Residence times of juvenile salmon and steelhead in …...ecology study in the same area where Sather et al. (2011) had...

ARTICLE

Residence times of juvenile salmon and steelhead inoff-channel tidal freshwater habitats, Columbia River, USAGary E. Johnson, Gene R. Ploskey, Nichole K. Sather, and David J. Teel

Abstract: We documented two life history strategies for juvenile salmonids as expressed in off-channel tidal freshwater habitatsof the Columbia River: (i) active migrations by upper river Chinook salmon (Oncorhynchus tshawytscha) and steelhead (Oncorhynchusmykiss) during the primary spring and summer migration periods and (ii) overwinter rearing in tidal freshwater habitats by cohosalmon (Oncorhynchus kisutch) and naturally produced Chinook salmon mostly from lower river sources. During spring–summer2007–2008, acoustic-tagged fish originating above Bonneville Dam (rkm 234) had short residence times in off-channel areas(rkm 192–203): median 2.5 and 2.6 h for yearling (mean lengths 134 and 158 mm) and 3.0 and 3.4 h for subyearling (104 and116 mm) Chinook salmon and 2.5 h for yearling steelhead (215 mm). The percentage of fish in off-channel areas out of the totalin the main- and off-channels areas was highest for yearling Chinook salmon (8.1% and 9.3% for 2007 and 2008, respectively) andlowest for steelhead (4.0% for 2008) and subyearling Chinook salmon (3.6% and 6.1% for 2007 and 2008, respectively). In lateJanuary and early February 2010, 2011, and 2012, we captured and tagged yearling Chinook and coho salmon occupyingoff-channel tidal freshwater habitats. Median residence times in off-channel areas were 11.6–25.5 days for juvenile Chinook (106,115, and 118 mm, respectively by year) and 11.2 days for coho salmon (116 mm). This study is the first to estimate residence timesfor juvenile salmonids specifically in off-channel areas of tidal fresh water and, most importantly, residence times for Chinooksalmon expressing a life history of overwintering in tidal fresh water. The findings support restoration of shallow off-channelhabitats in tidal freshwater portions of the Columbia River.

Résumé : Nous avons documenté les deux stratégies de cycle biologique de salmonidés juvéniles suivantes exprimées dans deshabitats hors chenal d’eau douce de marée du fleuve Columbia : (i) des migrations actives par les saumons quinnats (Oncorhynchustshawytscha) et arc-en-ciel (Oncorhynchus mykiss) du cours supérieur du fleuve durant les principales périodes de migration print-anière et estivale et (ii) la croissance hivernale dans des habitats d’eau douce de marée de saumons cohos (Oncorhynchus kisutch) etde saumons quinnats d’origine naturelle provenant principalement du cours inférieur du fleuve. Durant les printemps et étés de2007 et 2008, des poissons munis de radioémetteurs provenant d’en amont du barrage Bonneville (rkm 234) présentaient decourts temps de séjour dans des zones hors chenal (rkm 192–203), soit des séjours médians de 2,5 et 2,6 h et de 3,0 et 3,4 h,respectivement, pour les saumons quinnats d’un an (longueurs moyennes de 134 et 158 mm) et de moins d’un an (104 et 116 mm),et de 2,5 h pour les saumons arc-en-ciel d’un an (215 mm). Le pourcentage de poissons dans les zones hors chenal par rapport aleur nombre total dans le chenal principal et dans les zones hors chenal était le plus élevé pour les saumons quinnats d’un an(8,1 % et 9,3 % en 2007 et 2008, respectivement) et le plus faible pour les saumons arc-en-ciel (4,0 % en 2008) et les saumonsquinnats de moins d’un an (3,6 % et 6,1 % en 2007 et 2008, respectivement). À la fin de janvier et au début de février 2010, 2011 et2012, nous avons capturé et marqué des saumons quinnats et cohos d’un an occupant des habitats d’eau douce de marée horschenal. Les temps de séjour médians dans les zones hors chenal étaient de 11,6–25,5 jours pour les saumons quinnats juvéniles(106, 115 et 118 mm, respectivement, selon l’année) et de 11,2 jours pour les saumons cohos (116 mm). Il s’agit de la première étudea estimer les temps de séjour de salmonidés juvéniles dans des zones hors chenal d’eau douce de marée et, surtout, les temps deséjour de saumons quinnats présentant un cycle biologique qui comprend l’hivernation dans des zones d’eau douce de marée.Les résultats appuient la restauration d’habitats hors chenal peu profonds dans des portions d’eau douce tidales du fleuveColumbia. [Traduit par la Rédaction]

IntroductionThis study examined residence times of juvenile salmonids

(Oncorhynchus tshawytscha, Oncorhynchus kisutch, and Oncorhynchusmykiss) to assess differential occupation of off-channel habitats byfish with different life history strategies in the tidal freshwaterportion of the lower Columbia River and estuary (LCRE; Fig. 1).Early life history strategies for juvenile Chinook salmon emigra-tion have two broad themes: fish that reside in natal streams for a

year or more before migrating to the ocean as yearlings and fishthat migrate downstream as subyearlings and enter the oceanduring the first year after emergence (Myers et al. 2006). Oceanentry by yearlings occurs primarily in spring, followed by a peakoutmigration of subyearlings in summer (Healey 1991). The manyvariations on these themes in terms of migration timing, fish sizedistribution, estuarine residence time, and other factors result ina diversity of life histories (e.g., Bottom et al. 2005; Miller and

Received 17 February 2014. Accepted 22 December 2014.

Paper handled by Associate Editor Aaron Fisk.

G.E. Johnson. Pacific Northwest National Laboratory, 620 SW 5th Avenue, Portland, OR 97204, USA.G.R. Ploskey.* Pacific Northwest National Laboratory, P.O. Box 241, North Bonneville, WA 98639, USA.N.K. Sather. Pacific Northwest National Laboratory, 1286 Washington Harbor Road, Sequim, WA 98382, USA.D.J. Teel. NOAA Fisheries, Northwest Fisheries Science Center, P.O. Box 130, Manchester, WA 98353, USA.Corresponding author: Gary E. Johnson (e-mail: [email protected]).*Present address: P.O. Box 212, North Bonneville, WA 98639-0212, USA.

684

Can. J. Fish. Aquat. Sci. 72: 684–696 (2015) dx.doi.org/10.1139/cjfas-2014-0085 Published at www.nrcresearchpress.com/cjfas on 6 January 2015.

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

Sadro 2003; Rich 1920). For example, juvenile Chinook salmon arefound in the LCRE in autumn and winter, well after the subyear-ling summer outmigration period (Rich 1920; McCabe et al. 1986;Roegner et al. 2012). However, it is not known how long those fishremain in LCRE habitats. Information about such uncommon lifehistories is increasingly valued given that life history diversity isconsidered essential for the long-term viability of salmonid pop-ulations (Fleming et al. 2014).

In the LCRE, intensified interest in life history diversity andjuvenile salmonid migration characteristics was prompted by theEndangered Species Act (ESA 1973) listing of 13 salmonid stocks inthe Columbia River basin, all of which use this portion of the riverto reach the ocean. The ESA listings led regulators to include LCREhabitat restoration as a key mitigation strategy in ESA biologicalopinion documents on operation of the Federal Columbia RiverPower System (e.g., NMFS 2014). Accordingly, restoration manag-ers are aggressively pursuing restoration of floodplain wetlandsthat were historically connected to off-channel habitats in thetidally influenced freshwater stretch (rkm 58–234) of the 234 kmLCRE from Bonneville Dam to the ocean under the premise thatrestoration of shallow, off-channel habitats will improve survivaland increase life history diversity (BPA and USACE 2014). Under-standing early life history characteristics and realized functions(Simenstad and Cordell 2000) is essential to inform managementdecisions about habitat restoration to benefit ESA-listed fish.Thom et al. (2013) concluded that data on realized function forjuvenile salmonids were sorely lacking in the LCRE. To our knowl-edge, no studies in the LCRE or elsewhere have estimated resi-dence times for yearling and subyearling salmonids in off-channeltidal freshwater habitats during spring–summer and late winter –early spring. The latter period is particularly important to resto-

ration program managers because it includes relatively uncommonlife histories (Burke 2004; Fresh et al. 2005; Teel et al. 2014).

Our goal was to examine residency patterns of juvenile sal-monids in off-channel tidal freshwater habitats of the LCRE. Theobjectives during 2007–2008 were to estimate residence times andproportional use of off-channel pathways compared with the mainriver channel by juvenile Chinook salmon and steelhead duringthe primary spring and summer outmigration periods. Those fishwere actively migrating to the ocean from areas upriver of theLCRE. During 2010–2012, we studied juvenile Chinook and cohosalmon that we hypothesized were expressing a largely undocu-mented life history of LCRE overwintering. To address this hy-pothesis, we estimated residence times and also examined theexit timing and condition of juvenile Chinook and coho salmonduring late winter and early spring. Because the ESA listing andrecovery framework focuses on specific stocks (NMFS 2014), wealso used genetic stock identification to evaluate the origins of thejuvenile Chinook salmon during the 2010–2012 studies.

