321 



Abstract— We estimated the impact 

 of striped bass [Morone saxatilis) pre- 

 dation on winter-run chinook salmon 

 {Oncorhynchus tshawytscha) with a 

 Bayesian population dynamics model 

 using striped bass and winter-run 

 chinook salmon population abundance 

 data. Winter-run chinook salmon ex- 

 tinction and recovery probabilities 

 under different future striped bass 

 abundance levels were estimated by 

 simulating from the posterior dis- 

 tribution of model parameters. The 

 model predicts that if the striped bass 

 population declines to 512,000 adults 

 as expected in the absence of stocking, 

 winter-run chinook salmon will have 

 about a 28'7f chance of quasi-extinction 

 (defined as three consecutive spawning 

 runs of fewer than 200 adults) within 50 

 years. If stocking stabilizes the striped 

 bass population at 700.000 adults, the 

 predicted quasi-extinction probability 

 is 30%. A more ambitious stocking 

 program that maintains a population 

 of 3 million adult striped bass would 

 increase the predicted quasi-extinction 

 probability to 55'7c. Extinction prob- 

 ability, but not recovery probability, was 

 fairly insensitive to assumptions about 

 density dependence. We conclude that 

 winter-run chinook salmon face a seri- 

 ous extinction risk without augmenta- 

 tion of the striped bass population and 

 that substantial increases in striped 

 bass abundance could significantly 

 increase the threat to winter-run chi- 

 nook salmon if not mitigated by increas- 

 ing winter chinook salmon survival in 

 some other way. 



Modeling the effect of striped bass 

 (Morone saxatilis) on the population viability 

 of Sacramento River winter-run chinook salmon 

 (Oncorhiynchus tshawytscha) 



Steven T. Lindley 

 Michael S. Mohr 



Santa Cruz Laboratory 



National Marine Fisheries Service 



110 Shaffer Road 



Santa Cruz, California 95060 



E-mail address (for S. T Lindley) Steve Lindley^noaa gov 



Manuscript accepted 23 October 2002. 



Manuscript received 31 December 2003 

 at NMFS Scientific Publications Office. 



Fish. Bull. 101:321-331 (2003). 



Predation is a factor in the (decline of 

 many Pacific salmon populations (Nehl- 

 sen et al., 1991), and fisheries manag- 

 ers may need to evaluate the potential 

 benefits of predator control or the pos- 

 sible impacts of predator augmentation. 

 Such evaluations require estimates of 

 the current predation rate, how the 

 predation rate would change with 

 changes in predator abundance, and 

 how changes in predation rate affect 

 the prey population viability. Predation 

 rate can be estimated in at least three 

 ways. Coordinated studies of predator 

 and prey distribution and abundance, 

 combined with predator diet studies, 

 can provide direct estimates of preda- 

 tion rate (e.g. Rieman et al., 1991). This 

 approach, however, is time-consuming, 

 labor-intensive, and difficult because 

 of the typically patchy distribution of 

 predators and prey in space and time. 

 Another approach is to build highly 

 detailed, spatially explicit simulation 

 models of predator and prey popula- 

 tions (e.g. Jager et al., 1997; Petersen 

 and DeAngelis, 2000). Such models, 

 although biologically realistic, are 

 data-intensive, have many param- 

 eters, and have outputs that can be 

 sensitive to parameter values that 

 are not well constrained by data. An 

 alternative modeling approach is to 

 use simple models of predator and prey 

 population dynamics and estimate the 

 unknown parameters from time series 

 of predator and prey abundance within 

 a statistical framework (Walters et al., 

 1986; Berryman, 1991; Carpenter et al., 

 1994). This approach is based on readily 

 available data and is relatively quick to 

 implement, making it suitable for ini- 



tial assessments of predation effects. 

 The model, once its unknown param- 

 eters have been estimated, can also be 

 used to assess the impact of predator 

 population size changes on the prey 

 population. 



We took this statistical modeling ap- 

 proach to investigate how the proposed 

 augmentation of the Sacramento River 

 striped bass (Morone saxatilis) popula- 

 tion might increase the risk of extinc- 

 tion faced by the endangered winter- 

 run chinook salmon (Oncorhynchus 

 tshawytscha ). Striped bass prey on juve- 

 nile chinook salmon in the Sacramento 

 River system (Stevens, 1966; Thomas, 

 1967), as well as in other west-coast 

 rivers (Shapovalov, 1936), and striped 

 bass prey upon juvenile Atlantic salmon 

 in east-coast rivers were they co-occur 

 (Blackwell and Juanes, 1998). Although 

 winter-run chinook salmon juveniles 

 are not the primary prey of striped 

 bass and striped bass predation is only 

 one of many mortality sources affecting 

 winter-run chinook salmon, an increase 

 in striped bass abundance has the po- 

 tential to negatively impact winter-run 

 chinook salmon. This potential must 

 be assessed before the striped bass 

 population can be augmented because 

 winter-run chinook salmon are listed as 

 endangered under the U. S. Endangered 

 Species Act. 



Because few data are available on 

 the details of the interaction between 

 winter-run chinook salmon and striped 

 bass (e.g. functional response, role 

 of alternate prey), we explored the 

 simplest models that can capture the 

 predation effect to assess the effect of 

 striped bass population manipulations. 



