Adult Steelhead Monitoring Challenges in Cedar Creek, WA Josua Holowatz & Dan Rawding Species of Concern: Hatchery Strays and Natural Production Coastal Cutthroat Tule Fall Chinook Winter Steelhead Coho Salmon Grist Mill Site River Mile 2.5 Fishway constructed in 1958 Adult trap operated since 1999 Smolt Trap operated since 1998 Smolt Monitoring Goals •Unbiased estimates of coho, steelhead, & cutthroat smolt yield (CV < 10%). Estimates of juvenile Chinook outmigrants if funding is available •Contribution of coho from a remote site incubator program is based on otolith marks. •CWT natural origin coho to estimate marine survival and fishery contributions Cedar Creek Coho and Steelhead Smolt Estimates 60,000 Coho Smolts 50,000 STD Smolts 40,000 30,000 20,000 10,000 Year 20 09 20 07 20 06 20 05 20 04 20 03 20 02 20 01 20 00 19 99 0 19 98 Estimate 70,000 Cedar Creek Adult Monitoring Goals & Methods • Estimate the abundance, age, & origin for Tule Fall Chinook, coho, and winter steelhead populations • BioSample all adult salmonids trapped at the fishway (River Mile 2.5) • Tag all fish captured in the fishway for markrecapture abundance estimates • Monitor Petersen assumptions needed for an unbiased estimate • Recover adult salmonids through a variety of methods Recapture Events • Recapture event is a combination of carcass recoveries (salmon), along with hoop trap & resistance board weir recoveries (4 miles upstream of the fishway) • Recover steelhead kelts in the screw trap & through seining above the resistance board weir • Estimate trap efficiency for steelhead – Trapped fish are tagged and released below the fishway – Enumerate number of steelhead that successfully re-ascend the fishway – Estimate the proportion fish using the ladder 2001-03 Hoop Trap • limited success at recapturing coho and steelhead due to trap avoidance and large debris loads Pre High Water Event Post High Water Event 2006:Installed Resistance Board Weir at RM 6.1 Weir fishes efficiently until flows reach 850 CFS Fully submerged panels Leaf debris high water mark Weir is not 100% efficient due to high flows and debris Fishway is operational until flows exceed 1250 CFS Steelhead catch and flows 2007-2009 Average Steelhead Catch at Fishway and Weir 10.00 9.00 8.00 Catch/CFS 7.00 6.00 5.00 Fishway Catch 4.00 3.00 Weir Catch 2.00 1.00 0.00 CFS x 100 Cedar Creek Flows 2007-08 water year Thresholds Coho Steelhead Fishway Weir Sampling kelts Mark-recapture statistics & summary for unmarked adult coho salmon and winter steelhead 2009 Steelhead 26 3 24 1840 80 200 >100% 2009 Cedar Creek Adult Estimates 3500 3000 Abundance Category Marks Recaptures Captures U95%CI L95%CI N CV 2009 Coho 1244 321 508 2973 1463 2218 17% 2500 2000 1500 1000 500 0 Coho Steelhead Species Cedar Cr. Adult Steelhead Trap Effeciencies Likelihood 0.25 2002 2003 0.2 2004 0.15 2005 2006 2007 2008 0.1 Maximum Likelihood Estimates • trap efficiency • % fish using ladder •binomial distribution •skewed or long tailed •imprecise estimates 2009 0.05 0 0% 20% 40% 60% 80% Trap Efficiency Cedar Cr. Adult Steelhead Population Estimates 0.25 2002 2003 2004 2005 2006 2007 2008 2009 0.2 Likelihood Maximum Likelihood Estimates • Abundance •binomial approximation to hypergeometric distribution •skewed or long tailed •very imprecise estimates 0.15 0.1 0.05 0 0 200 400 Abundance 600 800 Rivot, E., and E. Prevost. 