We Ain't Afraid of No Ghosts: Tracking Habitat Interactions and Movement Dynamics of Ghost Tags under Differing Flow Conditions in a Sand‐Bed River

Abstract

The use of PIT tags has rapidly proliferated since their introduction, and new mobile detection methods have been developed. However, the presence of ghost tags (i.e., PIT tags left in the system after a fish dies) creates uncertainty about the status (live or dead) of tags detected. Herein, we describe our raft‐based mobile PIT tag antenna system, which was used to evaluate the movements of “seeded tags” (i.e., PIT tags that we placed in the river as ghost tag analogs) and their interactions with habitat features. We deployed 5,000 seeded tags in the San Juan River, a large sand‐bed river in the southwestern USA. Total distances moved by PIT tags ranged from 0.8 to 4,124 m, but 75% of movements were less than 100 m. Flow conditions causing the smallest to largest movements were (1) base flows, (2) spring runoff flows, (3) flash flood flows, and (4) a combination of spring runoff and flash flood flows. Based on Ivlev's electivity index, tags were more likely to be detected in riffles than in runs. These findings will help to classify mobile PIT tag detections as ghost tags or live fish, a critical data gap limiting the accurate estimation of demographic rates, population status metrics, and descriptions of the habitat use of fishes.

Keeping it classy: classification of live fish and ghost PIT tags detected with a mobile PIT tag interrogation system using an innovative analytical approach

“…We expected differences in movement between ghost tags and live fish would be important in classification. Razorback sucker, flannelmouth sucker, and Colorado pikeminnow are all known to move very large distances (McKinney et al 1999;Irving and Modde 2000;Zelasko et al 2010), and while ghost tags can move distances up to 4 km (Bond et al 2019;Stout et al 2019), we observed a large difference between the two classes. Fish carcasses can also move large distances after death (Havn et al 2017), but the average time for a white sucker carcass to shed an abdominally located tag was only 73.3 h (Muhametsafina et al 2014).…”

“…As expected, live fish generally moved greater distances than the ghost tags. Also, fish and other organisms exhibit behavioral changes to minimize the effect of high flows (Lytle and Poff 2004), and ghost tags respond to higher flows similarly to sediment (Bond et al 2019;Stout et al 2019). Therefore, we expected ghost tags and live fish to exhibit different responses to changes in discharge, and this expectation was demonstrated by the fifth variable influencing classification, monsoonal flows.…”

“…Larger total distances and upstream distances moved are more likely to belong to a live fish. The probability of being a live fish increased with larger upstream movements up to ϳ2000 m, at which point a larger distances moved upstream did not increase the probabil-ity of classification as a live fish (some ghost tags were observed to move upstream up to 99 m; Stout et al 2019). Distance moved upstream is an example of the limited interpretability of these plots.…”

“…In a previous study (Stout et al 2019), we describe the Passive Integrated Transponder Portable Antenna SystemS (PITPASS, a raft-based PIT tag antenna system with integrated GPS), our sampling method, the distribution and detection of known ghost tags, and the data recorded for each detection in greater detail. We used three PITPASS boats, and each sampling pass varied in length (115-264 rkm) and river discharge (11.3-130.3 m 3 •s −1 ).…”

“…Each reach received ϳ1250 tags per year distributed on the first sampling trip of the season. We randomly distributed onethird of the tags from each boat at a rate of one tag per minute across all habitat types (Stout et al 2019). Unknown tag refers to any PIT tag implanted in a fish at any time in the San Juan River.…”