Understanding, Quantifying, and Reducing Bias in Fisheries-independent Visual Surveys

Authors

James KilfoilFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Biology

First Advisor's Name

Yuying Zhang

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Michael Heithaus

Second Advisor's Committee Title

Committee member

Third Advisor's Name

Joel Trexler

Third Advisor's Committee Title

Committee member

Fourth Advisor's Name

Wei Zeng

Fourth Advisor's Committee Title

Committee member

Fifth Advisor's Name

Matthew Campbell

Fifth Advisor's Committee Title

Committee member

Keywords

Survey methods, underwater visual census, elasmobranch, biases, unmanned aerial vehicle, baited remote underwater video system

Date of Defense

2-14-2020

Abstract

Understanding spatiotemporal changes in populations is vital for conservation managers to assess current recovery efforts, determine future conservation priorities, and forms the basis to explore complex ecological questions. In fisheries, these data have traditionally been collected using fisheries-independent surveys that rely on extractive sampling practices (e.g., longlines, gillnets, trawls). However, with the growing availability of low-cost, high-definition cameras, researchers are increasingly using visual surveys as a non-invasive alternative. Camera surveys have a number of advantages including their archivable data, and offer insights into species habitat use and behavior. However, the use of cameras has a number of inherent biases. Understanding, quantifying, and mitigating against these biases is critical if camera systems are to be used to inform management and policy. In this dissertation, potential biases were explored for two commonly used visual survey methods; baited remote underwater videos (BRUV), and unmanned aerial vehicles (UAV). Specifically, our objectives were to answer: (1) Are metrics of relative abundance derived from BRUVs linearly related to true changes in abundance for elasmobranchs, (2) Are these same metrics sensitive to changes in density-independent factors, and (3) Can UAVs be used to replace or supplement traditional diver transects for marine invertebrate species? Using a combination of standard and full-spherical camera BRUV deployments, Chapter One found that tradition BRUVs likely undercount sharks in high density environments, while also having lower probability of detection than full-spherical cameras. Using a spatially-explicit, individual-based-model, Chapter Two revealed that metrics of relative abundance derived BRUVs are also highly sensitive to factors unrelated to changes in abundance (e.g., swimming speed, current strength, and movement patterns). Lastly, using paired snorkeler-UAV transect sampling Chapter Three found counts derived from UAV transects did not significantly differ from divers, and offered a number of advantages over this traditional technique (increased percision, larger surveyed area, and automation). Furthermore, we found that UAVs could be used to improve sampling design used to quantify invertebrates, by estimating their distribution within a study region prior to initiating transect sampling. Collectively, these works improve our understanding and interpretation of video survey results that are used for management across the globe.

Identifier

FIDC008896

ORCID

0000-0002-3794-0085

Previously Published In

Kilfoil, J. P., A. J. Wirising, M. D. Campbell, J. J. Kiszka , K. R. Gastrich, M. R. Heithaus, Y. Zhang, and M. E. Bond. 2017. Baited Remote Underwater Video surveys undercount sharks at high densities: Insights from full-spherical camera technologies. Marine Ecological Progress Series, 585: 113-121.

Recommended Citation

Kilfoil, James, "Understanding, Quantifying, and Reducing Bias in Fisheries-independent Visual Surveys" (2020). FIU Electronic Theses and Dissertations. 4445.
https://digitalcommons.fiu.edu/etd/4445