A statistical mechanistic approach including temperature and salinity effects to improve salmon lice modelling of infestation pressure

Abstract

Salmon lice Lepeophtheirus salmonis pose a major threat to the sustainable development of salmonid farming. To investigate effects of farm-origin salmon lice on wild salmonids, salmon lice dynamics are typically simulated using models that depend on experimentally determined rates of development, reproduction, mortality and infestation. Several recent studies provide new estimates of how these demographic rates depend on temperature and salinity. Here, we review and synthesize these studies and test if updating a salmon lice infestation model based on the new insights improves predictions of salmon lice infestations on salmon post-smolts in experimental cages in the sea. This model predicts spatiotemporal variation in infestation pressure based on weekly monitoring data of salmon lice and sea temperature in all salmonid fish farms in Norway, here supplemented by temperature and salinity data from a regional ocean model. Using data from 2012-2017 to select model formulation, we found the largest improvement in explanatory power by incorporating a salinity-dependent infestation rate. Updating functions for temperature-dependent egg production and infestation rates led to smaller improvements. Moreover, results suggest additional effects of temperature and a possible temperature-salinity interaction effect, not captured by the modelled processes. Out-of-sample predictions for experimental cage data from 2018-2020 confirmed that the uncertainty was realistically quantified, but also showed that associations of salmon lice infestations with salinity and temperature had changed. These results provide a field evaluation of experimental data and point to a knowledge gap regarding the combined effects of temperature and salinity on salmon lice infestations.