UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Season-specific survival and growth rates of coastal cutthroat trout across a gradient of stream sizes… Sheldon, Kim Antoinette 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2010_fall_sheldon_kim.pdf [ 1.36MB ]
Metadata
JSON: 24-1.0070993.json
JSON-LD: 24-1.0070993-ld.json
RDF/XML (Pretty): 24-1.0070993-rdf.xml
RDF/JSON: 24-1.0070993-rdf.json
Turtle: 24-1.0070993-turtle.txt
N-Triples: 24-1.0070993-rdf-ntriples.txt
Original Record: 24-1.0070993-source.json
Full Text
24-1.0070993-fulltext.txt
Citation
24-1.0070993.ris

Full Text

Season-specific survival and growth rates of coastal cutthroat trout across a gradient of stream sizes in southwestern British Columbia  by  KIM ANTOINETTE SHELDON B.Sc., University of British Columbia, 2007  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES  (Forestry)  THE UNIVERSTIY OF BRITISH COLUMBIA (Vancouver)  June 2010  © Kim Antoinette Sheldon, 2010  Abstract To have greater predictive abilities for informed resource management guidelines for freshwater species protection and management, we need to increase our limited knowledge of the many aquatic species that inhabit small coastal streams. Coastal cutthroat trout (Oncorhynchus clarkii clarkii) are common to small streams in the Pacific Northwest and are a species of concern or threatened throughout their native range. To address knowledge needs I (1) examined the seasonal variation in survival rates of coastal cutthroat trout across a size gradient of smaller streams in coastal BC; and (2) contrasted 10th year post-harvest seasonal trout body condition and relative abundances, and thermal and physical habitats with pre-harvest and 4th year post-harvest values in small streams. The survival study used a robust mark-recapture design stratified seasonally to estimate monthly survival rates; and the streamside harvest study used a multi-year, replicated, before-after-control-impact design. Within the size range of smaller streams studied (n = 7), availability of aquatic habitat (i.e., residual depth) at low-flows was the best predictor of monthly survival rates (p = 0.011), supporting my hypothesis that greater availability of habitat confers higher survival in trout. In addition, a fitted curve suggested an asymptotic relation between water depth and survival rates; where beyond 25 cm of water, greater depth did not confer greater benefits to trout survival. Survival estimates also showed that the summer season had the lowest monthly survival rates across all streams in our study area. Post-harvest effects were not detected in trout relative abundances in the 10th year; however body condition had significantly (p < 0.001) increased in both the control and treatment streams compared to previous periods. During 2008-2009 treatment streams had comparatively higher water temperatures (0.8 – 1.5°C) than control streams. Given the cooler conditions of shaded streams in our coastal mountain region (e.g., 8-10°C); the modest warming of treatment streams may have conferred a benefit to trout growth and condition. ii  Table of contents Abstract .................................................................................................................................. ii Table of contents ................................................................................................................... iii List of tables ......................................................................................................................... vii List of figures ...................................................................................................................... viii Acknowledgements ................................................................................................................ x Co-authorship statement ........................................................................................................ xi Chapter 1 : Season-specific survival and growth rates of coastal cutthroat trout (Oncorhynchus clarkii clarkii) across a gradient of stream sizes in southwestern British Columbia ..................... 1 Forest harvesting, stream size and concern for cutthroat trout ................................................. 3 Study overview ....................................................................................................................... 6 Literature cited ....................................................................................................................... 8 Chapter 2 : Season-specific survival rates and physical characteristics of resident cutthroat trout in small streams of coastal British Columbia. ........................................................................... 14 Introduction .......................................................................................................................... 14 Methods ............................................................................................................................... 18 Study area ......................................................................................................................... 18 Gradient of stream sizes .................................................................................................... 18 Stream habitat surveys ...................................................................................................... 19 Trout sampling design and tagging .................................................................................... 21 Data analysis ........................................................................................................................ 23 iii  Defining stream size.......................................................................................................... 23 Biological variables in trout populations by stream and season .......................................... 24 Estimating apparent survival and recapture probabilities ....................................................... 25 Candidate models .............................................................................................................. 27 Model structure and fit ...................................................................................................... 28 Model averaging and selection .......................................................................................... 30 Biological covariates ......................................................................................................... 31 Trout population estimates ................................................................................................ 31 Results .................................................................................................................................. 32 Defining stream size.......................................................................................................... 32 Biological attributes of trout populations ........................................................................... 33 Space and time: apparent survival and recapture probabilities ........................................... 36 Survival model covariates ................................................................................................. 38 Discussion ............................................................................................................................ 38 Trout survival and stream size ........................................................................................... 38 Seasonal trout survival and stream size ............................................................................. 41 Trout biological characteristics and survival ...................................................................... 42 Movement and survival estimates...................................................................................... 45 Conservation implications and persistence of cutthroat populations ................................... 47 Literature cited ..................................................................................................................... 62  iv  Chapter 3 :  A before-after study contrasting the effects of streamside harvesting on cutthroat  trout populations. ..................................................................................................................... 71 Introduction .......................................................................................................................... 71 Methods ............................................................................................................................... 75 Study area ......................................................................................................................... 75 Stream surveys .................................................................................................................. 76 Trout sampling .................................................................................................................. 76 Data analysis ........................................................................................................................ 78 Fish habitat assessment ..................................................................................................... 78 Environmental data ........................................................................................................... 78 Relative abundance of trout ............................................................................................... 78 Trout body condition ......................................................................................................... 80 Results .................................................................................................................................. 81 Physical attributes ............................................................................................................. 81 Trout abundance ............................................................................................................... 82 Trout body condition ......................................................................................................... 83 Discussion ............................................................................................................................ 84 Trout abundance ............................................................................................................... 84 Trout body condition ......................................................................................................... 88 Stream A: upper reach ....................................................................................................... 89 Literature cited ..................................................................................................................... 99 v  Chapter 4 : Summary .............................................................................................................105 Literature cited ....................................................................................................................110 Appendices .............................................................................................................................111 Appendix A .....................................................................................................................111 Appendix B ......................................................................................................................112  vi  List of tables Table 2.1: Physical characteristics of seven small (second and third order) streams in the southern portion of Malcolm Knapp Research Forest, British Columbia. Measurements were record during the period of May 2008 to June 2009 (SE± 1). .................................................... 50 Table 2.2: Estimated abundance (N) (SE ± 1) for cutthroat trout using passive trapping methods within four primary periods occurring during spring, late summer and winter conditions (i.e. 3-4 secondary trap sessions for each primary) for seven small streams of Malcolm Knapp Research Forest . An „appropriate‟ model for calculating estimates and standard error of N for each primary session and stream was selected using population estimate software CAPTURE (program MARK; White and Burnham 1999). (see Appendix 1 for raw data used to calculate estimates). ................................................................................................................................ 51 Table 2.3: Output of top recapture and survival models with effects constrained to test a priori candidate models for „best‟ model to estimate recapture and survival probabilities for all individual trout encounter histories combined into stream groups. A logit link function was applied to fit a general linear model to all estimates (White and Coach 2009). .......................... 52 Table 2.4: Based on structural differences in the top models likelihood ratio tests between models were used to identify the main independent variables contributing to estimates of trout survival. A P <0.05 indicates null model factors, compared to the paired alternative model, contribute significantly to trout survival. .................................................................................. 53 Table 3.1: Stream survey of fish habitat variables for post-impact period July 2008, the 10th year means are contrasted with mean values from pre-impact surveys assessed in June 1997. Signs indicate: (-) relative loss, (+) relative gain; and percentage (%) change from initial preharvest measures. ..................................................................................................................... 91 Table 3.2: Mean weather variables recorded at local climate station (Haney, Malcolm Knapp Research Forest) for the summer period (July 1 to August 31) 2006 to 2009. ........................... 92 Table 3.3: Stream temperature and flow variables (± 1 SE) collected during the summer period (July1 to August 31) at weir stations on East Creek and Stream A during the summer of 2008, and temperatures recorded by stream loggers for the summer period of 2009. Columns are organised by control first (East and Spring Creek) then logged treatments (Stream A and C) (Weir data courtesy of M. Feller, Department of Forest Sciences, UBC). ................................. 93  vii  List of figures Figure 1.1: Map of Canada and the study location at Maple Ridge (solid star), southwestern BC, and sites within Malcolm Knapp Research forest........................................................................ 7 Figure 2.1: Schematic representation of robust sampling design (based on Pollock 1982) for sampling cutthroat trout in seven small streams. Primary periods were one week, and within primaries, secondary bouts were one day each and trout were captured each day for 3 to 4 days (adapted from Krebs 1999). Intervals between primaries were 3 to 5.5 months. All fish were released in each bout. CJS¹ (Cormack-Jolly-Seber). ................................................................. 54 Figure 2.2: Density (estimated population size/m² ± 1 SE) of trout by stream size per season. Streams increase in size (i.e. residual depth) from left to right along x-axis. ............................. 55 Figure 2.3: Mean Fulton‟s condition index (FCI ±1 SE) for trout over four primary sampling periods. Streams increase in size from left to right along x-axis. The dashed line represents an FCI of 1 and is displayed for reference between seasons. For sample sizes see Appendix A. ... 56 Figure 2.4: Geometric mean (± 1 SE) of biomass of trout for each primary sample period and stream. Streams increase in size (measured as mean residual depth) from left to right along the xaxis. ......................................................................................................................................... 57 Figure 2.5: Specific growth rate in mass (grams) per season against cutthroat trout population densities (fish/m²) in seven small streams of Malcolm Knapp Research Forest. ........................ 58 Figure 2.6: Probability of recapture (± 1 SE) calculated from model averaged outputs. Since the best model held recapture as P (stream X winter), there are not separate probabilities for spring and summer (i.e. they are constant); whereas recapture was more influenced by winter and is time-dependant. Stream names are ordered by increasing size. ................................................ 59 Figure 2.7: Mean monthly survival estimates (± 1 SE) from program MARK (see Table 2.3 for models) for cutthroat trout for three seasons (summer, winter and spring; 2008-2009) against streams arranged in order of increasing size (i.e. predicted residual depth in habitats at base flows) as the predictor for survival parameters. Non-linear regression is plotted for winter and summer periods ........................................................................................................................ 60 Figure 2.8: Regression of initial mass and body condition (FCI) against seasonal survival rates for cutthroat trout in seven small streams of coastal BC............................................................ 61 Figure 3.1: Stream temperatures (Celsius) from weir stations downstream of sample sites for East Creek (control) and Stream A (log treatment) for the period May1, 2008 to July 1, 2009. . 94 Figure 3.2: Summer relative abundance (CPUE/m sample length ± 1 SE) of age 1+ year cutthroat trout per stream for two unlogged (control) and two logged treatments for BACI design; before-impact sample period (1997 and summer 1998), period of impact (winter 1998), and two after-impacts sample periods i.e., 2000 – 2002 inclusive, and 2008 – 2009.................. 95 viii  Figure 3.3: Mean summer relative abundance (CPUE/m ± 1SE) of age + 1 and older cutthroat trout in control streams and logged streams for the before-impact period (1997 and 1998) and two after-impact periods (2000-2002 inclusively, 2008 and 2009). Some standard error bars are hidden by means symbols. Control-before n = 4; control-four n = 6; control-ten n = 4; Loggedbefore n = 4; logged-four n = 6; logged-ten n = 4 (Ten-Stream C grey symbol * Stream A is removed, n =2). ........................................................................................................................ 96 Figure 3.4: Winter relative abundance of cutthroat trout in treatments: Streams A and C, Creeks Spring and East, during (winter 1998) and after logging (1999 – 2009). Data were not available for streams in 1997 or for Spring Creek in the winters of 1998-1999. For all streams and years only single winter samplings occurred (note break between 2002 and 2009 years). .................. 97 Figure 3.5: Least-squared-mean (LSM) summer body condition (± 1 SE) (i.e. log mass-log length) of cutthroat trout in two unlogged (control) streams and one logged stream before (19971998) and after (10th year; 2008 - 2009) logging impact in Malcolm Knapp Research forest. Note that symbols hide some error bars. ................................................................................... 98  ix  Acknowledgements I would like to thank my supervisor, Dr. John Richardson, for his long running interest in my academic pursuits, and for his mentoring of my scientific education from the undergraduate to the graduate level. I would also like to thank my committee, Drs. Scott Hinch and Tom Sullivan, for their academic support and words of wisdom in the planning, logistics, and the final delivery of my thesis; and a special acknowledgement to Dr. Michael Feller for freely providing his stream data. Thank you to Helene Boulanger for her friendship and excellent field work throughout my project - her loyalty and support were invaluable. Immeasurable thanks to my entire field crew for their considerable efforts and keen interest in this study. Without their enthusiasm in rain, snow, sleet and shine this project would not have been possible. A special thanks to Andrew Lotto for his training and supervision of my first trout tagging, for his knowledge of the study streams and happy disposition throughout our winter samplings; and to the staff at Malcolm Knapp Research Forest for their help, and interest in the little trout of the forest‟s streams. Many thanks to my friends and colleagues in the Richardson Lab for their help in the field, advise, long talks, many coffees, hugs and friendship. Funding for this research project was provided by Forest Investment Account (FIA), and I am very grateful for the financial assistance provided by entrance and internal scholarships, and NSERC. A special thanks to my partner, Patricia, for your love and patience as you stood beside me at either computer or fish scales, and to my parents for always believing in me and celebrating my sense of adventure. To the 1940 trout who participated in this study ... may the force be with you. x  Co-authorship statement This thesis is a component of a multi-disciplinary research program investigating the impacts of forestry and climate on stream trout populations. The sampling design was initiated by me, as was the analysis of the data, and writing of the chapters. Throughout this process Dr John S. Richardson gave considerable guidance and contributed to defining study objectives for the population survival study (Chapter 2). The broader objectives involving forestry effects (Chapter 3) on stream trout were initiated by Drs. John S. Richardson and Scott G. Hinch, with historic sampling and data analysis completed by Jenifer De Groot. Accordingly, these participants are listed as co-authors for the manuscript chapters. Ultimately, responsibility for the quality of the data analyses and writing was my own. Chapter 2:  Season-specific survival and growth rates of coastal cutthroat trout (Oncorhynchus clarkii clarkii) across a gradient of stream sizes in southwestern British Columbia.  Authors:  Kim A. Sheldon and John S. Richardson  Status:  Manuscript to be submitted (journal to be selected)  Comments  This study was conducted, analysed and written by KAS under the supervision of JSR. JSR provided conceptual guidance, intellectual feedback on analyses and written documents.  Chapter 3:  A before-after study contrasting the effects of streamside harvesting on cutthroat trout populations.  Authors:  Kim A. Sheldon, John S. Richardson, Jenifer D. De Groot, and Scott G. Hinch  Status:  Manuscript to be submitted (journal to be selected)  Comments  This study was conducted, analysed and written by KAS under the supervision of JSR and SGH. JSR provided conceptual guidance, intellectual feedback on analyses and written documents; SGH provided historic references to the study design and conceptual input for intended publication; JDG completed the first portion of the sampling and data analyses for this on-going study. xi  Chapter 1 : Season-specific survival and growth rates of coastal cutthroat trout (Oncorhynchus clarkii clarkii) across a gradient of stream sizes in southwestern British Columbia Within North America, and globally, freshwater ecosystems are the most imperilled natural systems on earth (Dudgeon et al. 2006; Ormerod 2009) with extinction rates of freshwater fishes estimated to be five times higher than in terrestrial fauna (Ricciardi and Rasmussen 1991; Malmqvist and Rundle 2002). In the past, threats to freshwater ecosystems have largely been attributed to anthropogenic activities such as land cover alteration, water pollution, invasion of introduced species, habitat destruction, over-harvesting of species, channelization and water extraction (e.g., Fausch et al. 2006; Williams et al. 2009). Currently, due to the accelerated change in world climate (IPCC 2007), shifting precipitation regimes and air temperature patterns may add to the multiple anthropogenic stressors already threatening and degrading freshwater ecosystems (e.g., Schindler 2001; Williams et al. 2009). Of special concern are small streams which are highly integrated with their landscape features and local climates (Gomi et al. 2002), making them sensitive to shifts in precipitation patterns and vulnerable to physical alteration, which can shift hydrological regimes. In turn, aquatic fauna may be threatened by stream disturbances due to habitat and connectivity loss, especially during periods of reduced surface flows. Changes in hydrological resources may cause extinctions or genetic isolation in local populations through the effects of loss of genetic exchange between adjacent populations, or via lack of seasonal habitat supply and increased inter- and intra-species competition as population ranges are compressed. For example, in a recent paper assessing the persistence of tree sub-species of cutthroat trout (O. c. lewisi (Girard 1856), O. c. Utah (Suckley 1874), and O. c. pleuriticus (Cope 1872)) in mountain streams, Williams et al. (2009) predicted that inland and Northern Rocky Mountain populations of trout 1  will be at increasing risk from environmental and demographic stochasticity due to shifts in hydrological regimes; such as, earlier snow-melts increasing the intensity and timing of spring spates, with negative effects during spawning; increases in frequency of wildfire due to higher summer air temperatures; and genetic isolation associated with loss of connectivity to cohorts and sub-populations. Within North America these current and future environmental issues create a challenge for policy makers and resource managers in our aims to achieve a balance between the use and protection of freshwater resources for present and future generations, and for the conservation of freshwater species (e.g., Lowe et al. 2006; Ormerod et al. 2010). For instance, in the coastal Pacific Northwest (PNW) region, a large percentage of the land is covered with temperate rainforest (Naiman et al. 2000); integrated into these forested ecosystems are stream networks which supply aquatic habitat and important energy subsidies for downstream productivity and fisheries (Wipfli et al. 2007) e.g., Pacific salmon populations (Wipfli and Gregovich 2002). It has been estimated that small streams comprise over 70% of the total channel length in freshwater networks in the PNW (Benda et al. 1992). A characteristic of small streams in this region is the high bankfull width to bank length ratio between the terrestrial interface and stream (Gomi et al. 2002), where the often dense overhanging canopy of streamside trees contributes important inputs of organic matter, seasonal litterfall and terrestrial insects provided from the terrestrial-riparian area for in-stream energy processes and nutrients (Richardson et al. 2005). Furthermore, the terrestrial area provides shading from radiation, stream bank stability via rooted vegetation, fine sediment and organic matter retention, and large wood inputs important in stream structure and complexity (Gregory et al. 1991; Fausch and Northcote 1992). Consequent to these linkages between the terrestrial-riparian and streams, the removal of streamside vegetation during forest harvesting has been observed to have varying effects on stream ecosystems and their populations (e.g., Mellina and Hinch 2009). 