time and time_uv aggregations (add lat_uv and lon_uv)

[dpsnowden]

We have two distinct time axes in each segment file. time will be multi valued (10000s of time steps) and most variables are dimensioned along time and time is the primary aggregation in THREDDS.

There is a second time_uv axis for segment averaged velocities.

Kyle will experiment with creating two aggregations for the same set of data files.

[kwilcox]

The two aggregations are up here: http://54.221.199.138/thredds/scratch.html

I'd recommend we clear out the variables that are not part of the aggregations. ie. Clear all variables that are not dimensioned by time_uv from the time_uv aggregation. The values represent a random dataset in the aggregation (you can also set it to do the latest or first) and are just confusing.

[kerfoot]

The time_uv aggregation is also a bit confusing to me:

http://54.221.199.138/thredds/dodsC/RU5MonthTimeUV.nc

The time_uv has 363 values, so the variables that use this dimension all show up as having the same number of values. Why to the non time_uv dimensioned variables (ie: temperature, etc.,) have 1344 values? Is this a random dataset aggregation as in the 'time' aggregation?

We'll need to remove all of the variables that don't use the time_uv dimension from this dataset as well.

[kwilcox]

The THREDDS aggregation is choosing a random file from the collection, and that random file has a time dimension of length 1344. I'll remove the unneeded variables from the aggregations with the new files you emailed out.

[kwilcox]

Aggs are up:

• Time: http://54.221.199.138/thredds/dodsC/RU5MonthTime.nc.html
• TimeUV: http://54.221.199.138/thredds/dodsC/RU5MonthTimeUV.nc.html