Percent Good

Percent Good (PG) indicates what fraction of the available pings were accepted for a given ensemble. It is calculated independently for each bin in a given ensemble (profile). The minimum for accepteance is 30PG.

If you ran quick_adcp.py to set up your processing tree, then you used a value of PG whose default is 50PG (see below). Setflags is run and that percent good value is applied to the database before you edit. You can adjust the PG value in the gautoedit editing interface, but TAKE NOTE:

• this only shows you the effect of changing PG on your data, it does not change the data
• PG can only be set in the database with setflags. You need to run setflags at the end of your editing session, so if you want to change it, change it then.
• You may need to select “do not show profile flags” to see data below the PG value applied by setflags

Percent Good defaults

In regions where data are good, PG will often stay very high (near 100 percent) for most of the profile, plummeting to a low value at the bottom. In regions where bubbles or other acoustic noise cause signal attenuation or blockage, percent good can have a more linear profile with depth, starting well below 100PG in the upper bins, decreasing slowly in deeper bins. Under these conditions, there are often dubious-looking tails in the velocity near the bottoms of the profiles.

Quick_adcp.prl uses 50PG as its default cutoff because that seems to be enough to o trim those dubious tails without losing too much data. In the former case (good data), only 1-2 bins are lost at the bottom, because the dropoff is so pricipitous. In the latter case, (bubbles or ship heaving during bad weather) the data at the bottom of the profiles is questionable and it seems like a good idea to trim them.

Near-surface editing (underway bias)

Underway bias can exist due to ringing or bubbles, and necessitates the removal of bins near the surface. It is not always easy to know how many bins to remove. The velocity signature of underway bias is a ship-speed-dependent bias in the top few bins which is in the direction of ship’s motion. Sometimes the “show amp and forward” option in gautoedit is useful. One way to remove near-surface bias is to simply flag the top N bins (eg. the top 2 bins removed)

A more selective criterion is to flag as bad, any bins near the surface which are lower than the maximum PG for that profile. An example is shown here (parameters for this example may be set in aflagit_setup.m: read and edit if needed)

(thumbnails are links to postscript files)

no editing	gautoedit defaults (PG50)	example