2  Forest harvesting, stream size and concern for cutthroat trout Removal of the riparian forest area during harvesting or large-scale clear-cutting of watersheds has incurred positive and negative responses in fish populations, such as salmonids. Salmonids are common to streams in the PNW and have been observed to have short-term increases in abundance in response to increases in primary consumers when canopy removal has increased solar radiation in streams (e.g., Murphy et al. 1986). However, these short-term benefits following logging may be negated when instream habitat degrades with increased fine sediment inputs, lethal elevation of water temperatures, stream flow alteration, or loss of structural complexity used as habitat by fish and other aquatic organisms. Furthermore, historically small streams in the PNW were not typically avoided during land conversion, or forest harvesting (Gomi et al. 2002; Lee et al. 2004), and degradation of habitat during the last decades of landscape change is one of the primary explanations of the decline of native stream fishes such as cutthroat trout (e.g., Haring and Fausch 2002, ). Coastal cutthroat trout (Oncorhynchus clarkii clarkii (Richardson, 1836)) belong to one of twelve extant-sub-species of cutthroat trout (Oncorhynchus clarkii (Richardson, 1836)) which are native to western North America (Trotter 2008). Coastal cutthroat trout (herein cutthroat trout) has an historical range extending from the lower Eel River in California, northward along the Pacific coastal ecoregion to Prince William Sound, Alaska (Behnke 1992), and inland to the Cascade crest (Trotter 2008). In the southern region of the British Columbia the fry emerge between March and June, and wild adult trout may live to an average age of 3 years (max. 6 yr), maturing during their age 1+ and 2nd years of life (Trotter 2008). Cutthroat trout are frequent inhabitants of low gradient, small streams (< 4% gradient; < 5m bankfull width) in the PNW for part of, or for all of their life history (Rosenfeld et al. 2000). Within the sub-species both resident and migratory life history strategies are exhibited. Unlike migratory-forms of trout that 3  display synchronized population-scale movement between their stream spawning grounds and rivers, lakes or sea environments, stream-resident forms of trout generally complete their entire life cycle in their natal streams (Trotter 2008), often moving only small distance (2 – 100 meters) throughout their life time. Due to the impacts of landcover change which can lead to instream habitat degradation and loss cutthroat are listed as threatened or as a species of concern throughout the PNW region (e.g., threatened species, U.S. Fish & Wildlife Service; blue-listed in British Columbia, COSEWIC; Coastal cutthroat trout symposium 2005). Further to these concerns are the evolutionary consequences of losing isolated populations of resident stream trout which may represent unique genotypes and an important source of genetic diversity in stream fishes (Northcote 1992). Even with recent changes in forest management guidelines (e.g., FRPA 2004) the degree of protection received by small streams in regards to limitations on riparian harvesting within forested watersheds is still ambiguous, with responsibility for small stream protection in many regions (e.g., coastal British Columbia, FRPA) dependent upon geo-political jurisdiction and management objectives for harvest volumes (Moore and Richardson 2003). In addition, changes in precipitation regimes or temperature during warmer periods of the year may add to the vulnerably of cutthroat trout in small stream environments. Resident-trout live in dynamic systems with a high degree of spatial and temporal heterogeneity. Seasonal precipitation patterns and land cover interact to define the flow regime of streams; the flow regime in turn defining the aquatic structure and function of a stream and its riparian areas (Poff et al. 2010). The timing and quality of food resources, large wood recruitment for physical habitat, sediment and nutrient supply, and hydrology may interact in producing varying magnitudes of response in abiotic and biotic stream variables. While it is assumed that resident trout existing in these streams are well-adapted to cope with the high 4  variability in conditions which they experience (Carlson and Letcher 2003), stochastic events, for example extreme low-flow periods or disturbances from forest harvesting, may change instream environmental conditions beyond past the levels to which individuals are adapted. Depending on their frequency and severity, these disruptive events may threaten the persistence of local populations of stream-dwelling trout. One of the most vulnerable periods for trout is during periods of low-flow (Harvey et al. 2006). For instance, in coastal BC, low late-summer flows can lead to a reduction in habitat volume, and extended periods of dry, warmer temperatures will further reduce habitat, limiting trout food supply, compressing habitats and decreasing connectivity between seasonal refugia. On a gradient of stream size, this could have a strongly negative effect on trout populations in smaller streams compared to populations in larger streams with predictable and sufficient annual flows (e.g., Gresswell and Hendricks, 2007). Narrower, small streams on the lower range of the size gradient are more strongly influenced by their surrounding landscape than are larger streams (Allan 2004). As such narrower streams may have less predictable or sufficient flows during some portions of the year, and along a gradient of stream size (e.g., 1-5 m bankfull width) resident trout in narrower streams are likely to be more vulnerable to stochastic events than are trout in wider streams (Gresswell et al. 2006; Rosenfeld et al. 2000, Cunjak 1996). Given the concern for cutthroat trout, a greater knowledge of their population demographics and interactions with the spatio-temporal heterogeneity of stream habitat templates is required to inform fisheries and forest managers for the conservation of this species (e.g., Southwood 1977; Schneider 1994). Studying populations of resident cutthroat trout across a gradient of stream size may help us isolate some of the physical factors affecting their persistence within the landscape.  5  Study overview The above concerns for stream-dwelling populations of trout in the PNW are considered in my study (Figure 1.1). To address some of our knowledge gaps for stream trout, the main objective of my second chapter was to quantify the season-specific survival rates of resident populations of cutthroat trout across a gradient of stream sizes. This would allow us a greater predictive ability to assess the risk of trout populations to local extinction in the event of extreme hydrological variation in small streams of BC. Secondly, these observations would give us greater knowledge of the influences of summer versus winter conditions on trout survival rates. In my third chapter, I report on the 10th year measurements of a long-term riparian manipulation experiment which examined whether removal of second-growth forest had incurred post-logging lag effects on the abundance and body condition of trout in treatment streams. My 10th year observations were contrasted and discussed in relation to pre-treatment and four years postharvest observations reported in De Groot et al. (2007).  6  Malcolm Knapp Research Forest Weir Perched culvert  Stream C  Stream A  Sample sites Scale 1: 32000 meters  Figure 1.1: Map of Canada and the study location at Maple Ridge (solid star), southwestern BC, and sites within Malcolm Knapp Research forest. 7  Literature cited Allan, J.D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems. Annual Review of Ecology and Evolution Systematics 35: 257-284.  Behnke, R. J. 1992. Native trout of western North America. American Fisheries Society Monograph 6:275.  Benda, L., T. J. Beechie, R. C. Wissmar, and A. Johnson. 1992. Morphology and evolution of salmonid habitats in a recently deglaciated river catchment, Washington State, USA. Canadian Journal of Fisheries and Aquatic Sciences 79:1246–1256.  Bilby, R. E., and P.A. Bisson. 1992. Allochthonous organic matter contributions to the trophic support of fish populations in clear-cut and old-growth streams. Canadian Journal of Fisheries and Aquatic Sciences 49: 540-551.  Boss, S.M., and J.S. Richardson. 2002. Effects of food and cover on the growth, survival, and movement of cutthroat trout (Oncorhynchus clarkii) in coastal streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 1044-1053.  Carlson, S. M., and B. H. Letcher. 2003. Variation in brook and brown trout survival within and among seasons, species, and age classes. Journal of Fish Biology 63: 780–794.  Coastal cutthroat trout symposium. 2005. http://www.fws.gov/columbiariver/cctsym/cctsymppt.htm. Accessed Nov 20, 2009.  Cole, M. B., K. R. Russell, and T. J. Mabee. 2003. Relation of headwater macroinvertebrate communities to in-stream and adjacent stand characteristics in managed second-growth forests of the Oregon Coast Range mountains. Canadian Journal of Forest Resources 33: 1433-1443.  Committee on the status of endangered wildlife in Canada (COSEWIC). http://www.cosewic.gc.ca/. Accessed Jan 21, 2010.  Cunjak, R. A. 1996. Winter habitat of selected stream fishes and potential impacts from landuse activity. Canadian Journal of Fisheries and Aquatic Sciences 53:267-282.  8  De Groot, J. D. 2004. Density, body condition, and movement of coastal cutthroat trout (Oncorhynchus clarkii clarkii) in logged and forested headwater streams of southwestern British Columbia. University of British Columbia press, MSc., thesis.  De Groot, J. D., S.G. Hinch, and J.S. Richardson. 2007. Effects of logging second growth forests on headwater populations of coastal cutthroat trout: a 6-year, multi-stream, before-and-after field experiment. Transactions of the American Fisheries Society 136: 211-226. Dudgeon D., A.H. Arthington, M.O. Gessner, Z. Kawabata, D.J. Knowler, C. Le´veˆque, R. J. Naiman, A. Prieur-Richard, D. Soto, M. L. J. Stiassny, and C. A. Sullivan. 2006. Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews 81: 163–182.  Fausch, K. D., and T. G. Northcote. 1992. Large wood debris and salmonid habitat in a small coastal BC stream. Canadian Journal of Fisheries and Aquatic Sciences 49: 682-693.  Fausch, K. D., B. E. Rieman, M. K. Young, and J. B. Dunham. 2006. Strategies for conserving native salmonid populations at risk from non-native fish invasions: tradeoffs in using barriers to upstream movement. General technical report RMS-GTR-174. U.S. Department of Agriculture Forest Service, Fort Collins, Colorado.  Fausch, K. D., C. E. Torgersen, C. V. Baxter, and H. W. Li. 2002. Landscapes to Riverscapes: Bridging the Gap between Research and Conservation of Stream Fishes BioScience 52: 483-498.  Forest and Range Practices Act (FRPA). 2004. http://www.for.gov.bc.ca/code/. Accessed Jan 21, 2010.  Gomi, T., R. C. Sidle, and J. S. Richardson. 2002. Understanding processes and downstream linkages of headwater systems. BioScience 52: 905–916.  Gregory, S.V., F.J. Swanson, W.A. McKee, and K.W. Cummins, 1991. An Ecosystem Perspective of Riparian Zones. BioScience 41:540-551.  Gresswell, R.E., and S.R. Hendricks. 2007. Population-scale movement of coastal cutthroat trout in a naturally isolated stream network. Transactions of the American Fisheries Society 136: 238-253. 9  Gresswell, R.E., C.E.Torgersen, D.S.Bateman, T.J. Guy, S.R. Hendricks, and J.E.B. Wofford. 2006. A spatially explicit approach for evaluating relationships among coastal cutthroat trout, habitat, and disturbance in small Oregon streams. American Fisheries Society Symposium 48: 457–471.  Harvey, B.C., J. L. White, and R. J. Nakamoto. 2005. Habitat-specific biomass, survival, and growth of rainbow trout (Oncorhynchus mykiss) during summer in a small coastal stream. Canadian Journal of Fisheries and Aquatic Sciences 62: 650-658.  Harvey, B.C., R. J. Nakamoto, and J. L. White. 2006. Reduced streamflow lowers dry-season growth of rainbow trout in a small stream. Transactions of the American Fisheries Society 135: 998–1005.  Holtby, L. B. 1988. Effects of logging on stream temperatures in Carnation Creek, British Columbia, and associated impacts on the coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Sciences 45: 502-515.  Intergovernmental Panel on Climate Change (IPCC). 2007. Climate Change: Synthesis Report. http://www.ipcc.ch/publications_and_data/publications_ipcc_fourth_assessment_report_s ynthesis_report.htm. Accessed Jan 21, 2010.  Lee, P., Smyth, C. and S. Boutin. 2004. Quantitative review of riparian buffer width guidelines from Canada and the United States. Journal of Environmental Management 70: 165-180.  Lowe, W. H., G. E. Likens, and M. E. Power. 2006. Linking Scales in Stream Ecology. Bioscience 56:591-597.  Malmqvist, B., and S. Rundle. 2002. Threats to the running water ecosystems of the world. Environmental Conservation 29: 134-153.  Mellina, E., and S. G. Hinch. 2009. Influences of riparian logging and in-stream large wood removal on pool habitat and salmonid density and biomass: a meta-analysis. Canadian Journal of Forest Research 39: 1280-1301.  Mellina, E., S.G.Hinch, and K.D. MacKenzie. 2005. Seasonal movement patterns of streamdwelling rainbow trout in north-central British Columbia, Canada. Transactions of the American Fisheries Society 134: 1021–1037.  10  Moore, R.D., and J.S. Richardson. 2003. Progress towards understanding the structure, function, and ecological significance of small stream channels and their riparian zones. Canadian Journal of Forest Research 33:1349-1351.  Murphy K. M., C. P Hawkins, and N.H. Anderson. 1981. Effect of canopy modification and accumulated sediment on [small] stream communities. Transactions of the American Fisheries Society 110: 469-478.  Murphy, M. L., J. Heifetz, S. W. Johnson, K. V. Koski, and J. F. Thedinga. 1986. Effects of clear-cut logging with and without buffer strips on juvenile salmonids in Alaskan streams. Canadian Journal of Fisheries and Aquatic Sciences 43:1521–1533.  Naiman, R.J., Bilby, R.E., and P.A. Bisson. 2000. Riparian ecology and management in the Pacific Coastal Rain Forest. BioScience 50: 996-1011  Northcote, T. G., and G. F. Hartman, editors. 2004. Fishes and forestry: worldwide watershed interactions and management. Blackwell Scientific Publications, Oxford, UK.  Northcote, T.G. 1992. Migration and residency in stream salmonids - some ecological considerations and evolutionary consequences. Nordic Journal of Freshwater Research 67: 5-17.  Ormerod, S. J. 2009. Climate change, river conservation and the adaptation challenge. Aquatic Conservation: Marine and Freshwater Ecosystems 19: 609–613.  Ormerod, S. J., M. Dobson, A. G. Hildrew, and C. R. Townsend. 2010. Multiple stressors in freshwater ecosystems. Freshwater Biology 55: 1-4.  Poff, L., B.D. Richter, , Arthington, A. H., Bunn, S.E., Naiman, R. J., Kendy, E., Acreman, M., Apse, M., Bledsoe, B. P., Freeman, M.C., Henriksen, J., Jacobson, R.B., Kennen, J. C., Merritt, D.M., O'Keeffe, J.H., Olden, J.D., Rogers, K., R.E. Tharme, and A. Warner. 2010. The ecological limits of hydrologic alteration (ELOHA): a new framework for developing regional environmental flow standards. Freshwater Biology 55:147-170.  Ricciardi, A., and J. B. Rasmussen. 1991. Extinction Rates of North American Freshwater Fauna. Conservation Biology 13: 1220-1222.  11  Richardson, J.S., R. J. Naiman, F. J. Swanson, and D. E. Hibbs. 2005. Riparian communities associated with Pacific Northwest headwater streams: assemblages, processes, and uniqueness. Journal of the American Water Resources Association 41: 935-947.  Rosenfeld, J., M. Porter, and E. Parkinson. 2000. Habitat factors affecting the abundance and distribution of juvenile cutthroat trout (Oncorhynchus clarkii) and coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science 57: 766-774.  Schindler, D. W. 2001. The cumulative effects of climate warming and other human stresses on Canadian freshwaters in the new millennium. Canadian Journal of Fisheries and Aquatic Sciences 58: 18-29.  Thedinga, J. F., M.L. Murphy, J. Heifetz, K.V. Koski, and S.W. Johnson. 1989. Effects of logging on size and age composition of juvenile coho salmon (Oncorhynchus kisutch) and density of pre-smolts in southeast Alaska streams. Canadian Journal of Fisheries and Aquatic Sciences 46: 1383-1391. Trotter, P. 2008. Cutthroat trout: native trout of the west 2nd Edition. University of California Press, Ca.  U.S. Fish & Wildlife Service (USFWS).http://www.fws.gov/endangered/wildlife.html. Accessed April 21, 2010.  White, J. L., and B. C. Harvey. 2007. Winter feeding success of stream trout under different stream flow and turbidity conditions. Transactions of the American Fisheries Society 136: 1187–1192. Williams, J. E., A. L. Haak, H. M. Neville, and W. T. Colyer. 2009. Potential Consequences of Climate Change to Persistence of Cutthroat Trout Populations. North American Journal of Fisheries Management 29: 533-548.  Wipfli, M.S. and D.P. Gregovich. 2002. Invertebrates and detritus export from fishless headwater streams in Southeast Alaska: Implications for downstream salmonid production. Freshwater Biology 47: 957-970.  Wipfli, M.S., J.S. Richardson, and R.J. Naiman. 2007. Ecological linkages between headwaters and downstream ecosystems: transport of organic matter, invertebrates, and wood down headwater channels. Journal of the American Water Resources Association 43. 12  Young, K. A. 2000. Riparian zone management in the Pacific Northwest: who‟s cutting what? Environmental Management 26: 131-144.  13  Chapter 2 : Season-specific survival rates and physical characteristics of resident cutthroat trout in small streams of coastal British Columbia1. Introduction An aspect of population ecology is to predict how environmental conditions can affect the dynamics of animal populations, and to what extent their abundances may differ spatially and temporally across their range (Brown 1984; Brown et al. 1995). Population abundance can be affected by the underlying mechanisms of recruitment, survival (or mortality), immigration and emigration acting within a population, and physical factors such as availability of favourable habitat (Elliott 1987; Krebs 2001). As such, contrasting and quantifying demographic parameters such as survival rates among populations, across a gradient of environmental conditions, may give us the ability to predict a population‟s viability and abundance relative to specific environmental factors (Hawkins et al. 1993, Poff et al. 1997). For instance, in coastal temperate regions, streams can be dynamic environments which have high variability in the habitat provided for stream-dwelling populations (Heggenes 2002). In particular, in regions such as the PNW, small streams which are tightly coupled with their surrounding watersheds can have a high degree of hydrological and morphological variability, creating spatially and temporally heterogeneous habitats for stream organisms (Poff and Allan 1995; Heggenes 2002; Winemiller et al. 2010). Stream ecology theory suggests that these spatial and temporal shifts in hydrological habitat may exert strong selective pressures on stream organisms, limiting their populations through the supply of favourable habitats (Chapman 1966;  1  A version of this chapter will be submitted for publication by Sheldon KA and JS Richardson. Titled as: Seasonspecific survival rates and physical characteristics of resident cutthroat trout in small streams of coastal British Columbia.  14  Elliott 1987; Schlosser et al. 1985; Power et al. 1988 Resh et al. 1988; Poff and Allen 1995; Quist and Hubert 1995; Heggenes et al. 1991; Gresswell and Hendricks 2007). Subsequently, to quantify how aquatic populations may be limited by their physical habitat, the heterogeneous environment of small streams provides a suitable template for contrasting the relative influences of physical factors on survival demographics among stream populations of the same species. Stream-dwelling salmonids, for example, are an important group of fishes which present a model opportunity to examine how variability in stream environments limits their populations and may predict their persistence. For instance, stream-resident trout often occur in small streams of the PNW region, tending to live within a short distance of a stream reach for their entire lives (e.g., 20 m², Heggenes 2002; > 100 m, Gresswell and Hendricks 2007), and because of their high site fidelity to natal tributaries are immediately affected by local changes in environment (Reeves et al. 1997; Behnke 2002). Furthermore, stream trout compete for space and are considered to be governed in part by habitat supply (Chapman 1966), where limited supply of preferred or favourable habitat may induce changes in trout numbers via demographic mechanisms (Elliott 1973, 1987). In addition, headwater populations often exist in allopatry, removed from competition from other fish species, and making identification of the physical factors affecting population abundance less complex (Gowen and Fausch 1996). Further to increasing our understanding of mechanisms limiting the abundance of stream species, there is currently substantial uncertainty in predictions of how low-flow conditions affect stream populations. If as predicted, climatic patterns vary beyond conditions to which populations have adapted; resource and conservation managers will need a greater knowledge of population responses to seasonal stream conditions (Bradford and Heinonen 2008). Recent studies on inland trout populations in NA have indicated that changes in climate are already having serious consequence for trout conservation in smaller streams (e.g., Williams 15  et al. 2009). The recent occurrence of early snow melts has led to winter floods, and drier summer conditions, increasing drought conditions and the frequency of wildfires in many regions of NA (Williams et al. 2009). Furthermore, a recent review on Canadian water resources concluded there is a large degree of uncertainty in predicting how stream biota will respond to site- and time-specific variation in habitat change during low flow periods (Bradford and Heinonen 2008). For example, within the PNW climatic change may lead to higher summer temperatures and lower precipitation levels, and relative to larger streams, small streams may be vulnerable to altered precipitation patterns and shifts in watershed hydrology (Gomi et al. 2002; Allan 2004). In particular, during hot, dry periods the compounding effect of increased evaporation through increased air temperatures and reduced base flows may shrink residual aquatic habitats (Harvey et al. 2005; Williams et al. 2009). Possibly, larger streams may have a greater buffering capacity to flow alteration than smaller streams with trout in larger streams having a greater supply of habitat during drier, low-water periods and therefore potentially having higher survival probabilities. Although it is assumed that trout inhabiting stream environments are locally adapted to survive environmental stochasticity (Heggenes 2002), trait selection for adaption to seasonal variation in habitat availability may not be suited to all conditions experienced by trout among seasons (Carslon and Letcher 2003). Selective mortality or survival within a local environment may be related to an individual‟s traits (Warren and Liss 1980), with the survival of individuals being reflected at a population level. Therefore, mark-recapture techniques which use uniquely marked individuals will allow us to estimate survival parameters at an individual-level and to infer how biological traits, such as body condition, may determine population-level survival rates in relation to habitat heterogeneity (e.g., Carlson and Letcher 2003; Berger and Gresswell 2009). Moreover, studies using individually marked fish to estimate survival probabilities over multiple 16  spatial and temporal scales are rare, with few studies having incorporated the spatial scale among streams into survivals models (e.g., Olsen and Vollestad 2001; Carlson and Letcher 2003; Berger and Gresswell 2009). Therefore, combining capture-mark-recapture (CMR) techniques using individually marked trout to contrast season-specific survival rates across a gradient of stream sizes would give us comprehensive knowledge of how spatio-temporal factors affect trout demographics in small streams (e.g., trout and pond size, Elliott 1997). This study examined seasonal variation in survival rates of coastal cutthroat trout across a size gradient of smaller streams in coastal BC. Based on the prediction that the streams with the narrowest bankfull widths would have the lowest availability of habitat during the summer lowflow period, I tested the hypotheses that summer survival rates would be lower relative to other seasons, and that survival would be positively related to stream size. The among-stream comparison was predicted to show that due to reduced habitat in the summer period, populations in narrower streams were more vulnerable to low-flow conditions than populations in wider streams.  