The Effects of Logging and Mass Wasting on Juvenile ... · The Effects of Logging and Mass Wasting on Juvenile Salmonid Populations in Streams on the Queen Charlotte Islands Land

The Effects of Logging and MassWasting on Juvenile Salmonid Populations in Streams on the Queen Charlotte Islands

Land ManagementReport NUMBER

Ministry of Forests

80ISSN 0702-9861

NOVEMBER 1992

1992

Ministry of ForestsResearch Program

The Effects of Logging and Mass Wastingon Juvenile Salmonid Populations in

Streams on the Queen Charlotte Islands

D.B. Tripp and V.A Poulin

FISH/FORESTRY INTERACTION PROGRAM

SEQ 6543 JOB FISH-000-011 PAGE-0001 FRONT MATTER REVISED 21SEP00 AT 08:17 BY BC DEPTH: 62 PICAS WIDTH 43 PICAS COLOR LEVEL 1

iii

SUMMARY

The effects of logging and mass wasting on juvenile coho salmon (Oncorhynchus kisutch), steelhead trout(O. mykiss, formerly Salmo gairdneri), and Dolly Varden char (Salvelinus malma) were assessed in streamson the Queen Charlotte Islands. Fish densities and habitat characteristics of 27−33 stream reaches weremeasured during summer and fall. Reaches sampled included undisturbed old-growth forest streams(unlogged), logged streams not directly affected by recent mass wasting (logged), and logged streamsdirectly affected by recent debris torrents and slides (mass wasted). Overwinter survivals and smolt yieldsin three mass wasted and three non-mass wasted streams (all logged) were also estimated in a down-stream spring fish trapping program, after determining the number of fish present in each stream theprevious fall.

Logged reaches had less undercut bank cover than unlogged reaches, but did not differ significantlyfrom unlogged areas in any other habitat variable measured. Mass wasted stream reaches, in contrast, hadeven less undercut bank cover, less large organic debris (LOD), fewer pools and glides, and more riffles.They also had shallower pools during summer, a smaller wetted stream width relative to rooted channelwidth, and less overwinter cover in the form of deep pools with undercut banks and abundant LOD.

With one exception, there was no relationship between summer and fall fish densities and any of thehabitat parameters measured in this study. The exception was the depth of gravel scour overwinter, whichappeared to determine the early summer abundance of coho fry in mass and non-mass wasted streams (alllogged). Logged reaches had significantly higher coho fry densities than unlogged or mass wasted reachesin summer and fall. Fish in mass wasted reaches exhibited faster growth rates and attained larger sizes, aslong as fry were not trapped in isolated pools when reaches ‘‘dewatered’’. In mass wasted streams, acombination of poor egg-to-fry survivals due to excess gravel scour, and poor juvenile overwinter survivalsdue to overwinter habitat loss, nullified any gains in production attributable to logging. It also nullified thehigh growth rates and large size achieved by fish in their first year in mass wasted streams. Juvenileoverwinter survivals for all species were 2.1−3.5 times higher in non-mass wasted streams than in masswasted streams; smolt yields were 1.5−3.3 times higher.

The overall impacts of mass wasting on juvenile fish, and coho salmon in particular, are seriousenough to jeopardize the continued existence of self-sustaining populations in directly affected reachesuntil stream conditions improve. Four out of 11 mass wasted reaches in 1982 and 2 out of 3 mass wastedreaches in 1984 had effectively no coho fry. Impacts on Dolly Varden and steelhead trout did not appear asserious, though they too showed declines in overwinter survivals and smolt yields. Impacts on other speciessuch as chum and pink salmon were not investigated, though presumably these species would benegatively stressed by increased gravel scour. Fish populations in otherwise normal (logged) reachesdownstream of major mass wasting events may also be adversely affected by mass wasting upstream, butthe problem requires further study.