Materials and methods

Study designWe conducted field work in different time periods because the

research was performed in collaboration with two separate,broader research efforts. During 2007–2008, we deployed acousticreceivers for collecting residence time data in off-channel areas atthe Sandy River delta (rkm 198; see Sather et al. 2011 for a descrip-tion of the study area) in the Columbia River to supplement acous-tic detection arrays for dam-passage survival studies by others inspring and summer (Ploskey et al. 2009). During 2010–2012, weimplemented residence time research as part of a juvenile salmon

Fig. 1. Map of the lower Columbia River and estuary (LCRE) showing location of Sandy River delta (SRD; rkm 198) and Cottonwood Island(COT; rkm 112) study areas. Sampling occurred during 2010 and 2012 at SRD and during 2012 at COT.

Johnson et al. 685

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

ecology study in the same area where Sather et al. (2011) hadidentified a bimodal size distribution for Chinook salmon in win-ter (Fig. 2a). Their finding of relatively large (�105 mm fork length)juvenile salmon during winter months prompted us to hypothe-size these fish were overwintering in tidal off-channel habitats. Toaddress this hypothesis, we sampled juvenile salmon in January2010 and February 2011 at the Sandy River delta and in February2012 at Cottonwood Island (Fig. 1). Sample sizes were much largerduring 2007–2008 than 2010–2012 (Table 1) because thousands offish from smolt bypass systems at mainstem dams during springand summer were necessary for the 2007–2008 dam-passage sur-vival studies, whereas only tens of fish could be obtained bybeach seining during winter 2010–2012 for residence time studiesin tidal fresh water.

Telemetry equipmentTo obtain data to estimate the residence times of juvenile sal-

monids, we used the Juvenile Salmon Acoustic Telemetry System(JSATS; described by McMichael et al. 2010) because it provided thesmallest acoustic tag available, which enabled us to tag juvenilesalmonids as small as 95 mm fork length (Brown et al. 2010). TheJSATS operating frequency was 416.7 kHz. Specific tag sizes, pulse-repetition intervals (PRIs), and tag lives used in this study varied

from year to year depending on equipment availability and, dur-ing 2007 and 2008, our collaborators’ study objectives (Table 1).We measured tag life for various PRI levels by annually monitor-ing a random sample of 5 to 50 transmitters in a tank, dependingon the availability of transmitters. Mean tag life ranged from 31 to100 days, depending on the PRI (Table 1).

Fish sources and genetic stock identificationFrom late April through July 2007 and 2008, we monitored ju-

venile Chinook salmon and steelhead that had been tagged as partof other studies at upriver locations (Bonneville Dam, rkm 234;John Day Dam, rkm 349; and Lower Granite Dam, Snake Riverrkm 173, 693 km from the Pacific Ocean). During tagging, re-searchers distinguished yearling and subyearling salmon bylength and body morphology (Tiffan et al. 2000). In 2007 and 2008,mean fork lengths (hereafter, all fish lengths are fork lengths)were 134 and 159 mm for yearling Chinook salmon, 104 and116 mm for subyearling Chinook salmon, and 215 mm for yearlingsteelhead, respectively (Table 2). Over 20 000 fish were taggedeach year. For context, percentages of hatchery-origin fish duringthe 2008 study, as indicated by a clipped adipose fin or fin abra-sions, varied among species tagged: 55% yearling Chinook salmon,68% steelhead, and 39% subyearling Chinook salmon (Ploskey

Fig. 2. Size frequency distributions: (a) beach seine samples during 2007–2010 at Sandy River delta (modified from Sather et al. 2011) and(b) smolt monitoring samples during 2008 at John Day Dam. CH0, subyearling Chinook salmon; CH1, yearling Chinook salmon;STH, steelhead.

Table 1. Summary of acoustic-telemetry and tagging methods.

Factor 2007 2008 2010 2011 2012

Study period 4 Apr.–18 Aug. 26 Apr.–25 July 27 Jan.–23 Apr. 3 Feb.–17 May 2 Feb.–31 MayTag manufacturera SC ATS ATS SC ATSTag dimensions (mm) 5.5×4.8×19 5.2×3.8×12 5.2×3.8×12 5.5×4.8×19 5.2×3.8×12Transmitting power (dB re:1 �Pa@1 m) 153 156 156 153 156Pulse repetition interval (s) 5 and 10 3 7 10 10Mean tag life (days) (range) 53 (5–74) 31 (9–48) 60b 108c 100 (91–109)No. of tagged fishd >23 000 23 340 41 CH1 12 CH1, 34 CO 14 CH1No. of release sitese 2f 6g 1 SRD 1 SRD 1 COT

Note: Tag masses are provided in Table 2.aSC, Sonic Concepts; ATS, Advanced Telemetry Systems.bTag life was not measured for the 2010 tagging effort. We approximated tag life of at least 60 days based on data from tag-life studies for

transmitters with 3 and 5 s repetition rates.cAll four tags were still transmitting when the 2011 tag life study was terminated after 108 days because data collection in the field had ended.dCH1, yearling Chinook salmon; CO, coho salmon.eSRD, Sandy River delta; COT, Cottonwood Island.fLower Granite and Bonneville dams.gLower Granite Dam, Arlington, John Day Dam, The Dalles Dam, Bonneville Dam, and Skamania.

686 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

2008). Genetic stock identification of 2007–2008 study fish was notconducted. However, because all of the tagged fish were collectedat or upstream of Bonneville Dam, the juveniles included fishfrom tributaries in the Columbia River gorge and from rivers eastof the Cascade Mountains, i.e., upper river fish.

To capture juvenile salmon for tagging during January 2010,February 2011, and February 2012, we used a beach seine in off-channel areas near McGuire Island (rkm 188; 2010), west of GaryIsland (rkm 198; 2011), west of Sandy Island (rkm 120; 2012), andeast of Cottonwood Island (rkm 112; 2012) (Fig. 1). Efforts to capturefish focused on off-channel habitats because these LCRE areashave higher densities of juvenile salmon during winter monthscompared with main-channel habitats (Johnson et al. 2013). Taggedjuvenile Chinook salmon ranged in length from 103 to 121 mm,with sample sizes of 41, 12, and 14 fish during 2010, 2011, and 2012,respectively (Tables 1 and 2). During 2011, we also captured, tagged,and monitored 36 juvenile coho salmon (mean size 116 mm). Alltagged fish were determined to be yearlings, based on fish size andcapture season (late winter – early spring). Juvenile steelhead werenot captured during beach seine efforts. We determined hatcheryorigin by observing whether the fish was marked, as indicated bya clipped adipose fin or presence of a coded wire tag or both. In2010–2012, adipose fin clipping rates in hatcheries were very highfor the four most common stocks of juvenile Chinook salmonoccupying tidal freshwater habitats (>99%), although the mark ratewas lower (72%) for Upper Columbia River summer–fall hatcheryjuveniles (Teel et al. 2014). Of the total 103 juvenile Chinook andcoho salmon we tagged during 2010–2012, only 7% were marked,indicating that most were of natural origin.

To determine the genetic stock of origin for juvenile Chinooksalmon sampled during 2010–2012, fin tissues from individual fishwere analyzed using microsatellite DNA data and genetic stockidentification methods (Teel et al. 2009, 2014). The genetic stockidentification computer program ONCOR (Kalinowski et al. 2007)was used to estimate the relative probability of stock origin ofeach sample. Descriptions of the genetic stocks included in theanalysis are presented by Teel et al. (2014). Genetic stock identitieswere not estimated for coho salmon.

Tagging and releaseTrained staff surgically implanted acoustic tags in fish using

methods similar to those described by Brown et al. (2010) andDeters et al. (2010). Before tagging, fish were anaesthetized usingfresh river water and MS-222 (tricaine methanesulfonate; 80 to100 mg·L–1) (Carter et al. 2011) and individually weighed and mea-sured. After tagging, we acclimated fish in river water for 18 to24 h before transporting and releasing them (Oldenburg et al.2011).

Release locations for tagged fish varied depending on studyyear. During 2007–2008, collaborators from other studies releasedfish at various sites upstream of Bonneville Dam (Table 1). During27–30 January 2010 and 3–5 February 2011, we released tagged fish

west of Gary Island (rkm 198; Fig. 1). During 3–7 February 2012, wereleased tagged fish east of Cottonwood Island in Carroll’s Channel(Fig. 1). We released tagged fish during daytime hours and re-corded the date, time, and location of the release for each taggedfish, except during 2010 when the release time was not recorded;it was arbitrarily set at 1100 h based on staff recollection.

Data collection, reduction, and analysisWe placed autonomous JSATS receivers in arrays at strategic

locations during each of the 5 study years (Fig. 3). Deployment andretrieval methods were similar to those described by Titzler et al.(2010). Generally, receivers were positioned in relatively deep ar-eas of off-channel habitats to maximize signal detectability, givena detection range for JSATS transmitters of about 300 m in openwater (McMichael et al. 2010). During 2007 and 2008, we placedsingle receivers at four locations in off-channel areas near Gary,Flag, and Reed islands (Figs. 3a, 3b), which were used in residencetime estimation along with receivers for other studies deployedupstream (entrance array; rkm 203) and downstream (exit array;rkm 192) of the Sandy River delta (Figs. 3a, 3b). During 2010 and2011, we deployed five receivers in the off-channel area west ofFlag and Gary islands (Figs. 3c, 3d), and during 2012, we deployedfive receivers east of Cottonwood Island (Fig. 3e). Except duringservicing to replace batteries and data storage media, receiverswere operational 24 h·day–1 to detect, digitize, and decode trans-mitter signals from tagged fish and write data to files on compactflash cards. We downloaded data from each receiver every 3 to4 weeks during the monitoring periods, which lasted severalmonths each study year (see Table 1).