17  Methods To test my hypotheses I uniquely marked individual trout across a range of seven different sized small streams (1st and 2nd order; Strahler 1952), using a robust mark-recapture design stratified seasonally to quantify the season-specific survival rates for each population. In addition, I used trout characteristics as covariates in survival models to determine how individual variation was reflected in population level survival rates. Study area The study area was located at the University of British Columbia‟s Malcolm Knapp Research Forest (MKRF), close to the town of Maple Ridge in south-western BC (122° 34‟ W, 49° 16‟ N) approximately 40 km from Vancouver. The climate for the region is cool mesothermal, characterized by warm, dry summers and wet, mild winters. The MKRF covers an area of approximately 5170 ha and lies within the Coastal Western Hemlock (CWH) biogeoclimatic zone (Meidinger and Pojar 1991). The dominant forest cover is a matrix of old growth (>250 years) and second- growth (40-80 years) Douglas-fir (Pseudotsuga menziesii, (Mirb.) Franco), western hemlock (Tsuga heterophylla (Raf.) Sarg.), and western red cedar (Thuja plicata, Donn). Secondary canopy layers within the riparian zone of study streams were dominated by red alder (Alnus rubra, Bong), big leaf maple (Acer macrophyllum Pursh), vine maple (Acer circinatum Pursh), salal (Gaultheria shallon Pursh) and several Vaccinium species associated with the soil nutrient and moisture characteristics of the very moist CWH zone, with soils composed of glacial till and glacio-marine deposits (Feller and Kimmins 1979). Gradient of stream sizes Seven first and second order streams with only resident populations of coastal cutthroat trout were selected to obtain a gradient of stream sizes based on increasing bankfull channel 18  width (BFW; defined in Johnson and Slaney 1996), and were sampled over a period of 13 months in the years of 2008 and 2009 (Table 2.1). Originally nine streams had being selected, however, two streams, Donegani Creek and Stream A, were not included in these results because Donegani Creek was found to have a migratory population of rainbow trout, and no trout were captured or observed in Stream A at any period during the study (see Chapter 3 on Stream A). The hydrology of the streams were rain-dominated with some snow-influence between the winter months (December to March) in streams at higher elevations (i.e. Blaney, Upper Spring and Upper East Creeks), and study reaches were either riffle-pool or step-pool morphologies (Montgomery and Buffington 1997), ranging from 3.7 m to 9.5 m bankfull channel width with sample reaches of 200 m to 436 m in length (Table 2.1). Stream habitat surveys Stream assessments were conducted at different periods throughout the study. Initially sample reaches were assessed for trout presence and physical stream characteristics in May 2008. Habitats where trout would be trapped (henceforth „habitat or trap units‟) were identified following the minimum size criteria in Johnson and Slaney (1996) where a trap unit was considered a separate unit if measures of bankfull channel width and minimum area (ma) met the following criterion, i.e., 0 - 2.5 m BFW, ma = 1 m²; 2.5 - 5.0 m BFW, ma = 2 m². If habitats were smaller than the criteria they were combined with an adjacent habitat unit that met the size criterion. Depth was also included in the criterion because traps were ineffective in less than 15 cm of water. Trap units were numbered at streamside or in overhanging trees along each stream reach with the distances between consecutive units recorded. In order to have comparable trapping effort across all seven streams approximately 45 trap units were identified for each stream, and sample reaches therefore differed in length because some streams had less habitat  19  available for trapping than others, and more stream length was required to set 45 traps (Table 2.1; Appendix A). The physical characteristics of each trap unit and sample reach were assessed based on an adaption of Johnson and Slaney‟s (1996) Level One Stream Habitat Assessment and Moore (2003). Gradient was measured along the sample reach and between trap units using a handheld clinometer and 1.5 meter rod; for each unit percentage riparian canopy cover was visually estimated and percentage composition of substrate characteristics were classified by size classes (i.e. sand < 2 mm, gravel 2-64 mm, cobble 65-256 mm, boulders 257-4000 mm, and bedrock > 4000); sample reach lengths, habitat widths and lengths, bankfull widths and bank depths were measured to 1 cm using a 30-meter tape; and habitat (trap unit) water depths were measured using meter sticks. Large wood pieces (>2-m length and 10-cm diameter) which influenced stream morphology and provide cover or shelter for fish were counted within each unit. Maximum and crest depths for each unit were measured during all sampling bouts and during the low-flow summer period when trapping did not occur (i.e. July). In addition, during the lowflow summer period each stream reach was visually assessed for habitat connectivity, i.e., a habitat was considered isolated if no surface flow connected it to another habitat unit. During the winter, spring and summer periods of 2009 one Hobo © temperature logger was deployed in a pool habitat in each stream to measure daily temperatures during the winter, spring and summer periods. Stream current velocities were measured during the spring and fall sampling periods when flows tended to be highest due to seasonal precipitation. Measurements were made before and after trapping using a portable Swoffer (model 2100) velocity meter. In general, stream velocities were < 0.2 meters/second (m/s) during samplings, except during three spring bouts when velocities were > 0. 6 m/s. Trapping did not occur at these times because  20  velocities exceeded trout preferences for feeding and fish were likely to be sheltering from higher flows (Heggenes et al. 1991). Trout sampling design and tagging To test my survival hypothesis, I used a robust capture-recapture design (Pollock 1982) which is a combination of the Cormack-Jolly-Seber (CJS) open population model for estimating survival probabilities, and the CAPTURE model for closed population size estimates. An advantage of using this mark-recapture sample design is that it is robust to unequal catchability issues inherent in marked animal population estimate studies (Krebs 1999). Trout populations were sampled across the gradient of streams, with samplings stratified into four primary sample occasions: (1) spring, May and June 2008; (2) late summer, September 2008; (3) winter, February and March 2009; and (4) 2nd spring, May and June 2009 (Figure 2.1). Trout were captured using baited (brined salmon roe) Gee-type traps with a standard mouth size of approximately 2.5 cm diameter. All cages were numbered to help reference trout to their habitat units and were placed along the stream length within each numbered unit. To estimate emigration of marked trout from the defined trap area, several fish cages were also placed outside of the study reaches (up to 30 m) during the final primary sampling session. The passive capture method of using Gee-type traps was selected in favour of electrofishing due to possible trout mortality associated with repeated electro-shocking, and the low conductivity levels in streams (Young et al. 1999). Low conductivity in stream water can lead to shock-induced mortality due to the high amperage levels required for stunning (pers. comm. S.G. Hinch, UBC, 2008). Furthermore, Gee-type traps had comparable capture efficiencies to electro-fishing in our study area (Young et al. 1999). However, passive trapping is not as efficient for young-of-year fish (YOY, < 4.5 cm) (Bloom 1976), therefore trout below 4.5 cm length were not included in survival or recapture estimates. 21  It has been observed that temperature affects trout feeding and movement rates, with quicker capture rates in warmer periods relative to cold, winter periods (e.g., Elliott 1973). Temperature-dependant behavioural heterogeneity was accounted for by applying a 3-hour trapping period for spring and summer samplings, and a 24-hour, overnight trap period for the winter samplings. The 3-hour period took advantage of the higher activity of trout in the warmer periods, when capture rates were high and longer trap periods did not increase catch-per-uniteffort (CPUE, De Groot et al. 2007). As well, the shorter interval reduced incidences of trapinduced cannibalism from trout interactions and cage interference by bears. For the winter period, the 24-hour period accounted for the greatly reduced movement rates, and overnight feeding behaviour of trout during the cold (< 4°C) winter months (e.g., Heggenes et al. 1993). Captured trout were anaesthetized with buffered (sodium bicarbonate) pre-mixed tricaine methanesulfonate (MS-222, 0.1 g per litre), weighed to the nearest 0.1 g, and fork length (FL) was measured to the nearest mm. Trout ≥ 5.5 cm FL were injected using a hypodermic needle with a passive integrated transponder (PIT) tag into their peritoneal cavity (i.e., body cavity). Trout < 5.5 cm fork length were marked with visible implant elastomer (VIE) dye where skin pigment was lightest on their lower sides or stomach. The individual characteristics of each trout, its unique identifying number or dye code, and habitat/trap unit were recorded. All trout were scanned with a portable digital Biomark® tag reader to identify tagged trout, and all trout were examined for dye marks or tagging scars. If a trout had a tagging scar yet no tag was detected, then the tag loss was recorded and the trout was then retagged. In some cases it was possible to match trout to lost tags due to the age of their scar, their individual measurements and their association with a particular habitat unit. Recaptured trout which had initially been dyed were tagged with a PIT tag if they had grown beyond 5.5 cm FL, and their dye code and new tag code were used to identify them. After inspection trout were placed into buckets of fresh stream 22  water containing vegetation cover to reduce possible stress until they had fully recuperated from handling. Trout were judged recovered when they were capable of swimming rapidly, i.e., were not sluggish or easily handled. Once recovered, trout were released back to the habitat unit where they were captured.  Data analysis Defining stream size Bankfull channel width (m) was initially used to define stream size, creating the gradient of stream sizes among which survival rates of trout would be compared. In addition, since this study compared the survival rates of trout across low-flow versus higher flow seasonal periods, estimates of residual depth within habitat units was also used as a physical measure of stream size. I reasoned that the depth of water remaining within a habitat unit in drier seasons (e.g., summer low-flows) may be a biologically meaningful measure of stream size when predicting trout survival rates and population demographics. Whereas bankfull channel width may be a less sensitive metric for predicting current trout population dynamics because this metric reflects the hydrological regime occurring during maximum flow events (e.g., winter and spring). Therefore, to determine a hydrologically meaningful measure of stream size, I used a linear regression to test if bankfull width predicted mean residual depth (as a proxy for minimum habitat). Secondly, I used a linear regression to test if mean residual depth of habitat was a stronger predictor of seasonal CPUE, seasonal densities and recapture probability rates compared with mean bankfull width. The better predictor was then used to define stream size for use in all remaining analyses and graphs.  23  Biological variables in trout populations by stream and season For comparison of trapping effort between streams and seasons CPUE was calculated for each secondary session by dividing the number of trout caught within each sample session by the number of traps per trap day. An ANCOVA (analysis of covariance) was used to test for mean differences in CPUE between streams using secondary sample bouts as a covariate to adjust for difference in capture rates within seasons. Trout densities were calculated using Equation 2.1 (below) where the estimated number of trout was derived from population models described further in “Trout population estimates”. Equation 2.1:  Estimated trout density (number/m²) = 𝑁/ (sample reach (m) * mean bankfull width (m)).  Where 𝑁 denotes estimated population size. Fulton‟s Condition Index (FCI; Ricker 1975) for fish was calculated to compare trout body condition between seasons and streams (Equation 2.2). An ANOVA (analysis of variance) was used to test for mean differences in FCI between streams and seasons, and a linear regression was used to examine the relation of trout condition with stream size within seasons. FCI was also used as a covariate in program MARK as a linear constraint on survival estimates to test if seasonal survival was a function of pre-season trout condition (see „Survival model covariates‟).  Equation 2.2:  Fulton‟s condition index K =  mass (g) fork length (mm )3  × 100  Geometric mean masses were calculated to describe trout populations and to examine the relationship between seasonal trout mass (size) and stream size using a linear regression and ANOVA. Linear regressions were used to identify significant biological trends between mean 24  trout geometric mean masses, seasonal densities and growth rates across streams. Mean seasonal growth rates (SGR) in trout mass were estimated using fish recaptured between seasons (Equation 2.3); an ANCOVA (stream * season) with initial mass as a covariate was used to test for differences in mean trout growth rates between seasons and streams.  Equation 2.3: Mean specific growth rate (SGR) =  ln M t − ln M 0 t  × 100  Where Mt is final mass (grams), M0 is the initial mass (grams), and t is the growth period in days (Harvey et al. 2005). Data were sorted and compiled using Excel (Microsoft© Corp. 2007, Redmond, Wa.), statistically analysed in Minitab® (2007, version 15.1) and SAS (SAS version 9.1, SAS Inc., Cary, NC, USA), and graphically displayed using SigmaPlot® (2004, version 9.0, Systat Software© Inc., Chicago, Il) software packages. All data were tested for normality with Anderson-Darling (n<20) or Kolmogorov-Smirnov (n>20), and equal variances (Bartlett and Levene‟s Tests) to meet assumptions of linear regressions and inferential analyses; data not meeting assumptions of normality or equal variance were natural log (ln) or log10 transformed. In addition, the distribution of residuals and probability plots of linearity were examined with each inferential analysis to ensure all data met assumptions of even distribution and spread of residuals. Means are reported as ± one standard error. An alpha level of 0.05 was used to test for statistical significance.  Estimating apparent survival and recapture probabilities Apparent survival probabilities were estimated between primary sampling occasions, when the population was demographically open. Over a period of one year there were four primary sampling occasions with 4, 3, 3, 4 respective secondary closed-capture samplings within 25  each primary, i.e., a total of 14 sample occasions. These four primary sessions bracketed three survival intervals: the seasons of summer 2008, winter 2009, and spring 2009. Individual encounter histories of marked (tagged or dye coded) trout were recorded in 14-cell binomial matrices of zeros or ones, where one indicated a trout was encountered, and zero indicated that a trout was not encountered during a trapping bout. Each individual encounter history of marked trout was attributed a group coding for population defined by stream size, and complied encounter histories for all streams were used as an input file to estimate apparent survival (φ) and recapture (P) probabilities in program MARK (White and Burnham 1999). The term „apparent‟ survival represents the probability a trout is either caught (i.e. known to be alive) within the study area at time + i, or was unavailable for capture due to immigration from the study area (i.e., alive but unseen) at time + i, and therefore apparent survival is an estimate of true survival (herein „survival‟). Program MARK uses the method of maximum likelihood (sensu Fisher 1922), a method which fits models to the observed data set by maximising an explicit likelihood function to estimate the probabilities of survival and recapture rates. Model fit is an optimal balance between precision and fit of the observed data by the model, and for this study was determined using Akaike Information Criterion (AIC; Equation 2.4) adjusted for small-sample bias (corrected, AICc; Anderson and Burnham 2002). AIC uses two components: negative loglikelihood (L) which measures lack of fit to observed data, and a correction factor which penalises the model for using an increasing number of parameters (K) and over-fitting the data to the model. A general interpretation of AICc is that the lower the AICc the better the model optimises a balance between fit and precision for the data relative to other candidate models in the model set.  26  Equation 2.4: AIC = −2 ln L) +2K Where L is the model likelihood and K is the number of parameters in the model (White and Burnham 1999). After each model within the set was fitted to the data, the best model was selected based on the relative Akaike weights of each model when normalised across the set of reduced candidate models in the full data set. Candidate models Following the methods suggested in Lebreton et al. (1992) I used a set of candidate a priori models based on the ecological reasoning underlying my specific hypothesis that stream size (stream + size dependence) and season (time-dependence) influenced survival rates of trout across the gradient of streams sampled. Competing candidate models were nested subsets of the fully parameterised global model which mathematically described how survival and recapture probabilities may differ within and among streams over time. Following the naming convention recommended in Lebreton et al. (1992) the syntax used to describe the fully structured global model was { φ (stream + time + stream x time), P(stream + time + stream x time)}, where (stream) denotes stream populations, and (time) is full-time dependence, i.e., population parameters vary over time. Unequal time intervals between samplings were accounted for in parameter estimates by making each time interval (the survival season) an exponent of each survival or recapture estimate for each related sample period (White and Burnham 1992). This allowed for comparison of survival and recapture probability estimates between periods, even though time intervals between samplings varied. The logit link function (log [x)] – log [1-x]) was used to constrain survival and recapture probability estimates as linear functions of independent factors expressed in each model, i.e., 27  survival of stream populations were constrained by habitat measures and season. Logit estimates were then re-constituted to real parameter estimates with 95% confidence intervals within MARK, and the relationship between survival (dependent variable), and the independent factors of stream size and season were then statistically compared through regression analyses. Model structure and fit While the sampling plan utilised a robust sampling design I did not analyse the data with the robust model structure available in program MARK, i.e., Pollock robust design with Huggins population estimator (Pollock-Huggins; White and Burnham 1999). The global model was initially tested using the Pollock-Huggins model structure, however, model deviances were high and assessment of c was > 53, where < 3 was desired; where c measures variance in model dispersion. Since initial tests indicated that the Pollock-Huggins structure did not fit the data well, it was decided to manually pool the secondary encounter occasions for each primary session (Krebs 1999: see CJS method) to estimate survival rates, and to separately estimate population sizes in program CAPTURE using appropriate closed-capture models. This method reduced model complexity and increased parsimony because computations of immigration, emigration and population size parameters were removed and only survival and recapture estimates were computed using CJS. Therefore, for survival estimates, encounter histories were compiled as 4-cell binominal recapture histories that recorded when a marked trout was last captured over the 3-5 day samplings bouts within each primary session, and analysed using the CJS live-capture open population model. The assumptions of the CJS model are that (i) all animals have the same probabilities ( 𝛼𝑡 ) of being captured in the t-th sample, regardless of whether marked or not, (ii) every animal which is marked had the same probability ( 𝜑𝑡 ) of surviving the interval between trappings from time of capture to sample time +i, (iii) marks used to identify animals are not missed, lost or 28  affecting the animal‟s survival in any manner, (iv) sampling occurs over a short period of time which is negligible relative to the interval between samples (v) animals are independent of each other without influence on catchability (Krebs 1999). This first assumption of equal catchability among marked and unmarked individuals is critical for reliable estimates of survival from the CJS model. The advantage of the robust design used for sampling was that it strengthened estimates from the CJS model because pooling encounter histories from the closed-secondary bouts de-sensitized the CJS model to the variations in catchability within the short time periods when trapping occurred (Krebs 1999). In addition, using the full 14-cell binomial recapture matrix, heterogeneity of capture histories was tested for using TEST 2 and TEST 3 available in RELEASE within program MARK. The second assumption was met since all trout appeared in stable condition when released. The third assumption was met with either most trout being marked with unique PIT tag numbers, or unique pre-defined dye colour codes. Trout who lost tags and whose encounter histories were unidentifiable were removed from the encounter history matrix and percentage tag loss was calculated. Tagging-related mortality was not noted after sampling and tagging bouts; in addition, past studies have shown that survival rates are between 95 - 99% in PIT tagged trout of >52 - 57 mm (FL) respectively (e.g., Acolas et al. 2007). The fourth assumption was met since secondary samplings were intensive and occurred over 3-5 days (see Trout sampling design and tagging in Methods). The fifth assumption of independence among trout was accounted for with a variance inflation factor (ĉ) estimated using the median ĉ approach suggested in MARK. Ĉ corrected for the potential lack of independence among trout data due to hierarchical or communal behaviour reflected in the capture data which may inflate model variances. The general model was tested for goodness-of-fit (GOF) to the CJS model structure using median ĉ and program RELEASE available in MARK. Separate GOF tests where run for: the 429  cell binomial data set (i.e. ĉ = 1.462; survival models), and for the 14-cell binomial data set (i.e. ĉ = 1.218; population size models) these values of ĉ were used to adjust the variance inflation factor for over-dispersion (i.e. a lack of independence so that extra binomial variation occurs) of the global model with the symbol Q („quasi‟) added to reported AIC values. The adjusted global model for survival was then used to build the candidate models where the random effects of time were reduced to specific seasonal periods, i.e., winter, summer or spring. Initially the full survival model was held constant (i.e. stream + time + stream x time) to determine the best model (lowest QAIC) for describing recapture (Lebreton at al. 1992). Then, the best recapture model was held constant to compare survival models from the set of a priori candidate models. Once identified, the best models for survival were then compared using QAIC weights to assess the relative likelihood of each model given the observed data. In addition, I applied likelihood ratio tests (LRT) to nested models to identify the influence of the main effects and interaction terms in model fit and parameter estimates. Model averaging and selection To account for model selection uncertainty due to several competing best models in the model set (i.e. Δ QAICc < 2; Equation 2.5), the method of model averaging (Burnham and Anderson 2002) was used for estimates of survival and recapture probabilities. Model averaging considers the uncertainty in model selection by using the QAICc weights of each model to calculate a weighted average across all candidate models. Averaging is accomplished by taking estimates from each candidate model and weighting them by their relative support in the model set using normalised QAICc weights. An average value with standard errors and 95 % confidence intervals is then calculated for each population parameter (sensu Buckland et al. 1997).  