Miller and Gotshall and Miller and Geibel datasets understanding

[brianlangseth-NOAA]

@melissamonk-NOAA Thanks for putting up the miller and gotshall and miller and geibel datasets. For my understanding (and for anyone else's who may use this data) Im adding this issue to discuss how to use this data properly.

I see there are 133 quillback caught across the board (see Table 2 or pivot.2 in the dataset) for Miller and Gotshall, but according to Table 3, there are only 33 length measurements, and these only occur in1959 and 1960.

For Miller and Geibel, I see 57 quillback lengths in 1959-1960. These dont look like the same data as Miller and Gotshall, so I plan to combine. I see a column "COAST_DIST" in the data, but this looks like coastal distance (from shore) instead of District. I plan to use our MRFSS lookup table to convert county to district.

1. Are you ok with combining these datasets?
2. Sample sizes are going to be tricky to obtain. I can derive number of trips from Miller and Gotshall, but there is no trip info for Miller and Geibel. Is there trip info for Miller and Geibel?
3. Do you know whether these are fork lengths or total lengths?

[brianlangseth-NOAA]

Im assigning ports/counties to districts in these datasets. Is the following correct?

For Miller and Gotshall (port to district)
Shelter Cove = Wine (Or in 1960 was this part of redwood?)
Fort Bragg = Wine
Bodega Bay = Bay
Princeton = Bay

For Miller and Geibel (county to district)
DelNorte and_Humboldt = Redwood
Mendocino and Sonoma = Wine
San Francisco Bay = Bay
Santa Cruz and Monterey = Central

[melissamonk-NOAA]

Miller and Gotshall - there are 33 length measurements from 59-60. The 133 quillback are the catches, but not all were measured. This study focused on blue rockfish and lingcod.

1. Yes, we can combine the two datasets.
2. I agree. We can look to see what we've used in past assessments.
3. Both measured total length

Miller and Gotshall = assign Shelter Cover to Wine.
Miller and Geibel = Mendocino and Sonoma - this depends since Sonoma is in the Bay district We don't have a lot of info on where these fish were actually caught. The only port in Sonoma is Bodega Bay. My guess is Wine is probably ok.

[brianlangseth-NOAA]

Thanks Melissa! I'll apply a conversion to fork length.

2. Regards number of trips, I think it will actually be impossible to determine this with certainty given the data we currently have for either dataset. The Miller-Gotshall data have total number of trips and total number of fish (across many species) within a month-port-year-mode combination. Thus, we cannot know the number of trips that caught only a quillback. Rather, we can only know the maximum (whatever is smaller between either the total number of quillback or total number of trips for that month-port-year-mode combo) or minimum (1) number of trips.

For example
-13 trips for April 1960 in Bodega Bay yet only 2 quillback. Where those quillback caught on 1 trip or 2? No way to know.
-15 trips for August 1960 in Bodega Bay yet 6 quillback. Thus, N trips could be between 1 and 6.

I looked at each year-month-port-mode combo and the number of fish measured is never greater than the number of total trips for that combo.

For Copper N (page 17 of the report) "A similar methodology was applied for the CRFS and Miller and Gotshall data, where county, water area, interview site, and mode were used to determine the number of unique trips." though I dont know how you did that because I dont see that level of detail for Miller and Gotshall.

I see a couple options:
A. Assume each fish accounts for a separate trip. Ntrip would be number of fish.
B. Assume each year-month-port-mode combination as the number of trips. Ntrip would be the number of combos of that within a year.

For Miller and Geibel, Ntrip would be year-county-mode since month is not provided.

My preference is to use B. Its not absolute, but accounts for some measure of correlation