Modeling fish habitat: model tuning, fit metrics, and applications

Knowledge of the habitat associations and spatial–temporal distributions of wild animals is essential for successful ecosystem management, and effective analytical approaches are key to develop accurate models of these relationships. We explore the influence of several modeling techniques, tuning parameters, and assignment thresholds on a variety of model fit metrics to characterize habitat associations and make spatial–temporal predictions of species distribution based on a nine-year acoustic telemetry fish tracking dataset from a freshwater system. Unweighted generalized linear mixed models (GLMM) and random forests (RF) had the highest prediction accuracy of fish occupancy (> 84%) and precision (positive predictive value accuracy), but because the data were imbalanced (> 70% absences), predictions had low sensitivity (accuracy of true presences, < 45%), and therefore, low accuracy balance. Model weighting to prioritize presences and lowered presence probability thresholds both produced more balanced models, but RF exhibited low sensitivity to alterations in probability thresholds. Model weighting presents a straightforward approach to balance class accuracy in imbalanced datasets, which are common in species distribution samples. However, there is a wide range of weighting options and an important trade-off between model sensitivity and precision, either of which may be favoured depending on the research question or management application.

[1]  G. A. Goodchild.  Fish habitat is everyone's business, Canada's fish habitat management programme , 2004 .

[2]  Jacob Cohen,et al.  Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. , 1968 .

[3]  Euan S. Harvey,et al.  Potential of video techniques to monitor diversity, abundance and size of fish in studies of Marine Protected Areas , 2003 .

[4]  Bogdan E. Popescu,et al.  PREDICTIVE LEARNING VIA RULE ENSEMBLES , 2008, 0811.1679.

[5]  Francis Juanes,et al.  Listening to fish: Applications of passive acoustics to fisheries , 2006 .

[6]  A. Fisk,et al.  Biotelemetry informing management: case studies exploring successful integration of biotelemetry data into fisheries and habitat management , 2019, Canadian Journal of Fisheries and Aquatic Sciences.

[7]  Karl A. Lamothe,et al.  Characterizing species co‐occurrence patterns of imperfectly detected stream fishes to inform species reintroduction efforts , 2019, Conservation biology : the journal of the Society for Conservation Biology.

[8]  D. R. Cutler,et al.  Utah State University From the SelectedWorks of , 2017 .

[9]  David A. Mann,et al.  Listening to Fish , 2006 .

[10]  Brandon M. Greenwell.  pdp: An R Package for Constructing Partial Dependence Plots , 2017, R J..

[11]  Richard J. Hobbs,et al.  Habitat Restoration—Do We Know What We’re Doing? , 2007 .

[12]  Michael B. Morrissey,et al.  Multiple Regression Is Not Multiple Regressions: The Meaning of Multiple Regression and the Non-Problem of Collinearity , 2018, Philosophy, Theory, and Practice in Biology.

[13]  Mollie E. Brooks,et al.  Generalized linear mixed models: a practical guide for ecology and evolution. , 2009, Trends in ecology & evolution.

[14]  Robin Freeman,et al.  Emerging Network-Based Tools in Movement Ecology. , 2016, Trends in ecology & evolution.

[15]  Bette A. Loiselle,et al.  Avoiding Pitfalls of Using Species Distribution Models in Conservation Planning , 2003 .

[16]  Christopher M Holbrook,et al.  Acoustic telemetry and fisheries management. , 2017, Ecological applications : a publication of the Ecological Society of America.

[17]  Brendan A. Wintle,et al.  Is my species distribution model fit for purpose? Matching data and models to applications , 2015 .

[18]  S. Richards,et al.  Prevalence, thresholds and the performance of presence–absence models , 2014 .

[19]  Zhan Wang.  The Applications of Deep Learning on Traffic Identification , 2015 .

[20]  A. Townsend Peterson,et al.  Novel methods improve prediction of species' distributions from occurrence data , 2006 .

[21]  J Elith,et al.  A working guide to boosted regression trees. , 2008, The Journal of animal ecology.

[22]  Joshua J. Millspaugh,et al.  High-Tech Behavioral Ecology: Modeling the Distribution of Animal Activities to Better Understand Wildlife Space Use and Resource Selection , 2001 .

[23]  Scott M. Williams,et al.  A balanced accuracy function for epistasis modeling in imbalanced datasets using multifactor dimensionality reduction , 2007, Genetic epidemiology.

[24]  A. Adams,et al.  A practical method to account for variation in detection range in acoustic telemetry arrays to accurately quantify the spatial ecology of aquatic animals , 2019, Methods in Ecology and Evolution.

[25]  J. Elith,et al.  Determinants of reproductive success in dominant pairs of clownfish: a boosted regression tree analysis. , 2011, The Journal of animal ecology.