30  Equation 2.5: Average (Φ) =  R i=1  (ωi Φi )  Where ωi reflects the Akaike weight for model i Once estimated, survival rates were used to predict life expectancy of trout in each stream (i.e. 1/–LN [survival probability]) and examine the difference ratio between the contributions of stream size versus season in influencing population survival rates. Biological covariates The initial mass and FCI of all trout at first capture were used as covariates with individual encounter histories to examine if the mass or condition of a trout influenced its survival rate relation to stream or time. These models were run separately from the grouped stream and time models in order that the effects of each covariate could be individually identified using program MARK, i.e., real covariates were used in models to determine if an individual‟s biological attributes influenced their survival. This technique constrained survival to be a function of trout biological characteristics and the physical environment allowing for statistical inferences of trout survival using the slope of the relationship between mass and FCI versus stream and season. Trout population estimates Using individual encounter histories from mark-recapture studies has been widely used for estimating fish population sizes since 1896 (Ricker 1975). As mentioned earlier, population size and densities of cutthroat trout were estimated for each stream and season using the closedsecondary sampling session histories and model selection procedures available in program CAPTURE (White et al. 1982). All 14 encounter histories for each captured trout were used for the input file for program CAPTURE, i.e., each primary session contained three or four secondary sampling occasions. Once input into CAPTURE a model for estimating population 31  size (𝑁) was selected using the „Appropriate‟ model available in program CAPTURE. Using „appropriate‟ enabled the selection of the best model to estimate N based on the patterns reflected in the recapture data (e.g., behaviour responses or heterogeneity in capture) for each primary session for each stream. This was deemed a suitable approach because each stream had different capture patterns, and the most appropriate model selection reflected these differences in catchability across streams and seasons, i.e., population models were selected based on which model best fit the data set.  Results Defining stream size Among streams Lower East Creek had the narrowest mean bankfull channel width (BFW) (3.7 m ± 0.2 BFW), and Lower Spring Creek had the widest with the highest variance (9.5 m ± 1.16 BFW) (Table 2.1). Both Upper and Lower Spring Creeks had broad, low gradient sections (approximately 15 m and 20 m BFW, respectively) along approximately 40 m of their sample reaches. It was observed that these areas extended secondary habitat to trout during high flow periods (spring) and were either swampy or disconnected dry habitats during mean and lower flow periods. Sample reaches ranged from 200 m (Stream C) to 436 m (Stream H) in length with an overall mean trap density of one trap per 7.1 m of sample reach. The greatest mean distance between habitat units was in Stream H (i.e., 11-m mean distance) due to shallower (< 15 cm depth; not trappable) riffle habitats occurring in the lower reaches of the stream. Mean depth of residual habitat (henceforth “residual depth”), which was measured to predict the relative amount of habitat available between streams for trout populations at base flow, ranged from 14.1 cm ± 0.4 (Stream C) to 29.2 cm ± 1.07 (Blaney Creek) of depth at low-flow (Table 2.1).  32  Across the range of stream widths used in this study, BFW stream did not predict the depth of habitat (i.e., residual depth) available in at base flow (Linear regression, R² = 0.011, P = 0.821, df = 6). In addition, BFW was a weak predictor of mean CPUE (Linear regression, R² = 0.208, P = 0.015, df = 27) and a poor predictor of trout density (fish/m²) (R² = 0.06, P = 0.209, df = 27), and recapture probabilities (R² = 0.014, P = 0.606, df = 20). In comparison, residual depth was a similar predictor of trout CPUE (Linear regression, R² = 0.229, P = 0.003, df = 27), and a better predictor of trout density (R² = 0.277, P = 0.004, df = 27) and recapture probabilities (R² = 0.266, P = 0.017, df = 20). Therefore, to examine the hypothesis that trout populations with less habitat are more limited than populations with a great supply of aquatic habitat, the following results use the hydrological measure of residual depth as the independent variable (termed „stream size‟) for predicting population demographics. Biological attributes of trout populations A total of 3129 trout (new and recaptured) were captured, measured and released over the entire study (Appendix A); 1940 trout were uniquely marked across all streams and sessions. Tag loss, determined by tag scars which remained visible during the course of the study, occurred in 4.8% of trout. CPUE differed among streams (ANCOVA: stream [secondary bouts], F6, 69 = 4.77, R² = 0.464, P < 0.001), but not between seasons (ANCOVA: season [secondary bouts], F3, 69 = 1.02, R² = 0.464, P = 0.388) with no significant interaction effect between the independent factors of stream and season. Within each season, closed-secondary sample bouts showed a significant difference in CPUE between short-term trap sessions, where capture numbers tended to decrease after the first day of sampling (ANOVA: secondary bouts, F3,94 = 17.01, R² = 0.352, P < 0.001). Population size, estimated from recapture histories (i.e., CAPTURE), varied within streams between seasons and among streams (Table 2.2). On average Upper East had the smallest estimated trout population (mean 𝑁 86, range 33-149 33  between seasons), and Blaney Creek had the largest trout population (mean 𝑁 307, range 142591 between seasons). Trout densities (fish/m²) differed significantly among streams (ANOVA, stream size, F6,21 = 7.65, R² = 0.681, P < 0.001; Figure 2.2). Blaney Creek had the highest overall mean trout densities and variance (0.238 fish/m², ± 0.08, n = 4) with the lowest densities occurring in Stream H (0.034 fish/m², ± 0.008, n = 4). However, trout densities were not significantly different among seasons across streams (ANOVA, season, F3,21 = 0.81, R² = 0.09, P = 0.500). Although not significantly different, mean spring 2008 densities were the highest (0.166 fish/m², ± 0.053, n = 7) and ranged from 0.055 fish/m² to 0.470 fish/m² with Stream H having the lowest density and Blaney Creek the highest within the spring 2008 period (Figure 2.2). Across all seasons and relative to the other streams, Stream H (the 2nd shallowest stream), had the lowest mean trout densities (0.034 n/m², 0.0083 SE, n = 4). In comparison, Blaney Creek (the deepest stream) had the highest mean densities across all seasons (0.239 fish/m², ± 0.083, n = 4) with the spring 2008 value (0.470 fish/m²) largely contributing to Blaney Creek‟s higher mean density. Mean trout condition (FCI) varied between streams and seasons, showing a similar pattern along the gradient of stream sizes across seasons (Figure 2.3). However, while patterns remained similar, there were significant differences in mean FCI between streams across seasons (ANOVA, stream, F6,21 = 3.95, R² = 0.530, P = 0.008). On average, Upper East (3rd shallowest stream, 18.7 cm residual depth) had significantly lower trout condition (0.958 FCI, ± 0.010, n = 4) than the two shallowest streams, Stream C and H, with Stream H having the highest mean FCI among streams and across seasons (1.08 FCI, ± 0.031, n = 4). Seasonally, trout condition was significantly lower in the winter (early 2009) period (0.975 FCI, ± 0.014, n =4) (ANOVA, season, R² = 32.8%, F = 3.90, P = 0.021, df = 27) and higher in the spring periods of 2008 (1.05 FCI, ± 0.022, n = 4) and 2009 (1.05 FCI, ± 0.021, n = 4). There was no significant relation 34  between residual depth and FCI (R² = 0.019, P = 0.489, df = 27), or with density predicting body condition across streams within primary samplings (R² = 0.07, P = 0.670, df = 27). Trout geometric mean mass (grams) did not differ significantly among seasons (ANOVA, season, F3,21 = 0.27, R² = 0.033, P = 0.844), but did differ among streams (ANOVA, stream size, F6,21 = 5.68, R² = 0.619, P < 0.001; Figure 2.4) across seasons. Stream C, the shallowest stream (14.1 cm residual depth) had the overall lowest mean and highest variance in trout mass (5.35 g, ± 1.67, n = 4), and Lower East Creek showed the highest mean mass across seasons (11.93 g, ± 0.364, n = 4). In addition, the highest frequency of young (i.e., < 1yr. age approximated by < 7 cm FL; De Groot et al. 2007) fish occurred in Stream C with approximately 60% of the population being in the age class 0+ to 1 years over the study period. There were no strong trends between stream size (depth) and trout mass within seasons, however, the late summer 2008 period did show a slightly positive relation of increasing trout mass with increasing stream size (i.e., residual depth) (late summer trout mass vs. stream size, R² = 0.272, P < 0.001, df = 6; Figure 2.4). In addition, there was a slight negative trend between mass (grams) and seasonal densities but the relation was not significant (seasonal density vs. seasonal trout mass, R² = 0.132, P = 0.105, df = 20). Mean seasonal daily growth rates in trout mass (adjusted for initial mass) ranged from 0.00137g/day to 0.00487 g/day. There was a significant interaction between streams and seasons (ANCOVA: [PROC MIXED] stream*season [initial mass], F12,477 = 5.30, R² = 0.577, P < 0.001). All streams other than Blaney Creek and Stream C showed a declining trend in mean daily growth rates from spring (0.00297 g/day, ±0.00046, n = 7) to summer (0.00147 g/day, ±0.00028, n = 7) to winter (0.00082 g/day, ±0.0003, n = 7). Blaney Cr. and Stream C showed different trends in growth rates over the summer period; for Blaney Cr. the mean trout growth rate was highest in the summer period, whereas within Stream C growth rates were lowest during 35  the summer compared with other seasons. There was a weak negative relation between increasing stream size and daily growth rates in trout mass (specific growth rate (SGR grams) vs. residual, R² = 0.096, P <0.001, df = 498). Relations between season SGR and pre-season trout densities were not significant (R² = 0.031, P = 0.446, df = 20; Figure 2.5). Space and time: apparent survival and recapture probabilities All trout captured and marked were used in the analysis of survival between streams and seasons, with the exception of Y-O-Y trout captured after the spring 2008 sampling, i.e., only trout known to have been living at the start of the study were used in the survival analysis. The best model predicting recapture probabilities was an interaction between stream and the winter season, i.e., P (stream x winter) (Table 2.3). Recapture probabilities varied among streams with lower recapture in the winter period for most streams except for Stream C where recapture rates were similar among seasons, and Blaney Cr., where recapture rates were higher in the winter period (Figure 2.7). Recapture rates ranged from 16.6% to 60.6% recapture with a mean of 39.0% across all streams and seasons. There were several models with competing QAICc weights, or relative support, for explaining trout survival probabilities (Table 2.3). The summer season, where survival was constrained to be a linear function of a subset of time variation, was highly represented in the top models (cumulative QAICc weight 33.1%). Spatially, „stream‟ as an independent factor was supported in model selection with nested models (Δ QAICc ≤2) containing both an interaction effect between stream and summer (model QAICc 2553.845), and marginally greater support for an additive effect, i.e., stream + summer (QAICc 2553.752). However, the interaction effect in the φ (stream x summer) model did not significantly contribute to variation in trout survival when assessed using likelihood ratio tests (LRT, χ² = 12.21, P = 0.0576, df = 6; Table 2.3). Since statistical support of an interaction effect was not strong, the main effects of the additive model 36  where then tested. The results for the LRT concurred with the cumulative QAICc weights (above) that the summer season appeared to be the strongest contributing variable in trout survival estimates (Table 2.3). To account for model uncertainty in estimations of survival and recapture probabilities the top models (Δ QAICc ≤2) were averaged; model averaging accounted for 88.6% of total model weight. Survival estimates reflected the influence of the summer season of trout survival across the seven streams. All streams showed lower mean monthly survival rates in the summer period compared to other seasons (Figure 2.7). Among streams, Stream C and H, the two shallower streams, had the lowest mean monthly survival rates, and the highest variance during the summer period compared to all other streams. To test for a linear relationship between low-flow conditions as a measure of stream size which may predict summer trout survival, the re-constituted summer survival probabilities generated from logit transformations in program MARK were cube transformed to meet assumptions of normality and compared against the independent variable mean residual depth of habitat for all streams (i.e. interpolation, Gotelli and Ellison 2004). Regressions showed that the mean residual depth at base flow had a positive relation with mean predicted summer trout survival rates (R² = 0.296, P = 0.011, df = 19). When the relation between seasonal survival rates (summer and winter) and residual depth was extrapolated, the model showed a strong curvilinear trend where survival was predicted to reach an asymptote with diminishing benefits to apparent survival with incremental increases in depth (Nonlinear regression, summer survival, y = 0.812 (1 – 0.925x), R² = 0.653, P = 0.0278, df = 5; Figure 2.7). In comparison with the summer season, the winter season showed  37  higher survival rates, and a similar relation to residual depth (Nonlinear regression, winter survival, y = 0.729 (1 – 0.939x), R² = 0.884, P = 0.0083, df = 5; Figure 2.7). Survival model covariates Mass (grams) and FCI were modelled separately with stream and season as independent factors to estimate the slope of the relationship between biotic characteristics and physical factors such as stream size and seasonal influence on survival. Mass with the interaction of stream and season was the top biotic covariate in survival models indicating variable relationships between body size (mass), season and stream size; in general mass alone showed a positive relation to survival (linear on the logit scale; ß parameter for the slope = 1.84 ± 0.42). Alternatively, FCI also showed an interaction between stream and season with streams demonstrating positive and negative relations; in general FCI showed a negative relation to monthly survival rate (ß parameter for the slope = - 0.259 ± 0.053) (Figure 2.8).  Discussion Trout survival and stream size In this study top a priori survival models supported the hypothesis that trout populations in smaller streams had lower survival than populations inhabiting larger streams. However, interpretation of these results depended on which metric was used to express stream size: bankfull width or predicted depth of available habitat at low-flows. While „stream‟ was a main contributing factor in model selection, supporting the prediction that trout survival differed among populations, when survival was constrained as a function of bankfull width no relation was detected. Similarly, directional trends in trout densities were not predicted by bankfull width as a measure of stream size.  38  This lack of relation between bankfull width, and trout demographics was unexpected because other empirical studies have found a relation between bankfull width and trout densities. For example, Rosenfeld et al. (2000) found that at a reach scale anadromous juvenile cutthroat trout in BC disproportionally selected smaller streams (< 7 m channel width) for rearing, and densities showed a negative relationship with increasing bankfull width and percentage pools, i.e., streams ranged from 1.2 m to 11.2 m channel width. In addition, Kozel and Hubert (1989) also found that brook trout (Salvelinus fontinalis Walbaum) were negatively associated with increasing stream size due to loss of preferred habitat with distance away from bank, and competition with other trout. Variability in densities in relation to stream size may be a reflection of the differences in ontology and life history traits expressed by different trout species. In this current study of allopatric resident cutthroat trout, all age classes were present in streams, with trout densities and distributions reflecting age-related habitat preferences with occupancy of specific habitat-types or stream areas influenced by hierarchical behaviour (e.g., Lonzarich and Quinn. 1995). The hierarchical behaviour of age-classes within the population may have influenced the density patterns of trout within stream reaches. In contrast, juvenile sea-run forms of cutthroat trout, often co-exist with juvenile coho (Oncorhynchus kitsch Mitchell; Rosenfeld et al. 2000), and compete with other fish species for preferred habitats (i.e., pools or riffles; Rosenfeld and Boss 2000; Gresswell and Hendricks 2007). In comparison to allopatric populations which can exploit a range of habitat-types, competitive relations between species may potentially limit the spatial distribution of each species within a stream (e.g., Gatz et al. 1987). Furthermore, because trout are often associated with stream margins, a higher width to length ratio in sample reaches would reduce the proportion of bank and margin habitat preferred by stream trout (e.g., Kozel and Hubert 1989); as such we may expect to see relatively low densities of trout in wider 39  streams. However, the gradient of smaller stream widths used in this study may not have spanned the spatial range necessary to detect these negative trends between trout densities and bankfull width. In contrast to bankfull width, residual depth of habitat did show a positive relation with trout densities and survival rates, supporting the hypothesis that a greater availability of habitat during low flow conditions can positively influence survival rates, and hence trout persistence. Although model inferences should not be over interpreted, the curvilinear relation of survival rates against residual habitat predicted that trout with less than 20 cm mean depth of habitat may have decreasing probabilities of survival. In comparison, trout populations with greater than 20 cm of mean residual habitat displayed an asymptotic increase in survival rates with increasing depth. Such an asymptotic relation, where greater depth beyond some maximum value may not incur greater benefits, may be expected since depth of water has been related to optimal feeding opportunities with a trade-off between trout energetics in relation to food delivery via optimal water velocity, and effective depth of water as cover from predators (e.g., Harvey et al. 2005). A positive relation of trout with habitat depth is consistent with other studies, where pool depth and width have been observed to be strong predictors of trout persistence and densities (e.g., Heggenes et al. 1991; Gowen and Fausch 1996; Harig and Fausch 2002; Harvey et al. 2005). Harig and Fausch (2002) observed that low pool availability and depth limited the translocation success of cutthroat trout in small mountain streams in Colorado; and Heggenes et al. (1991) showed that in coastal BC, cutthroat trout larger than 9 cm (total length) had strong habitat preference for pools with mean depths > 25 cm (coupled with > 40% in-stream cover from overhead vegetation or large wood). Furthermore, Gowan and Fausch (1996) demonstrated that increasing pool depths in experimentally manipulated streams in Colorado increased adult trout abundance as fish re-distributed themselves into the deeper habitats. A habitat-scale, 40  Harvey et al. (2005), showed that rainbow trout (O. mykiss) densities were positively related to water depth. Moreover, Heggenes (2002) found that while brown trout showed high versatility in habitat use based on changes in water velocities, they displayed more stable behaviour in relation to water depth, with strong preference for deeper pools in the summer period. Accordingly, comparing trout demographics in relation to the depth of available habitat during periods of low-flow may be at an appropriate spatial and temporal scale (e.g., Schlosser and Angermeier 1995). As such, it may also be a suitable hydrological measure of stream size for the hypothesis being tested in this study, where availability habitat is within the ecological context of the species being examined (e.g., Harvey et al. 2005), and is therefore applicable for predicting stream-dwelling cutthroat trout densities and persistence in streams on the lower range of bankfull width. Seasonal trout survival and stream size Seasonality contributed significantly to variation in monthly survival rates. Coupled with spatial measures, this indicated that both time-dependence and population-dependence were contributing to variation in survival estimates, i.e., cumulative weightings 88.6%. The summer period contributed the most to seasonal variation (33.1% of the cumulative weighting vs. 15.0% winter and 8.4% spring) with estimated monthly survival rates decreasing in all streams over the summer period as compared to the spring and winter seasons. These results supported the hypothesis that monthly summer survival rates were higher in larger streams, using residual depth as a measure of stream size. Few other studies have directly estimated survival rates of trout populations in small streams using mark-recapture techniques and individual trout characteristics (e.g., Olsen and Vollestad 2001; Carlson and Letcher 2003; Berger and Gresswell 2009). From these studies, similar observations of variance in trout survival responses to either seasonality, and/or low-flow 41  conditions have been observed. In a 3-year mark-recapture study of cutthroat trout populations in two Oregon headwater streams, Berger and Gresswell (2009) observed that trout survival rates were consistently lowest during the autumn period, with evidence of a negative association with low-flow periods. They also inferred that adult trout were negatively affected by seasonal abiotic conditions which induced an annual survival bottle neck. In western Massachusetts, Carlson and Letcher (2003) observed that brown trout (Salmo trutta Linnaeus; >1yr. age) and brook trout had their lowest survival during the early summer and the autumn periods. In southeastern Norway, Olsen and Vollestad (2001) found that populations of brown trout had their lowest survival rates in the summer period. Conversely, other studies have found little or no relation between habitat and trout survival. For example, Harvey et al. (2005) could find no conclusive support between habitat variables (e.g., depth, cover, low-flow discharge) and rainbow trout survival in an experimental manipulation of flow rates during the summer period in California, and cited the challenges of reliably quantifying habitat-specific survival rates due to the confounding influence of other factors, such as the trade-offs in habitat use between optimal cover from predation versus optimal habitat for foraging. Trout biological characteristics and survival Although mean monthly survival probabilities did not differ greatly between stream populations (means 0.87 – 0.94 per month), even small difference in survival rates among streams can reflect a biological meaningful difference in the life expectancies and annual rates (White and Burnham 1999). For example, over the summer period estimates of life expectancies in Stream C ranged from several months to several years, whereas in Lower East Creek, where habitat availability was greater, life expectancies were almost 30% longer than in Stream C (i.e., between 5 months and 3 years). Shorter life expectancies in one population compared to another could reflect several processes occurring within a population. For example, age-selective 42  mortality in younger fish, could lead to an overall lower mean survival rate within the population. Age-selective mortality can occur at various life stages in fish populations (e.g., Elliott 1987; Carlson et al. 2004; Berger and Gresswell 2009), and juvenile salmonids densities can be regulated during egg stages, emergence or 0+ age through density-dependence mechanisms acting on specific life stages (Elliott 1987; Elliott1989; Schindler 1999) and depending on the favorability of environmental conditions experienced by different populations (Elliott 1987). In this study, covariates used in survival models and patterns in descriptive data did indicate that populations with a higher proportion of younger/smaller fish, such as Stream C, had a positive slope for the relation between survival and mass, inferring possible size selection against smaller fish. However, while age-selective mortality may explain some of the variation in survival rates between streams, there was no relation between seasonal trout densities and growth rates or fish condition that would suggest the mechanisms responsible for greater mortality in younger fish, such as negative density-dependant competitive interactions between younger trout (Elliott 1987; Schindler 1999). In fact, populations with relatively higher proportions of smaller fish still tended to have high adjusted- least- squared-means growth rates, and the highest condition of any age class, suggesting that these populations were not density limited, and that the costs of rapid growth were low (Carlson et al. 2004). Similarly, Jenkins et al. (1999) and Harvey et al. (2005) also found that neither survival nor maximum growth rates of fish were linked to density. Furthermore, Elliot (1987) found that resident brown trout in a low density population which experienced unfavourable stream conditions was likely regulated by both positive density-dependent and density-independent factors. Therefore, in this current study mechanisms other than density-dependence are likely contributing to mortality in smaller fish in our study streams. 