search

SOLV-Code / Open-Source-Env-Cov-PacSalmon

Open-source, Human-centered Data Management for Environmental Covariates in Pacific Salmon Models
GNU General Public License v3.0
0 stars 0 forks source link

PDO: Sum vs. Mean? #37

Open SOLV-Code opened 6 months ago

SOLV-Code commented 6 months ago

The DFO Fraser River forecasts and the NOAA ocean conditions both use the PDO data, but the DFO forecasts use mean Nov-Mar while the NOAA index uses sum Dec-March and sum May-Sep.

Why mean vs. sum?

Have started a note with some diagnostics

Schgrant commented 6 months ago

We always averaged for the forecast but I'll ask Dr. Charles Hannah, maybe that is inappropriate with an indices if NOAA doesn't do this, maybe we are missing something.

SOLV-Code commented 6 months ago

Given the way the scatterplot comes out, it doesn't look like it makes much of difference whether you use mean or sum. There may be some years where including/excluding Dec changes the picture a bit, but that's a separate discussion.

A more important question might be whether it makes sense to use the de-trended version of the PDO Index (as per the note in one of the source files). Maybe the long-term underlying change should be part of the signal used as an indicator, or included as a separate indicator?

More detailed discussion of comparison in this note