SEQ 6544 JOB FISH-000-011 PAGE-0002 FRONT MATTER REVISED 21SEP00 AT 08:17 BY BC DEPTH: 62 PICAS WIDTH 39 PICAS COLOR LEVEL 1

iv

ACKNOWLEDGEMENTS

Sincere appreciation is extended to the many individuals who contributed to this study. Special thanksespecially to Jeff Cederholm who worked closely with the project in its initial stages and who reviewed anumber of preliminary reports. Thanks also to those who reviewed the present manuscript and providedmany valuable comments: Al Cowan, Dionys DeLeeuw, Dr. Tom Northcote, Bill Pollard, Charles Scrivener,and Pat Slaney.

Ted Harding, with the assistance of Brian Eccles, Dave Davies and Mary Morris, carried out thesynoptic survey, while technical assistance for the detailed studies was provided by Ted Bellis, DonCadden, Charles Rally, and Glen Kendall. The study was undertaken as part of the Fish/Forestry InteractionProgram, an interdisciplinary research program funded by the Government of Canada Department ofFisheries and Oceans, British Columbia Ministry of Forests, and British Columbia Ministry of Environment,Lands and Parks. A Government of Canada Fisheries Employment Bridging Assistance Program providedadditional field staff in 1982 and 1983.

SEQ 6545 JOB FISH-001-009 PAGE-0001 TOC REVISED 21SEP00 AT 08:17 BY BC DEPTH: 62 PICAS WIDTH 43 PICAS COLOR LEVEL 1

v

TABLE OF CONTENTS

SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 STUDY DESIGN AND STREAM CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.1 Study Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Synoptic Survey Stream Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3 Detailed Study Stream Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.1 Reach Selection and Habitat Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2 Fish Population Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.3 Spring Downstream Trapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.4 Habitat Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.5 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4.1 Stream Habitats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94.2 General Distribution and Abundance of Fish Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124.3 Habitat Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134.4 Effects of Logging and Mass Wasting on Fish Abundance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154.5 Upstream/Downstream Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.6 Effects on Coho Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.7 Effects on Biomass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.8 Smolt Yields and Overwinter Survivals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234.9 Coho Fry Recruitment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

5 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

5.1 Effects of Logging and Mass Wasting on Rearing Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275.2 Effects on Coho Salmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

5.2.1 Logging impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285.2.2 Mass wasting impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285.2.3 Downstream effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

5.3 Effects on Other Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315.4 Management Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.5 Recommendations for Further Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

LITERATURE CITED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

SEQ 6546 JOB FISH-001-009 PAGE-0002 TOC REVISED 21SEP00 AT 08:17 BY BC DEPTH: 62 PICAS WIDTH 39 PICAS COLOR LEVEL 1

vi

TABLES

1 Synoptic survey study reach characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Stream habitat characteristics in the full and reduced sets of unlogged, logged and masswasted study reaches in the synoptic survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Habitat characteristics of detailed study streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4 General distribution characteristics, abundance and age composition of juvenile salmonidpopulations in 1982 synoptic survey reaches on the Queen Charlotte Islands . . . . . . . . . . . . . . . 12

5 Mean juvenile salmonid densities in the full set of unlogged, logged and mass wasted(logged) stream reaches sampled on the Queen Charlotte Islands in 1982 . . . . . . . . . . . . . . . . . . 16

6 Juvenile coho salmon, Dolly Varden char and steelhead trout densities (number of fish permetre of stream) in the detailed study streams, fall (Sept. 12−Oct. 6, 1983) and summer(July 13−23, 1984) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

7 Mean juvenile salmonid densities in the reduced set of unlogged, logged and mass wasted(logged) stream reaches sampled on the Queen Charlotte Islands in 1982 . . . . . . . . . . . . . . . . . . 17

8 Coho fry densities (number of fish per metre of stream) and sample dates in logged streamreaches, with and without mass wasting upstream . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

9 Mean coho fry fork length, coho fry density, total salmonid density and total salmonidbiomass in selected streams on the Queen Charlotte Islands, 1982 . . . . . . . . . . . . . . . . . . . . . . . . 20

10 Annual variation in age composition and fry growth of coho salmon in detailed studystreams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

11 Salmonid biomass (grams of fish per metre of stream) in unlogged, logged and mass wasted(logged) stream reaches on the Queen Charlotte Islands in 1982 . . . . . . . . . . . . . . . . . . . . . . . . . . 22

