review exp_n and survey_shannon

[andybeet]

We have been using exp_n and survey_shannon to try to understand the shifts in species composition from year to year. I am not sure we can use these two in tandem to help explain this.

In addition, the survey_shannon index for each YEAR ,EPU, SEASON is being calculated as the mean of the CRUISE6, STATION, STRATUM level shannon indices, which is quite different from pooling all STATION level data over the year and calculating the index. I guess it depends what we want to say. Overall system diversity (along with true number of species caught*) or a function of STATION level diversity?

Another reason we may want to shift to an overall system level diversity is that we can easily explain changes in the index by looking at the abundance of the species within the YEAR. This is not so when the diversity index is a index of indices

* We might want to adjust for effort (which is kind of what exp_n is trying to do)

This would be the system level shannon index

And the true number of species

@sgaichas @jcaracappa1 @BBeltz1