Integration of hydrodynamics into a statistical model on the spread of pancreas disease (PD) in salmon farming

Abstract

Pancreas disease (PD) is an emerging disease in salmon farming caused by the salmonid alphavirus (SAV). SAV is evidently spread horizontally between neighbouring salmon farms, but whether such transmission occurs by passive drift in the water current or via fomites is not known. We tested whether hydrodynamic modelling contributes to explain the spread of PD, in which case SAV is likely to spread by passive drift. We present a simple logistic regression model that accounts for the effect of PD in the neighbourhood on the probability of acquiring PD in cohorts of farmed salmonids from an area on the west coast of Norway between 2005 and 2008. For a given cohort, we calculated infection pressure (IP) based on Euclidean distance, seaway distance or estimated water contact to sites with PD, and compared the amount of variance explained in the regression model by the different variants of IP. Water contact between a discharging farm site and a receiving site was calculated by simulating particle discharge using a hydrodynamic model. IP estimated by water contact was the best predictor of PD cases and controls in the model, which performed significantly better than IP estimated by seaway distance or Euclidean distance. Since the spread of PD in the study area was best explained by modelled water velocity, we conclude that PD is likely to be spread by passive drift of SAV in the water current.