Preliminary data - Part 2

I called the point goofy. Sometimes I'll say a data point is wonky. Technically it is what a statistician would call an outlier. It is a goofy & wonky outlier. It is perfect for me because it gave me opportunity to learn. Why is it so high?? To answer this I went back to the raw data. Here I'm showing the 123 kHz data as an example.

I found that this particular 100 m segment had very high numbers of very small scatterers consistent in size with large-bodied zooplanktors like Bythotrephes or Leptodora (see upper left panel). The mask applied to find these small scatterers seemed to work fine (upper right panel). The code to replace the small scatterers (we take them out of the data because they are too big to be eaten by a larval fish ) with nearby data looks to have worked (left hand bottom), but looks can be deceiving. In the last step where these scatterers are replaced with nearby data (bottom right), we can find evidence that the scatterers are still there (barely so, but trust me). So I think I have an issue to resolve. When there are very high densities of large-bodied zooplanktors it seems that the small-bodied zooplankton Sv estimates can be biased high. So you see there is a lot one can learn from a goofy data point. Fixing this issue will have to wait for another day. We say in America, "Close but no cigar."