Computational protocol: Using underwater video to evaluate the performance of the Fukui trap as a mitigation tool for the invasive European green crab (Carcinus maenas) in Newfoundland, Canada

Similar protocols

Protocol publication

[…] We built two linear mixed-effects models using the nlme package () in R () in order to test whether the presence of the camera had an effect on green crab catch. We analyzed a subset of the green crab catch data which included only Fair Haven, NL and Little Port Harmon, NL. All other sites were excluded from our subset due to either zero green crab catch, or low mean catch rates (). We did not see any meaningful relationship between deployment duration and catch ( and ). However, the soak times were not consistent between Fair Haven (range = 21.8–24.3 h; mean = 22.9 h; SD = 0.8) and Little Port Harmon (range = 7.4–14.1 h, mean = 11.1 h; SD = 2.8) (). To account for this, we created a separate model for each location because the underlying effect of soak time on catch was potentially unique to each site (). These two models tested the fixed effects of camera presence (i.e., camera present, camera absent) and duration on catch-per-deployment. We did not include deployment type (i.e., daytime, overnight, mixed) in our final models as the term was insignificant in the Little Port Harmon model, and lacked sufficient factor levels in the Fair Haven model (mixed deployments only). Due to the paired nature of our design we designated each camera and non-camera pair as a single deployment, which was included in each model as a random effect. The residuals for both the Fair Haven and Little Port Harmon models met the assumptions for homogeneity, normality, and independence. […]

Pipeline specifications

Software tools lme4, nlme
Application Mathematical modeling