Reduction and analysis of raw detection data involved determi-nation of detection events, categorization of migration pathwaysthrough main- and off-channel areas (2007–2008 only), and esti-mation of residence times. A detection event was determined byat least four valid acoustic signal receptions with a PRI matchingthe temporal pattern of a properly functioning JSATS tag within atime window defined by the PRI specific to each transmitter. Thewindow duration was 47.8 s for 3 s tags, 79 s for 5 s tags, 110.2 s for7 s tags, and 157 s for 10 s tags. We filtered tag code receptions toremove multipath (nondirect) signals from a transmitter by delet-ing all that occurred within 0.156 s after an identical tag codedetection, under the assumption that closely lagging signals weremultipath. Initial tag code receptions were retained. We matcheddetected fish with release codes and developed time-of-detectionhistories for the detection events for each tagged fish. For a giventagged fish detected by the receiver(s), the primary results afterdata reduction were time and location of first detection event,time and location of last detection event, and its history of detec-tion events by all receivers.

Only fish detected on both the entrance and exit arrays wereincluded in the 2007–2008 residence time analyses. Using detec-tion histories, we categorized migration in off- and main-channelareas during 2007–2008 as follows (see Fig. 3): “Main only”: de-

Table 2. Characteristics of tagged fish.

Species Year SeasonMean fishlength (mm)

Mean fishmass (g)

Mean tagmass (g)

CalculatedFulton’s K

Calculatedtag burden

STH 2008 Spring 215 75.1 0.43 0.7567 0.006CH1 2007 Spring 134 26.3 0.63 1.1038 0.024

2008 Spring 158 37.2 0.43 0.9396 0.012CH0 2007 Summer 104 12.2 0.63 1.0712 0.052

2008 Summer 116 14.7 0.43 0.9418 0.029CH1 2010 Winter 106 11.8 0.43 0.9908 0.036

2011 Winter 115 15.7 0.63 1.0323 0.0402012 Winter 118 15.7 0.43 0.9556 0.027

Coho 2011 Winter 116 16.1 0.63 1.0315 0.039

Note: Fulton’s K represents fish condition (K = 100 × W/L3, where W is fish mass in grams and L is fish fork length incentimetres). Tag burden was calculated as tag mass divided by fish mass. STH, steelhead; CH1, yearling Chinook salmon;CH0, subyearling Chinook salmon.

Johnson et al. 687

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

tected only by main channel receivers; “Off Channel”: detected bya receiver(s) west of Gary and other delta islands or the receivernorth of Reed Island; “Gary only”: detected only by a receiver(s)west of Gary and other delta islands; “Reed only”: detected only bythe receiver north of Reed Island; “Main and Gary”: detected bythe entrance array and the receivers west of Gary Island and otherdelta islands; “Main and Reed”: detected by the entrance array andthe receiver north of Reed Island.

In all study years, we estimated arithmetic mean and medianresidence times. During 2007–2008, residence times for individualfish were calculated as the time between the last detection by theexit array and the first detection at the entrance array (Fig. 3).Overall, 83% of the tagged fish detected on the entrance arraywere detected on the exit array. During 2010–2012, residence timefor individual fish was calculated by the difference between thedate and time of last detection at any receiver in the study areaand the date and time of release. The residence time estimates for2010–2012 are minimums because the amount of time a fish wasin the general area before we captured it for tagging was un-known, and the last detection could have been due to tag batterydepletion, not necessarily emigration out of the study area.

We used Kruskal–Wallis and Mann–Whitney tests to comparemedian residence times among the tagged species and betweenmain- and off-channel areas, respectively, for the 2007–2008 data-set. To examine temporal variation, we calculated the off-channelpercentage (out of the total for main- and off-channel areas) forsuccessive 3-day blocks of time, except during the start or end ofsome runs when more days had to be pooled to obtain adequatenumbers of fish for precise estimates. Minimum, mean, and maxi-

mum sample sizes used to calculate the percentage of fish detectedoff-channel were as follows: yearling Chinook salmon 285, 522, and822; subyearling Chinook salmon 107, 470, and 966; and steelhead126, 196, and 240, respectively. The off-channel percentage for eachrun of fish studied in 2007 and 2008 was plotted by the mid-date ofeach block of pooled days along with river discharge and water tem-perature data measured at Bonneville Dam and obtained from DataAccess in Real Time (www.cbr.washington.edu/dart). We used non-parametric statistics to test for differences in residence times be-cause the data were not normally distributed.

Results

2007–2008Median residence times for acoustic-tagged fish migrating

through the study area during spring–summer 2007–2008 wereshort — 2.5 to 3.4 h — and varied among species (Table 3). Regard-less of detection location in main- or off-channel areas, medianresidence times for steelhead (STH) were 22.8% shorter than thoseof subyearling Chinook salmon (CH0) and 10.8% shorter thanthose of yearling Chinook salmon (CH1). Median residence timeswere significantly different among the three groups of fish (CH0,CH1, and STH; P < 0.001, Kruskal–Wallis). Dunn’s multiple pair-wise comparisons of the medians (CH1 versus STH, CH1 versusCH0, and CH0 versus STH) were all significant (P < 0.05). For eachrun of fish (CH0, CH1, and STH), the median residence times forfish detected in off-channel areas were significantly longer thanthe times for fish detected only in main-channel areas (Mann–Whitney rank sum test; P < 0.001). However, differences between

Fig. 3. Locations of acoustic receivers for studies during 2007, 2008, 2010, 2011, and 2012.

688 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

off- and main-channel median residence times were short (1.43 hfor CH1, 0.87 h for STH, and 0.38 h for CH0) and may not bebiologically significant.

Of the thousands of acoustic-tagged fish of each species de-tected on receivers in the study area during spring–summer 2007–2008 (Table 4), most were migrating in the main channel (90.7% to96.4% out of the total of main- and off-channel pathways, by spe-cies and year). The percentage of fish in off-channel areas out ofthe total in the main- and off-channel areas was highest for CH1salmon (8.1% and 9.3% for 2007 and 2008, respectively), and lowestfor STH (4.0% for 2008) and CH0 salmon (3.6% and 6.1% for 2007and 2008, respectively). The off-channel percentages of CH1 andCH0 salmon varied (Fig. 4) and were negatively associated withwater temperature (P < 0.01), whereas percentages of off-channelfor STH were positively associated with water temperature (P < 0.01).We observed no relation between the percentage of CH1 salmonin off-channel areas and river discharge (P = 0.76), but percentagesof off-channel STH and CH0 salmon were positively correlatedwith river discharge (P < 0.01; Fig. 4).

2010–2012Median residence times observed in winter and early spring

studies (2010–2012), which are minimum estimates because it wasunknown how long a given fish was in the study area before it wascaptured and tagged, varied depending on year and species(Table 5). Median times were 25.5, 11.6, and 18.4 days for the juve-nile Chinook salmon during 2010, 2011, and 2012, respectively, but

a Kruskal–Wallis one-way analysis of variance (ANOVA) on ranksof residence times was not significant (P = 0.351). Yearling cohosalmon tagged in 2011 had a median residence time of 11.2 days.Median residence times of juvenile Chinook salmon sampled inoff-channel habitats in winter and early spring 2010–2012 (531.8 h)were much longer and significantly different than those for juve-niles sampled in the same habitats during spring–summer 2007–2008 (3.6 h; Mann–Whitney rank sum test; P < 0.001).

Tagged fish exited the study areas during February, March, andApril of 2010–2012 (Fig. 5). In general, we observed peaks in exitingtiming of tagged yearling salmon during February and April, andsmaller peaks in mid-March. Coho and Chinook salmon had sim-ilar exit timing patterns. Additionally, the exit timing for Chinooksalmon at the Sandy River delta was similar to the timing at theCottonwood Island study area.

The genetic stock composition of the juvenile Chinook salmoncaptured and tagged during winter 2010, 2011, and 2012 variedamong years (Table 6). Nearly all (93%) of the tagged juveniles didnot have clipped adipose fins and were therefore likely of naturalorigin. The Willamette River spring stock was most prevalentduring 2010 (24 of 41 fish; 59%) and 2011 (8 of 12 fish; 67%). Most ofthese juveniles were captured in the area of the Sandy River deltaand likely originated in the nearby Sandy River, which had beenextensively stocked with Willamette River hatchery Chinooksalmon during the 20th century (Myers et al. 2006; Teel et al. 2014).

Table 3. Residence times (h) of acoustic-tagged fish in the vicinity of the Sandy River delta during spring and summer 2007 and 2008.

