Pre-treatment and post-treatment model selection statistics.

Models were ranked by Akaike weight, and the collection of best-supported models corresponding to a 95% cumulative weight is shown. Models were generated for the intrinsic sea louse population growth rate (r) and all subsets of the covariates cohort surface area (A), temperature (T), salinity (S), and number of months since treatment (M), with the exception that the covariate for months since treatment (M) is excluded from the pre-treatment models. Columns show the negative log likelihood values (NLL), Akaike Information Criterion (AIC), the AIC differences (ΔAIC), and the Akaike weights.