Integrated analysis of marked and count data to characterize fine-scale stream fish movement

Abstract

Immigration and emigration are key demographic processes of animal population dynamics. However, we have limited knowledge on how fine-scale movement varies over space and time. We developed a Bayesian integrated population model using individual mark-recapture and count data to characterize fine-scale movement of stream fish at 20-m resolution in a 740-m study area every two months for 28 months. Our study targeted small-bodied fish, for which imperfect capture was accounted for (bluehead chub Nocomis leptocephalus, creek chub Semotilus atromaculatus and mottled sculpin Cottus bairdii). Based on data from 2021 individuals across all species, we found that proportions of immigrants in 20-m sections averaged 30-42% among the study species, but they varied over space and time. Creek chub immigrants increased during warmer intervals when individuals grew more and transitioned between body size classes, suggesting that immigration was due to ontogenetic habitat shifts. There was a weak pattern across the species that individuals were more likely to leave 20-m sections when flow was higher. Water-column species (bluehead chub and creek chub) were more likely to immigrate into and stay in deeper sections with more pool area. Across all species and occasions, number of immigrants to stream sections did not decrease with number of individuals that survived and stayed in the same sections. Thus, the habitat did not appear saturated, and our data provided no evidence that intra-specific interactions affected fine-scale movement at our fish densities. In conclusion, high turnover rates characterized fish movement among stream sections and their variation was associated with temporal and spatial shifts in abiotic conditions.

Keywords: Animal behavior; Cormack–Jolly–Seber models; Dispersal; Freshwater fish; Integrated population models; State-space models.