Heuristics for identifying Vector components

[what_nick]

Hi All,

We have internal conventions in CSIRO to mark the components of 3D vectors in currents/winds. This convention however is not universally adopted and we have had to make adjustments for data from local agencies such as Bureau of Met.

What conventions/heuristics for Vector component identification are we going to use in iGlobe (other than say a huge look-up table or ask the user everytime) ? Are there any conventions at NOAA ?

Cheers,

whatnick

[what_nick]

Hi All,

We are coming to a point where we should consider consolidating volume visualisation work Jim did (http://people.eecs.ku.edu/~miller/WorldWind/VectorFieldVis.jar) and I did - https://github.com/whatnick/glob3_csiro, into iGlobe. I will start making a series of commits on this bend. I will add vectorfieldvis.jar into the repo unless jim objects or he is willing to commit the sources and massage them in.

We have 3 choices as outlined before:

1) Vertex Arrays - used in Glob3 CSIRO
2) Display Lists - used in Vector Field Vis
3) VBO's - included experimentatlly in WorldWind

Projects like Glyph Sea - http://visservices.sdsc.edu/projects/scec/vectorviz/glyphsea/ , perform the component to tensor glyph conversion using vertex shaders and generate the glyphs procedurally. This might be pathway to pursue for higher end video cards.

Cheers,

whatnick.

[what_nick]

Hi All,

I started integrating Jim's old demo code and realized that it is compiled against the older (0.6?) version of worldwind and I need to maintain it for 1.2. Could we please get a source code drop from Jim.

Cheers,

whatnick.

[what_nick]

The vector variables in SHOC (http://emg.cmar.csiro.au/www/en/emg/software/EMS/hydrodynamics.html) (our internal model) are labelled as below:

ncdump -h mackenzie_depth_out_cf.nc | grep components
wind_u:vector_components = "wind_u wind_v" ;
wind_v:vector_components = "wind_u wind_v" ;
u:vector_components = "u v" ;
v:vector_components = "u v" ;

I will start by implementing this as the heuristics for automatic vector variable identification.

Cheers,

whatnick.

[what_nick]

More vector component detection heuristics as captured by our legacy code. I am considering creating a JSON config document to capture these, alternative suggestions welcome. Note that there are only 2 components since our legacy renderer is limited to 2D.


Name-----------Description-------U-------------V
---------------------------------------------------------------------Coards
current----------Current-----------u-------------v
wind-------------Wind--------------u_avg-------v_avg
wind_avg10m---Wind_avg10m---av10u-------av10v
wind10-----------Wind_10m------ u10----------v10
wind--------------Wind-------------ueavg-------veavg
wind--------------Wind-------------zonal_wnd--merid_wnd
----------------------------------------------------------------------CF
current-----------Current Velocity
--------------------in the Water
---------------------Column----------vozocrtx-----vomecrty
current-----------Current-----------u-------------v
current_avg------Current Averaged--ue---------ve
wind--------------Wind----------------u-----------v
wind--------------Wind----------------wind1------wind2
wind--------------Wind----------------wind_u-----wind_v
wind--------------Wind----------------zonal_wnd--merid_wnd
-----------------------------------------------------------------------MECO
current-----------Currents-------------u1----------u2
avg_current------Depth
averaged currents---u1av--------u2av
bottom_current--Near
-------------------bottom currents--bottom_u-----bottom_v
wind_shoc--------Wind shoc--------wind1----------wind2
wind_rams------- Wind rams-------wind_v---------wind_u
wind_gasp--------Wind GASP-------merid_wnd----zonal_wnd
umean------------Mean velocity-----u1mean-------u2mean
ofam_vel---------Ofam velocity-----ovelu---------- ovelv

[what_nick]

Hi All,

Gary has pointed me to this master list of conventions - http://www.unidata.ucar.edu/software/netcdf/conventions.html . I am including a link here for posterity.

Cheers,

Tisham.