Welcome to the interactive web schedule for the 2018 Midwest Fish & Wildlife Conference! For tips on how to navigate this site, visit the "Helpful Info" section. To return to the main Conference website, go to: www.midwestfw.org.

View analytic
Wednesday, January 31 • 8:40am - 9:00am
FISH HABITAT & GENETICS: Using GIS to Predict Habitat in Lakes: An Example Using Nearshore Substrate Categories

Sign up or log in to save this to your schedule and see who's attending!

AUTHORS. Douglas Zentner, University of Arkansas-Pine Bluff; Joshua Raabe, University of Wisconsin - Stevens Point, Timothy Cross, Minnesota Department of Natural Resources; Peter Jacobson, Minnesota Department of Natural Resources; Benjamin Schall, South Dakota State University

ABSTRACT. Nearshore habitat is an integral component of lake ecosystems that is often threatened by anthropogenic alterations. Quantifying habitat can allow for evaluation of potential factors limiting species and identification of areas for habitat restoration or protection projects. However, quantifying habitat is typically labor intensive and spatially limited (i.e., transects), or requires specialized field equipment combined with computer analyses (i.e., side scan sonar). Furthermore, quantification via these methods does not always result in an understanding of the processes influencing habitat distributions. We developed a framework that uses Geographic Information Systems (GIS) to answer habitat-based questions and allows for an evaluation of factors influencing habitat distributions. To asses this framework, we derived layers for nearshore areas to determine if we could accurately predict nearshore substrate categories in lakes. We used exposure, riparian height, maximum nearshore depth, wind power, and bathymetric aspect as model predictor variables. To develop and validate our model, we collected nearshore substrate data along transects in 28 lakes across Minnesota during the summers of 2014-2016. Prior to modeling, K-means clustering algorithm was used to group substrate data into three general categories: muck, sandy gravel, and coarse substrates. We randomly selected 85% of data to develop a Classification Tree and used the remaining 15% of data to validate the model. Using repeated sampling with replacement, we determined this model to be 62.0 – 71.0% accurate. We used Belle Lake as an example to determine spatial sources of error within our model. Our framework and methodology have the potential to increase our understanding of habitat distributions and structuring factors. We envision numerous applications for this GIS model, such as understanding factors that influence substrate distributions, identifying critical habitat (e.g., fish spawning habitat), and assisting management decisions on shoreline development or habitat protection, restoration, or enhancement.

Wednesday January 31, 2018 8:40am - 9:00am

Attendees (28)