Skip to main content
  • EN

I have a sample file from a gridded wave model for a small section on the South coast of England, ie it is a lat/long grid covering a few hours with 11 variables (wave height, wave period, that sort of thing) for each hour. I can view the file with Panoply.

While the larger spatial overview is great, I would love to extract a time series for a couple of grid cells with all parameters so i can for example compare the data with measured wave buoy data.

It should be quite easy as the the output should just be a table with the 11 variables as columns and each hour as a row (in a date/time format), but given the netcdf format this does not seem quite so straight forward - at least at my level of FME understanding 🙂 .

Not urgent but any pointers would be very welcome.

Thanks!

It is essentially a number of rasters on top of each other. What I would do is start with one of the rasters (Sampler) then create polygons from the raster (RasterCellCoercer), determine how the raster number / time interval relation works, determine the rasterband / attribute relation. Then relate you buoys with the polygons.

Many thanks Niels, that is exactly how it worked. The tricky thing is finding out about the number format, especially as one of the files i had was somehow corrupted, but it all makes sense now.

After reading the netcdf file first step is the attribute exposer to get to some of the netcdf attributes. The RasterBandSeparator then splits the file into the hourly time data, and the CasterCellCoercer creates the raster cell polygons which contain all the wave parameters. Some follow up manipulation then extracts the actual values using the integer value and factor for each parameter and composes the date/time to then export.

Reply


Community Stats

29,739
Posts
113,001
Replies
37,384
Members
Terms & ConditionsCookie settings