2007 2008

Migrationpathway Statistic CH1 CH0 CH1 STH CH0

All combined Mean (SE) 4.0 (0.1) 3.8 (0.1) 6.6 (0.3) 2.7 (0.1) 3.3 (0.0)Median 2.6 3.4 3.0 2.5 3.0Range 1.3–255.7 1.6–565.0 0.8–373.2 0.8–211.0 1.3–55.6n 3567 5222 3748 1525 4157Cum. %

Main only Mean (SE) 3.6 (0.1) 3.7 (0.1) 5.5 (0.3) 2.7 (0.1) 3.2 (0.0)Median 2.6 3.4 3.0 2.5 3.0Range 1.4–255.7 1.8–565.0 0.8–373.2 0.8–211.0 1.6–26.5n 3231 5028 3347 1468 3932Cum. %

Off-channel Mean (SE) 8.4 (0.7) 4.8 (0.3) 15.2 (1.6) 3.7 (0.3) 4.4 (0.3)Median 3.2 4.0 4.5 3.4 3.0Range 1.3–107.0 1.7–23.2 0.9–278.1 0.9–12.6 1.3–55.6n 336 194 401 57 225Cum. %

Note: The x-axis scale is from 0 to 30 h, which includes 90%–100% of observed residence times. The y-axis scale is cumulative percentage from 0% to 100%.

Johnson et al. 689

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

During 2012, the West Cascades fall stock from the lower ColumbiaRiver evolutionarily significant unit (described by Myers et al.2006) dominated the catch (10 of 14 fish; 71%). Upper ColumbiaRiver summer–fall stock (five fish) was present only in 2010. Over-all, 75% of the fish analyzed were assigned to a genetic stock with

high relative assignment probabilities (P ≥ 0.90). Two fish in 2010were estimated to come from Snake River stocks but both hadrelative assignment probabilities less than 0.50.

We observed weak negative relationships among residencetime and fork length and mass measurements for all tagged fish

Table 4. Number (No.) and proportion (p) of acoustic-tagged fish detected by migration pathway in the vicinity of the Sandy Riverdelta during spring and summer 2007 and 2008.

2007 2008

CH1 CH0 CH1 STH CH0

Migration pathway No. p No. p No. p No. p No. p

Main only 4626 0.919 5558 0.964 6437 0.907 2477 0.960 5576 0.939Gary only 9 0.002 0 0.000 8 0.000 3 0.001 6 0.001Reed only 193 0.038 35 0.006 332 0.047 33 0.013 208 0.035Main and Gary 73 0.015 9 0.002 215 0.030 57 0.022 69 0.012Main and Reed 131 0.026 165 0.028 103 0.016 11 0.004 81 0.013

Total off-channel 406 0.081 209 0.036 658 0.093 104 0.040 364 0.061Grand total 5032 1.000 5767 1.000 7095 1.000 2581 1.000 5940 1.000

Fig. 4. Trends in water temperature, river discharge, and the percentage of each run of fish in off-channel habitats of the lower ColumbiaRiver during fish runs in 2007 (left) and 2008 (right). Three days of observations were pooled to calculate the percentage of each fish rundetected in off-channel areas, except in a few cases at the beginning or end of some runs when more days had to be pooled to obtain adequatesample sizes of at least 100 fish. Numbers above or next to some points indicate the number of days that had to be pooled when more than3 days were required.

Table 5. Residence times (days) of acoustic-tagged fish during late winter and early spring 2010, 2011, and 2012.

2010 2011 2012

Statistic CH1 CH1 Coho CH1

Mean (SE) 33.4 (4.0) 24.7 (8.1) 28.6 (5.4) 26.2 (6.8)Median 25.5 11.6 11.2 18.4Range 0.3–77.6 0.1–73.7 0.02–89.8 0.9–83.7n 41 12 36 14Cum. %

Note: The x-axis scale is from 0 to 90 days, which includes 99% of observed residence times. The y-axis scale is cumulative percentagefrom 0% to 100%. CH1, yearling Chinook salmon.

690 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

over 2010, 2011, and 2012 combined (Fig. 6), although neither re-lationship was significant (P = 0.29 for length–residence time andP = 0.15 for mass–residence time).

DiscussionShort residence times of only several hours (Table 3) in the

11 km study area in LCRE tidal fresh water during spring–summer2007–2008 by stocks upriver of Bonneville Dam indicate yearlingand subyearling life history strategies that feature active migra-tion toward the ocean mostly down the main river channel. Forjuvenile salmonids, we define “active migrants” as fish exhibit-ing net downstream movement with short residence time in anyparticular location. On the other hand, during late winter – earlyspring 2010–2012, median residence times in off-channel areasranged from 11 to 26 days for CH1 salmon (mean yearly lengths106, 115, and 118 mm; Table 5). These fish, mostly naturally pro-duced juveniles from tributaries in the lower Columbia and Wil-lamette rivers, presumably emigrated from their natal streamsduring fall or winter as subyearlings to rear in tidal fresh water,

followed by resumption of seaward migration in late winter –early spring as yearlings. Juvenile coho salmon residence periods(median 11.2 days) and fish size (mean 116 mm) were similar tothose of Chinook salmon, indicating similar patterns of migrationand rearing during this period.

Our study has several important assumptions and caveats. Weassumed that the tagging process from capture to release did notaffect residence time estimates. This assumption is sensible be-cause methods for collecting, tagging, and releasing fish are wellestablished, were applied consistently throughout our study and,based on other evaluations (e.g., Weiland et al. 2013), were un-likely to bias residence time estimates. Tag burdens in our study(Table 2) were always below the 7% threshold recommended byBrown et al. (2010). Also, effects of an acoustic tag on subsequentbehavior were likely minimal (Anglea et al. 2004; Brown et al.2010). Another assumption was that residence time estimateswere not affected by the translocation of fish from the capturelocation to the release location 10 and 8 km away during 2010 and2012, respectively, and during spring–summer studies 2007–2008.Translocation could have biased residence time estimates posi-tively because of possible acclimation time after release or nega-tively because of potential after-handling escape behavior at therelease location, even after tagged fish were acclimated for 18–24 h in river water prior to release. Effects of translocation onresidence time estimates are unknown, but many studies of juve-nile salmon necessarily involve translocating tagged fish (e.g.,Sommer et al. 2001). We also assumed that the last detection of atagged fish represented exit from the study area by the taggedfish. This assumption is reasonable because mean tag lives wereusually longer than residence times (Tables 1, 3, and 5). Underlyingthis is the assumption is that detections were attributed to fishmovements and not to predators that may have consumed taggedfish. A caveat to the study was that using fish size to categorize lifehistory type (yearling versus subyearling) must be qualified by thefact that migration distances from natal streams to the ocean,water temperature, and other environmental factors affect size atmigration to sea (Taylor 1990).

Fig. 5. Exit timing represented by date of last detection event for tagged Chinook salmon (CH) during 2010, 2011, and 2012 and for cohosalmon (CO) during 2011. Release dates were 1–2 days before the first animal was detected exiting the study area.

Table 6. Genetic stock identification of yearling Chinook salmontagged during 2010, 2011, and 2012. Fish capture dates are presentedbelow each year.

2010 2011 2012

25, 26, 28 Jan. 1, 2, 3, 14 Feb. 1, 2 Feb.

Snake, fall 1 0 0Snake, spring 1 0 0Upper Columbia,

summer–fall5 1 0

West Cascades, fall 4 2 10West Cascades, spring 4 1 2Willamette, spring 24 8 2Unknown 2 0 0

Total 41 12 14Unmarked 40 11 11

Johnson et al. 691

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

Residence times and migration characteristicsJuvenile salmon can display a range of migration behaviors and

residence times during seaward migration (e.g., Moser et al. 1991;Connor et al. 2005). Residence times can vary among habitats innatal streams, reservoirs, tidal fresh water, or estuarine water(McCleave 1978; Pinnix et al. 2013; Tytler et al. 1978). In the LCRE,previous telemetry studies of juvenile salmonids (>95 mm) re-vealed relatively short residence times during the spring and earlysummer study periods. For example, CH1 and CH0 salmon cap-tured at mainstem dams and tagged with acoustic transmittersexhibited mean residence times in the 234 km LCRE of 3.4 and4.1 days, respectively (McMichael et al. 2010). Harnish et al. (2012)reported that median residence times of 1–2 days between rkm 8and 86 were similar for the CH1 and CH0 salmon and yearling STHthey studied. By recapturing at rkm 74 fish that were released atvarious locations in the Columbia River basin, Dawley et al. (1986)found that smaller subyearling salmon (<120 mm) using shallow-water habitats tended to spend more time in the LCRE than largerjuvenile migrants (>120 mm). In our study, the relationships be-tween fish size and exit timing (positive trend) and fish size andresidence time (negative trend), although not statistically signifi-cant, were similar to observations by Dawley et al. (1986).

During 2010–2012, we examined the latter portion of the over-wintering period, as shown by the protracted exit timing duringFebruary, March, and April (Fig. 6). Our exit timing data are con-sistent with Dawley et al. (1984), who observed that fall-releasedfish migrated past Jones Beach (rkm 74) in February, March, andApril. Tagged fish exited our off-channel tidal freshwater studyareas during late winter and early spring, presumably migratingtoward the ocean. Such timing would indicate ocean entry earlierin spring than the peak for yearling fish in May (Daly et al. 2014;Weitkamp et al. 2012) and subyearling fish in June (Weitkampet al. 2012). Ocean entry spread over time among various stocks andlife history strategies may contribute to fish resilience (Bottom et al.2009), i.e., the ability to persist in a changing environment (Holling1973).

