search

rBatt / trawl

Analysis of scientific trawl surveys of bottom-dwelling marine organisms
1 stars 0 forks source link

Increase strata, reduce replicates (K) #85

Closed rBatt closed 8 years ago

rBatt commented 8 years ago

A suggestion from Morgan: 9 replicates (substrata, K) is a lot, but 14 strata (sites) per region-year is not a lot. Maybe define strata on a 0.5º grid, and substrata on a 0.25º grid, which could lead to as many as 4x's as many strata (but probably not quite that high, because each 1º isn't sampled that thoroughly),

rBatt commented 8 years ago

This is becoming more and more of an ill-formed objective/ question. 1) we will use many years in 1 model now, so our sample size increases that way; 2) Covariates for Psi and Theta (presence and detection) actually exist on the per-sample resolution. So the distinction between a replicate and site are becoming fuzzier and fuzzier.