12 Densities, smolt yields and apparent overwinter survivals of juvenile coho salmon, DollyVarden char and steelhead trout in each detailed study stream, October 1983−July 1984 . . . . 23

FIGURES

1 Location of synoptic survey (•) and detailed study streams (∗) on the Queen CharlotteIslands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Length frequency distributions of coho salmon, steelhead trout and Dolly Varden char insynoptic survey streams on the Queen Charlotte Islands, 1982−1984 . . . . . . . . . . . . . . . . . . . . . . . 13

3 Habitat preferences and densities of juvenile salmonids in pools, glides and riffles . . . . . . . . . . . . 14

4 Seasonal variation in mean fork length for coho salmon juveniles in unlogged, logged andmass wasted stream reaches on the Queen Charlotte Islands, June 9−September 30, 1982 19

5 Seasonal variation in juvenile coho salmon growth for the 1983 and 1984 year classes inmass and non-mass wasted streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

6 Downstream movements of juvenile coho salmon in mass and non-mass wasted studystreams, April 10−June 29, 1984 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

7 Correlation between mean scour depth and coho fry density (spring outmigrants plus Julyresident fish combined) in detailed study streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

SEQ 6573 JOB FISH-008-005 PAGE-0001 CHAP 1

1 INTRODUCTION

Widespread mass wasting in the form of debris slides, flows and torrents (Varnes 1978) is evident on thesides of virtually every steep watershed on the Queen Charlotte Islands (Gimbarzevsky 1988). Whenclearly unstable areas such as these are logged, the volume of soil and debris mobilized by mass wasting isincreased 34 times over the volume observed in forested areas, while the volume introduced into fishbearing streams is increased 43 times (Rood 1984). Open slope debris slides are the most common type ofslope failure events on the Queen Charlotte Islands, but they rarely enter streams and therefore rarelycause immediate damage to fish or fish habitats. Most of the sediment that enters streams directly istransported by debris flows or torrents that can damage fish habitat extensively through excessive streamscouring and the deposition of large quantities of sediment and debris.

On the Queen Charlotte Islands, debris torrents in small and medium size streams have the greatestimpacts in stream reaches with a gradient less than 7% (Tripp and Poulin 1986a). In these zones, pooldepth, pool area, LOD (large organic debris) cover and stable undercut bank area are reduced, while rifflearea, channel width and the degree of ‘‘dewatering’’ that occurs during low flow periods are all increased.Fish overwinter habitat especially — in the form of deep pools with abundant LOD cover and stableundercut banks — is lost in the mainstem portions of affected streams, a critical factor that may severelylimit juvenile survival and smolt yield if protection from high flows is lacking elsewhere. Gravel scouring inspawning areas is also increased (Tripp and Poulin 1986b), which may further limit production if the numberof eggs or alevins that survive to emerge from the gravel is too low to fill or ‘‘seed’’ the available rearinghabitat.

The effects of debris torrents on juvenile fish survivals and smolt yields in mass wasted streams on theQueen Charlotte Islands have yet to be measured directly. Though the damage caused by torrents in thedirectly affected reaches of small streams is obviously great, studies on the effects of logging elsewhereindicate that a number of processes can help offset or compensate for declines in the quantity or quality offish habitat present. Higher intragravel water temperatures during winter, for example, can accelerate fryemergence and extend the growing season, resulting in larger fish in fall with a better chance of survivingthe winter (Scrivener and Anderson 1984; Holtby 1988). Increased light in the stream after the surroundingforest is removed may also stimulate fish growth through increased periphyton and macroinvertebrateproduction (Hawkins et al. 1983; Bisson and Sedell 1984; Murphy et al. 1986). Presumably similarprocesses occur in streams on the Queen Charlotte Islands, but it is not known whether they are sufficientto offset the damage caused by debris torrents on the scale common in the steeper regions of the QueenCharlotte Islands.

