Revealing migration and reproductive habitat of invasive fish under an active population suppression program

Abstract

Endemic species face a variety of threats including predation from non‐native invaders. In some cases, however, invasive species can be managed by directly suppressing populations, and tracking technologies that allow researchers to identify movement patterns and aggregations representative of the population can facilitate suppression activities. In Yellowstone Lake (Yellowstone National Park, Wyoming), invasive lake trout (Salvelinus namaycush) have been the target of a population suppression program for over two decades. For this form of management, the reproductive period is particularly important because fish migrate to and from spawning grounds. From 2011 to 2014, adult lake trout (n = 317) in Yellowstone Lake were tracked using acoustic biotelemetry. After controlling for spatial and temporal dependency in the data, total abundance of unique individuals was estimated where migratory trajectories occurred at confirmed spawning sites. Aggregations and migratory trajectories were further estimated at locations where spawning had not previously been observed. Across years, the greatest number of individuals was observed along a migration corridor in the southwestern area of the lake. Novel strategies for analyzing acoustic telemetry data provided insights into the behavior of an invasive fish species. By betraying the positions of conspecifics, tagged fish revealed potentially important reproductive habitats and migration corridors that warranted further investigation as possible sites for population suppression.

Quantifying the Spatial Structure of Invasive Lake Trout in Yellowstone Lake to Improve Suppression Efficacy

“…Lake Trout spawning activity across the species' range in North America typically increases during the evening, immediately after sunset (Martin and Olver 1980). However, during previous studies of broad-scale movement patterns of Lake Trout in Yellowstone Lake, there was no evidence of diel patterns in spawning behavior (Gutowsky et al 2020). Therefore, tracking surveys were conducted from 0600 to 1600 hours with boats traveling at a maximum speed of 9.7 km/h.…”

Invasive Lake Trout Reproduction in Yellowstone Lake under an Active Suppression Program

“…Passive and active tracking of acoustically telemetered Lake Trout have also been used to identify movement corridors to spawning locations and to target aggregations that are proximal or moving towards these grounds through a “Judas fish” technique (Gutowsky et al. 2020; Williams et al. 2020).…”

“…In Yellowstone Lake, Lake Trout suppression has been conducted throughout the lake during the ice-free season but fall gillnetting effort has been heavily concentrated on known spawning locations, using larger mesh sizes to target spawning capable (>500 mm) females (Syslo et al 2011). Passive and active tracking of acoustically telemetered Lake Trout have also been used to identify movement corridors to spawning locations and to target aggregations that are proximal or moving towards these grounds through a "Judas fish" technique (Gutowsky et al 2020;Williams et al 2020). Based on assessment modeling results, annual exploitation levels of Lake Trout fully selected to suppression gill nets (≥4 years of age) has exceeded 50% since the mid-2010s (Syslo et al 2020).…”

“…Through a combination of direct observation and telemetry studies, some aspects of Lake Trout spawning in Yellowstone Lake have been well-documented. For example, littoral habitat in Yellowstone Lake is abundant (Bigelow 2009), and numerous Lake Trout spawning locations have been identified (Ruzycki et al 2003;Gutowsky et al 2020;Williams et al 2020). Lake Trout spawning in Yellowstone Lake occurs from September to early October, with the greatest proportion of spawning-capable (a fish capable of spawning within the current reproductive cycle, Brown-Peterson et al 2011) females generally captured in the last week of September.…”

A Novel Framework to Predict Relative Habitat Selection in Aquatic Systems: Applying Machine Learning and Resource Selection Functions to Acoustic Telemetry Data From Multiple Shark Species

“…Whereas RF models are unable to easily incorporate, RSF GLMMs can explicitly include individual ID as a random effect (Gillies et al, 2006). Further, generalized models can incorporate autocorrelation dependency structures (Zuur et al, 2017;Winton et al, 2018a;Griffin et al, 2019;Gutowsky et al, 2020), however, it is worth noting that defining the correct correlation structure still remains challenging within a use/availability (presences/pseudo-absences) sampling design (see Koper and Manseau, 2009;Fieberg et al, 2010). In this study, while BBMMs highlighted individual variation in space use, simplified GLMMs indicated including individual as a random effect contributed relatively less to explaining overall variance than the fixed effects alone.…”