Ecological Modelling 221 (2010) 582–589 Contents lists available at ScienceDirect Ecological Modelling journal homepage: www.elsevier.com/locate/ecolmodel Using multistate mark-recapture methods to model adult salmonid migration in an industrialized river Rebecca A. Buchanan ∗ , John R. Skalski Columbia Basin Research, School of Aquatic & Fishery Sciences, University of Washington, 1325 Fourth Avenue, Suite 1820, Seattle, WA 98101-2509, USA article info Article history: Received 3 August 2009 Received in revised form 6 November 2009 Accepted 10 November 2009 Available online 16 December 2009 Keywords: Chinook salmon Memory effects Multistate model Radiotelemetry Release-recapture abstract A multistate mark-recapture (MSMR) model of the adult salmonid migration through the lower Columbia River and into the Snake River was developed, designed for radiotelemetry detections at dams and trib- utary mouths. The model focuses on upstream-directed travel, with states determined from observed fish movement patterns indicating directed upstream travel, downstream travel (fallback), and use of non-natal tributaries. The model was used to analyze telemetry data from 846 migrating adult spring- summer Chinook salmon (Oncorhynchus tshawytscha) tagged in 1996 at Bonneville Dam on the Columbia River. We used the model to test competing hypotheses regarding delayed effects of fallback at dams and visits to tributaries, and to define and estimate migration summary measures. Tagged fish had an average probability of 0.755 (  SE = 0.018) of ending migration at a tributary or upstream of Lower Granite Dam on the Snake River, and a probability of 0.245 (  SE = 0.018) of unaccountable loss (i.e., mortality or mainstem spawning) between the release site downstream of Bonneville Dam and Lower Granite Dam. The highest probability of unaccountable loss (0.092;  SE = 0.012) was in the reach between Bonneville Dam and The Dalles Dam. Study fish used the tributaries primarily as exits from the hydrosystem, and visits to non-natal tributaries had no significant effect on subsequent movement upriver (P = 0.4245). However, fallback behavior had a small effect on subsequent tributary entry and exit (P = 0.0530), with fish using tributaries as resting areas after reascending Bonneville Dam after fallback. The spatial MSMR model developed here can be adapted to address additional questions about the interaction of migrating organisms with their environment, or for the study of migrations in other river systems. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Migratory species play a crucial role in ecological communi- ties, connecting distant ecosystems through transfers of organic matter. From salmon returning to their natal streams to spawn, to songbirds traveling between breeding and wintering grounds, such species bring significant energetic contributions to numer- ous ecosystems. For example, salmon accrue much of their biomass during the ocean life stage but spawn in freshwater. Their decay- ing carcasses transfer rich marine-derived nutrients to their inland spawning habitat, thereby increasing productivity of the freshwa- ter ecosystem. Although the migratory life stage is necessary for completing the life cycle, populations often incur high mortality during migration. As society works to protect migratory species from the effects of habitat loss, pollution, overharvest, and climate change, detailed information is needed regarding the migratory life stage: What regions and habitats are utilized and how? Where is survival most threatened during migration, and what are the domi- ∗ Corresponding author. Tel.: +1 206 685 2793; fax: +1 206 616 7452. E-mail address: rabuchan@u.washington.edu (R.A. Buchanan). nant mortality risks? How do human actions and habitat alteration help or hinder migration? Pacific salmonids (Oncorhynchus spp.) from the Columbia and Snake river basins in the Pacific Northwest region of the United States pass up to nine large hydroelectric dams on both their juve- nile migration to the ocean and their adult migration to inland spawning grounds (Fig. 1). With 12 of these populations protected under the Endangered Species Act of 1973 (U.S. Code, Title 16, Chapter 25, sections 1531–1544), much research attention has focused on the migratory life stage through the hydrosystem. Biol- ogists have hypothesized that spawning success is affected by experiences during the adult upriver migration, such as non-natal tributary use and fallback over a dam (descending the dam after ascending it). Dam fallback and other downstream travel during the adult migration have been observed among numerous popula- tions (Dauble and Mueller, 2000). Reischel and Bjornn (2003) and Boggs et al. (2004) hypothesized that fallback hinders migration success by depleting a fish’s energy reserves or causing injury or direct mortality. Several researchers have studied non-natal trib- utary use by migrating adult salmonids. Quinn (1993) reported varying levels of straying to non-natal tributaries among wild and hatchery-produced salmonids, with high straying rates possibly 0304-3800/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolmodel.2009.11.014