From reactive towards anticipatory fishing agents

ABSTRACT

Governing human-environmental ecosystems is a complex problem. Rule-based fisheries models are faced with several challenges. First, for large geographical problems like oceans, they require considerable time to find satisfactory solutions. Second, they tend to be reactive rather than anticipatory. Behavioural assumptions directly impact fishers’ capacity for adaptation and behaviour, which influences possible management strategies. To capture style and speed of adaptation to changes in the environment, coupled human-environment models must progress toward cognitively and socio-culturally realistic representations of fisher decision-making. In this paper, we implement the erotetic decision-making model in the POSEIDON fisheries model. The agents replicate observed behaviours such as fishing the line of a Marine Protected Area, using Individually Tradable Quotas, and returning to favoured fishing locations, and learning to break rules given harsh constraints. This provides a principled proof that reasons-based cognitive structures allow for anticipatory behavioural adaptation rather than reactive behavioural adaptation.

KEYWORDS:

• Agent-based models
• reasons-based decision making
• coupled human-environment systems
• anticipatory behaviour
• POSEIDON

Acknowledgments

The work reported in this paper was undertaken as part of the Oxford Martin Programme on Sustainable Oceans, funded by the Oxford Martin School, and with financial support and other valuable contributions from the Ocean Conservancy. Many thanks to Aarthi Ananthanarayanan and Chris Dorsett for their constructive and helpful comments.

Disclosure statement

The authors have no competing interests to declare.

Notes

1. POSEIDON is open source on GPL-3 licence and available at https://github.com/CarrKnight/poseidon; the scenario files of all the simulations are present in the folders inputs/erotetic_paper.

2. This tests the appropriateness of the simulations as pattern-oriented modelling (Grimm et al., 2005).

3. Of course, there are significant socio-cultural differences between and within fisheries. For example, highly competitive fisheries may be characterised by economic considerations while small-scale fisheries may be more pro-social and consider adherence to cultural norms and legal frameworks to be critical. In the current paper, we explore the erotetic theory in principle and have thus sought to simplify the decision model to explore anticipatory rather than reactive agents.

4. If the agent visits an area, which is subsequently deemed unsatisfactory, the agent cannot return to that plot in the future. This is due to the fact that the agents in the current framework have perfect memory.

5. Under circumstances in which an MPA is introduced as well as a quota for bycatch (such that bycatch quotas are tradable) on a particular species, fleets have been known to avoid the line rather than to fish along it in order to avoid the protected species (Miller & Deacon, 2014; see also Toft et al., 2011).

6. The quantitative experiments below do not rely on empirically grounded parameter settings. Rather, they provide a proof of principle for the quantitative use of the erotetic ABM.

Additional information

Funding

This work was supported by the David and Lucile Packard Foundation; Gordon and Betty Moore Foundation; Laces Trust; Ocean Conservancy; Oxford Martin School, University of Oxford; Walton Family Foundation.