The main objective of this study was to assess the effects of mass wasting on the growth, abundanceand overwinter survivals of juvenile salmonids in directly affected streams on the Queen Charlotte Islands.Because many mass wasted stream reaches on the Queen Charlotte Islands are also logged, the studyalso assessed the effects of logging to distinguish the effects of mass wasting. This is the third of threerelated reports on logging and mass wasting under the Fish/Forestry Interaction Program (Poulin 1984) inBritish Columbia. The first report described the impacts of mass wasting on juvenile fish rearing habitat(Tripp and Poulin 1986a); the second report described the effects of logging and mass wasting on salmonidspawning habitats and egg-to-fry survivals (Tripp and Poulin 1986b).

SEQ 6547 JOB FISH-002-016 PAGE-0002 CHAP. 2 REVISED 21SEP00 AT 08:17 BY BC DEPTH: 60 PICAS WIDTH 42.10 PICAS COLOR LEVEL 1

2

2 STUDY DESIGN AND STREAM CHARACTERISTICS

2.1 Study Design

To assess the effects of mass wasting on juvenile fish, the densities of coho salmon (Oncorhynchuskisutch), steelhead trout (O. mykiss, formerly Salmo gairdneri), and Dolly Varden char (Salvelinus malma)were first measured in a synoptic survey of 33 stream reaches in 29 small and medium sized streams on theQueen Charlotte Islands (Figure 1). This project was followed by a second, more detailed study onoverwinter survivals in six streams previously sampled in the synoptic survey. In the synoptic survey, threetypes of stream reaches were compared: undisturbed old-growth forest streams (unlogged), loggedstreams not directly affected by recent mass wasting (logged), and logged streams directly affected byrecent debris torrents and slides (mass wasted). In the detailed study, only logged streams with a range ofmass wasting conditions were compared.

The synoptic survey portion of the study was conducted in 1982, during which all 33 stream reacheswere sampled once for fish and habitat in summer (June 15−August 23), and 27 were sampled again forfish in fall (September 1−30). Streams in the overwinter survival study were sampled in 1983−1984, once infall (September 12−October 5, 1983) to determine the number of fish present before winter, once in spring(April 16−June 29, 1984) to enumerate smolts and fry migrating downstream, and again in summer (July13−22, 1984) to measure habitat and estimate the number of fish present after downstream movementshad ceased.

2.2 Synoptic Survey Stream Characteristics

The location, drainage area, gradient characteristics and dimensions of each study reach in the synopticsurvey portion of the study are presented in Table 1, together with a brief description of the logging andmass wasting conditions within and above each reach.

Study streams were first to third order (Strahler 1957) gravel and cobble bedded streams, 17 of whichwere located within the Queen Charlotte Ranges and 12 in the Skidegate Plateau (Figure 1, inset). Averagedrainage area of the study streams was 10.5 km2 (range 1.5−47.5 km2 ); average drainage area of the studyreaches was 8.4 km2 (range 1.5−20.0 km2). Average channel gradient of the study reaches was 2.4%(range 0.8−6.1%), average length 530 m (range 215−1100 m), and average width 22.5 m (range 6.4−55.1 m).

Most study reaches started at the mouth of the stream. Some, however, were 5−9 km upstream. Averagedistance upstream was 760 m (range 0−8600 m). Almost all the study reaches contained juvenile coho salmonand Dolly Varden; about two-thirds also contained steelhead trout. Most of the reaches are used for spawning bychum salmon (O. keta), and about a third, mainly the larger third order streams, are used by pink salmon (O.gorbuscha).

None of the streams are used by sockeye (O. nerka) or chinook (O. tshawytscha) salmon, and only two(Phantom and Sachs creeks) had cutthroat trout (O. clarki, formerly Salmo clarki). The latter appeared to beresident populations, inasmuch as they were restricted in their distribution to the headwaters of each stream, wellupstream of the study reaches. Sculpins (Cottus aleuticus and C. asper) were common but were not included in theanalysis.