43  Several studies have inferred that larger stream trout do less well than smaller fish during low-flow periods (e.g., Rosenfeld and Boss 2002, Harvey et al. 2006; Berger and Gresswell 2009). Harvey et al. (2006) found that trout growth rates were 8 times less in habitats where flows were diverted. Rosenfeld and Boss (2002) concluded that larger trout did less well than smaller fish because of their higher energetic needs not being met by reduced food delivery, and Boss and Richardson (2002) found that larger trout were potentially at risk from predation in streams with reduced water depths. Whereas, in contrast, the overall observations in this study found that on average larger trout (> 8 cm FL) had a positive relation to survival. A positive relation of mass to survival may be due to differing supplies of hydrologic habitat among streams because models of trout mass and survival inferred an interaction between stream and season in relation to survival when mass was included as a covariate. Similarly, Quinn and Peterson (1996) found that larger coho smolts had better survival than smaller fish, but the relationship was also complicated by covariates such as habitat complexity, distance between habitat units and site-specific locale of emergence. On average, larger trout (1+ yr. age) trout tended to occupy streams, or habitat-units, with deeper depths during the summer period, possibly suggesting that these trout found sufficient habitat for their survival. If larger trout found sufficient depth of habitat and avoided stranding, their greater size could have incurred them the advantage of defending their higher-quality habitats from other sub-dominant fish (e.g., Fausch and White 1981). This would have helped them to avoid intra-specific cannibalism by smaller, gape-limited trout, or predation from piscivorous birds or mammals (e.g., Parker 1971). In turn, smaller trout may have been competitively excluded from pools by hierarchical behaviour or threats of cannibalism by larger trout (Fausch and White 1981). These territorial interactions could have displaced smaller subdominant trout from deeper habitats, increasing their occupancy of shallow riffle-habitats, 44  exposing them to higher mortality risks such as predation or stranding during low-flow periods (e.g., Riley et al. 2009). Predation of fish in shallower habitats is a plausible explanation for some of the mortality occurring in stream populations. For example, in a low-flow experimental study using brown trout in chalk streams of Wales, Riley et al. (2009) found that mortality rates increased in 0+ age brown trout, with higher juvenile mortality likely being attributed to competitive displacement of smaller fish into sub-optimal habitat, leading to death via predation. In an experimental study examining food and cover limitations of cutthroat trout in two streams also used in this study, Boss and Richardson (2002) noted that a single predator could cause very high mortality (> 50%), with physical cover improving survival probabilities of larger trout. For this current study, predators were noted to be present (e.g., black bears) in some stream reaches. For example, in Stream H where survival was lowest in the summer months, less fortunate fish who had not found refuge in larger pools were found dead, or trapped within shallow pockets (< 2 cm depth habitat) of the stream reach, and small mammal, black bear and birds‟ prints were observed during several visual inspections of the lower reaches. Plus, black bears were regularly sighted in the lower reaches of Streams H and C during the spring and summer months. Therefore, predation is a plausible explanation for some fish mortality, especially in areas where movement was greatly restricted, with isolated pools being hotspots for predation. Movement and survival estimates Immigration from study reaches may also have contributed to lower apparent survival estimates in some streams. In the past it was generally thought that adult resident trout led fairly sedentary lives, only travelling short distances within a stream reach (20 – 50 m) over a lifetime (i.e. restricted movement paradigm; Gerking 1959). This paradigm has been challenged over the last 20 years, and it is now realised that some stream resident trout will move large distances 45  (e.g., > 1 km; Harvey 1998; Rodriguez 2002; Mellina et al. 2005; Gresswell and Hendricks 2007), while others will move to a lesser extent, and for some populations it has been found that 70 – 80% of trout may only move 20 -100 m (Gresswell and Hendricks 2007). However, even short distance movers can bias apparent survival estimates in mark-recapture studies, and it is likely that movement outside of the study area accounted for some of the trout which were never recaptured (Rodriguez 2002). Northcote (1978) termed these reach-scale migrations or movements by cutthroat trout as trophic or refuge migrations, where movement was occurring in response to environment cues, and not as synchronized mass migrations as in anadromous trout species. Small or large scale movement was likely occurring in this study because while densities fluctuated within streams, there was no significant difference among seasons, indicating that emigrating fish (1+ y. age) were replacing fish that had either died or left the study area. This assumption was supported by between population comparisons, where in some streams (e.g., Stream H) the proportions of unmarked apparent immigrant adult trout caught in later capture sessions was significantly higher than the number of marked recaptures, which suggested that movement rates were relatively higher in some streams (e.g., Riley et al. 1992; Gowen and Fausch 1996; Rodriguez 2002). Trout may move for several reasons; seasonal changes in habitat use and stream discharge, spawning, and the influence of stream productivity and competition for resources among trout may explain why trout move, and how dependant survival is on movement rates for stream resident populations (e.g., Trotter 2008). Movement may depend on the age of an individual fish and/or on seasonal changes in food supply, critical habitat and stream flow (e.g., Hilderbrand and Kerschner 2000; Mellina et al. 2005; Gresswell and Hendricks 2007). If habitat resources are below optimal levels, then trout are likely to move within their reach. For example, 46  Keeley et al. (2001) found that when allowed the possibility of emigration, poorer condition trout would emigrate presumably in search of more optimal conditions. This result suggested that while food and space are important factors in regulating the growth and condition of individuals within a population (e.g., Boss and Richardson 2002), behavioural responses such as movement between habitats within a stream matrix, and the ability to move are also key governing factors for trout in stream systems. Gresswell and Hendricks (2007) found that trout in shallower habitats (< 70 cm) moved to a greater extent than trout occupying deeper, preferred habitats. Furthermore, Kahler et al. (2001) stated that movement was common for trout, finding that cutthroat trout (and coho) would move in search of more optimal habitat, and that individual movers „thrived‟ in comparison to non-movers who stayed in shallower habitat. They also concluded that movement was not necessarily due to competitive exclusion by other trout but possibly due to habitat choice. Given the compression of habitat during the low-flow summer period in the two shallowest streams, C and H streams, it is possible that these trout moved to a greater extent than trout in streams where availability of preferred habitat was not limited, or lost, during different periods of the year, e.g., Lower East Creek. In addition, similar to observations by Kahler et al. (2001), the patterns of higher movement rates may also be linked to the higher mean body condition observed in these shallower streams. Conservation implications and persistence of cutthroat populations In conclusion, this study links the local demographics of seven trout populations to the temporal and spatial processes occurring across a size gradient of small streams. Studies such as this are fundamental for providing insights into the processes determining the persistence of trout populations (Lowe et al. 2006). Identifying the mechanisms limiting populations is one of the fundamental questions in ecology (Krebs 2002) and is necessary so we may understand the impacts of our management activities. However, while environmental variability is a strong 47  mechanistic contender to explain the dynamics at work within the populations of this study, inferences need to be made with care. There are likely many limiting factors acting on trout populations, such as food and habitat supply, movement between sub-populations, and competition and predation. All of these variables have some degree of influence on defining the carrying capacity for each trout population (Kiffney and Roni 2007). However, water is a critical component for trout survival, and estimates from this study suggested that streams providing more than 20 cm of mean residual habitat would have higher population-level survival rates than streams with fewer or shallower habitats, especially during the summer months. Survival rates in smaller streams were lower in all seasons compared to streams with greater habitat availability. In particular, summer low-flows appeared to lead to a reduction in functional aquatic habitat and loss of connectivity, reducing survival rates in trout. Mechanisms limiting trout populations are difficult to isolate in empirical studies, yet patterns in this study suggest that density and survival fluctuated within and among streams with temporal variation acting as a dominant component of trout population dynamics (Gowen and Fausch 1996; Berger and Gresswell 2009). Spatially, similarity among trout populations has been shown to be higher as distance between streams decreases (e.g., Platts and Nelson 1988, Gowan and Fausch, 1996; Kocovsky and Carline 2006; Berger and Gresswell 2009). Therefore, the close spatial proximity of streams in this study, and the high degree of concordance in trout density fluctuations within and among streams, may indicate that regional factors such as climate were influencing trout populations (e.g., Gowen and Fausch 1996). Similarly to Berger and Gresswell (2009), temporal variability suggested that sitespecific mechanisms, such as low-flows occurring during the summer period, were affecting cutthroat trout survival. As such, the differing effects of available habitat during the summer season on survival of trout in small versus large streams may be inferred to reflect the overriding 48  influence of the density-independent factors, seasonality and stream habitat, in this study (e.g., Gowan and Fausch, 1996; Jackson et al. 2001, Quist et al. 2005; Berger and Gresswell 2009). As mechanisms for mortality, it is likely that predation accounted of some of the loss of trout in smaller streams. However, while biological factors such as density, condition and growth rates varied, it was not possible to attribute these factors to lower survival rates in trout populations. In streams where flows were greatly reduced during the summer period, movement was high and results indicated that emigration and immigration were linked to growth and condition of trout as well as to their survival. As a conservation measure, Hilderband and Kreshner (2000) predicted that to support lower numbers of viable trout populations (0.1 fish/m), similar to numbers in this study, cutthroat trout would need a minimum of 25 km of stream length. Given the distance of reach needed to support viable trout populations, and the observations that trout moved to a greater extent in some streams in this study because they were shifting between suitable habitats, connectivity is proposed to be a vital component for ensuring persistence of these small local populations of stream dwelling trout. Based on the findings of this one year study of trout in the Malcolm Knapp Research Forest it is recommended that fisheries or forest resource management consider the vulnerability of small streams (<10 m bankfull width) to loss of surface flows during dry periods of the year. Secondly the need to ensure that 2-way connectivity between reaches is maintained where possible, with deeper refugia provided along trout-bearing reaches so trout may find safe and functional habitat during extreme periods of low-flow.  49  Table 2.1: Physical characteristics of seven small (second and third order) streams in the southern portion of Malcolm Knapp Research Forest, British Columbia. Measurements were record during the period of May 2008 to June 2009 (SE± 1).  Stream  Elevation (range)  Water-shed area (m²)  Degree Slope (range)  Length reach (m)  Mean BFW (m)  Mean Trap dist. (m)  Mean residual pool depth  Trap Area (m²)  Mean % LWD cover  Mean % Boulder  C-Stream  110 - 285 *  89*  1-4  200  4.0 (0.27)  4.9  14.1 (0.40)  804.0  24.0  14.5  Upper Spring  135 - 500*  35 *  3-4  230  5.0 (0.59)  5  25.4 (1.43)  1150.5  44.4  <1  Lower Spring  108 - 126  38  0.5 - 4  236  9.5 (1.16)  6.9  22.5 (1.49)  2478.0  32.3  11.8  Blaney  300-340  29  1-4  242  5.2 (1.18)  5.5  29.2 (1.05)  1256.8  23.7  13.9  Upper East  288-491  44 *  1-7  246  4.6 (0.33)  7.6  18.7 (2.24)  1131.6  39.9  24.3  Lower East  135-147  44  1–2  355  3.7 (0.26)  8.1  24.0 (1.97)  1312.8  10.4  1.0  H-Stream  265-285¹  55.4¹  1-2  437  6.2 (0.38)  11.2  15.8 (0.68)  2707.5  17.2  11.5  *Kiffney et al. (2003); ¹Gomi et al. (2006)  50  Table 2.2: Estimated abundance (𝑵) (SE ± 1) for cutthroat trout using passive trapping methods within four primary periods occurring during spring, late summer and winter conditions (i.e. 3-4 secondary trap sessions for each primary) for seven small streams of Malcolm Knapp Research Forest . An „appropriate‟ model for calculating estimates and standard error of N for each primary session and stream was selected using population estimate software CAPTURE (program MARK; White and Burnham 1999). (see Appendix 1 for raw data used to calculate estimates).  Seasons trout captured February & September March 2009 2008  Stream name  June 2008  C-Stream  110 (41.5) 144 (9.66)  96 (3.46)  June 2009  99 (4.67)  Upper Spring 225 (76.72) 155 (14.72)  174 (15.96) 123 (5.57)  Lower Spring 248 (18.67) 238 (9.49)  148 (11.78) 285 (18.97)  Blaney  591 (34.19) 142 (8.37)  159 (74.0) 377 (3.40)  Upper East  105 (12.98)  56 (3.46)  33 (4.24) 149 (19.00)  Lower East  144 (22.94) 204 (34.16)  142 (55.35) 341 (79.71)  H-Stream  150 (16.97)  108 (10.10)  51 (4.90)  64 (8.49)  51  Table 2.3: Output of top recapture and survival models with effects constrained to test a priori candidate models for „best‟ model to estimate recapture and survival probabilities for all individual trout encounter histories combined into stream groups. A logit link function was applied to fit a general linear model to all estimates (White and Coach 2009).  *QAICc  Δ QAICc  AICc Weight  Number Parameters  Deviance (Q)  2560.74  7.5927  0.00338  35  44.1974  Recapture (P) Models P (Stream X Winter) P (Stream X Summer) P (Stream + Winter) P (Stream + Time) P (Stream X Time)  2547.21 2558.71 2558.90 2559.51 2560.83  0 11.504 11.695 12.306 13.620  0.98998 0.00315 0.00286 0.00211 0.00109  28 28 22 23 35  45.018 56.522 69.011 67.577 44.199  Survival (Φ) Models Φ (.) Φ (t) Φ (Summer) Φ (Stream + Summer) Φ (Stream X Summer) Φ (Winter) Φ (Spring) Φ (Stream + Winter) Φ (Stream)  2553.23 2553.28 2553.29 2553.75 2553.85 2554.15 2554.16 2554.64 2555.02  0 0.0490 0.0571 0.5177 0.6111 0.9171 0.9245 1.4062 1.7866  0.13374 0.13050 0.12998 0.10324 0.09853 0.08455 0.08424 0.06621 0.05474  15 16 16 22 28 16 16 22 21  77.602 75.620 75.628 63.862 51.657 76.488 76.496 64.750 67.174  Biotic covariates Φ (Stream X Mass X Season) Φ (Stream X FCI X Season)  5149.37 5172.42  0 23.05  29 29  5090.94 5114.00  Models Full model Φ(Stream X time) P(Stream X time)  Notes: With survival (φ) held constant (stream x time) the best recapture model was determined. Next, the top recapture model was held constant (stream x winter) to identify the best candidate models for estimating survival for all streams. Best model fits are ranked by descending order in the table. „+‟ denotes an additive model testing the main effects; „X‟ denotes an interaction including the main effects; „.‟ denotes a constant effect without time or group dependent effects; „time‟, „season‟ or „stream‟ indicates specific-dependence on the stated factor(s) in the model (notation as per Lebreton et al. 1992). *QAIC = Quasi Akaike Information Criterion; variance inflation factor applied; Δ „Delta‟ = change in the QAIC between the top model and the next best model.  52  Table 2.4: Based on structural differences in the top models likelihood ratio tests between models were used to identify the main independent variables contributing to estimates of trout survival. A P <0.05 indicates null model factors, compared to the paired alternative model, contribute significantly to trout survival.  Alternate Model  Null Model  χ²  df  P-value  Φ (Summer)  (stream + summer)  11.766  6  0.0674  Φ (Summer)  (stream X summer)  23.971  12  0.0205  Φ (stream + summer)  (stream X summer)  12.205  6  0.0576  53  Primary sample period Population open between primary periods (i.e. 4 over one year)  Survival (Φ) estimated using CJS¹ open population model  3 to 5.5 months  Next primary sample period (k)  Population closed between secondary bouts . . . trap day 1  1 or 2 days . . . trap day 2  Population size (N) estimated across short time period, i.e., closed CAPTURE model  . . . trap day t (i.e. 3 to 4 trap days per primary)  Figure 2.1: Schematic representation of robust sampling design (based on Pollock 1982) for sampling cutthroat trout in seven small streams. Primary periods were one week, and within primaries, secondary bouts were one day each and trout were captured each day for 3 to 4 days (adapted from Krebs 1999). Intervals between primaries were 3 to 5.5 months. All fish were released in each bout. CJS¹ (Cormack-Jolly-Seber).  54  0.5  Spring 2008  Late Summer 2008  0.4  0.3  Density (number/m2)  0.2  0.1  0.5  Spring 2009  Winter (Feb.) 2009  0.4  0.3  0.2  0.1  0.0 14  16  18  20  22  24  26  28  30  14  16  18  20  22  24  26  28  30  Residual depth (cm)  Figure 2.2: Density (estimated population size/m² ± 1 SE) of trout by stream size per season. Streams increase in size (i.e. residual depth) from left to right along x-axis.  55  1.20  Spring 2008  Late Summer 2008  1.15 1.10  Fulton's Condition Index  1.05 1.00 0.95  1.20  Spring 2009  Winter (Feb.) 2009  1.15 1.10 1.05 1.00 0.95  14  16  18  20  22  24  26  28  30  14  16  18  20  22  24  26  28  30  Residual depth (cm)  Figure 2.3: Mean Fulton‟s condition index (FCI ±1 SE) for trout over four primary sampling periods. Streams increase in size from left to right along x-axis. The dashed line represents an FCI of 1 and is displayed for reference between seasons. For sample sizes see Appendix A.  56  16  Spring 2008  Late Summer 2008  14 12 10  Geometric mean mass of trout (grams)  8 6 4 2 0 16  Winter (Feb.) 2009  Spring 2009  14 12 10 8 6 4 2 0 14  16  18  20  22  24  26  28  30  14  16  18  20  22  24  26  28  30  Residual depth (cm)  Figure 2.4: Geometric mean (± 1 SE) of biomass of trout for each primary sample period and stream. Streams increase in size (measured as mean residual depth) from left to right along the xaxis.  57  0.005  Spring 0.004 0.003 0.002  SGR in mass (g) over seasonal period  0.001  0.005  Summer 0.004  14.1 cm (C) 15.8 cm (H) 18.7 cm (Upper East) 22.5 cm (Lower Spring) 24 cm (Lower East) 25.6 cm (Upper Spring) 29.2 cm (Blaney)  0.003 0.002 0.001  0.005  Winter 0.004 0.003 0.002 0.001 0.000  0.10  0.20  0. 50  Start of season densities (est. N/m2)  Figure 2.5: Specific growth rate in mass (grams) per season against cutthroat trout population densities (fish/m²) in seven small streams of Malcolm Knapp Research Forest.  58  Recapture probability for seasonal interval  Summer and Spring Winter 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 14  16  18  20  22  24  26  28  30  Residual depth (cm)  Figure 2.6: Probability of recapture (± 1 SE) calculated from model averaged outputs. Since the best model held recapture as P (stream X winter), there are not separate probabilities for spring and summer (i.e. they are constant); whereas recapture was more influenced by winter and is time-dependant. Stream names are ordered by increasing size.  59  Summer (error bars heavy cap) Winter (error bars medium cap) Spring (error bars small cap) Regression for summer period Regression for winter period 0.98  Survival rate (monthly)  0.96 0.94 0.92 0.90 0.88 0.86 0.84 0.82 0.80 0.78 10  15  20  25  30  Residual depth (cm)  Figure 2.7: Mean monthly survival estimates (± 1 SE) from program MARK (see Table 2.3 for models) for cutthroat trout for three seasons (summer, winter and spring; 2008-2009) against streams arranged in order of increasing size (i.e. predicted residual depth in habitats at base flows) as the predictor for survival parameters. Non-linear regression is plotted for winter and summer periods  60  Summer 2008 0.94  0.92  0.90  Mean monthly apparent survival of trout  0.88  Winter 2008 0.96  0.94  Blaney C Stream H Stream Lower East Lower Spring Upper East Upper Spring  0.92  0.90  0.88  Spring 2009 0.96  0.94  0.92  0.90  0.88  0.90  0.95  1.00  1.05  1.10  1.15  Pre-season mean FCI  1.20  2  4  6  8  10  12  14  16  Pre-season mean mass (grams)  Figure 2.8: Regression of initial mass and body condition (FCI) against seasonal survival rates for cutthroat trout in seven small streams of coastal BC.  61  Literature cited Acolasa, M.L., J.M. Roussela, J.M. Lebelb, and J.L. Baglinièrea. 2007. Laboratory experiment on survival, growth and tag retention following PIT injection into the body cavity of juvenile brown trout (Salmo trutta). Fisheries Research 86: 280-284.  Behnke, R. E. 2002. Trout and salmon of North America, First Edition. The Free Press, New York.  Berger, A.M., and R.E. Gresswell. 2009. Factors influencing coastal cutthroat trout (Oncorhynchus clarkiii clarkiii) seasonal survival rates: a spatially continuous approach within stream networks. Canadian Journal of Fisheries and Aquatic Sciences 66: 613632.  Bloom, A. M. 1976. Evaluation of minnow traps for estimating populations of juvenile coho salmon and Dolly Varden. The Progressive Fish-Culturist 38: 99-101  Boss, S.M., and J.S. Richardson. 2002. Effects of food and cover on the growth, survival, and movement of cutthroat trout (Oncorhynchus clarkii) in coastal streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 1044-1053.  Bradford, M. J., and J. S. Heinonen. 2008. Low flows, instream flow needs and fish ecology in small streams. Canadian Water Resources Journal 33: 165-180.  Brown, J. H., D. W. Mehlman, and G.C. Stevens. 1995. Spatial variation in abundance. Ecology 76: 2028-2043.  Brown, J.H. 1984. On the relationship between abundance and distribution of species. The American Naturalist 124: 255-279.  Buckland, S. T., K. P. Burnham, and N. H. Augustin. 1997. Model Selection: an integral part of inference. Biometrics 53: 603-618.  Burnham, K. P., and D. R. Anderson. 2002. Model Selection and Multimodel Inference: a practical information-theoretical approach, Second Edition. Springer, New York.  62  Carlson, S. M., A. P. Hendry, and B. H. Letcher. 2004. Natural selection acting on body size, growth rate and compensatory growth: an empirical test in a wild trout population. Evolutionary Ecology Research 6: 955–973.  Carlson, S. M., and B. H. Letcher. 2003. Variation in brook and brown trout survival within and among seasons, species, and age classes. Journal of Fish Biology 63: 780–794  Carlson, S. M., E. M. Olsen, and L. A. Vøllestad. 2008. Seasonal mortality and the effect of body size: a review and an empirical test using individual data on brown trout. Functional Ecology 22: 663-673.ckwell  Chapman, D.W. 1966. Food and space as regulators of salmonid populations in streams. The American Naturalist 100: 345 -357.  Cunjak, R. A. 1996. Winter habitat of selected stream fishes and potential impacts from landuse activity. Canadian Journal of Fisheries and Aquatic Sciences 53:267-282  Cunjak, R. A., and G. Power. 1987. The feeding energetic of stream-resident trout in winter. Journal of Fish Biology 31: 493-511.  De Groot, J. D. 2004. Density, body condition, and movement of coastal cutthroat trout (Oncorhynchus clarkii clarkii) in logged and forested headwater streams of southwestern British Columbia. University of British Columbia press, MSc., thesis.  De Groot, J. D., S.G. Hinch, and J.S. Richardson. 2007. Effects of logging second growth forests on headwater populations of coastal cutthroat trout: a 6-year, multi-stream, before-and-after field experiment. Transactions of the American Fisheries Society 136: 211-226.  Dunham, J. B., and B. E. Rieman. 1999. Metapopulation structure of bull trout: influences of physical, biotic, and geometrical landscape characteristics. Ecological applications 9: 642-655.  Elliottt, J. M. 1989. The critical time for survival: mechanisms responsible for population regulation in young migratory trout (Salmo trutta). Journal of Animal Ecology 58: 9871001.  63  Elliottt, J. M. 1973. The food of brown and rainbow trout (Salmo trutta and S. gairdneri) in relation to the abundance of drifting invertebrates in a mountain stream. Oecologia 12: 329-347.  Elliottt, J. M. 1987. Population Regulation in Contrasting Populations of Trout Salmo trutta in Two Lake District Streams. Journal of Animal Ecology 56: 83-98.  Fausch, K.D., and R.J. White. 1981. Competition between brook trout (Salvelinus fontinalis) and brown trout (Salmo trutta) for positions in a Michigan stream. Canadian Journal of Fisheries and Aquatic Science 38: 1220-1227.  Feller, M. C., and J. P. Kimmins. 1979. Chemical characteristics of small streams near Haney in southwestern British Columbia. Water Resources Research 15: 247-58.  Fisher , R. A. 1922. On the Mathematical Foundations of Theoretical Statistics. Philosophical Transactions of the Royal Society of London. Series A, 222: 309-368.  Gatz, A.J., M.J. Sale, and J.M. Loar. 1987. Habitat shifts in rainbow trout: competitive influences of brown trout. Oecolgia 74: 7-19.  Gerking, S.D. 1959. The restricted movement of fish populations. Biological Review Cambridge Philosophy Society 34: 221–242.  Gomi, T., R. D. Moore, and A. S. Dhakal. 2006. Headwater stream temperature response to clearcut harvesting with different riparian treatments, coastal British Columbia, Canada. Water Resources Research 42: 1-11.  Gowan, C., and K. D. Fausch. 1996. Long-Term Demographic Responses of Trout Populations to Habitat Manipulation in Six Colorado Streams. Ecological Applications 6: 931-946.  Gresswell, R.E., and S.R. Hendricks. 