96-26 Multistate Modeling Approach for Analyzing Survival and Movement of Radio Frequency Identification (RFID) PIT Tagged Trout in Rivers

Eric R. Fetherman , Fish, Wildlife and Conservation Biology, U.S.G.S. Colorado Cooperative Fish and Wildlife Research Unit, Colorado State University, Fort Collins, CO
Dana L. Winkelman , Colorado Cooperative Fish and Wildlife Research Unit, Colorado State University, Fort Collins, CO
George J. Schisler , Colorado Division of Wildlife, Fort Collins, CO
Larissa Bailey , Fish, Wildlife and Conservation Biology, Colorado State University, Fort Collins, CO
William L. Kendall , Fish, Wildlife and Conservation Biology, U.S.G.S. Colorado Cooperative Fish and Wildlife Research Unit, Colorado State University, Fort Collins, CO
Passive Integrated Transponder (PIT) tags allow individual identification, have an infinite life, are easily applied and well retained, and have minimal effects on growth and survival.  Currently, stationary Radio Frequency Identification (RFID) PIT antennas are widely used to detect PIT tagged fish; however, estimating population parameters of interest, such as survival and movement, from RFID detections can be challenging.  We describe a multistate modeling approach used to obtain estimates of apparent survival (S) and movement probabilities (ψ) for brown trout and rainbow trout populations in the Cache la Poudre, Colorado.  This approach was used to estimate the impact of brown trout removal (management treatment) on survival and retention probability for introduced whirling disease (WD) resistant rainbow trout.  In one section of the river, we removed brown trout (removal section; 0.6 mile), and in a second section, located five miles downstream, we did not (control section; 0.8 mile).  Brown trout upstream and downstream of the removal section were Radio Frequency Identification (RFID) PIT tagged during removal operations.  Brown trout upstream and downstream of, as well as within, the control section were also RFID PIT tagged.  After removal, two thousand PIT tagged, WD resistant rainbow trout were stocked into both the removal and control sections.  Paired RFID flat-bed loop antenna arrays were deployed at the upstream and downstream ends of both the removal and control sections (8 antennas) and were used to monitor movement of the PIT tagged fish.  Antennas were paired to determine directionality of movement.  We empirically estimated detection probability at several water depths and velocities using PIT tagged fish in hatchery raceways; detection probability was also estimated weekly for each antenna loop in the river.  Using estimates of movement direction and assuming no fish passed the paired antenna without detection, a detected individual could be assigned with certainty to a given state (upstream, downstream, or within a section).  However, situations where these assumptions are not met and individual state uncertainty exists will also be discussed.