Notes and paper outline

[ericward-noaa]

Should we be using environmental variables? Examples could be heating degree days (which is complicated because the mechanism would dictate which window to focus on). Or ROMS?

Some distributions are constant over time -- LB: one thing to calculate would be pairwise similarity metric of overlap between sectors. Possible plot: (1) calculate seasonal overlap between sectors, for each year (2) make pairwise correlation style plot with each overlap metric in each year as a different point.

Paper Outline: Importance of quantifying individual differences versus aggregates

• Contrast fisheries impact (in space/time) with the aggregate versus changes in individual behavior
• Which are most useful for predicting change?
• Metrics to focus on:
• 1. Spatial inertia
• 2. Spatial footprint (this would be good for a map -- but could be hard to visualize trends)
• 3. Seasonality / Effective days fished (p 13) shows catch shares effect for WC in 2010
• 4. Distance from port?

Scale: We could aggregate summaries across sectors, or look at each individually. It probably doesn't make sense to include

• should we delete nearshore and OA altogether because of low coverage rates? Coverage is spatially stratified -- but may have some confounding effects. Will re-do with LE / CS Trawl, LE Sablefish, At Sea hake
• also focusing on these 3 helps tell a more unified story with AK data

Alaska: eastern GOA or western GOA for sablefish could be a comparison Could use ESR lines for non-trawl sectors (sablefish, halibut), right in the middle of PWS (NMFS areas line up). For example, Sitka/Yakutat would be EGOA ports, and Kodiak/Sand Cove would be WGOA ports LB: 147 is cutoff for central / eastern GOA for pollock fishery NMFS area 610 = western, 620/630 = central, 640/650 Eastern. Is there trawling in 640?

[ericward-noaa]

sub areas in WC data more limited: N/S. Sectors also smaller (3) simpler to have a more basic null model?

JW: General theme: is fishing ground available to each port changing over time? KR: hypotheses that divide by subarea vs port could be also exploring divisions by ecological factors (subarea = fishers tracking fish) vs ports (economic/management/social factors) that are driving more fine scale change. pre-post/IFQ as binary indicator variable? indicators by sector: AK longline / pollock / rockfish / misc groundfish WC: hake, LE sablefish / CS worth including COG / GIC / other? co-location across ports? pairwise as simple metric?