Schlitzer has kindly thought of everything, so the user enjoys as many display options as
the underlying analytical tools. First, in any display mode, the user can instantly switch any variable along any
axis; this permits rapid scanning through all sorts of multiple-property graphs with a view toward quick quality
control, or water mass assessment. Second, the user can switch into a graphical layout screen that permits
moving, re-sizing, adding, or deleting graphs. Third, the color-scale for the Z-variable can be expanded or
contracted to suit the data better. An eerily consistent logic is at work, so that any changes to the color scale or to
the range of any axis cascade through the other relevant graphs, all automatically.
This all sounds very tempting, but how much trouble is it to enter the data in which I was really interested?
Several templates for the ODV "data collection" exist, such as the variable set in the World Data Center A’ s new
World Ocean Database 98 (WOD98; http://www.nodc.noaa.gov/OC5/index.html) or the variable set in the
traditional SD2 exchange format (http://www.nodc.noaa.gov/NODC-Archive/sd2.htm). Even the regrettable
"mass-per-mass" WOCE data units (http://whpo.ucsd.edu/manuals.htm) are on the list, an act of supreme charity
one would think, considering WOCE’s apostasy in these matters. [The reviewer is an ICES true believer. ] Into
any template, the user can load data from several or many different recognized formats. But, be careful you don’t
use a template that omits an analyte of interest; for instance, don’t use the SD2 template if you want to load
chlorophyll data. The recognized import formats are: World Ocean Atlas 94, WOD98, SD2, WOCE, ODV’s own
hierarchical ASCII format, and ODV’s own ASCII spreadsheet format. There is an ingenious function that
matches up the incoming variables with the template variables (based on the field names ), and even performs
necessary conversions when needed. The troublesome mass-per-mass variables can be handled in this way.
Data importing is aided by a well-designed set of filters to impose time, geographic, quality, source, or type
limits on incoming data. This same filter is, however, always available during the graphing process, so the user
can safely import the maximum amount of data initially. Veteran users of ODV who may wonder why I have not
mentioned unzipping the WOD98 files will be happy to know that the current version does this automatically,
which brings up a good point: Schlitzer has posted several updates to the program with little fanfare, so sign up
for his mail list to receive update notices.
So how good is Ocean Data View, really?
From the very parochial perspective of the Ocean PC software project (http://www.ices.dk/ocean/ocean-pc.htm),
where we desperately need a well-documented, Windows graphical engine to replace ATLAST, ODV offers a
wide range of attractive analytical tools that are immediately grasped by students in developing countries. We
can take regional data straight from WOD98, and have the students making rather sophisticated graphs in
minutes. The next step, adding local data for national-and institutional-level projects, is obvious, and -what’s
more important -made attractive by the simplicity of the import routines. One attractive route would be raw data
entry with OceanPC, then use of OceanPC’s new ODV export routine. We hope eventually in the International
Oceanographic Data and Information Exchange (IODE) community to reap a large fraction of these additional
data, because they have already necessarily been converted to a "standard" format during the process. From this
point of view, ODV is the Great Carrot, thanks to the good Professor Schlitzer.
From the broader perspective of students and professionals everywhere, ODV is definitely the best freebie
around. Although there are some tricks you need to learn to get CD datasets into the system (for instance, there
are various routes from WOD98 to ODV), the compatibility with other information models is inspiring. Author
Schlitzer has incorporated important existing formats, and kept his own format-craft to a necessary minimum, a
commendable restraint that we could use more of. Further, he has achieved a reputation for near-instant upgrades
in response to user requests and comments. One very recent upgrade was the addition of the WOD98 format;
another was explicit logging of user corrections to the data (yes, edits are possible too!). Carefully working
within the context of existing widgets when they are good, but imaginatively creating new ones when they are
lacking, is the crowning achievement here. That, plus the very professional "look and feel" of the program
screens, menus, and documents. Don’t miss a very good thing ... add Ocean Data View to your software
collection ASAP.