2007. Population-scale movement of coastal cutthroat trout in a naturally isolated stream network. Transactions of the American Fisheries Society 136: 238-253.  Gresswell, R.E., Torgersen, C.E., Bateman, D.S., Guy, T.J., S.R. Hendricks, and J.E.B. Wofford. 2006. A spatially explicit approach for evaluating relationships among coastal cutthroat trout, habitat, and disturbance in small Oregon streams. American Fisheries Society Symposium 48:457–471. 64  Harig, A.L., and K.D. Fausch. 2002. Minimum habitat requirements for establishing translocated cutthroat trout populations. Ecological Applications 12: 535-551.  Harvey, B.C., J. L. White, and R. J. Nakamoto. 2005. Habitat-specific biomass, survival, and growth of rainbow trout (Oncorhynchus mykiss) during summer in a small coastal stream. Canadian Journal of Fisheries and Aquatic Sciences 62: 650-658.  Harvey, B.C., R. J. Nakamoto, and J. L. White. 2006. Reduced streamflow lowers dry-season growth of rainbow trout in a small stream. Transactions of the American Fisheries Society 135: 998-1005.  Hawkins, C. P., J. L. Kershner, P.A. Bisson, M. D. Bryant, L. M. Decker, S.V. Gregory, D. A. McCullough, C. K. Overton, G.H. Reeves, R. J. Steedman, and M. K. Young. 1993. A Hierarchical Approach to Classifying Stream Habitat Features. Fisheries 18: 3-12.  Heggenes, J. T. 2002. Flexible summer habitat selection by wild, allopatric brown trout in lotic environments. Transactions of the American Fisheries Society 131: 287-298.  Heggenes, J. T., O. M. W. Krog, O. R. Lindås, and J. G. Dokk. 1993. Homeostatic Behavioural Responses in a Changing Environment: Brown Trout (Salmo trutta) become nocturnal during winter. Journal of Animal Ecology 62: 295-308.  Heggenes, J., T. G. Northcote, and A. Peter. 1991. Seasonal Habitat Selection and Preferences by Cutthroat Trout (Oncorhynchus clarkii) in a Small Coastal Stream. Canadian Journal of Fisheries and Aquatic Science 48: 1364-1370.  Hicks, B. J., J. D. Hall, P. A. Bisson, and J. R. Sedell. 1991. Responses of salmonids to habitat changes, pages 483–518 in W. R. Meehan, editor. Influences of forest and rangeland management on salmonid fishes and their habitats. American Fisheries Society, Special Publication 19, Bethesda, Maryland.  Hilderbrand, R. H., and J. L. Kershner. 2000. Conserving inland cutthroat trout in small streams: How much stream is enough? North American Journal of Fisheries Management 20: 513–520.  Holtby, L. B. 1988. Effects of logging on stream temperatures in Carnation Creek, British Columbia, and associated impacts on the coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Sciences 45: 502-515.  65  Hurst, T. 2007. Review Paper: Causes and consequences of winter mortality in fishes. Journal of Fish Biology 71: 315–345.  Huusko, A., L. Greenberg, M. Stickler, T. Linnansaari, M. Nyka¨ nen, T. Vehanen, S. Koljonen, P. Louhi, and K. Alfredsen. 2007. Life in the ice lane: the winter ecology of stream salmonids. River Research and Applications 23: 1–23.  Jackson D.A., P.R., Peres-Neto, and J. D. Olden. 2001. What controls who is where in freshwater fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences 58: 157–170.  Jenkins T.M., Jr. Sebastian, D., Kratz, K.W. and S.D. Cooper. 1999. Effects of population density on individual growth of brown trout in streams. Ecology 80: 941–956.  Johnston, N.T., and P.A. Slaney. 1996. Fish habitat assessment procedure. Watershed Restoration Tech. Circ. 8. Queen‟s Printers,Victoria, B.C.  Kahler, T. H., P. Roni, and T. P. Quinn. 2001. Summer movement and growth of juvenile anadromous salmonids in small western Washington streams. Canadian Journal of Fisheries and Aquatic Science 58: 1947–1956.  Keeley, E.R. 2001. Demographic responses to food and space competition by juvenile steelhead trout. Ecology 82: 1247–1259.  Kiffney, P. M., and P. Roni. 2007. Relationships between Productivity, Physical Habitat, and aquatic invertebrate and vertebrate populations of forest streams: an informationtheoretic approach. Transactions of the American Fisheries Society 136: 1088–1103.  Kiffney, P. M., J. S. Richardson, and J. P. Bull. 2003. Responses of periphyton and insects to experimental manipulation of riparian buffer width along forest streams. Journal of Applied Ecology 40:1060–1076.  Kocovsky, P. M, and R. F. Carline. 2006. Influence of landscape-scale factors limiting brook trout populations in Pennsylvania streams. Transactions of the American Fisheries Society 135: 76-88. Kozel, S. J., and W. A. Hubert. 1989. Factors Influencing the Abundance of Brook Trout (Salvelinus fontinalis) in Forested Mountain Streams Journal of Freshwater Ecology 5: 113-122. 66  Krebs, C. J. 2002. Two Complementary Paradigms for Analysing Population Dynamics Philosophical Transactions: Biological Sciences 357: 1211-1219. Krebs, C. J. 1999. Ecological methodology 2nd Ed. Addison-Wesley Educational Publishers, Inc., Don Mills, Ontario p: 49-70.  Krebs, C. J. 2001. Ecology: the experimental analysis of distribution and abundance 5th edition. Benjamin Cummings, San Francisco, California.  Lonzarich D. G., and T.P. Quinn. 1995. Experimental evidence for the effect of depth and structure on the distribution, growth, and survival of stream fishes. Canadian Journal of Zoology 73: 2223–2230.  Meidinger, D., and J. Pojar (Eds.). 1991. Ecosystems of British Columbia. Ministry of Forests, Victoria, British Columbia, Special Report Series No. 6.  Mellina, E., Hinch, S.G., and K.D. MacKenzie. 2005. Seasonal movement patterns of streamdwelling rainbow trout in north-central British Columbia, Canada. Transactions of the American Fisheries Society 134: 1021–1037.  Milner, N. J., J. M. Elliottt, J. D. Armstrong, R. Gardiner, J.S. Welton, and M. Ladle. 2003. The natural control of salmon and trout populations in streams. Fisheries Research 62: 111– 125.  Minitab. 2007. Statistical software for Windows, Release 15.1. Minitab, Inc., State College, Pennsylvania.  Montgomery, D.R., and M. J. Buffington. 1997. Geological Surveyors of America Bulletin 109:596–611.  Moore, K., K. Jones, and J. Dambaucher. 1997. Methods for stream habitat surveys: aquatic inventory project. Oregon Department of Fish and Wildlife, Corvallis.  Northcote, T. G. 1978. Migratory strategies and production in freshwater fishes. Pages 326-359 in S. D. Gerking, ed. Ecology of Freshwater Fish Production. Blackwell Sci-entific, Oxford, England.  67  Parker, R.R. 1971. Size selective predation among juvenile salmonid fishes in a British Columbia inlet. Journal of Fisheries Resources Board Canada 28: 1503–1510.  Peterson, D. P., K. D. Fausch, and G. C. White. 2004. Population Ecology Of An Invasion: Effects of brook trout on native cutthroat trout. Ecological Applications 14: 754-772.  Platts, W. S., and R. L. Nelson. 1988. Fluctuations in trout populations and their implications for land-use evaluation. North American Journal of Fisheries Management 8: 333- 345.  Poff, N. L. 1997. Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. Journal of the North American Benthological Society 16: 391-409.  Poff, N. L., and J. D. Allan. 1995. Functional Organization of Stream Fish Assemblages in Relation to Hydrological Variability. Ecology 76: 606-627.  Pollock, K. H. 1982. A capture-recapture design robust to unequal probability of capture. Journal of Wildlife Management 46: 752-757.  Power, M. E., R. J. Stout, C. E. Cushing, P. P. Harper, F. R. Hauer, W. J. Matthews, P. B. Moyle, B. Statzner, and I. R. Wais De Badgen. 1988. Biotic and abiotic controls in river and stream communities. Journal of the North American Benthological Society 7: 456479.  Quinn, T. P., and N. P. Peterson. 1996. The influence of habitat complexity and fish size on over-winter survival and growth of individually marked juvenile coho salmon (Oncorhynchus kisutch) in Big Beef Creek, Washington. Canadian Journal of Fisheries and Aquatic Sciences 53: 1555-1564.  Quist, M.C. & Hubert, W.A. 2005. Relative effects of biotic and abiotic processes: a test of the biotic–abiotic constraining hypothesis as applied to cutthroat trout. Transactions of the American Fisheries Society 134: 667-686.  Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fisheries Research Board of Canada Bulletin 191.  Riley, S. C., K. D. Fausch, and C. Gowan. 1992. Movement of brook trout (Salvelinus fontinalis) in four small subalpine streams in northern Colorado. Ecology of Freshwater Fish 1: 112122. 68  Riley, W. D., D. L. Maxwell, M. G. Pawson, and M. J. Ives. 2009. The effects of low summer flow on wild salmon (Salmo salar), trout (Salmo trutta) and grayling (Thymallus thymallus) in a small stream. Freshwater Biology 54: 2581–2599.  Rosenfeld, J., M. Porter, and E. Parkinson. 2000. Habitat factors affecting the abundance and distribution of juvenile cutthroat trout (Oncorhynchus clarkii) and coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science 57: 766– 774.  Schlosser, I. J. 1985. Flow regime, juvenile abundance, and the assemblage structure of stream fishes. Ecology 66: 1484-1490.  Schlosser, I. J. 1991. Stream Fish Ecology: A Landscape Perspective. BioScience 41: 704-712.  Schlosser, I.J., and P.L Angermeier. 1995. Spatial variation in demographic processes of lotic fishes: conceptual models, empirical evidence, and implications for conservation. American Fisheries Society Symposium 17: 392-401.  Strahler, A. N. 1952. Dynamic basis of geomorphology. Geological Society of America Bulletin 63: 923-938. Trotter, P. 2008. Cutthroat trout: native trout of the west 2nd Edition. University of California Press, Ca.  Warren, C. E., and W. J. Liss. 1980. Adaptation to aquatic environments. In Fisheries management. Lackey, R.T., and L. Nielsen, eds. Blackwell Scientific, Oxford, UK. pp 15 -40. White, G. C., and Cooch. 2008. Program MARK: a gentle introduction. 8th Edition. Available online at http://www.phidot.org/software/mark/docs/book/. Accessed June 2008.  White, G.C., and K.P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46: 120-139.  White, G.C., D.R. Anderson, K.P. Burnham, and D. L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. In: LA-8787-NERP, Los Alamos National Laboratory, Los Alamos, New Mexico, USA.  69  Williams, J. E., A. L. Haak, H. M. Neville, and W. T. Colyer. 2009. Potential Consequences of Climate Change to Persistence of Cutthroat Trout Populations. North American Journal of Fisheries Management 29: 533-548.  70  Chapter 3 : A before-after study contrasting the effects of streamside harvesting on cutthroat trout populations2. Introduction Population abundance can be related to the quality of a species‟ habitat, and the magnitude of difference in abundance between populations may be attributed in part to relative differences in habitat quality (Krebs 2001). In the case of stream environments, population abundances are strongly linked to the supply of food resources and geomorphic units provided by the stream environment (Chapman 1966). In turn, the supply of energy resources and the geomorphic characteristics of a stream, and its quality of aquatic habitat, are linked with the surrounding landscape and local climate (Hynes 1975, Elliott 1985, Gomi et al. 2002; Benda et al. 2004). Concomitant with the spatial variation of stream environments, stream-dwelling species often exhibit variation around a mean abundance in synchrony with temporal fluctuations in seasonal conditions (e.g., Fausch et al. 1988). These temporal fluctuations in a population‟s abundance can be impacted by either natural or human-induced disturbance events which can disrupt structural and functional elements within the stream environment (Schlosser 1991), or its riverscape (Ward et al. 2002). Consequently, these physical impacts may lead to changes in habitat availability and quality through alteration of in-stream structure, and subsequent changes in population abundances over different spatial and temporal scales (Schlosser 1991). Forest harvesting can be a large-scale landscape disturbance which impacts small-stream ecosystems, with population responses varying across geographical regions, depending on the  2  A version of this chapter will be submitted for publication by Sheldon KA, JS Richardson, JD De Groot, and SG Hinch. Titled as: A before-after study contrasting the effects of streamside harvesting on cutthroat trout population: a 12 year multi-stream experiment.  71  level of disturbance, and the sensitivity of the species (Resh et al. 1988; Schlosser 1991). In the Pacific Northwest (PNW) studies of forestry-stream related activities have observed variable responses to forest harvesting (Meehan 1991; Murphy et al. 1986; De Groot et al. 2007; Mellina and Hinch 2009). For instance, salmonids which commonly inhabit small streams (< 5 m bankfull channel width; Rosenfeld 2000) of the region, have been observed to have both positive and negative responses to harvest-related activities, especially when harvesting alters the interface between the terrestrial and aquatic environments (e.g., Hicks et al. 1991; Meehan 1991; Northcote and Hartman 2004). Furthermore, succession of streamside vegetation, alteration of sediment retention and organic matter inputs, and loss of bank stability following the removal of streamside timber may incur differing magnitudes of response in primary and secondary production. Directional responses may continue to change along a time-scale from 0 to 250 years, as the lag effects from timber harvesting can persist decades after the initial disturbance (e.g., Gregory et al. 1991; Beechie et al. 2000; Zhang et al. 2009). Solar flux to the stream bed can increase immediately after the removal of streamside timber, with allochthonous subsidies from the terrestrial-riparian area simultaneously decreasing (Hawkins et al. 1983; Bilby and Bisson 1992). Stream temperatures may increase (Moore et al. 2005), and elevated photosynthetically active radiation (PAR) levels may shift basal energy resources from allochthonous to autochthonous, stimulating in-stream primary and secondary production (e.g., Kiffney et al. 2003). In relation to these resource shifts stream-dwelling trout, for example, may benefit because they are often food limited (Boss and Richardson 2002), and an increase in production of primary consumers may lead to increased trout abundance and condition (Bilby and Bisson 1992). However, initial gains in biological productivity may be negated as stream environments and their riparian areas undergo further physical and functional changes. After several years, 72  regenerating streamside vegetation may create a dense canopy, increasing stream shading and decreasing light levels (e.g., Young et al. 2000) and in-stream primary productivity. Bank stability and sediment retention may decrease, with bank erosion increasing fine sediment inputs (Gomi et al. 2005) which may contribute to diminishing pool depths and numbers via in-filling (Benda et al. 2005). In addition, during the wet season peak stream flows may increase in magnitude, and displace existing in-stream wood structures (Benda et al. 2005) that may not be replaced for up to several decades following harvesting because of the removal of mature streamside timber during harvesting (Hicks et al. 1991). Subsequently, the loss of large wood, important for physical structure, sediment trapping (Bisson et al. 1987) and pool formation, may result in reduced channel complexity and habitat quality, leading to decreased habitat for fish and other stream organisms (e.g., deep pools for cover and shelter, Fausch and Northcote 1992). After a lag time of one to several decades fish populations may become habitat-limited, and initial gains in food supply observed in the first few years post-logging may be negated (e.g., Murphy et al. 1986). Many past forestry-fish studies have used comparatively short-term post-treatment (1-6 years post-logging) or chronosequence designs to examine the effects of riparian harvesting on stream ecosystems and fish populations. However, these comparisons have often been among streams running through old-growth (>250 years) forest and clear-cut reaches (e.g., Holtby 1988; Thedinga et al. 1989; Bilby and Bisson 1992). Whereas, studies rarely compare the effects of second-growth riparian harvesting on stream variables by contrasting initial conditions with responses over extended post-harvest time periods (e.g., De Groot et al. 2007). Furthermore, predictions based on short-term post-logging studies which compared old-growth forests and clear-cut sites may not be predictive of the effects of removing 80-100 year old second-growth forests commonly been felled today or subject to newer forestry practices.  73  To partially remedy this knowledge gap, De Groot et al. (2007) investigated the responses of cutthroat trout populations to removal of second-growth streamside forest in small streams of the PNW. They compared population responses along a timeline from 2-years preharvest to 4-years post-logging using a before-after-control-impact (BACI) study design. Their findings indicated that relative abundances of trout in the summer period were unchanged in logged streams, yet had decreased in control streams relative to pre-harvest levels. Their relation with stream and climatic variables suggested that trout in logged streams may have initially benefited from the removal of dense second-growth forest canopy with modest increases in stream temperatures during an otherwise cooler climatic phrase in the Pacific Decadal Oscillation (PDO), whereas, by contrast control streams were cooler by 0.5 – 1 °C (De Groot et al. 2007). This positive response of fish populations to canopy removal was not unexpected when compared to past studies (e.g., Murphy et al. 1981), and may have been partially due to the careful harvesting practices employed in the study and mild air temperatures of the preceding summer, especially in cooler, higher elevations or latitudes. However, in the time period of 4years post-logging, trout responses may not have reflected the potential lag effects predicted to occur in association with riparian succession and loss of large wood inputs (see review Meehan 1991; meta analysis Mellina and Hinch 2009). In this paper, I contrast mean seasonal trout abundances and body condition for the 10th year post-logging period relative to the pre-harvest and 4th year post-logging values reported in De Groot et al. (2007). Based on responses reported in the forestry-stream literature I predicted that relative increases in primary production may be providing benefits for trout (e.g., Gregory et al. 1987, Kiffney 2008). In particular, inputs of deciduous leaves from early seral succession could be contributing higher food resources for primary consumers (e.g., shredders and collectors) relative to pre-harvest coniferous inputs, i.e., higher food quality (Wipfli and  74  Musselwhite 2004). Conversely, trout may have become habitat-limited (e.g., Murphy and Koski 1989) due to the loss of large wood, pool numbers and depths. Loss of these structures may have negated the benefits associated with increased autochthonous and deciduous-type allochthonous inputs. Based on these predictions I hypothesised that trout populations 10-years post-logging would have good body condition in response to higher productivity, yet have declined in abundance compared to pre- harvesting and 4th year post-logging levels due to habitatlimitations. To test this hypothesis I compared seasonal mean differences and directional trends in relative abundance and body condition by contrasting 2-years pre-impact, 4-years and 10years post- impact values. To examine changes in physical variables and habitat I related trout biological responses with proportional changes in habitat availability between the pre-impact period and 10th year post-impact, and trends in mean stream water and air temperatures from 2006 - 2009.  Methods Study area Study sites were located in the Malcolm Knapp Research Forest (MKRF), near the town of Maple Ridge, British Columbia (122° 34´ W, 49° 16´ N) (see Chapter 2 for study area descriptions). This study was part of an experimental large-scale spatially and temporally replicated before-after-control-impact (BACI) design (details in De Groot et al. 2007). My sampling sites were the same as in De Groot et al. (2007) and I attempted to re-sample and survey the same study reaches examined in past sampling and surveys, i.e., historic markers were evident in most reaches.  75  Stream surveys Habitat assessments were conducted in July 2008 to quantify habitat and to compare relative changes in stream habitat with past surveys, i.e., June 1997 and June 2002. Surveys identified habitat units as pools, riffles, cascades, glides or steps based on gradient, surface water characteristics and substrate composition (Montgomery and Buffington 1998). Each identified fish habitat unit was marked and distance between units, and mean and residual depths were measured with a measuring tape/depth stick. Percentage overhanging vegetation providing stream cover was visually assessed, percentage gradient between habitat units was measured using a Suunto™ clinometer and meter stick, large wood pieces (length, > 2m: diameter, >10 cm) were tallied within habitat units, and percentage substrate composition was visually assessed for each habitat unit and classified by size classes: fines < 2 mm, gravels 2 - 64 mm, cobbles 65 - 256 mm, boulders 257 - 4000 mm, and bedrock > 4000 mm (Johnson and Slaney 1996; as per De Groot 2007). All sample reach lengths were measured longitudinally down the centre of the channel. Trout sampling Trout abundance and body condition were calculated using three trapping periods from the summer 2008 period (i.e. May, June and early September ), and one sampling period in June 2009 (i.e., termed summer sampling for 10th year post-logging period, n = 4/ treatment). During the winter period one sampling event occurred in each stream (February/March 2009). Trout were captured using baited (brined salmon roe) Gee-type traps with a standard mouth size of approximately 2.5 cm diameter. The passive capture method of using Gee-type traps was selected in favour of electro-fishing due to possible trout mortality associated with repeated electro-shocking, and the low conductivity levels in streams (Young et al. 1999), i.e., low conductivity in stream water can lead to shock-induced mortality due to the high amperage levels  76  required for stunning (S. G. Hinch, UBC, personal comm. 2008). Furthermore, Gee-type traps have comparable capture efficiencies to electro-fishing in our study area (Young et al. 1999). All cages were numbered and were placed along the stream length within each numbered habitat unit. It has been observed that temperature affects trout behaviour and movement rates, with quicker capture rates in warmer periods relative to during cold, winter periods. Temperature-dependant behavioural heterogeneity was accounted for applying a 3-hour trapping period for spring and summer samplings, and a 24-hour, overnight trap period for the winter samplings. The 3-hour period took advantage of the higher activity of trout in the warmer periods, when capture rates were rapid and longer trap periods did not increase CPUE (De Groot et al. 2007). As well, the shorter interval reduced incidences of trap-induced cannibalism from trout interactions and cage interference by bears. For the winter period, the 24-hour trap day accounted for the greatly reduced movement rates and overnight feeding behaviour of trout during the cold (<4°C) winter months (e.g., Heggenes et al. 1993) . Captured trout were anaesthetized with buffered pre-mixed tricaine methanesulfonate (MS-222, 0.1 g per litre), weighed to the nearest 0.1 g, and fork length (FL) was measured to the nearest mm. Trout were marked with unique elastomer dye codes or passive integrated transponders to identify when they were last measured. After inspection trout were placed into buckets of fresh stream water containing vegetation cover to reduce possible stress until they had fully recuperated from handling. Trout were judged recovered when they were capable of swimming rapidly, i.e., were not sluggish or easily handled. Once recovered, trout were released back to the habitat unit where they were captured.  