Shark-management strategies effect-size analyses

Analyses to assess probability of detecting differences in the number of shark bites on humans with/without shark management strategies in place.

See also the Australian Shark-Incident Database (ASID).

Prof Corey J. A. Bradshaw
Global Ecology | Partuyarta Ngadluku Wardli Kuu, Flinders University, Adelaide, Australia
July 2023/updated October 2023
e-mail

contributors: Charlie Huveneers

Paper

Huveneers, C, C Blount, CJA Bradshaw, PA Butcher, MP Lincoln Smith, WG MacBeth, DP McPhee, N Moltschaniwskyj, VM Peddemors, M Green. 2023. Shifts in the incidence of shark bites and efficacy of beach-focussed mitigation in Australia. Marine Pollution Bulletin 198:115855. doi:10.1016/j.marpolbul.2023.115855

Abstract

Shark-human interactions are some of the most pervasive human-wildlife conflicts, and their frequencies are increasing globally. New South Wales (Australia) was the first to implement a broad-scale program of shark-bite mitigation in 1937 using shark nets, which expanded in the late 2010s to include non-lethal measures. Using 196 unprovoked shark-human interactions recorded in New South Wales since 1900, we show that bites shifted from being predominantly on swimmers to 79 % on surfers by the 1980s and increased 2–4-fold. We could not detect differences in the interaction rate at netted versus non-netted beaches since the 2000s, partly because of low incidence and high variance. Although shark-human interactions continued to occur at beaches with tagged-shark listening stations, there were no interactions while SMART drumlines and/or drones were deployed. Our effectsize analyses show that a small increase in the difference between mitigated and non-mitigated beaches could indicate reductions in shark-human interactions. Area-based protection alone is insufficient to reduce shark-human interactions, so we propose a new, globally transferable approach to minimise risk of shark bite more effectively.

Scripts

• SMSeffectSizeAnalysis.R: main R code for analysis
• new_lmer_AIC_tables3.R: source code for information-theoretic algorithms
• r.squared.R: source code for calculating goodness-of-fit for linear models (including mixed-effects models)

Data

• beachmesh.csv: data of shark bites at beaches with and without beach-mess protection (all interactions)
• beachmeshonlybites.csv: data of shark bites at beaches with and without beach-mess protection (bites only)
• sms.csv: data describing shark bites across different beaches before and after shark-management strategies

Required R packages

• performance
• sjPlot
• lme4
• doSNOW
• snow
• iterators
• foreach
• parallel