Visualization of Ocean Data Sources Websites and Software

Visualization of Ocean Data Sources, Websites and Software for Non-Genius Oceanographers without a Resident Staff of Wizards & Gurus SCOR 2006 London

Visualization n Working with earth science datasets IS visualization n Both fundamental and complex analyses of the data (and the quality control procedures that precede them) must be viewed graphically in order to understand patterns and relationships – or to identify problems

Credo n. A robust, pre-compiled program that you install with a click and run without handholding & grief is a GOOD THING n Loose bits and pieces of source code, requiring licensed platforms, additional libraries, compiling, UNIX system adjustments, de-bugging, INI/BAT file editing, and script writing is BAD

I’m Giving This Talk, and You Aren’t n PC>Mac>UNIX/Linux is the natural order of the universe n Windows is insulting and degrading, but it works n Scientists at the cutting edge of ocean data visualization probably don’t need to hear this: – It may not highlight your favorite software – It is oriented toward capacity building

The Background Paper n n n n Format types (15) Relations between formats and visualization (3) Specific formats of interest (6) Integration diagrams (3) Format conversion (20+) Programs & projects (8) Data (500/5000) Software (27)

The Visualization Paradigm n SOURCE->MANAGER->VISUALIZER->YOU – Source – datasets “out there” and on hand – Manager – dynamic methods to get, subset, reformat and deliver – Visualizer – user and his/her software n Basis for IODE training in Ostend, Belgium, using Ocean. Teacher n This simple paradigm comes in 3 main flavors:

MODE 1: Like Mother Used to Make n Source Manager Visualizer/You n The main paradigm without frills n Sources are typical CDs/DVDs and basic, online collections n Managers deliver desired data in basic formats n Visualizers are just about any ocean software with graphics

MODE 1 Examples n WOD 01 WOD Select ODV, JOA n NVODS LAS/FERRET GIS, nc. Browse n ARGO CORIOLIS nc. Browse, ODV, JOA n WMO/WWW GTS GRADS n EOS POET-ESIP HDFView n VMAP 0 FTP VPFImport/GIS n Public GIS data HTML GIS

Classic ODV Example

MODIS L 2 Data in ODV

Operational Surface Vectors February currents from averaged GDP surface trajectories* *But see below

MODE 1 Issues Format issues predominate n Climatologies & atlases n – Often simple grids – Metadata poor or missing n Operational Data – GRIB: modelers love, oceanographers hate § High priests & gurus § Worst website interfaces on the Web – Net. CDF: cottage industry of “versions” – HDF: 4 major versions in NASA format wars

MODE 2: Peekaboo n Source/Manager/Visualizer------- You n Mainly in the GIS area n Essentially, the user simply gets pictures of the data from fancy black boxes n Limited control over output, as provided by controls & menus n Vast majority of IMSs published today do not provide data

MODE 2 Examples n IMS/WMS---- PNG/JPG graphics – cannot be captured; underlying shapes and grids (“features”) are not exposed n Dynamic “Operational GIS” interfaces based on OGC – SEACOOS – Open. IOOS prototypes n Many java applets that draw maps from underlying, server-side databases

WMS: A Tale of 2 Attitudes Geological data (image only) from a typical IMS; Argo drifters (annotated point shape) from the Argo Information Center

Sea. COOS: OGC at the Limit Selected Models Combined Forecasts

MODE 2: Issues Data access issues predominate n Dominance of a single commercial firm n More copyright hurdles than any other data type n – $$$$$ Slow emergence of community “marine GIS” n De facto marine GIS n – Open. IOOS – images only – Sea. COOS – images only

Related GIS Issues n Sudden appearance of GOOGLE/KML – “Wrapper” for images – Limited shape capability n Direct Net. CDF and HDFx import – Nothing implemented yet in open-source – GDAL possibility for NC not yet explored n IOC training Saga for images & gridding; Map. Window for general use

Map. Window for Earth Data Coastal deserts of Namibia, visualized from SRTM data

MODE 3: Son of OPe. NDAP n Source------- Manager/Visualizer/You – OPe. NDAP plus really good client end n All the action is on the client side – nontrivial learning curve n Full potential of OPe. NDAP used n Levels the playing field for PCs and work stations, due to stride/step controls n Data storage and platform capacity issues for model output

MODE 3 Examples n OPe. NDAP Catalog Items-------- IDV n THREADS Catalog Items---- nc. Browse n GCMD Catalog Items--------- ODC n In other words, many sources of the special URLs required by these clients

Near Real-Time Argo Data in nc. Browse TS Plot for Namibia’s EEZ

IDV for Oceanography

MODE 3 Issues n Skill and infrastructure issues predominate n OPe. NDAP Data Connector – never completed; some features in IDV, but still needed n Catalog maintenance is poor because it’s voluntary n Lights are on but nobody’s home…system components frequently broken n Software solutions tend to be big, complex

Remaining Technical Issues n n n Need for a basic vector-drawing utility for simple U and V grids Continued heavy use of GRIB by modelers for output……time to move on OPe. NDAP cataloging has stalled and NVODS maintenance is unfunded now The principal NASA site for color and SST data (L 2, L 3) uses HDF 4 and not HDF 4 -EOS Absence of biological data from this talk is criminal and needs attention ASAP

Remaining General Issues n Divide between high-end technical applications in leading laboratories and the visualizations needs of compiler- or library-challenged ordinary folks – The Microsoft model is not completely evil n Even greater divide between visualization capabilities in advanced institutions and in the developing world – The WWW provides a good model to follow – GRIB isn’t easy, but it sure does work

Conclusions n n Format proliferation has slowed down in oceanography, but is not dead Visualization methods and tools will continue to be scattered and disjoint until we focus on an even smaller family of formats OPe. NDAP-type approaches provide good solutions to MODE 1 and MODE 3 issues OGC-related work must continue to solve MODE 2 issues

Recommendations n n n n Common visualization solutions for the global ocean community Targeting a reasonable computer platform for visualization development OPe. NDAP-type protocol robustly supported for all data types Small family of common formats Non-proprietary path, even at the expense of early progress Inclusion of “formal” data visualization methods in all CB activities Involvement of all SCOR-sanctioned programs

Thank You • murraybr@bellsouth. net • oceanportal. org • oceanteacher. org