[netCDFJava #WIK-413864]: Reading forecast dates from grib2

> 
> it says its ensemble data, but all have perturbation = 0. So an ensemble 
> dimension of 1 is created.
> 
> im seeing 2 variables, Temperature and "Total Precipitation rate".
> 
> For Temperature:
> 
> 1) multiple runs (refDate), 6 hours apart.
> 2) for each refDate, 6 hour forecasts out to 360 hours.
> 
> since these overlap, I end up with 664 unique forecast times. (this is in 4.3 
> sorry you cant see it yet).
> 
> For Total Precipitation rate:
> 
> 1) multiple runs (refDate), 6 hours apart.
> 2) for each refDate, 6 hour forecasts out to 360 hours
> 3) each forecast is for the interval 0-foreast (0 = refDate), so one sees 
> 0-6, 0-12, 0-18 ,...0-360.
> 
> these constitute unique coordinates (no overlap) so you get 36844 of them. 
> The only thing I can think of if I am misunderstanding what the intervals 
> are. (?)
> 
> what would you expect ?
> 
> 
> 
> 
> Actually I would expect a run dimension , e.g. :
> 
> Total_precipitation_rate (run=600, time=60, ......)
> 
> since the file contains approximately 600 forecast runs - first 
> refDate=01/01/2011 00:00, last refDate=05/31/2011 18:00,6 hours apart, and
> the runs are 6 hour forecasts out to 360 hours (this info is provided by  
> NDFD DeGrib)

I see, so you actually want this to be an FMRC with a two dimension time 
coordinate ?

Ticket Details
===================
Ticket ID: WIK-413864
Department: Support netCDF Java
Priority: Normal
Status: Open