Thanks for all the help...
Below is a summary of the suggestions I received for dealing with the
problems.

With regards to using nibble...it was pointless.  Because I am dealing with
DEM's, and the nibble works on getting the most frequently appearing
neighbor values, nothing really ever happened.  However, the response to my
question of what to use as a mask grid was that i use the same grid as the
no-data values were present.  A better option for actually doing some
processing to the grid was to derive the mean of all neighbors:

[CoastalGrid].IsNull.Con([CoastalGrid].FocalStats(#GRID_STATYPE_MEAN,NbrHood
.Make,FALSE),        [CoastalGrid])

The  problem with using the original grid as the mask was that the area
outside the actual area of interest is also no-data so in the end the area
of interest/boundary is extended by one cell.

However, in the end I realized that i don't really need to do any analysis
of the small isolated no-data pockets insofar as doing neighbor analysis or
nibble etc.  Because in reality these areas fall at sea level, once
identified and isolated from the full no-data area outside the DEM's area of
interest I forced the values to 0m and merged them back with the original
grid.

Thanks for all the suggestions and help though.

Original Question:
I have a question regarding the clean up of a Grid.
The problem I have is with small clusters of no-data values being located up
inlets, along the coastal area of my grids.  They are the result of the grid
cells extending slightly beyond the actual coastal boundary used in
generating the DEM's.
What I would like to do is be able to identify these "pockets" of no data in
the original grid and use them as a mask for the nibble request to clean up
these small annoying pockets in the DEM.

If anyone has any clue how to go about selecting these no data values it
would be a great help.  Or, alternatively if there are other ideas on how to
clean up the coastal areas with other GRID based clean-up options that i'm
not familiar with i'd love to hear any suggestions.

__________________________________________________
Ryan Preston.

GIS Research Technician
Teagasc/Compass Informatics
19 Nassau St
Dublin 2, Ireland
rpreston@compass.ie