Abstract
Interest in marine tidal turbines, particularly in coastal waters, raises concerns about collisions between marine wildlife and underwater turbine blades. Prediction methods for collisions are necessary to evaluate possible consequences for marine animal populations. Existing collision risk models, based on analytical solutions, assume simplistic non-behavioural traits. This paper seeks to advance these collision models to represent real behaviours of marine species by extending an existing numerical Agent-Based Model (ABM) to include predictions of collisions.
The ABM successfully reproduced the results of the Collision Risk Model [1]. The ABM offers the advantage that the distribution of marine animals around the turbine does not need to be specified a priori, but arises from the swimming behaviours of the individuals within the model.
The ABM was applied to predict the impact of different swimming behaviours on collision rates for migrating silver eels passing a tidal turbine in Strangford Narrows. Just 1.1% of eels passing the Narrows were predicted to collide with the turbines. Different vertical swimming behaviours influenced the number of eels leaving the estuary and the number of predicted collisions. Other behaviours (e.g. active turbine avoidance) could be included in the ABM making this a valuable method for assessing turbine interactions.
Highlights:
- A collision risk model for marine animals and underwater turbines was added to an agent-based model of fish movement.
-
Agent-based model results were comparable to an analytical collision risk model in benchmark tests.
-
In a real world test, turbine collision rates were predicted to be relatively low (<1.1%).
-
Including diel behaviours for eels in the ABM affected the likelihood of collision.