Unlogged study reaches (N = 11) were all in old-growth forests of hemlock, cedar and Sitka spruce, and, withthe exception of three reaches (Cache 1 and 2, Phantom 1; Table 1), unlogged reaches were also unloggedupstream. Cache Creek had the upper one-third of its drainage logged 1−7 years before the study, but no obviouseffects of the logging showed in the study reaches downstream. Phantom Creek, in contrast, had several activegravel deposits within the study reach, which may have been started by a recent slide out of a logged area 800 mupstream. Phantom Creek and three other unlogged reaches also had recent (<4 years old) debris torrents inunlogged areas upstream (Table 1). Located within 120−400 m of the study reach, two of the torrents (Windy BayCreek tributary and Government Creek) may have indirectly influenced fish populations in the study reachdownstream. The other two torrents (in Phantom Creek and Matheson Side Creek) were 2.5−4.5 km upstream ofthe study reaches.

3

FIGURE 1. Location of synoptic survey (•) and detailed study streams (∗) on the Queen Charlotte Islands.


4

TABLE 1. Synoptic survey study reach characteristics.

PeriodReach

Basin Distancea Reach ChannelGrad.

% Periodb

upperno.

area upstream length width(%)

Basin reachreaches Remarks

(km2) (m) (m) (m) logged loggedlogged

Unlogged Reaches

Cache 1 1.5 0 450 6.5 2.5 33 Unlogged H1975-81Cache 2 1.3 450 280 6.4 5.1 38 Unlogged H1975-81Government 1 6.4 730 630 24.5 2.0 0 Unlogged Unlogged Headwaters torrented pre-1940, 1972, 1978Hangover 1 20.1 250 490 29.5 1.5 0 Unlogged Unlogged Large 155-year-old rock slide upstreamInskip 1 13.0 0 490 55.1 2.6 0 Unlogged Unlogged Old slides upstream (undated)Marshall 1 2.3 0 270 17.7 3.9 0 Unlogged UnloggedMatheson Head 1 6.7 0 500 18.8 1.1 0 Unlogged UnloggedMatheson Side 1 5.0 0 490 25.5 1.8 0 Unlogged Unlogged Headwaters torrented 1978−79Phantom 1 18.3 800 500 32.1 1.2 3 Unlogged H1978-81 Headwaters torrented 1979; slides in logged areasSalmon 1 10.7 0 600 38.9 1.5 0 Unlogged UnloggedWindy Bay Trib. 1 2.7 0 640 15.8 2.9 0 Unlogged Unlogged Headwaters torrented 1978−79

Logged Reaches

Landrick 1 2.3 0 450 11.0 3.8 43 A36 H78-82Peel Road 1 11.6 0 550 18.0 1.4 15 H50’s H50’sPiper 1 4.2 0 600 7.7 2.2 10 A55 H76-79 1983 summer flood damage extensiveSachs 1 17.8 0 1100 24.4 0.8 62 pre 56 H74-82 Headwaters torrented 1974Sachs 2 16.9 1100 800 ND 1.7 60 pre 56 H74-82 Headwaters torrented 1974Schomar 1 6.6 0 400 17.6 2.1 36 H65-67 H70-76 Torrented upstream 1979; study area

impacted 1984Talunkwan 1 3.9 0 360 17.8 2.3 36 A45 H69-73 Torrented upstream 1978−79Tarundl 1 11.0 0 540 29.0 1.1 37 H64 H64-82Tarundl 2 9.6 1600 500 11.6 4.0 28 H72-78 H78-82 5 m debris jam at end of reachThurston 1 6.1 10 700 8.0 2.0 72 A,S39-48 H68-75 Torrented upstream 1978−79Riley 1 13.6 5150 300 29.3 1.8 12 H74 H77-80 Torrented upstream 1978−79