[26]  J. Rosenfeld,et al.  Information needs for assessing critical habitat of freshwater fish , 2006 .

[27]  J. Lobo,et al.  Threshold criteria for conversion of probability of species presence to either–or presence–absence , 2007 .

[28]  Max Kuhn,et al.  caret: Classification and Regression Training , 2015 .

[29]  C. Krueger,et al.  Seasonal habitat-use differences among Lake Erie’s walleye stocks , 2020, Journal of Great Lakes Research.

[30]  John Bell,et al.  A review of methods for the assessment of prediction errors in conservation presence/absence models , 1997, Environmental Conservation.

[31]  G. Pierce,et al.  A comparison of approaches for modelling the occurrence of marine animals , 2008, Hydrobiologia.

[32]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[33]  M. Heupel,et al.  Ghosts in the data: false detections in VEMCO pulse position modulation acoustic telemetry monitoring equipment , 2015, Animal Biotelemetry.

[34]  J. Evans,et al.  Modeling Species Distribution and Change Using Random Forest , 2011 .

[35]  Daniel Boisclair,et al.  Fish habitat modeling: from conceptual framework to functional tools , 2001 .

[36]  T. Dawson,et al.  Selecting thresholds of occurrence in the prediction of species distributions , 2005 .

[37]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[38]  J. Andrew Royle,et al.  ESTIMATING SITE OCCUPANCY RATES WHEN DETECTION PROBABILITIES ARE LESS THAN ONE , 2002, Ecology.

[39]  Hugh P Possingham,et al.  Current practices in the identification of critical habitat for threatened species , 2015, Conservation biology : the journal of the Society for Conservation Biology.

[40]  Simon Ferrier,et al.  Evaluating the predictive performance of habitat models developed using logistic regression , 2000 .

[41]  Darren J. Bender,et al.  Assessing critical habitat: Evaluating the relative contribution of habitats to population persistence , 2010 .

[42]  N. Klinard,et al.  Living until proven dead: addressing mortality in acoustic telemetry research , 2020, Reviews in Fish Biology and Fisheries.

[43]  Steven J. Cooke,et al.  Conducting and interpreting fish telemetry studies: considerations for researchers and resource managers , 2019, Reviews in Fish Biology and Fisheries.

[44]  John T. Finn,et al.  Applying network methods to acoustic telemetry data: Modeling the movements of tropical marine fishes , 2014 .

[45]  A. Zuur,et al.  Mixed Effects Models and Extensions in Ecology with R , 2009 .

[46]  D. Bates,et al.  Fitting Linear Mixed-Effects Models Using lme4 , 2014, 1406.5823.

[47]  J. Kocik,et al.  Aquatic animal telemetry: A panoramic window into the underwater world , 2015, Science.

[48]  Angela B. Collins,et al.  Variation in the performance of acoustic receivers and its implication for positioning algorithms in a riverine setting , 2008 .

[49]  R. Real,et al.  AUC: a misleading measure of the performance of predictive distribution models , 2008 .

[50]  J. Kie,et al.  A Comparison of Two Modeling Approaches for Evaluating Wildlife–Habitat Relationships , 2009 .

[51]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[52]  Kim Whoriskey,et al.  Current and emerging statistical techniques for aquatic telemetry data: A guide to analysing spatially discrete animal detections , 2019, Methods in Ecology and Evolution.

[53]  Navinder J. Singh,et al.  Linking Movement Ecology with Wildlife Management and Conservation , 2016, Front. Ecol. Evol..

[54]  Brian J. Smith,et al.  Space use and relative habitat selection for immature green turtles within a Caribbean marine protected area , 2020, Animal Biotelemetry.

[55]  M. Termansen,et al.  The use of genetic algorithms and Bayesian classification to model species distributions , 2006 .

[56]  G. Tang,et al.  Indian Hedgehog: A Mechanotransduction Mediator in Condylar Cartilage , 2004, Journal of dental research.

[57]  P. Levin,et al.  Habitat triage for exploited fishes: Can we identify essential “Essential Fish Habitat?” , 2005 .

[58]  G. Fay,et al.  A spatial point process model to estimate individual centres of activity from passive acoustic telemetry data , 2018, Methods in Ecology and Evolution.

[59]  Bernard Bobée,et al.  A review of statistical methods for the evaluation of aquatic habitat suitability for instream flow assessment , 2006 .

[60]  Eric Hervet,et al.  Applications for deep learning in ecology , 2019, Methods in Ecology and Evolution.

[61]  Hugh P Possingham,et al.  Tradeoffs of different types of species occurrence data for use in systematic conservation planning. , 2006, Ecology letters.

[62]  Scott A. Bonar,et al.  Standard Sampling of Inland Fish: Benefits, Challenges, and a Call for Action , 2002 .