AUTHORS: Ben C. Neely, Kansas Department of Wildlife and Parks; Susan F. Steffen, Kansas Department of Wildlife and Parks; Jeff D. Koch, Kansas Department of Wildlife and Parks
ABSTRACT: Harvest regulation is a common tool used by fisheries managers to structure fish communities. However, anglers must comply with harvest restrictions for expected outcomes to occur. Dynamic angling motivations have led to supposition that different, unexplored factors influence contemporary angler decisions to harvest captured fish. We used recursive partitioning (i.e., decision trees) to model fate of angled fish (harvested or released) from creel data collected throughout Kansas from 1997 to 2024. A suite of 20 variables that related to the captured fish (e.g., fish length, legal status), location (e.g., surface area, rural/urban), time (e.g., year, day of week), and angling party (e.g., number of anglers, proportion male) were associated with each fish encounter. The global model indicated that legal status of an individual fish (i.e., vulnerable to harvest based on length), species, and fish length were the most important variables and correctly classified fate in 83.8% of test data records. Submodels developed individually for 12 species or species groups correctly classified fate in 67% (Lepomis spp.) to 96% (Smallmouth Bass Micropterus dolomieu) of test data records. These models highlight how variables influencing harvest differed by species or species group, and specified fish length, legal status, rural/urban characterization, and fishery longitude as the most important variables impacting fate. Our models demonstrate that angler decision to harvest a fish can be approximated from a few important variables commonly captured during standard creel surveys. Although these variables may differ based on fish taxa, generalities within can be useful for informing harvest regulation for fisheries management.