Over study years 2007–2008, 3.6%–9.3% of total passage wasthrough off-channel routes in our Sandy River delta study area atrkm 192–203; the bulk of the tagged fish migrated in the mainchannel of the river (Table 4). Weitkamp et al. (2012) reported thatyearling salmon were distributed primarily in the main channelnear the mouth of the Columbia River. Harnish et al. (2012) foundthat tagged fish detected at rkm 50 downstream of Puget Island inthe Columbia River were detected mostly in the main channel(84%–88%), although some (12%–16%) used Clifton Channel, an off-

channel route to the ocean via Cathlamet Bay. Dawley et al. (1986)noted that the larger, faster migrating juvenile salmon were moreclosely associated with the main channel of the river than thesmaller juvenile salmon, which were more laterally dispersed inoff-channel areas. There was no relationship between use of off-channel pathways at the Sandy River delta and vicinity during2008 and subsequent use of off-channel routes in the estuarineportion of the LCRE (G. Johnson, unpublished data). Given the verylarge number of individuals that were tagged each year (>20 000),our results are representative of limited (3.6% to 9.3% of totalpassage) use off-channel tidal freshwater habitats relative to themain channel during spring and summer by most juveniles orig-inating in the interior Columbia River basin. These percentages,however, are not inconsequential in the context of the millions ofjuvenile salmon migrating through the LCRE each year.

The timing of migrations by juvenile salmon is linked to a vari-ety of environmental conditions, including water temperaturesand discharge level (Hoar 1958; Koski 2009). Bjornn and Reiser(1991) report the downstream migration by juvenile salmon intonon-natal habitats during winter was a result of low water tem-perature and lack of instream habitat. In our study, the relation-ship between water temperature (as measured at Bonneville Dam)and the percentage of off-channel fish out of the total in main- andoff-channel was negative for CH1 and CH0 and positive for STH.The spring results are likely spurious because the trends wereopposite for CH1 and STH, which have similar temperature toler-ances. In addition, spring water temperatures and the range intemperatures seem to be too low (11.5 to 14.4 °C) to have had abiological impact on behavior and distributions. However, thesignificant inverse correlation of the percentage of CH0 off-channeland water temperature may reflect a valid biological responsebecause water temperatures rose to 21 °C in July when off-channelpercentages for CH0 were lowest. Moderate to high river dis-charge (>11 000 m3·s–1) was positively related to the percentage ofSTH in off-channel areas (Fig. 5). The strongest positive relation-ship between discharge and migration we observed was for CH0 inoff-channel areas during discharge levels of 3000–10 000 m3·s–1.Perry et al. (2010) also noted a correlation between the migrationtiming of juvenile Chinook salmon and river discharge in theSacramento–San Joaquin River delta, where fish released coinci-dent to relatively low river discharges had longer travel times andhigher survival rates compared with fish released during rela-tively high discharge.

Distribution of juvenile salmon in LCRE off-channel areas mayalso have been driven by resource partitioning among different

Fig. 6. Relationships between residence time and (a) fork length and (b) mass for juvenile Chinook and coho salmon tagged during winter2010, 2011, and 2012 for residence time studies.

692 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

size classes of fish and different habitat types. This was indicated,for example, by the short residence times and low percentagesoff-channel for steelhead compared with Chinook salmon. Giventhat approximately 100 million juvenile Chinook salmon are re-leased from Columbia basin hatcheries each year (HSRG 2009),off-channel distribution may also reflect density-dependent allo-cation of available space and resources. At a minimum, off-channelhabitats provide a migration route and, at a maximum, they pro-vide access to prey and other benefits, even though the taggedjuvenile salmonids we studied were actively migrating down-stream and not residing in off-channel areas for long. Short resi-dence times in off-channel areas by actively migrating salmonidsin spring and summer do not necessarily infer lack of importanceof these habitats.

Life history strategies and overwinteringOur study documented two life history strategies for juvenile

salmonids in off-channel habitats of LCRE tidal fresh water:(i) active migrations and short off-channel residence by upperriver Chinook salmon and steelhead during the primary springand summer migration periods and (ii) overwinter rearing in tidalfreshwater habitats by coho salmon and naturally produced Chi-nook salmon mostly from lower river sources. The brief residencyand minimal use of off-channel pathways by fish expressing theactive migration life history strategy are consistent with previousresearch documenting rapid migrations through the LCRE bylarge CH1 and CH0 salmon and STH originating in upriver areas(McMichael et al. 2010; Harnish et al. 2012; Weitkamp et al. 2012).

In contrast, the extended tidal freshwater rearing in late winterand early spring is a relatively undocumented life history, thoughprevious studies have suggested this life history may have beenmore common in the past (Rich 1920; Burke 2004). Our data revealthat the overwintering behavior is currently expressed by natu-rally produced fish from a diverse suite of lower river geneticstocks. Moreover, our results illustrate that the full range of theChinook salmon juvenile life histories in the Columbia River isnot easily described using a simple subyearling–yearling dichot-omy. We suggest that the juveniles in our late winter – earlyspring 2010–2012 study entered the LCRE as subyearlings and mi-grated to the ocean as yearlings.

The 105 mm size mode for Chinook salmon that Sather et al.(2011) found in winter in the same Sandy River delta study area asours was not observed in spring and summer samples (Fig. 2a),implying these fish had exited shallow water in the study area.The spring and summer length frequency data from the beachseine sampling by Sather et al. (2011) are consistent with the pat-tern of relatively few large-sized, actively migrating yearling andsubyearling fish using off-channel habitats during spring and sum-mer (Table 4). This pattern is supported by Ploskey et al. (2009),who documented a corresponding 115 mm mode for Chinooksalmon sampled at Bonneville Dam during summer 2008 (Fig. 2b).

Coho salmon lengths, masses, and condition factors were sim-ilar to those for the tagged Chinook salmon (Table 2). The 36 cohosalmon collected off-channel in winter 2011 averaged 116 mm inlength, which suggests that they were overwintering yearling fish.This 116 mm mean length during January sampling comports withDawley et al. (1986), who reported a mean length of �130 mm foractively migrating coho salmon in spring.

Our data strongly suggest that juvenile Chinook and coho salmonwere overwintering in off-channel tidal freshwater habitats of theLCRE. Minimum residence times of 11.6–25.5 days for juvenileChinook and 11.2 days for coho salmon (Table 5) imply the fishwere not active migrants. In addition, the fish lengths of �115 mmof our yearling salmon captured and tagged during winters 2010–2012 (Table 2) indicate these fish likely migrated to LCRE tidalfresh water in fall and winter as subyearlings (Reimers and Loeffel1967). The relatively large CH1 salmon that we tagged were clearlydifferentiated from the smaller and younger CH0 salmon in the

off-channel LCRE habitats we studied (Fig. 2a). Our findings con-firm the early work of Rich (1920), who used scale analysis to inferthat some subyearlings remain in the Columbia River estuary overthe winter.

Winter rearing in the LCRE by juvenile Chinook salmon has alsobeen indicated in several previous investigations, although noneof those studies assessed the duration of residency in off-channeltidal freshwater habitats. Spring run Chinook salmon from theWillamette River basin in Oregon include a subyearling fall–winter-migrant life history (Keefer and Caudill 2010) that is pres-ent in winter in LCRE tidal freshwater habitats (Teel et al. 2014).Reimers and Loeffel (1967) analyzed scale samples of lower Colum-bia River fall-run Chinook salmon and found that juveniles hadextended residence times in freshwater habitats, which the au-thors suggested may have included tidal areas. Dawley et al. (1986)studied marked (hatchery) juvenile Chinook salmon that werereleased from Columbia River tributaries during late summer andautumn by recapturing fish downstream at Jones Beach (rkm 75).Some fish apparently overwintered in upstream areas, possiblyincluding tidal freshwater habitats, before emigrating to sea thefollowing spring.

Overwintering by juvenile salmon in tidal fresh water has beensuggested in other regions as well. Cunjak et al. (1989) surmisedthat some juvenile Atlantic salmon (Salmo salar) moved from anestuary on the Newfoundland coast back to fresh water for over-wintering. Juvenile Chinook salmon migrating from upstreamlocations in the Fraser River were found to occupy non-natal trib-utaries, many of which were tidally influenced (Levings et al. 1995;Murray and Rosenau 1989). Overwintering is indicated from thebimodal size distribution (�50 and �125 mm) data that Kjelsonet al. (1982) reported for fall-run Chinook salmon sampled in SanFrancisco Bay during January–March 1981. In Oregon coastal rivers,juvenile coho salmon move in autumn and winter from natalstreams downstream to tidal areas for rearing before entering theocean the following spring (Miller and Sadro 2003; Jones et al.2014).

The juvenile salmon we studied during 2010–2012 were likelyusing shallow tidal freshwater habitats to feed and grow. Al-though we did not sample the stomach contents of tagged fish,concurrent investigations in the Sandy River delta study area dur-ing 2010 suggested that dipterans and hemipterans were impor-tant components of the diet of Chinook salmon during late winterand early spring (Storch and Sather 2011). Bioenergetics modelingbased on juvenile salmon diet and environmental conditions dur-ing the 2010 study period indicated fish growth was generallypositive, except during temperature extremes (A. Storch, unpub-lished data). In our study, fish condition during winter was good(Table 2), which provides another indication supporting the eco-logical importance of off-channel, tidal freshwater habitats. Gen-erally, rearing in non-natal habitats is important to the survivaland productivity of salmon populations, as Murray and Rosenau(1989) concluded for the Fraser River.

