Development of a GIS Model to Predict Muskellunge Spawning Habitat in Northern Wisconsin Lakes

GIS Modeling;Spawning Muskellunge;Habitat Preferences of Spawning Muskellunge in Northern Wisconsin Lakes;School of Natural Resources and Environment

Nohner, JoelBreck, James ;

University of Michigan

关键词: GIS Modeling;    Spawning Muskellunge;    Habitat Preferences of Spawning Muskellunge in Northern Wisconsin Lakes;    School of Natural Resources and Environment;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/62090/NohnerThesis2009.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

This study determined the habitat preferences of spawning muskellunge in northern Wisconsin lakes and used these preferences to create two GIS-based models that predict the location of muskellunge spawning habitat. This information will enable efficient conservation of muskellunge spawning habitat, which has been implicated in declining natural reproduction. Muskellunge spawning sites were identified using spotlighting surveys and verified by the presence of muskellunge eggs. Aquatic vegetation and substrate maps were created using visual surveys to determine habitat preference and train the models. Vegetation was categorized structurally, and muskellunge preferred to spawn over emergent sedges and rushes as well as submersed short grasses and mat-forming vegetation. Muskellunge preferred sand, cobble, and coarse benthic organic matter substrates, areas with high potential groundwater flow, and areas adjacent to wetlands. Moderate to steep slopes were preferred for spawning, as were locations near bays and points. While shorelines facing east to north-east and south to south-west were slightly preferred, the biological connection to this pattern is likely tenuous. Muskellunge Spawning Habitat Models (MSHM) 1 and 2 were created using the Maxent modeling program. The models utilized the difference between characteristics of spawning sites and available habitat to assign probabilities of spawning across each variable. These probabilities were in general agreement with the spawning habitat preferences documented in this and other studies. While MSHM1 uses only data which can be obtained remotely in Wisconsin, MSHM2 utilizes low-cost habitat surveys to slightly improve model performance. MSHM1 and MSHM2 were tested by withholding 25% of the spawning sites from model training for testing. Both models performed significantly better than random at predicting spawning locations using a binomial test, and the area under the curve analyses are evidence that each model possesses reasonable efficiency. The models assign a probability of muskellunge spawning to cells in a raster grid, and these values can be used to rank the best spawning habitat in each lake. For example, using either MSHM1 or MSHM2, a manager could identify the best 10% of available habitat and protect approximately half of the muskellunge spawning sites. MSHM2, which includes variables from habitat surveys, appears to outperform MSHM1 in identifying the top 10% of available habitat. The muskellunge spawning habitat preferences identified by this study can inform habitat conservation and restoration. The spawning habitat models identify the locations of likely spawning habitat, allowing managers to efficiently protect these critical areas from the removal of vegetation and woody debris which muskellunge preferred for spawning.