77  Data analysis Fish habitat assessment Physical measures from stream habitat assessments (i.e., mean depth of pool, pool number per 100 m, percentage fine sediments, percentage canopy cover of channel, and pieces of LW per 100 m) were compared for proportional gains or losses within sites between the beforeimpact period and the 10th year post-logging period. Environmental data Average, maximum and minimum daily air temperatures for the summer period (July 1 to August 31) were calculated from the Malcolm Knapp Research Forest data records for the years 2006 to 2009. For the summer period of 2008 mean stream discharge (Q mean), water temperature (Celsius) and measured dissolved oxygen (DO) were calculated using continuous data collected from weirs positioned downstream of two sampling sites: East Creek (control) and Stream A (logged) (M. Feller, Department of Forest Sciences, University of British Columbia, unpublished data). Due to equipment failure at these weirs for the summer period of 2009 no discharge or temperature data were available from these stations. Therefore, for the 2009 summer period temperature data were collected using a Hobo ™ temperature logger placed in each of Stream C, Spring Creek and East Creek. Relative abundance of trout To compare the impacts of logging on trout relative abundance with pre-harvest and 4th year post-logging values, 10th -year post-logging averages for trout measures were calculated using age 1+ year (based on FL ≥ 7.5 cm; as per De Groot et al. 2007) and older trout. Young of the year (Y-O-Y) data were not used in past analyses for this long-term study design because trapping methods used have being ineffective at capturing YOY trout, therefore trout smaller than 7.5 cm FL were also excluded from the 10th year data set. For consistency of analyses and 78  to make among-stream contrasts more comparable, relative abundance was calculated as CPUE per stream segment length to adjust for slight changes occurring among years to study segment lengths within streams (De Groot et al. 2007). Unless otherwise stated the pre-harvest period refers to the sample years of the summer 1997 to 1998, the 4th-year period refers to the postlogging sample years of 1999-2002 inclusive, and the 10th year refers to the sample periods of 2008 to 2009. In addition, area calculations were not applied to calculated relative abundance because no stream width measures were taken in two of the six trapping years per meter of stream length trapped. Mean summer relative abundances for each sample year were used for comparison amongstreams and before and impact periods. In the 10th year after-impact sampling period trapping occurred three times between the early to late summer period of 2008, and once in the early summer of 2009; these data were used to calculate average relative abundance of trout in the summer period for the tenth year. Differences between logged and unlogged treatments were then contrasted with before-harvest, fourth-year post- harvest (2000 – 2002) and tenth-year postharvest measures of trout abundances using a mixed model analysis of variance (ANOVA). The model contained the main effects of treatment (logged, unlogged control), impact period (before [1997, 1998]; after [2000 – 2002] and [2008 – 2009]) with site nested within treatment, and an interaction between treatment and period (i.e. 3 levels of time). Post hoc pairwise mean differences were compared using Tukey‟s least squares means (LSM) comparisons. If no interaction effect was detected then streams were inferred to be responding similarly with no statistical evidence of a harvest effect on trout abundance; alternatively, a significant interaction may be due to harvesting effects on trout abundance. Contrasts using the same mixed model-ANOVA approach were not possible for amongstream and period comparisons of interannual winter trout densities. No winter data were 79  available for Spring Creek in 1997 (before-impact), and only a single sampling event occurred over the winter periods after-impact. Therefore, mean winter densities and their 95% confidence intervals were calculated for each stream to assess if pre-logging winter means (i.e., 1998) were within or outside of post-logging winter relative abundance confidence limits (i.e., 1999 – 2002, and 2009). Trout body condition Analysis of covariance (ANCOVA: GLM) was used to compare log10 mass against log10 length relationships for control and logged streams for the 10 th year period, with the body condition (LSM) been used to contrast summer body condition with before - impact mean values derived from De Groot et al. (2007). Comparisons between means were made for beforecontrol, before-logging, after-control and after-logging groups using 2-sample t-tests. Null hypotheses of equal means were rejected and significant differences among treatment means and time periods were inferred if 95% confidence intervals did not contain zero. Data were sorted and compiled using Excel (Microsoft© Corporation 2007), statistically analysed in Minitab® (2007, version 15.1, Redmond, Wa) and graphical displayed using SigmaPlot® (2004, version 9.0, Systat Software© Inc., Chicago, Il). All data were tested for normality and homogeneity of variance with Anderson-Darlings tests for normality, and Levene‟s and Bartlett‟s test of equal variance to ensure data meet assumptions of inferential analysis, if needed data were transformed using ln (x + 0.001) or log10 transformations. Means are reported ± one standard error.  80  Results Physical attributes Canopy cover increased in both treatment streams over the last 10 years since time of logging. In Stream A cover was similar to 1997 levels and in Stream C cover was 23% higher than pre-logging levels; and on average treatment and control streams showed a 20% to 28% respective increase in canopy cover since 1997. Fine sediments showed decreasing trends in all streams over the last 10 years, similar to 4th year observations by De Groot et al. (2007), with the highest loss occurring in Stream C (- 93 %) and lowest decrease occurring in Spring Creek (control, - 34%). Mean pool depth decreased in all streams since 1997 by 33 - 44%, however, these differences likely reflect variation in stream discharge across survey years. On average, pool numbers and pieces of large wood decreased in treatments compared to controls in relation to pre-logging 1997 values (Table 3.1). Average air temperatures for the 2008 summer period were lower than 2009 by 1.6 °C, and the same as 1999 average values. The range in air temperatures was also less with maximum values in the summer of 2008 being 4.5 °C lower than 2009, and minimum values being 0.5 °C less (Table 3.2). In addition, precipitation varied greatly between the two sample years, with 2008 having 254 cm of rain compared to 124 cm in the summer of 2009. Total maximum and minimum stream temperatures ranged from < 0 to 8.63 °C in Stream A, and from 0.21 – 7.65 °C in East Creek during the winter period (December 2008 - March 2009; values were not available for other streams). During the 2008 summer period water temperatures were cooler by 0.8 – 1.5 °C among all streams compared to average stream temperatures in the summer of 2009 (Table 3.3). In the summer of 2008 the control treatment was 1.2 °C cooler than the logged treatment (Figure 3.1); and in the summer of 2009 control treatments were 1 – 1.64 °C cooler than logged streams (Table 3.3). 81  Trout abundance A total of 607 trout were caught in the summer periods of 2008-2009, with no detection of trout in Stream A; of trout captured, 395 (65.1%) were ≥ 7.5 cm FL (age 1+ yr). Stream C had the lowest percentage of age 1+ yr trout (43.3%) in contrast to East Creek and Spring Creek, which each had higher percentage of age 1+ yr fish, i.e., 76.5% and 72.0%, respectively. Summer relative abundance (CPUE/m) of trout varied between streams, years and time periods. Similar to the pre-harvest and post-logging 4th year periods, Spring Creek had the highest trout abundance for 2008 and 2009 sample periods relative to other streams; with the Stream C logged treatment having the second highest abundance (Figure 3.2). In addition, in the 2009 summer sampling year Stream C trout abundance was the highest recorded value over the 12 years of the study. The 10th year post-logging comparison showed lower mean trout abundance in both the logged and control treatments relative to the pre-logging summer period (Figure 3.3). Relative to pre-harvest values, control streams showed a 63.7 % decrease in trout relative abundance (preimpact and 10th year, respectively; 0.0113 ± 0.0041 CPUE/m, 0.0041 ± 0.0001 CPUE/m); and mean values for logged streams 10 years post-harvest showed a decrease in trout abundance with a 31.8 % decrease compared to pre-logging values (pre-impact and 10th year, respectively; 0.0022 ± 0.0002 CPUE/m, 0.0015 ± 0.001). Relative summer abundance of trout differed significantly between treatments with control streams having the highest mean abundance (ANOVA: treatment, F1, 20 = 19.39, P <0.001), but not between streams nested within treatments (ANOVA: stream, F2, 20 = 1.09, P = 0.354), nor between before-after time periods (ANOVA: time period, F2, 20 = 3.23, P = 0.061) with no interaction between treatment (control, logged) and time period (before, 4th year and 10th  82  year) (ANOVA: treatment*time period, F2, 20 = 0.63, P = 0.542), indicating no significant harvest effect on summer trout abundances over time. The exclusion of Stream A (zero fish in the 10th year sampling) from the analyses did not alter the outcome of results (ANOVA: treatment*time period, F2, 18 = 1.63, P = 0.224). Winter trout abundance (CPUE/m) varied between streams, years and time periods (Figure 3.4). For mean winter abundance comparisons between pre-harvest and post-10th year logging, the mean relative abundances of trout before-impact (i.e. in winter 1998) for all streams were inside the 95% confidence intervals (CI) of the after-impact 10th year period, indicating no significant difference between winter trout abundances in pre-harvest and post-impact years. Respective mean winter densities with their 95% CI for the 10th year period and 1998 means were: Stream A 0.0019 fish/trap/m (0.00041 – 0.0035; 1998 0.0026), Stream C 0.0027 (0.0016 – 0.0039; 1998 0.0018), and East Creek 0.0039 (0.0016 – 0.0063; 1998 0.0046). Trout body condition The logged treatment (Stream C) in the 10th year period (after-impact) had significantly heavier trout per fork length during the summer period than the control streams (ANCOVA; treatment, F1, 388 = 19.61, P < 0.001; Figure 3.4). Mean summer trout body condition also significantly differed between the before-impact and after-impact 10th year period for both the control (2-sample T; P < 0.001, df = 328) and logged streams (2-sample T; P < 0.001, df = 105) (Figure 3.5). In addition, it should be noted that fish condition in East Creek was the lowest relative to others studied in 2008 and 2009, and the weighting of this lower mean may result in lowering the average fish condition for the control treatment.  83  Discussion Trout abundance There is no evidence that a temporal lag effect from streamside harvesting may have reduced fish habitat or induced a decrease in relative trout abundances by the tenth year post-logging. Although trout abundance did fluctuate within streams between sampling years (1997 to 2002, and 2009), the time period and treatment interaction was not significant. In part, the lack of statistical difference in mean trout abundance between treatments over time may reflect the loss of spatial replication in our study because of no trout present in Stream A in the 10th year. This does, however, raise ecological questions as to the reasons for the disappearance of trout in the upper reaches of Stream A. Prior to 2008 trout populations in both treatment streams (Streams C and A) were observed to have potentially benefited from removal of second-growth timber relative to populations in shaded control streams. De Groot et al. (2007) attributed this benefit to a relative 0.5 -1°C increase in stream temperatures in un-shaded logged streams during a cool phase in the PDO, compared to a decline in temperature in the shaded control streams. Subsequently the removal of streamside timbers may have benefited trout populations during a cooler climatic phase in the PDO, as their abundance did not decrease to the same proportion as did trout abundance in controls. Moreover, in the 10th year period, trout abundance for the summer of 2009 in Stream C was the highest observed across all samplings since the census began in 1997. During the 10th year period the PDO shifted into a cool phase in 2008, and was the most negative since 1925 (National Ocean and Atmospheric Agency, 2010), then in 2009 shifted back to a warm phase. During this period, study streams varied by at least 0.8 °C between years, and 2008 had temperatures and precipitation levels were similar to the 1999 summer sample year immediately following harvest impact. By the 10th year streamside succession was at a young 84  seral stage in logged treatments, and mean canopy cover had increased past pre-harvest levels in both the control and treatments streams. Although, while treatment and control streams showed a similar mean canopy cover, treatment stream vegetation differed in age-structure with over head canopy composition being dominated by low shrub layers and young alder trees. Therefore, intercepted solar radiation was still higher in treatment streams compared to coniferdominated mature stands in control. In addition, lower flow rates and shallow water depths in treatment streams also likely led to warmer stream temperatures compared to the control streams during 2008-2009 (Hetrick et al. 1998). Trout abundance did not appear to be negatively affected by warmer temperatures in the treatment. This may be attributed to maximum temperatures ranges not exceeding lethal limits for cutthroat trout (greater than 20°C for more than several days, Bjornn and Reiser 1991), the geographic location of study streams, and sufficient flow rates for cooling, pool numbers and large wood for fish cover and shelter during the low-flow summer periods (e.g., Hetrick et al. 1998). In a recent paper Merten et al. (2010) observed similar trends in a streamside harvested northern Minnesota stream, where variation over ten years in summer air temperatures explained temporal patterns in fish communities to a greater extent than spring precipitation or fine sediments. Furthermore, in cool mountain streams of the mid to high latitudes in North America, such as in southern BC, trout may benefit from a modest warming of several degrees in streams with normally cooler (e.g., 8-12 °C) temperatures. Consequently, in this study, warmer water temperatures in the treatment may have conferred a seasonal advantage for trout growth and relative abundance compared to control streams, which showed lower trout abundance in contrast to past sample periods. While cutthroat trout can be negatively affected by high water temperatures (>23°C , Bjornn and Reiser 1991) they can also be limited by cool water temperatures which incur 85  metabolic deficits leading to lower growth and survival rates, limiting recruitment and reducing persistence probabilities of cutthroat trout (Cunjak and Power 1987; Harig and Fausch 2002; Coleman and Fausch 2007). For example, Harig and Fausch (2002) found that cold summer temperatures, narrow stream width with a low supply of physical habitat were critical factors negatively affecting the translocation success of cutthroat trout in mountainous streams. In addition, lack of degree days and cool temperature periods during swim-up and early fry development can create bottle-necks in recruitment of cutthroat trout in higher elevation streams (Coleman and Fausch 2007). Furthermore, the optimal metabolic temperature for cutthroat trout is 18°C (MacNutt et al. 2004), and during warmer periods trout can obtain more meals/day, through increased feeding behaviour during the daylight hours. Having high energy intakes allow trout to meet their resting metabolic needs, and allocate energy to growth or mass (Elliott 1987). When energy trade-off costs are low, small fish can grow rapidly, obtain good body condition and are conferred the benefits of avoiding predation (or cannibalism) through decreasing the size difference between themselves and larger con-specifics (Carlson et al. 2004). Therefore, the warmer stream temperatures which did not exceed sub-lethal or lethal limits over our study period may account for the comparative lack of proportional difference between time periods in the logged treatment relative to the higher proportional decrease in abundance in the control streams for the 10th year comparison. Furthermore, in contrast to historic harvesting practices, the logging techniques used in this current study were “gentle”. Streamside vegetation in our treatment streams was carefully removed by felling logs away from the streams, avoiding excessive slash contributions to streams and limiting machine use to within 5 m of the stream margin. Added to these best management practices, existing large wood was not removed from streams, leaving structure for fish habitat. In addition, by the 10th year large wood pieces had increased in Stream C compared  86  to 1997 levels, with minimal decreases in pool numbers (by 18% or 2 pools). These newer logging practices and the geographic location of our study streams in the cooler foothills of the coastal PNW may explain why our results are similar to other studies which also found no-net loss or slightly positive response in trout abundances post-logging in high elevations or latitudes (e.g., Murphy et al. 1981; De Groot et al. 2007). For example, Murphy et al. (1981) found that salmonid parr abundance increased in the summer period in narrow, higher elevation clear-cut streams after the removal of second-growth forest. Moreover, a recent meta-analysis by Mellina and Hinch (2009) examining the effects of streamside logging on stream fish populations using 37 fish-harvest studies (in North America) found that the strongest post-logging responses in salmonid biomass and densities were related to the thorough removal of large wood during the logging process, with the magnitude of responses post-logging being largely independent of time since logging or stream size. Had the removal of large in-stream wood been employed in our study area, with streams scarified and a larger percentage of the watershed (>21%) cut, air temperatures may have been several degrees warmer over the study period, and we may have seen trout abundance decrease in our streams. For example, a study by Young et al. (1999) on the effects of second-growth clear-cut logging in the lower East Creek section of MKRF found that maximum stream temperatures reached 30 °C immediately after logging (i.e., section A) in their scarified study reach during 1973-1975, and summer temperatures remained in the range of 5-8 °C higher than controls for 4 years after-harvest. Trout abundance was ten-fold lower than up-stream reference levels until ten years post-logging, when densities in the logged reach were similar to reference levels.  87  Trout body condition Post- logging in 1999, an increase primary production in the treatment streams was associated with higher inputs of solar radiation, and biomass of some trout prey species (Chironomidae) did increase compared to controls (Kiffney et al. 2003; Gomi et al. 2006). However, these increases in primary consumer biomass did not appear in confer higher body condition in trout by the 4th year post-logging. In comparison, by the 10th year, summer trout condition demonstrated a significant increase for both the control and logged streams populations compared to pre-harvest. These increases in both the control and treatment stream may be due to various factors. Trout in the study area have been observed to be food-limited (Boss and Richardson 2002) and it is assumed that increases in primary consumers via higher inputs from terrestrial subsides or in-stream productivity would benefit trout condition or abundance through release of food limitations (e.g., Johnson et al. 1986, Thedinga et al. 1989). Although direct measures of primary consumer productivity were not available for the 10th year period, notable amounts of periphyton growth were observed in treatment streams during trout sampling, and compared to the conifer-dominated control streams, treatment streams had a high percentage of the (>70%) deciduous vegetation overhanging stream surfaces. Higher proportions of alder and a dense shrub layer in the understory of young-growth riparian vegetation have been observed to contribute a high proportion of terrestrial-derived invertebrates to juvenile salmonid diets, with potential benefits for their abundance and metabolic needs for growth (Wipfli 1997). In addition, young-growth forests can have higher exports of deciduous litterfall, such as red alder, with significantly greater production of aquatic invertebrates which contribute higher proportions for trout prey compared with denser old-growth or mature second-growth conifer forests (Wipfli and Musselwhite 2004). Therefore, the potential of higher seasonal inputs from young-growth  88  vegetation and warmer temperatures may give some explanation for trout abundance and body condition in treatment streams. It was interesting to observe that the control streams also showed an increasing trend in fish condition yet a decrease in abundance. While the mean condition was lower than treatment streams, gains were proportionally similar to trout in logged streams. Because control streams are mainly conifer-dominated and more heavily shaded than treatment streams, the food supply hypothesis cannot be inferred for these streams. Therefore, this may suggest that other regional or population regulating factors may be contributing to trout condition and population limitation, aside from increases in basal production due to canopy regeneration in logged streams. Stream A: upper reach While the absence of trout in the upper reaches of Stream A is ecologically significant, we cannot infer that their disappearance was only due to streamside harvesting. We can, however, pose factors which may explain their absence. Trout were sighted in the upper reaches of Stream A in the summer of 2006 (T. M. Hoover, UBC, pers. comm.), but were not detected in 2008 and 2009. However, chemical conditions in stream A did not indicate that the environment became lethal or deleteriously for trout habitation, and they were caught in 2008 in the lower reaches immediately below a weir constructed downstream of the clear-cut upper reach site. Therefore, trout were present within the stream network, and the upstream absence vs. downstream presence at the weir station suggests that trout are not passing upstream of the weir. Limited connectivity can affect trout migrations between habitat patches (Brown and Kodric-Brown 1977; Speirs and Gurney 2001), especially during seasonal periods when trout use different habitats for refugia, and this may explain why trout have not recolonized the upper reaches. In addition, during the fall of 2003 and 2005 stream discharge was the highest recorded since the mid-1970‟s (M. Feller, UBC, pers. comm.), and during high flows some trout could 89  have been flushed over the weir or possibly killed. Therefore, the compounded effects of connectivity loss and high flow events may have contributed to the decrease in trout in the upper reaches. The careful logging practices and retention of large wood for pool creation and maintenance employed in this study may have contributed to the no-net effects observed in trout relative abundances over the last 10 years post-impact. Without the loss of habitat complexity or pool numbers through infilling, trout may have had similar habitat to past conditions. In addition, slight increases in stream temperatures over the past decade do not appear to have adversely effected trout populations is logged treatments, and removal of the second-growth canopy may have conferred benefits in cooler years. This effect is possibly region-specific because while canopy removal in this riparian study did show modest increases in stream temperatures, stream location in the coastal mountains of the moist Pacific Northwest may have moderated the negative effects of increased water temperature sometimes observed in lower elevation or latitude streams. Increases in trout condition in the study area may be due to overriding regional variables such as increased air temperatures or increased productivity in the streams. While the construction of the weir in our study design was not intended to create a barrier to trout migration, it is possible trout were affected by this aspect of our design.  90  Table 3.1: Stream survey of fish habitat variables for post-impact period July 2008, the 10th year means are contrasted with mean values from pre-impact surveys assessed in June 1997. Signs indicate: (-) relative loss, (+) relative gain; and percentage (%) change from initial pre-harvest measures. 2008 Mean (SE) Stream name Variable  Stream C  East  Spring  45 (0.5)  71.7 (0.05)  68.6 (0.04)  49.0 (2.95)  Fine Sediments (%)  5 (0.1)  4.0 (0.60)  4.8 (1.54)  Depth of pools (cm)  18.8 (1.4)  14.1 (0.40)  Bankfull width (m)  3.5 (0.2)  Canopy cover (%)  Number of pools (/100 m) Number of LW (/100 m)  Stream A  Difference from pre-logged 1997 (% change) A  C  East  -2.4 (5)  + 23.6 (49)  + 30.5 (78)  -3.2 (7)  31.2 (3.6)  -25 (83)  -55.7 (93)  -39.5 (89)  -17.5 (34)  18.7(2.24)  25.4 (1.43)  -9.1 (33)  -9.8 (43)  -12.9 (41)  -19.4 (43)  4.0 (0.27)  4.6 (0.33)  5.0 (0.59)  -0.2 (5)  + 0.8 (25)  + 0.2 (4.5)  + 1.8 (49)  7  9  8  9  -1 (12)  -2 (18)  + 3 (60)  -3 (25)  25  68  62  91  -60 (75)  + 17 (33)  -9 (13)  + 23 (34)  ¹ surveyed July 2008 ² mean difference between pre-impact June 1997 survey and 2008 post-impact survey Notes: Stream survey data was interpreted from De Groot et al. 2007 for pre-impact data; see same for June 2002 means.  Spring  91  Table 3.2: Mean weather variables recorded at local climate station (Haney, Malcolm Knapp Research Forest) for the summer period (July 1 to August 31) 2006 to 2009. Variable  2006  2007  2008  2009  Maximum  35.5  37.5  32.5  37  Minimum  8.5  7  7.5  8  Average  18.09  18.06  17.35  18.91  47.8  138.9  254.9  126.4  Mean daily air temperature (°C)  Monthly rainfall (cm)  Note: data available at http://climate.weatheroffice.gc.ca/ (Accessed April 20, 2010).  92  Table 3.3: Stream temperature and flow variables (± 1 SE) collected during the summer period (July1 to August 31) at weir stations on East Creek and Stream A during the summer of 2008, and temperatures recorded by stream loggers for the summer period of 2009. Columns are organised by control first (East and Spring Creek) then logged treatments (Stream A and C) (Weir data courtesy of M. Feller, Department of Forest Sciences, UBC). Stream Year  Upper East 2008 2009  Spring 2009  A 2008  C 2009  Water temperature (°C) Average  12.56 (0.012)  13.64 (0.03)  14.25 (0.04)  13.73 (0.016)  15.28 (0.04)  Minimum  10.45  9.96  11.33  11.08  10.85  Maximum  15.58  18.14  18.48  16.80  30.56  20.66 (0.765)  ND  ND  6.08 (0.002)  ND  Mean summer stream discharge Q (l/s)  93  Stream water temperature (Celsius)  Date Figure 3.1: Stream temperatures (Celsius) from weir stations downstream of sample sites for East Creek (control) and Stream A (log treatment) for the period May1, 2008 to July 1, 2009.  94  Summer relative abundance (CPUE/m)  Summer relative abundance 0.025 East - control Spring - control Stream A - logged Stream C - logged  0.020 Impact winter 1998  0.015  0.010  0.005  09 20  08 20  02 20  01 20  00 20  98 19  19  97  0.000  Year Figure 3.2: Summer relative abundance (CPUE/m sample length ± 1 SE) of age 1+ year cutthroat trout per stream for two unlogged (control) and two logged treatments for BACI design; before-impact sample period (1997 and summer 1998), period of impact (winter 1998), and two after-impacts sample periods i.e., 2000 – 2002 inclusive, and 2008 – 2009.  