Future researchThe 2010–2012 component of this study focused on the larger

class of the bimodal size distribution of juvenile Chinook salmonpresent during winter in shallow tidal freshwater habitats of theLCRE (Sather et al. 2011; Fig. 2a). Such focus was necessitated bythe minimum size of fish (>95 mm fork length) that could betagged with an acoustic transmitter given available technology. Asacoustic-telemetry technology advances and smaller transmitterswith long-life batteries (6–12 months) become available, researchshould emphasize tagging smaller size ranges of fish across mul-tiple seasons to further elucidate the role of tidal freshwater hab-itats to juvenile fishes. While juvenile salmon >95 mm constituteimportant components of life history strategies, the majority ofjuvenile salmon occupying shallow tidal freshwater habitatsare smaller than 95 mm (Roegner et al. 2010, 2012; Sather et al.

Johnson et al. 693

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

2011). Quantifying the spatial and temporal extent of residencetime by juvenile salmon in the LCRE, especially at newly restoredwetlands, will provide a better understanding of the potentialbenefits derived from the different life history strategies of fish intheir respective habitats. Based on Dawley et al. (1986), we hypoth-esize that juvenile Chinook salmon smaller than 95 mm will havelonger residence times and potentially derive greater direct ben-efits from access to restored shallow water habitats comparedwith larger-sized fish. Future research should involve capturingand tagging upriver juveniles during late summer and fall, thenmonitoring them in the LCRE during later fall, winter, and earlyspring to further elucidate life histories as expressed by overwin-tering in off-channel, tidal freshwater habitats, such as in theColumbia River gorge, Multnomah channel, and behind mains-tem islands in LCRE tidal fresh water. Studies applying advancedtechnology for acoustic transmitters and receivers could allow forsurvival estimation based on methods developed by Perry andSkalski (2008) for juvenile salmon in restoring and reference LCREhabitats to quantify the effects of habitat restoration. Habitat-specific survival estimates would provide a definitive measure ofrealized function of particular habitats (Simenstad and Cordell2000). To substantiate our results, additional investigation of ju-venile salmonid residence times in different off-channel areas intidal fresh water is warranted. Finally, results from studies ofother realized functions for juvenile salmonids of restoring wet-lands besides residence time, such as fish growth and prey export,would be useful to restoration managers.

Restoration implicationsThe residence times we observed have implications for fish ac-

cess to and restoration of off-channel, tidal freshwater habitats.The winter residence times observed in this study were not par-ticularly long compared with the 3–6 months that residualizingfish can spend in fresh water (Connor et al. 2005), but overwinter-ing fish likely use more than one off-channel habitat in the234 km expanse of the LCRE before entering the ocean. Theamount of off-channel habitat available, therefore, is importantto support a diversity of life history strategies expressed by juve-nile salmon; there are approximately 80 km2 of off-channel areas,including distributary channels, in the mainstem LCRE comparedwith the potential recoverable area of 344 km2 in the LCRE flood-plain (Diefenderfer et al. 2013). In fact, lateral connectivity andaccess to such habitats is a fundamental mechanism for juvenilesalmon to derive direct and indirect benefits from such habitats(Simenstad and Cordell 2000). For example, a large number ofactively migrating juvenile salmonids could benefit directly(through access) because the numbers of these fish using off-channel areas are not inconsequential, even though their resi-dence times are short (�hours). Linkages between reconnectedwetlands in off-channel areas and the main stem also have indi-rect benefits through prey export, as demonstrated by Diefenderferet al. (2013). Off-channel areas generally provide structural andfunctional linkages between wetland restoration areas and themainstem river that benefit both actively migrating and overwin-tering or rearing juvenile salmonids.

Reconnecting diked areas of the floodplain to the mainstemriver to increase juvenile salmon access and export of prey andother materials from shallow tidal wetlands is a key strategy forLCRE habitat restoration (BPA and USACE 2014; NMFS 2014). Usinginformation about residence time, environmental conditions,and life history variability helps to elucidate patterns of habitatuse and advance informed decision-making by restoration pro-gram managers. Access to shallow-water, off-channel habitatsmay be especially important for juvenile salmon with subyearlinglife histories that rear in tidal areas (Jones et al. 2014), especiallythose that eventually emigrate to the sea as yearlings. Further-more, the juvenile salmon residence in shallow, tidal freshwaterwetland areas during late winter and early spring that we ob-

served and general year-round presence in such habitats noted byothers (Roegner et al. 2012; Sather et al. 2011; Thom et al. 2013)support including, in current restoration strategy, the consider-ation of ecological conditions across multiple seasons at prospec-tive restoration sites and landscapes. Our genetic data, togetherwith those of previous studies (Teel et al. 2009, 2014; Roegner et al.2012; Sather et al. 2011), indicate that such a strategy will benefitan array of distinct genetic stocks. Multiple life history strategiesincrease the chances of survival across stocks and populationsusing different habitats during different seasons (Bottom et al.2009; Elmqvist et al. 2003; Healey 1991). This study is the first toestimate residence times for juvenile salmonids specifically inoff-channel areas of tidal fresh water and, most importantly, res-idence times for Chinook salmon expressing a life history of over-wintering in tidal fresh water. These findings support restorationof shallow off-channel habitats in the LCRE, ultimately leading toimproved life history diversity and population resilience of sal-monids in the Columbia River basin.

AcknowledgementsThis research was funded by the Bonneville Power Administra-

tion (BPA) and the US Army Corps of Engineers Portland Dis-trict (USACE). We are grateful to Blaine Ebberts, Brad Eppard,Mike Langeslay, and Cindy Studebaker (USACE) and Tracey Yerxa(BPA) for supporting the study; David Kuligowski of the NationalMarine Fisheries Service (NMFS) for analyzing the genetic sam-ples; Earl Dawley (NMFS, retired) for summarizing migration char-acteristics; and Adam Storch and Tucker Jones of the OregonDepartment of Fish and Wildlife, and Amanda Bryson, DanielDeng, Susan Ennor, Eric Fischer, Matt Hennen, James Hughes,Ron Kaufman, Geoff McMichael, Mark Weiland, Christa Woodley,and Shon Zimmerman of the Pacific Northwest National Labora-tory for helping conduct the study. We sincerely appreciate theinsightful comments from three anonymous peer reviewers.

ReferencesAnglea, S.M., Geist, D.R., Brown, R.S., Deters, K.A., and McDonald, R.D. 2004.

Effects of acoustic transmitters on swimming performance and predatoravoidance of juvenile Chinook salmon. N. Am. J. Fish. Manage. 24(1): 162–170.doi:10.1577/M03-065.

Bjornn, T., and Reiser, D. 1991. Habitat requirements of salmonids in streams, InInfluences of forest and rangeland management on salmonid fishes and theirhabitats. Edited by W.R. Meehan. American Fisheries Society, Bethesda, Md.pp. 83–138.

Bottom, D.L., Jones, K.K., Cornwell, T.J., Gray, A., and Simenstad, C.A. 2005.Patterns of Chinook salmon migration and residency in the Salmon Riverestuary (Oregon). Estuar. Coast. Shelf. Sci. 64: 79–93. doi:10.1016/j.ecss.2005.02.008.

Bottom, D.L., Jones, K.K., Simenstad, C.A., and Smith, C.L. 2009. Reconnectingsocial and ecological resilience in salmon ecosystems [online]. Ecol. Soc.14(1): 5. Available from http://www.ecologyandsociety.org/vol14/iss1/art5/.

BPA and USACE. 2014. Columbia Estuary Ecosystem Restoration Program: 2014Restoration and Monitoring Plan. Final report, prepared by the BonnevillePower Administration and US Army Corps of Engineers, Portland, Ore.

Brown, R.S., Harnish, R.A., Carter, K.M., Boyd, J.W., Deters, K.A., andEppard, M.B. 2010. An evaluation of the maximum tag burden for implanta-tion of acoustic transmitters in juvenile Chinook salmon. N. Am. J. Fish.Manage. 30: 499–505. doi:10.1577/M09-038.1.

Burke, J.L. 2004. Life histories of juvenile chinook salmon in the Columbia RiverEstuary, 1916 to the present. Masters thesis, Oregon State University, Corval-lis, Ore.

Carter, K.M., Woodley, C.M., and Brown, R.S. 2011. A review of tricaine meth-anesulfonate for anesthesia of fish. Rev. Fish Biol. Fish. 21(1): 51–59. doi:10.1007/s11160-010-9188-0.

Connor, W.P., Sneva, J.G., Tiffan, K.F., Steinhorst, R.K., and Ross, D. 2005. Twoalternative juvenile life history types for fall Chinook salmon in the SnakeRiver basin. Trans. Am. Fish. Soc. 134: 291–304. doi:10.1577/T03-131.1.

Cunjak, R.A., Chadwick, E.M.P., and Shears, M. 1989. Downstream movementsand estuarine residence by Atlantic salmon parr (Salmo salar). Can. J. Fish.Aquat. Sci. 46(9): 1466–1471. doi:10.1139/f89-187.

