Introduction

SLIM is an FOSS sea lice lifecycle simulation tool. Its intended goal is to simulate sea lice pandemics in salmonid farms.

However, it is much more than a simple epidemic simulator. The tool is intended to answer What if? questions in terms of treatment strategies. For example, What will be our loss if we apply treatment as soon as the AF aggregation ratio reaches x? How long does it take before treatment resistance becomes significant? (regardless of what we mean by that)

Rationale

Sea lice (singular sea louse) are a type of parasitic organisms that occur on salmonid species (salmon, trout, char). We differentiate between farm salmon (in the sea cages) and wild fish (wild salmonid species in the reservoir, that is external environment). Salmon in sea cages can be infected when the sea lice crosses into the sea cage from the reservoir where it occurs on the wild fish.

Chemical or biological treatment can be applied to the sea cages against the sea lice. Currently, the only chemical treatment modelled in here is Emamectin Benzoate or EMB. Regardless of the adopted pesticide, there is extensive proof that genetic resistance to treatment developed after a few years, making each treatment cycle less and less effective. Therefore, farmers have resorted to a wide range of alternatives.

In the past few years, biological treatment via cleaner fish (typically lumpfish and ballan wrasse) has been introduced with mixed results. Another solution deployed is a time break between the farming cycles, typically known as fallowing, in order to reduce the amount of lice surrounding the cages before repopulating them again.

All treatments have a nonnull cost both in economical terms, collateral damage (intoxication, increased stress levels etc.) or ecological impact on other cultures. While farmers typically belong to cooperatives or consortiums there are no obligations to apply treatment simultaneously to everyone else. Therefore, defection and egoistical behaviour are the norm.

SLIM is the Sea Lice Model associated with a funded BBSRC project on the evolution to resistance to treatment in sea lice (BBR009309). The project kicked off at the University of Stirling, Scotland and then is being co-developed by the University of Stirling and University of Glasgow. We reputed that a large number of sea lice models have been developed on geographical contests different from the Scottish one, and the lack of readily available open source models gave us a motivation to start this project.

Features of this Program

SLIM is, in essence, a statistical model of sea lice epidemics that aims to be the opposite of this. That is, we aim for this project to provide a battery-ready developing and usage experience.

Overall, this project offers the following tools:

There are thus two designed workflows:

  1. Use the simulator to generate a session dump, then run the GUI to visualise such data.

  2. Use the strategy optimiser to obtain the best treatment parameters (more on this later), save the resulting configuration artifact.

The simulator is typically executed as a standalone command, but it provides a relatively easy Python API for embedding purposes.

The simulator is also battle-tested (see tests).

What makes this simulator enticing, however, is the underlying model. For a longer description of what and how we simulate please visit Model Overview.