2002. Hierarchical Bayesian analysis of capturemark-recapture data. Can. J. Fish. Aquat. Sci. 53:2157-2165. • Application of Rivot et al. 2002 to Cedar Creek adult steelhead estimate • Posterior probability = prior X likelihood • Non-informative priors & hyperpriors (hierarchical models) – priors have little influence on posterior probabilities • Common distributions for trap efficiency & population estimate are the beta and normal probability distributions, respectively • Other distributions could be used Hierarchical Modeling • all annual trap efficiencies Cedar Cr. Adult Steelhead Trap Effeciencies 0.25 2002 2003 2004 Likelihood 0.2 2005 2006 2007 0.15 2008 2009 Hierarchical 0.1 0.05 0 0% 20% 40% 60% 80% Trap Efficiency Hierarchical Modeling • shrink estimates toward the Cedar Cr. Adult Steelhead Population Estimates 0.25 2002 2003 2004 2005 2006 2007 2008 2009 Hierarchical 0.2 Likelihood mean, which yields improved precision • can be used for both trap efficiency & population abundance • compromise between individual and fully pooled estimates come from a common distribution of trap efficiencies and their ordering does not affect the model (exchangeable) • borrow strength from other trap efficiencies because they are similar 0.15 0.1 0.05 0 0 200 400 Abundance 600 800 I T r T r nd_ ap ap 1 &N_1 99 _199 9 9 In 999 T r T r d_ ap ap 2 &N_2 00 _200 0 000 I n T r T r d_ 0 ap ap 2 &N_2 00 _200 1 001 I n T r T r d_ 1 ap ap 2 &N_2 00 _200 2 002 I n T r T r d_ 2 ap ap 2 &N_2 00 _200 3 3 In 003 T T r r d_ ap ap 2 &N_2 00 _200 4 4 In 004 T T r r d_ ap ap 2 &N_2 00 _200 5 005 I n T r T r d_ 5 ap ap 2 &N_2 00 _200 6 006 I n T r T r d_ 6 ap ap 2 &N_2 00 _200 7 007 I n T r T r d_ 7 ap ap 2 &N_2 00 _20088 I 0 Tr Tr nd_ 08 ap ap 2 &N_2 00 _200 9 009 9 Population Estimate Cedar Creek Adult Steelhead Estimates (median) with 95%CI by Year and Method 2000 1500 1000 500 0 Method_Year I T r T r nd_ ap ap 19 &N_1 99 _1999 9 I T r T r nd_ 99 ap ap 2 &N_2 00 _20000 00 I T r T r nd_ 0 ap ap 20 &N_2 01 _2001 In 001 T T r r d_ ap ap 2 &N_2 00 _20022 00 I T r T r nd_ 2 ap ap 20 &N_2 03 _2003 0 I T r T r nd_ 03 ap ap 2 &N_2 00 _20044 In 004 T T r r d_ ap ap 20 &N_2 05 _2005 0 I T r T r nd_ 05 ap ap 2 &N_2 00 _20066 In 006 T T r r d_ ap ap 20 &N_2 07 _2007 0 I T r T r nd_ 07 ap ap 2 &N_2 00 _20088 00 I T r T r nd_ 8 ap ap 20 &N_2 09 _2009 00 9 Coefficient of Variation CV for Adult Steelhead Estimates by Year and Method 100% 75% 50% 25% 0% Year_Method Summary • Annual steelhead smolt estimates are an order of magnitude lower than coho. Mean steelhead smolts are 2,900 compared to 38,000 for coho salmon. • Our adult steelhead estimates are also an order of magnitude lower than coho. For example, in 2009 the steelhead abundance was 200 fish compared to 2,200 for coho. • Cedar Cr. Steelhead program should be designed to provide more precise abundance estimates. • Hierarchical approach is a good approach when dealing with sparse data and common distributions. Cedar Creek Adult Steelhead Recommendations • Mark all wild steelhead captured at the fishway trap. • Release all marked steelhead below the fishway trap to estimate ladder use. • The recapture events should be designed to obtain at least 10 recaptures through the resistance board weir, seining and the screw trap catch of kelts. • If recaptures are sparse, recaptures can be supplemented through snorkeling.