Daly, E.A., Scheurer, J.A., Brodeur, R.D., Weitkamp, L.A., Beckman, B.R., andMiller, J.A. 2014. Juvenile steelhead distribution, migration, feeding, andgrowth in the Columbia River Estuary, plume, and coastal waters. Mar. Coast.Fish. Dyn. Manage. Ecosyst. Sci. 6(1): 62–80. doi:10.1080/19425120.2013.869284.

694 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

Dawley, E.M., Ledgerwood, R.D., Blahm, T.H., Kirn, R.A., and Rankis, A.E. 1984.Migrational characteristics and survival of juvenile salmonids entering theColumbia River estuary during 1983 [online]. Bonneville Power Administra-tion, Portland, Ore. Available from http://www.cbfish.org/Report.mvc/SearchPublications/ [June 2014].

Dawley, E.M., Ledgerwood, R.D., Blahm, T.H., Sims, C.W., Durkin, J.T., Kirn, R.A.,Rankis, A.E., Monan, G.E., and Ossiander, F.J. 1986. Migrational characteris-tics, biological observations, and relative survival of juvenile salmonids en-tering the Columbia River estuary, 1966–1983 [online]. Bonneville PowerAdministration, Portland, Ore. Available from http://www.cbfish.org/Report.mvc/SearchPublications/ [June 2014].

Deters, K.A., Brown, R.S., Boyd, J.W., Eppard, M.B., and Seaburg, A.G. 2010. Per-formance assessment of suture type, water temperature, and surgeon skill injuvenile Chinook salmon surgically implanted with acoustic transmitters.Trans. Am. Fish. Soc. 139: 888–899. doi:10.1577/T09-043.1.

Diefenderfer, H.L., Johnson, G.E., Thom, R.M., Borde, A.B., Woodley, C.M.,Weitkamp, L.A., Buenau, K.E., and Kropp, R.K. 2013. An evidence-based as-sessment of the cumulative effects of tidal freshwater and estuarine ecosys-tem restoration: early life-stage habitat functions for endangered salmonids[online]. Final report prepared by Pacific Northwest National Laboratory andNOAA Fisheries for the US Army Corps of Engineers, Portland District, Port-land, Ore. Available from http://www.pnl.gov/main/publications/external/technical_reports/PNNL-23037.pdf [June 2014].

Elmqvist, T., Folke, C., Nyström, M., Peterson, G., Bengtsson, J., Walker, B., andNorberg, J. 2003. Response diversity, ecosystem change, and resilience. Front.Ecol. Environ. 1: 488–494. doi:10.1890/1540-9295(2003)001[0488:RDECAR]2.0.CO;2.

ESA. 1973. Endangered Species Act of 1973. 7 U.S.C. § 136, 16 U.S.C. § 1531 et seq.Fleming, I.A., Bottom, D.L., Jones, K.K., Simenstad, C.A., and Craig, J.F. 2014.

Resilience of anadromous and resident salmonid populations. J. Fish Biol. 85:1–7. doi:10.1111/jfb.12429. PMID:24980458.

Fresh, K.L., Casillas, E., Johnson, L., and Bottom, D.L. 2005. Role of the estuary inthe recovery of Columbia River Basin salmon and steelhead: an evaluation oflimiting factors. NOAA Technical Memorandum NMFS-NWFSC-69.

Harnish, R.A., Johnson, G.E., McMichael, G.A., Hughes, M.S., and Ebberts, B.D.2012. Effect of migration pathway on travel time and survival of acoustic-tagged juvenile salmonids in the Columbia River estuary. Trans. Am. Fish.Soc. 141(2): 507–519. doi:10.1080/00028487.2012.670576.

Healey, M.C. 1991. Life history of Chinook salmon (Oncorhynchus tshawytscha). InPacific salmon life histories. Edited by C. Groot and L. Margolis. UBC Press,Vancouver, B.C. pp. 311–393.

Hoar, W.S. 1958. The evolution of migratory behaviour among juvenile salmonof the genus Oncorhynchus. J. Fish. Res. Bd. Can. 15: 391–428. doi:10.1139/f58-020.

Holling, C.S. 1973. Resilience and stability of ecological systems. Annu. Rev. Ecol.4: 1–23. doi:10.1146/annurev.es.04.110173.000245.

HSRG. 2009. Columbia River hatchery reform system-wide report [online].Hatchery Scientific Review Group. Available from http://hatcheryreform.us/hrp/reports/system/welcome_show.action [November 2013].

Johnson, G.E., Sather, N.K., Storch, A.J., Johnson, J., Skalski, J.R., Teel, D.J.,Brewer, T., Bryson, A.J., Dawley, E.M., Kuligowski, D.R., Whitesel, T., Mallette,C. 2013. Multi-scale action effectiveness research in the lower Columbia Riverand Estuary, 2012. PNNL-22481, final annual report submitted to the US ArmyCorps of Engineers, Portland District, Portland, Ore., by Pacific NorthwestNational Laboratory, Richland, Wash.

Jones, K.K., Cornwell, T.J., Bottom, D.L., Campbell, L.A., and Stein, S. 2014. Thecontribution of estuary-resident life histories to the return of adult Oncorhynchuskisutch. J. Fish Biol. 85: 52–80. doi:10.1111/jfb.12380. PMID:24766645.

Kalinowski, S.T., Manlove, K.R., and Taper, M.L. 2007. ONCOR: A computer pro-gram for genetic stock identification. Department of Ecology, Montana StateUniversity, Bozeman, Mont.

Keefer, M.L., and Caudill, C.C. 2010. A review of adult salmon and steelhead lifehistory and behavior in the Willamette River basin: Identification of knowl-edge gaps and research needs. Technical Report 2010-8, final report submit-ted to the US Army Corps of Engineers, Portland District, by the University ofIdaho, Moscow, Ida.

Kjelson, M.A., Raquel, P.F., and Fisher, F.W. 1982. Life history of fall-run juvenileChinook salmon, Oncorynchus tshawytscha, in the Sacramento-San Joaquin es-tuary, California. In Estuarine comparisons. Edited by V. Kennedy. AcademicPress, Inc., New York. pp. 393–411.

Koski, K.V. 2009. The fate of coho salmon nomads: the story of an estuarine-rearing strategy promoting resilience [online]. Ecol. Soc. 14(1):4. Availablefrom http://www.ecologyandsociety.org/vol14/iss1/art4/.

Levings, C.D., Boyle, D.E, and Whitehouse, T.R. 1995. Distribution and feeding ofjuvenile Pacific salmon in freshwater tidal creeks of the lower Fraser River,British Columbia. Fish. Manage. Ecol. 2: 299–308. doi:10.1111/j.1365-2400.1995.tb00121.x.

McCabe, G.T., Jr., Emmett, R.L., Muir, W.D., and Blahm, T.H. 1986. Utilization ofthe Columbia River estuary by subyearling Chinook salmon. Northwest Sci.60: 113–124.

McCleave, J.D. 1978. Rhythmic aspects of estuarine migration of hatchery-rearedAtlantic salmon (Salmo salar) smolts. J. Fish Biol. 12: 559–570. doi:10.1111/j.1095-8649.1978.tb04202.x.

McMichael, G.A., Eppard, M.B., Carlson, T.J., Carter, J.A., Ebberts, B.D.,

Brown, R.S., Weiland, M.A., Ploskey, G.R., Harnish, R.A., and Deng, Z. 2010.The juvenile salmon acoustic telemetry system: a new tool. Fisheries, 35(1):9–22. doi:10.1577/1548-8446-35.1.9.

Miller, B.A., and Sadro, S. 2003. Residence time and seasonal movements ofjuvenile coho salmon in the ecotone and lower estuary of Winchester Creek,South Slough, Oregon. Trans. Am. Fish. Soc. 132: 546–559. doi:10.1577/1548-8659(2003)132<0546:RTASMO>2.0.CO;2.

Moser, M.L., Olson, A.F., and Quinn, T.P. 1991. Riverine and estuarine migratorybehavior of coho salmon (Oncorhynchus kisutch) smolts. Can. J. Fish. Aquat. Sci.48(9): 1670–1678. doi:10.1139/f91-198.

Murray, C.B., and Rosenau, M.L. 1989. Rearing of juvenile Chinook salmon innonnatal tributaries of the lower Fraser River, British Columbia. Trans. Am.Fish. Soc. 118: 284–289. doi:10.1577/1548-8659(1989)118<0284:ROJCSI>2.3.CO;2.

Myers, J.M., Busack, C., Rawding, D., Marshall, A.R., Teel, D.J., Van Doornik, D.M.,and Maher, M.T. 2006. Historical population structure of Pacific salmonids inthe Willamette River and lower Columbia River basins. NOAA TechnicalMemorandum NMFS-NWFSC-73, National Marine Fisheries Service, Seattle,Wash.

NMFS. 2014. Endangered Species Act Section 7(a) (2) Consultation Supplementalbiological opinion consultation on remand for operation of the Federal Co-lumbia River Power System [online]. National Marine Fisheries Service(National Oceanic and Atmospheric Administration Fisheries) – NorthwestRegion, Seattle, Wash. Available from http://www.salmonrecovery.gov/BiologicalOpinions/FCRPSBiOp.aspx.