95  Logged Control  0.014 0.012 0.010 0.008 0.006  a  Summer realtive (CPUE/m) Abundance (CPUE/m) Relative abundance  0.016  0.004  * 0.002 0.000 Before  Four  Ten Ten-Stream C only  Impact Period  Figure 3.3: Mean summer relative abundance (CPUE/m ± 1SE) of age + 1 and older cutthroat trout in control streams and logged streams for the before-impact period (1997 and 1998) and two after-impact periods (2000-2002 inclusively, 2008 and 2009). Some standard error bars are hidden by means symbols. Control-before n = 4; control-four n = 6; control-ten n = 4; Loggedbefore n = 4; logged-four n = 6; logged-ten n = 4 (Ten-Stream C grey symbol * Stream A is removed, n =2).  96  Relative winter abundance (CPUE/m)  0.007  Stream A Stream C East Creek Spring Creek  0.006 0.005 0.004 0.003 0.002 0.001 0.000 1998  1999  2000  2001  2002  2009  Figure 3.4: Winter relative abundance of cutthroat trout in treatments: Streams A and C, Creeks Spring and East, during (winter 1998) and after logging (1999 – 2009). Data were not available for streams in 1997 or for Spring Creek in the winters of 1998-1999. For all streams and years only single winter samplings occurred (note break between 2002 and 2009 years).  97  Trout summer body condition (log mass/length ratio)  Control Logged 0.99 0.98 0.97 0.96 0.95 0.94 0.93 0.92 Before  Four  Ten  Impact Period  Figure 3.5: Least-squared-mean (LSM) summer body condition (± 1 SE) (i.e. log mass-log length) of cutthroat trout in two unlogged (control) streams and one logged stream before (19971998) and after (10th year; 2008 - 2009) logging impact in Malcolm Knapp Research forest. Note that symbols hide some error bars.  98  Literature cited Beechie, T.J., G. Pess, P. Kennard, R.E. Bilby, and S. Bolton, 2000. Modeling recovery rates and pathways for woody debris recruitment in North Western Washington streams. North American Journal of Fisheries Management 20:436-452.  Benda, L., M.A. Hassan, M. Church, and C.L. May. 2005.Geomorphology of steepland headwaters: the transition from hillslopes to channel. Journal of American Water resources Association. 41: 835-851.  Benda, L.E., K. Andras, D. Miller, and P. Bigelow. 2004. Confluence effects in rivers: interactions of basin scale, network geometry, and disturbance regimes. Water Resources Research, Vol.40.  Bilby, R. E., and P.A. Bisson. 1992. Allochthonous organic matter contributions to the trophic support of fish populations in clear-cut and old-growth streams. Canadian Journal of Fisheries and Aquatic Sciences. 49: 540-551.  Bisson, P. A., R. E. Bilby, M. D. Bryant, C. A. Dolloff, G. B. Grette, R. A. House, M. L. Murphy, K. V. Koski, and J. R. Sedell. 1987. Large woody debris in forested streams of the Pacific Northwest: past, present, and future. Pages 143-190 in E. O. Salo and T. W. Cundy, eds. Streamside Management: Forestry and Fishery Interactions. Contribution no. 57, Institute of Forest Resources, University of Washington, Seattle.  Bjornn, T. C., and D. W. Reiser. 1991. Habitat requirements of salmonids in streams: influences of forest and rangeland management on salmonid fishes and their habitats. American Fisheries Society Special Publication 19: 83-138.  Boss, S.M., and J.S. Richardson. 2002. Effects of food and cover on the growth, survival, and movement of cutthroat trout (Oncorhynchus clarkii) in coastal streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 1044-1053.  Brown, J. H., and A. Kodric-Brown. 1977. Turnover Rates in Insular Biogeography: Effect of Immigration on Extinction. Ecology 58: 445-449.  Chapman, D.W. 1966. Food and space as regulators of salmonid populations in streams. The American Naturalist 100: 345 -357.  99  Coleman, M.A., and K.D. Fausch. 2007. Cold summer temperature regimes cause a recruitment bottleneck in age-0 Colorado River cutthroat trout reared in laboratory streams. Transactions of the American Fisheries Society 136: 639-654.  Cunjak, R. A., and G. Power. 1987. The feeding energetic of stream-resident trout in winter. Journal of Fish Biology 31: 493-511.  De Groot, J. D., S.G. Hinch, and J.S. Richardson. 2007. Effects of logging second growth forests on headwater populations of coastal cutthroat trout: a 6-year, multi-stream, before-and-after field experiment. Transactions of the American Fisheries Society 136: 211-226.  Elliottt, J. M. 1985. Population dynamics of migratory trout, Salmo trutta, in a Lake District stream, 1966–83, and their implications for fisheries management. Journal of Fish Biology 27: 35-43.  Elliottt, J. M. 1987. Population Regulation in Contrasting Populations of Trout Salmo trutta in Two Lake District Streams. Journal of Animal Ecology 56: 83-98.  Fausch, K. D., and T. G. Northcote. 1992. Large wood debris and salmonid habitat in a small coastal BC stream. Canadian Journal of Fisheries and Aquatic Sciences 49: 682-693.  Fausch, K. D., C. L. Hawkes, and M. G. Parsons. 1988. Models that predict the standing crop of stream fish from habitat variables: 1950–85. General Technical Report PNW-213. USDA Forest Service, Portland, Oregon, USA.  Gomi, T., R. C. Sidle, and J. S. Richardson. 2002. Understanding processes and downstream linkages of headwater systems. BioScience 52: 905–916.  Gomi, T., R. D. Moore, and A. S. Dhakal. 2006. Headwater stream temperature response to clearcut harvesting with different riparian treatments, coastal British Columbia, Canada. Water Resources Research 42: 1-11.  Gomi, T., R. D. Moore, and M. A. Hassan. 2005. Suspended sediment dynamics in small forest streams of the Pacific Northwest. Journal of the American Water Resources Association 41: 877-898.  Gregory, S. V., F. J. Swanson, W. A. McKee, and K. W. Cummins. 1991. An Ecosystem Perspective of Riparian Zones. BioScience 41: 540-551. 100  Gregory, S. V., G. A. Lamberti, D. C. Erman, K. V. Koski, M. L. Murphy, and J. R. Sedell. 1987. Influence of forest practices on aquatic production. Pages 233-256 in E. 0 .Salo and T. W. Cundy, editors. Streamside management: forestry and fishery interactions. Contribution No. 57.College of Forest Resources, University of Washington, Seattle,Washington, USA.  Harig, A.L., and K.D. Fausch. 2002. Minimum habitat requirements for establishing translocated cutthroat trout populations. Ecological Applications 12: 535-551.  Hawkins, C. P., J. L. Kershner, P.A. Bisson, M. D. Bryant, L. M. Decker, S.V. Gregory, D. A. McCullough, C. K. Overton, G.H. Reeves, R. J. Steedman, and M. K. Young. 1993. A Hierarchical Approach to Classifying Stream Habitat Features. Fisheries 18: 3-12.  Heggenes, J. T., O. M. W. Krog, O. R. Lindås, and J. G. Dokk. 1993. Homeostatic Behavioural Responses in a Changing Environment: Brown Trout (Salmo trutta) become nocturnal during winter. Journal of Animal Ecology 62: 295-308.  Hetrick, N. J,. M. A. Brusven, W. R. Meehan, and T. C. Bjornn. 1998. Changes in solar input, water temperature, periphyton accumulation, and allochthonous input and storage after canopy removal along two small salmon streams in Southeast Alaska. Transactions of the American Fisheries Society 127: 859-875.  Hicks, B. J., J. D. Hall, P. A. Bisson, and J. R. Sedell. 1991. Responses of salmonids to habitat changes, pages 483–518 in W. R. Meehan, editor. Influences of forest and rangeland management on salmonid fishes and their habitats. American Fisheries Society, Special Publication 19, Bethesda, Maryland.  Holtby, L. B. 1988. Effects of logging on stream temperatures in Carnation Creek, British Columbia, and associated impacts on the coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Sciences 45: 502-515.  Hynes, H.B.N. 1975. The Stream and its valley. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie 19: 1-15.  Johnson, S.W., Heifetz, J., and K.V. Koski. 1986. Effects of logging on the abundance and seasonal distribution of juvenile steelhead in some southeastern Alaska streams. North American Journal of Fish. Management 6: 532-537.  Johnston, N.T., and P.A. Slaney. 1996. Fish habitat assessment procedure. Watershed Restoration Tech. Circ. 8. Queen‟s Printers,Victoria, B.C. 101  Kiffney, P. A. 2008. Response of lotic producer and consumer trophic levels to gradients of resource supply and predation pressure. Okios 117: 1428 – 1440.  Kiffney, P. M., J. S. Richardson, and J. P. Bull. 2003. Responses of periphyton and insects to experimental manipulation of riparian buffer width along forest streams. Journal of Applied Ecology 40:1060–1076.  Krebs, C. J. 2001. Ecology: the experimental analysis of distribution and abundance 5th edition. Benjamin Cummings, San Francisco, California.  MacNutt, M. J., S. G. Hinch, A. P. Farrell, and S. Topp. 2004. Effects of thermal acclimation period on the repeat swimming performance of cutthroat trout (Oncorhynchus clarkii clarkii). Journal of Fish Biology 65:342–353.  Meehan, W. R., editor. 1991. Influence of forest and rangeland management on salmonid fishes and their habitats. American Fisheries Society, Special publication 19, Bethesda, Maryland.  Mellina, E., and S. G. Hinch. 2009. Influences of riparian logging and in-stream large wood removal on pool habitat and salmonid density and biomass: a meta-analysis. Canadian Journal of Forest Research 39: 1280-1301.  Mellina, E., Hinch, S.G., and K.D. MacKenzie. 2005. Seasonal movement patterns of streamdwelling rainbow trout in north-central British Columbia, Canada. Transactions of the American Fisheries Society 134: 1021-1037.  Merten, E.C., N.A.Hemstad, S.L. Eggert, L.B. Johnson, R.K. Kolka, R.M. Newman RM, and B. Vondracek. 2010. Relations between fish abundances, summer temperatures, and forest harvest in a northern Minnesota stream system from 1997 to 2007. Ecology of Freshwater Fish 19: 63-73.  Minitab. 2007. Statistical software for Windows, release 15.1. Minitab, Inc., State College, Pennsylvania.  Montgomery, D.R., and M. J. Buffington. 1997. Geological Surveyors of America Bulletin 109: 596-611.  102  Moore, R. D., P. Sutherland, T. Gomi, and A. Dhakal. 2005. Thermal regime of a headwater stream within a clear-cut, coastal British Columbia, Canada. Hydrological Processes 19:2591-2608.  Murphy, M. L., and K. V. Koski. 1989. Input and depletion of woody debris in Alaska streams and implications for streamside management. North American Journal of Fisheries Management 9:427-436.  Murphy, M. L., J. Heifetz, S. W. Johnson, K. V. Koski, and J. F. Thedinga. 1986. Effects of clear-cut logging with and without buffer strips on juvenile salmonids in Alaskan streams. Canadian Journal of Fisheries and Aquatic Sciences 43:1521-1533.  Murphy, M.L., and J. D. Hall. 1981. Varied effects of clear-cut logging on predators and their habitat in small streams of the Cascade Mountains, Oregon. Canadian Journal of Fisheries and Aquatic Sciences 38: 137-145.  Northcote, T. G., and G. F. Hartman, editors. 2004. Fishes and forestry: worldwide watershed interactions and management. Blackwell Scientific Publications, Oxford, UK.  Resh, V.H., A.V. Brown, A.P. Covich, M.E. Gurtz, H.W. Li, G.W. Minshall, S.R. Reice, A.L. Sheldon, J.B. Wallace, and R. Wissmar. 1988. The role of disturbance in stream ecology. Journal of the North American Benthological Society 7:433-455.  Rosenfeld, J., M. Porter, and E. Parkinson. 2000. Habitat factors affecting the abundance and distribution of juvenile cutthroat trout (Oncorhynchus clarkii) and coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science 57: 766-774.  Schlosser, I. J. 1991. Stream Fish Ecology: A Landscape Perspective. BioScience 41: 704-712.  Sedell, J.R. and Swanson, F.J., 1984. Ecological characteristics of streams in old-growth forests of the Pacific Northwest. In: W.R. Meehan, T.R. Merrell, Jr. and T.A. Hanely (Editors), Fish and Wildlife Relationships in Old-Growth Forests Proceedings Symposium. 12-15 April 1982, at Juneau, AK. American Institute of Fishery Research Biologists p. 9-16.  Speirs, D. C., and W. S. C. Gurney. 2001. Population persistence in rivers and estuaries. Ecology 82: 1219–1237.  Thedinga, J. F., M.L. Murphy, J. Heifetz, K.V. Koski, and S.W. Johnson. 1989. Effects of logging on size and age composition of juvenile coho salmon (Oncorhynchus kisutch) 103  and density of pre-smolts in southeast Alaska streams. Canadian Journal of Fisheries and Aquatic Sciences 46: 1383-1391.  Ward, J. V., K. Tockner, D. B. Arscott, and C. Claret. 2002. Riverine landscape diversity. Freshwater Biology 47: 517–539.  Wipfli, M. S. 1997. Terrestrial invertebrates as salmonid prey and nitrogen sources in streams: contrasting old-growth and young-growth riparian forests in southeastern Alaska, USA Canadian Journal of Fisheries and Aquatic Sciences 54: 1259-1269.  Wipfli, M. S., and J. Musslewhite. 2004. Density of red alder (Alnus rubra) in headwaters influences invertebrate and detritus subsidies to downstream fish habitats in Alaska. Hydrobiologia 520: 153-163. Young, K. A. 2000. Riparian zone management in the Pacific Northwest: who‟s cutting what? Environmental Management 26: 131-144.  Young, K. A., S. G. Hinch, and T. G. Northcote. 1999. Status of resident coastal cutthroat trout and their habitat twenty-five years after riparian logging. North American Journal of Fisheries Management 19:901-911.  Zhang, Y., J.S. Richardson, and X. Pinto. 2009. Catchment-scale effects of forestry practices on benthic invertebrate communities in Pacific coastal streams. The Journal of Applied Ecology 46: 1292-1303.  104  Chapter 4 : Summary To meet our aims to conserve or protect freshwater species and resources the importance of small coastal streams, and their contributions to the local and regional biodiversity, should not be under appreciated at a landscape or local scale. This study increased knowledge of the population dynamics of a threatened species of stream-dwelling trout which commonly resides in small streams within the Pacific Northwest region of North America. While extrapolation of observations from empirical studies should be made with caution, with consideration for differences in geographic region, topographic, climatic and socio-economic conditions, the observations from this study can contribute to applied management of trout-bearing streams. Measures of stream size (i.e., bankfull width) are commonly used in stream protection guidelines, e.g., Forest and Range Practices Act (FRPA). However, the results of this study show that stream size is likely relevant to the life history and ecology of the animal of interest. Furthermore, while readily obtained criterions for stream protection are desirable, such as stream width, they may not be ecologically applicable for all conservation objectives. For this study, measures of bankfull width did not predict the amount of habitat available to trout, or their probability of survival across seasons. Whereas, many studies have shown that water depth is a critical feature for trout in small (< 6 m bankfull) streams. Hence, for resident trout living in the smaller size range of streams (1st and 2nd order) common to the coastal PNW landscape, the physical width of streams may not be as important an indicator for population size and viability as the depth of habitat provided by a stream. Given that population responses to flow alteration, human induced or climatic, are of current concern for many fisheries and water resource managers, we cannot assume that stream width is the best criterion for protecting, and predicting trout abundance or persistence.  105  Seasonality, in relation to availability of habitat, had a significant influence on trout demographics, with the summer period having higher mortality than the winter period. In contrast to my observations, other studies have inferred that winter mortality can be high for stream salmonids (Cunjak and Power 1987; Cunjak 1996). This may indeed be so for fish in streams and rivers of eastern and central North America. However, along the Pacific ecoregion of western North America climatic patterns are influenced by the proximately of the Pacific Ocean and coastal topography, with mild winters and warm summer periods. Therefore, within this region trout populations in small streams may experience the highest or equally high mortality in the summer period as compared to the winter period. Consequently, management for both summer and winter habitat should be considered for small streams in coastal British Columbia. For example, large wood inputs provide the dual service of pool creation from the summer period and shelter from high flows during the winter period. In manipulated or managed streams, where flows may be regulated or altered by land use, the creation or maintenance of a series of pools (> 25 cm depth at low flows) throughout trout bearing reaches may be advisable. In addition, while such management activities may provide habitat, it should also be noted that pool systems may need monitoring for pool volume as flows diminish. Without ground water inputs pool volumes may decrease with reduced surface flows, resulting in the unintended stranding of fish in pool refugia. Therefore, the connectivity between reaches, ground water inputs, and monitoring of surface flows and pool volumes should be considered in highly managed or regulated systems. Historically forest harvesting has been observed to have both positive and negative effects on salmonid populations. Due to these observations forest harvesting practices have been altered, with best practices incurring less in-stream disturbance during harvesting operations. 106  Retention of large wood for cover, and creation of aquatic habitat appears to have had benefits or mediating effects for trout in this study, as well as in others (e.g., Mellina and Hinch 2009). The observed responses of trout in our study streams were attributed to the cooler, temperate climate of the study region. The removal of streamside vegetation in our large-scale experimental stream manipulation did not appear to lead to lethal or sub-lethal stream temperatures for trout populations. However, decreased shading did see temperature increases of a few degrees Celsius. In the case of our study area a few degree increase in water temperature may be beneficial for increased growth rates in trout, possibly translating to higher survival of young, rapid-growth fish. For example, during the spring and summer periods fish growth rates can be significantly higher than during the colder winter period. These higher growth rates can be related to the benefits of increased seasonal inputs of food and warmer stream temperatures. Yet it should be cautioned that these results are region-specific. Conversely, in lower elevation or lower latitude regions where climates may be hotter, increased stream temperatures (above 23°C) could have negative effects on these fish. In hotter climates streamside removal of shadeproviding vegetation may lead to increased metabolic demands in fish, increased occurrence of disease, decreased dissolved oxygen and lethal thermal regimes. Considered together results from this study would indicate that movement, connectivity and seasonal refugia are key factors for trout conservation in relation to geographic location. While not directly quantified, there were indications in our study that human-induced isolation of sub-populations in some stream reaches may have contributed to decreased population persistence in altered research streams. Trout populations in streams with impassable barriers and low supply of upstream habitat area showed patterns of decline over time relative to other populations (i.e., Upper East Creek and Stream A). For example, while Upper Spring and Stream H do have man-made barriers, Stream H drains from a lake and Upper Spring has a  107  supply of marshlands and secondary habitat. Therefore, trout populations appear able to persist above barriers when they are have sufficient habitat to maintain their local populations, or when they have the possibility of connectivity to other populations via intermittent stream networks and lakes. These observations suggest that fisheries managers should consider local scales of habitat, reach-scale movement and landscape scale connectivity for trout population management strategies. It is recommended that if in-stream structure must be placed within stream reaches, that fish ways or ladders are installed when connectivity may be decreased. In addition, culverts can be barriers to fish and should be inspected to ensure fish passage is maintained over time, and during spawning season. These recommendations are of specific importance when trout are known to exist in the upper extents of a stream network, or movement rates have begin estimated as high. For example, high movement rates may indicate a high frequency of refuge or trophic migrations, or movement between seasonal habitats and spawning grounds (e.g., Gresswell and Hendricks 1997). Mark-recapture studies are influenced by recapture frequencies which may be affected by tag loss, movement rates and behavioural heterogeneity. These effects can contribute to uncertainty or bias in survival estimates derived from capture histories. An assumption of the CJS model is that marks will not be lost, and during this study it was observed from scar marks that 4.8% of fish had lost their marks. While undetected tag loss can bias survival estimates toward slightly lower mean survival rates, all trout with lost tags in this study were re-marked and interval-specific recapture histories were adjusted to account for the 4.8% tag loss. Therefore, recapture rates based on marks are likely to be reasonably accurate. Secondly, between-habitat movement was expected in our wild fish populations and is a known aspect of their ecology (e.g., Berger and Gresswell 2009). This was accounted for in the spatial scale of the sampling design. In comparison, undetected permanent movement outside of the study could  108  have biased survival estimates resulting in lower mean apparent survival rates. Since trout were able to move up- and downstream in 3 out of 8 of the study streams, it is feasible that some trout were alive but had left the study area and were unavailable for capture. However, throughout the time period of this study movement rates in most streams were low (i.e., estimated mean movement ranged from 5 – 20 m over 12 months), and movement outside of the study area was observed to be absent or minimal i.e., 0.05% (1 out of 20) were found to have moved up to 30 m outside of study boundaries. Thirdly, behavioural heterogeneity can affect patterns in survival estimates. Homogeneity of behaviour is an assumption in the CJS model, and using primary and secondary capture periods in the study design accounted for this assumption by providing robust recapture estimates (Krebs 1999). In addition, capture frequencies were statistically tested with contingency tables and there was no indication that individual heterogeneity within the population estimates were affecting recapture probabilities. In conclusion, the results of this study suggest that forestry and fisheries managers should be aware of connectivity and availability of fish habitat at a landscape and local level when making decisions for aquatic conservation. In the PNW seasonality can influence habitat, especially during low flow periods, and movement rates of fish may increase during these periods. As such riparian forestry, stream channel alterations, or water use which affects stream hydrological processes through sudden changes in flow that do not reflect natural flow regimes should be approached with consideration for ensuring a minimum depth criteria, and assessment of spatial and temporal habitat connectivity within the stream network.  109  Literature cited Berger, A.M., and R.E. Gresswell. 2009. Factors influencing coastal cutthroat trout (Oncorhynchus clarkiii clarkiii) seasonal survival rates: a spatially continuous approach within stream networks. Canadian Journal of Fisheries and Aquatic Sciences 66: 613632.  Cunjak, R. A., and G. Power. 1987. The feeding energetic of stream-resident trout in winter. Journal of Fish Biology 31: 493-511.  Cunjak, R. A. 1996. Winter habitat of selected stream fishes and potential impacts from landuse activity. Canadian Journal of Fisheries and Aquatic Sciences 53:267-282  Gresswell, R.E., and S.R. Hendricks. 2007. Population-scale movement of coastal cutthroat trout in a naturally isolated stream network. Transactions of the American Fisheries Society 136: 238-253. Krebs, C. J. 1999. Ecological methodology 2nd Ed. Addison-Wesley Educational Publishers, Inc., Don Mills, Ontario p: 49-70.  Mellina, E., and S. G. Hinch. 2009. Influences of riparian logging and in-stream large wood removal on pool habitat and salmonid density and biomass: a meta-analysis. Canadian Journal of Forest Research 39: 1280-1301.  110  Appendices Appendix A: Number of total captures and recaptures for each primary and secondary sampling period for each of seven streams in Malcolm Knapp research forest (between May 20, 2008 to May 20, 2009). T = Total marked R = recapture. Session Stream  Spring 2008 Summer 2008 Winter 2009 Spring 2009 # traps T R # traps T R # traps T R # traps T R 45 8 0 45 73 9 43 60 18 43 50 17 45 3 0 45 33 4 43 25 7 42 23 12 C 45 2 0 45 27 4 43 11 4 42 25 7 45 33 4 42 14 9 43 10 0 45 33 6 44 42 15 43 20 7 45 20 0 45 6 1 44 22 11 43 13 3 H 45 25 1 43 7 3 44 19 5 43 9 3 45 28 4 43 7 4 45 26 0 42 74 36 42 18 13 42 51 18 45 15 1 41 21 11 42 17 9 43 22 11 Lower East 45 13 3 43 30 16 42 12 9 43 24 8 41 49 17 42 37 18 45 29 0 45 46 22 45 17 13 43 21 11 45 15 0 45 12 9 45 5 4 43 2 2 Upper East 45 9 4 43 3 1 45 3 1 43 6 2 45 19 5 42 29 24 44 81 0 45 145 46 44 36 14 45 130 66 44 22 0 45 57 26 44 49 27 45 21 9 Lower Spring 44 32 3 41 25 14 44 26 13 45 20 9 45 39 10 45 44 27 45 101 0 45 99 43 45 58 42 45 90 57 45 7 2 45 32 19 45 20 12 45 40 29 Upper Spring 45 24 9 45 38 29 45 22 18 45 24 20 45 38 16 45 37 21 46 73 0 44 102 31 45 82 45 44 41 26 45 17 1 44 13 6 45 59 44 44 7 5 Blaney 45 28 4 44 17 11 45 37 27 46 51 10 41 53 7 46 48 31  111  Appendix B: A University of British Columbia Research Ethics Board certificate of approval was issued to Professor Scott Hinch (a co-author and advisor) which permitted me to conduct research on wild trout. Document is attached following.  112  113  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0070993/manifest

Comment

Related Items