Live vs. dead fish landing by area

[chantelwetzel-noaa]

I thought this would be a good item for an example issue post. Posting issues here can help us track ongoing items and facilitate team discussion between meetings.

In the 2021 assessment of copper rockfish north of Point Conception, there was a clear shift in the length compositions coming from the commercial fishery starting around 2008. At that time the condition code (landed live vs. dead) was not populated in PacFIN for California so we were unable to explore whether this change in the data was due to changes in the condition of landed fish. This issue has been corrected in PacFIN which allows us to potentially answer this question.

It does appear that the majority of commercial landings occurring north of Point Conception are fish being landed live which does align with a shift in the length samples in this area are arising primarily from fish landed live:

I am sure we will want to dig into this further.

To do:

• [x] Calculate the proportion of fish landed live vs dead by year and year and then compare that to the length sampling frequency.

[chantelwetzel-noaa]

I calculated the proportion of length samples coming from fish landed live in the commercial fishery by area:

And then calculated the proportion of catch that is landed live by area from the commercial fishery:

It is clear that the proportion of length measurements is not proportional to the catch from live fish for either area. I don't think this is too surprising given the challenges of sampling fish being landed live but it is good to know.

[chantelwetzel-noaa]

Based on a suggestion made during our meeting last week I simulated data for alternative proportions of live versus dead lengths for each area. I calculated the mean and standard deviation of lengths in the north and the south for two periods of time for live and dead lengths: all years (1994+) and recent years (2010+). The frequency of fish being landed live has increased over the last decade in the northern area (north of Point Conception) with also the length of fish being landed decreasing compared to the average length calculated since 1994. Additionally, since the proportion of lengths coming from live fish does not equal the proportion of fish being landed live I simulated an additional two scenarios for each area: one that simulates the number of lengths from each condition type based on the proportion of length samples for each corresponding to each time period and a second call "perfect sampling" where the proportion of lengths coming from live versus dead aligns with the landing proportions.

The distribution of the composition data and the mean length (combined = black dashed line, light blue solid = live, light red solid = dead) has a difference of approximately 2 cm or greater, particularly for the northern area based on the proportion of fish being sampled live versus dead for recent years (2010+). The difference is less in the south for both periods but particularly in recent years where the difference between the two means is ~ 2 cm.