Oldenburg, E.W., Colotelo, A.H.A., Brown, R.S., and Eppard, M.B. 2011. Holding ofjuvenile salmonids for surgical implantation of electronic tags: a review andrecommendations. Rev. Fish Biol. Fish. 21: 35–42. doi:10.1007/s11160-010-9186-2.

Perry, R., and Skalski. J.R. 2008. Evaluating wetland restoration projects in theColumbia River Estuary using hydroacoustic telemetry arrays to estimatemovement, survival, and residence times of juvenile salmonids. Design andanalysis of salmonid tagging studies in the Columbia Basin [online]. Vol. XXII.Available from http://pisces.bpa.gov/release/documents/documentviewer.aspx?doc=P107849.

Perry, R.W., Skalski, J.R., Brandes, P.L., Sandstrom, P.T., Klimley, A.P.,Ammann, A., and MacFarlane, B. 2010. Estimating survival and migrationroute probabilities of juvenile chinook salmon in the Sacramento–SanJoaquin River Delta. N. Am. J. Fish. Manage. 30(1): 142–156. doi:10.1577/M08-200.1.

Pinnix, W.D., Nelson, P.A., Stutzer, G., and Wright, K.A. 2013. Residence time andhabitat use of coho salmon in Humboldt Bay, California: an acoustic–telemetrystudy. Environ. Biol. Fish. 96: 315–323. doi:10.1007/s10641-012-0038-x.

Ploskey, G.R. 2008. Report of take of listed species for PNNL-40. Report submit-ted by the Pacific Northwest National Laboratory, Richland, Wash., to theNational Marine Fisheries Service, Portland, Ore.

Ploskey, G.R., Weiland, M.A., Faber, D.M., Deng, Z., Johnson, G.E., Hughes, J.S.,Zimmerman, S.A., Monter, T.J., Cushing, A.W., Wilberding, M.C., Durham, R.E.,Townsend, R.L., Skalski, J.R., Buchanan, R.A., Kim, J., Fischer, E.S., Meyer,M.M., McComas, R.L., and Everett, J. 2009. Survival rates of juvenile sal-monids passing through the Bonneville Dam and spillway in 2008 [online].PNNL-18748, Pacific Northwest National Laboratory, Richland, Wash. Avail-able from http://www.pnl.gov/main/publications/external/technical_reports/PNNL-18748.pdf [June 2014].

Reimers, P.E., and Loeffel, R.E. 1967. The length of residence of juvenile fallChinook salmon in selected Columbia River tributaries. Fish Commission ofOregon Research Briefs, 13: 5–19.

Rich, W.H. 1920. Early history and seaward migration of Chinook salmon in theColumbia and Sacramento rivers. U.S. Bur. Fish. Bull. 37: 2–73.

Roegner, G.C., Dawley, E.W., Russell, M., Whiting, A.H., and Teel, D.J. 2010.Juvenile salmonid use of reconnected tidal freshwater wetlands in GraysRiver, lower Columbia River basin. Trans. Am. Fish. Soc. 139: 1211–1232. doi:10.1577/T09-082.1.

Roegner, G.C., McNatt, R., Teel, D.J., and Bottom, D.L. 2012. Distribution, size,and origin of juvenile Chinook salmon in shallow-water habitats of the lowerColumbia River and estuary, 2002-2007. Mar. Coast. Fish. Dyn. Manage. Eco-syst. Sci. 4: 450–472. doi:10.1080/19425120.2012.675982.

Sather, N., Teel, D., Storch, A., Johnson, G., Van Dyke, E., Dawley, E., Kuligowski,D., Jones, T., Bryson, A., and Sobocinski, K. 2011. Juvenile salmon and fishcommunity characteristics. In Ecology of juvenile salmon in shallow tidalfreshwater habitats of the lower Columbia River, 2007–2010 [online]. Edited byG.E. Johnson, A.J. Storch, J.R. Skalski, A.J. Bryson, C. Mallette, A.B. Borde,et al. PNNL-20083, Pacific Northwest National Laboratory, Richland,Wash. pp. 2.1–2.35. Available from http://www.pnl.gov/main/publications/external/technical_reports/PNNL-20083.pdf [June 2014].

Simenstad, C.A., and Cordell, J.R. 2000. Ecological assessment criteria for restor-ing anadromous salmonid habitat in Pacific Northwest estuaries. Ecol. Eng.15: 283–302. doi:10.1016/S0925-8574(00)00082-3.

Sommer, T.R., Nobriga, M.L., Harrell, W.C., Batham, W., and Kimmerer, W.J.2001. Floodplain rearing of juvenile Chinook salmon: evidence of enhancedgrowth and survival. Can. J. Fish. Aquat. Sci. 58(2): 325–333. doi:10.1139/f00-245.

Storch, A.J., and Sather, N.K. 2011. Feeding ecology. In Ecology of juvenile salmonin shallow tidal freshwater habitats of the lower Columbia River, 2007–2010[online]. Edited by G.E. Johnson, A.J. Storch, J.R. Skalski, A.J. Bryson, C. Mallette,

Johnson et al. 695

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.

A.B. Borde, et al. PNNL-20083, Pacific Northwest National Laboratory, Rich-land, Wash. pp. 4.1–4.20. Available from http://www.pnl.gov/main/publications/external/technical_reports/PNNL-20083.pdf [June 2014].

Taylor, E.B. 1990. Environmental correlates of life-history variation in juvenilechinook salmon, Oncorhynchus tshawytscha (Walbaum). J. Fish Biol. 37: 1–17.doi:10.1111/j.1095-8649.1990.tb05922.x.

Teel, D.J., Baker, C., Kuligowski, D.R., Friesen, T.A., and Shields, B. 2009. Geneticstock composition of subyearling chinook salmon in seasonal floodplainwetlands of the lower Willamette River, Oregon. Trans. Am. Fish. Soc. 138:211–217. doi:10.1577/T08-084.1.

Teel, D.J., Bottom, D.L., Hinton, S.A., Kuligowski, D.R., McCabe, G.T., McNatt, R.,Roegner, G.C., Stamatiou, L.A., and Simenstad, C.A. 2014. Genetic identifica-tion of Chinook salmon in the Columbia River estuary: stock-specific distri-butions of juveniles in shallow tidal freshwater habitats. N. Am. J. Fish.Manage. 34: 621–641. doi:10.1080/02755947.2014.901258.

Thom, R.M., Sather, N.K., Roegner, G.C., and Bottom, D.L. 2013. Columbia Estu-ary Ecosystem Restoration Program: 2012 synthesis memorandum. Final re-port prepared for the US Army Corps of Engineers, Portland, Oregon byPacific Northwest National Laboratory, Sequim, Wash., and National MarineFisheries Service, Seattle, Wash.

Tiffan, K.F., Rondorf, D.W., Garland, R.D., and Verhey, P.A. 2000. Identification

of juvenile fall versus spring Chinook salmon migrating through the lowerSnake River based on body morphology. Trans. Am. Fish. Soc. 129: 1389–1395.doi:10.1577/1548-8659(2000)129<1389:IOJFVS>2.0.CO;2.

Titzler, P.S., McMichael, G.A., and Carter, J.A. 2010. Autonomous acoustic re-ceiver deployment and mooring techniques for use in large rivers and estu-aries. N. Am. J. Fish. Manage. 30: 853–859. doi:10.1577/M09-143.1.

Tytler, P., Thorpe, J., and Shearer, W. 1978. Ultrasonic tracking of the movementsof Atlantic salmon smolts (Salmo salar L.) in the estuaries of two Scottishrivers. J. Fish Biol. 12: 575–586. doi:10.1111/j.1095-8649.1978.tb04204.x.

Weiland, M.A., Woodley, C.M., Ploskey, G.R., Hughes, J.S., Hennen, M.J., Kim,J.A., Deng, Z., Fu, T., Skalski, J.R., Townsend, R.L., Wagner, K.A., Fischer, E.S.,Duncan, J.P., Batten, G., Carlson, T.J., Carpenter, S.M., Cushing, A.W., Elder,T., Etherington, D.J., Johnson, G.E., Khan, F., Miracle, A.L., Mitchell, T.D.,Prather, K., Rayamajhi, B., Royer, I., Seaburg, A., and Zimmerman, S.A. 2013.Acoustic telemetry evaluation of juvenile salmonid passage and survival atJohn Day Dam, 2011 [online]. PNNL-22266, Pacific Northwest National Labo-ratory, Richland, Wash. Available from http://www.pnl.gov/main/publications/external/technical_reports/PNNL-22266.pdf [June 2014].

Weitkamp, L.A., Bentley, P.J., and Litz, M.N. 2012. Seasonal and interannualvariation in juvenile salmonids and associated fish assemblage in open wa-ters of the lower Columbia River estuary, U.S.A. Fish. Bull. 110: 426–450.

696 Can. J. Fish. Aquat. Sci. Vol. 72, 2015

Published by NRC Research Press

Can

. J. F

ish.

Aqu

at. S

ci. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y B

AT

TE

LL

E P

AC

IFIC

NW

NA

TL

LA

B o

n 04

/29/

15Fo

r pe

rson

al u

se o

nly.