Mass Wasted Reaches

Bonanza 1 20.0 8600 550 24.7 1.3 3 H80-82 Unlogged 1982 torrent deposition zone; channel aggradedMacmillan 1 6.2 0 400 12.3 3.1 77 H46 H76-82 Torrented 1978−79Mosquito Trib. 1 5.2 400 740 23.4 2.5 17 H60’s H60’s Banks eroded, channel wide and heavily aggradedMountain 1 12.8 0 515 34.9 1.1 9 S58 H65-67 Banks eroded, channel wide and heavily aggradedRiley 2 13.3 5600 400 44.1 2.5 8 H74 H77-80 Torrented 1978−79; extensive aggradationRiley Trib. 1 3.8 300 500 11.0 4.1 3 H74 Unlogged Banks eroded; channel torrent-likeSaltspring 1 6.3 0 240 41.9 2.1 13 H65 Unlogged Torrented 1978−79, 1981−82, 1983−84Shelley 1 5.4 0 495 19.9 2.9 17 H72 H73 Torrented 1978−79Southbay 1 4.0 0 420 24.7 3.6 82 H60 H61-80 Torrented 1978−79, 1983; new LOD added 1982Southbay 2 3.7 610 280 28.5 6.1 80 H61-67 H61-80 Torrented 1978−79, 1983Two Torrent 1 3.8 150 750 9.9 4.0 20 H65-66 H65-66 Torrented 1978−79

a Distance from stream mouth to start of reach.b A = A-frame logging; S = skidder logging; H = highlead logging.

Logged study reaches (N = 11) had a range of logging histories dating back between 4 and 46 years.Five reaches were A-frame or skidder-logged before the mid-1950’s; the rest were highlead-logged fromthe late 1950’s to the present. All logged reaches were more recently logged upstream with highleadsystems. Average age after logging within each study reach was 24 years, while the average age since thestart and end of logging upstream was 10 and 5 years, respectively. Two reaches (Riley 1 and Tarundl 2)had narrow buffer strips left alongside them after they were logged, but only one reach (Tarundl 2) hadsome of these trees still standing during the survey. Because dense stands of red alder dominated theriparian zone of all logged reaches, those reaches with no mass wasting were usually heavily shadedduring summer. Four reaches also had recent (<4 year old) debris torrents that stopped an average of 550m (range 0−1240 m) above the study reach.

Eight of the 11 mass wasted stream reaches selected for study had been directly affected by one ormore debris torrents within the past 3−4 years; 3 others had been affected by a combination of torrents,slides and accelerated bank erosion over the past 3−24 years (Roberts and Church 1986). All mass wastedreaches were in previously logged areas, and all except three were logged upstream. Average age sincethe end of logging in each reach was 16 years, and average age since logging started and stoppedupstream was 12 and 6 years, respectively. As for logged streams, dense stands of red alder dominated the


5

riparian zone of each study reach; however, unlike logged streams, most of the alder was too young and toofar back from the wetted channel to provide much shade during summer.

2.3 Detailed Study Stream Characteristics

The synoptic survey streams selected for further study on juvenile overwinter survivals in 1983−1984(Macmillan, Piper, Saltspring, Schomar, Southbay Dump and Tarundl creeks) were first and second orderstreams in the Skidegate Inlet — Skidegate Channel region between Moresby and Graham islands (Figure 1).Schomar Creek was a tributary of the Deena River; the others all flowed into the ocean. In Tarundl Creek,the study reach started 1.6 km upstream of the mouth and extended 600 m farther upstream to the base ofa large debris jam, above which there were no coho salmon. In the rest of the streams, the study reachesstarted at the mouth and extended as far upstream as there were coho. None of the streams, therefore, hadany coho that could move down into the study reaches, and only Schomar and Tarundl creeks had cohothat could move up into the study reach. Reach length ranged from 475 to 870 m, while channel widthranged from 7.7 to 41.9 m. Wetted width ranged from 2.6 to 7.1 m, channel gradient from 2.1 to 4.0 %, anddrainage area from 4.0 to 9.6 km2.

All six detailed study reaches were logged and, with the exception of Saltspring Creek, this loggingextended upstream (Table 1). Macmillan Creek (torrented in 1979) and Saltspring Creek (torrented in 1979with successive failures in 1982 and 1983) were considered to be the most seriously disturbed streams.Southbay Dump Creek was also torrented in 1979, but because new LOD was re-introduced in the upper315 m of the study reach in 1982 (Tripp 1986), it was considered to be less disturbed than Macmillan orSaltspring creeks. Although much of the new LOD added to Southbay Dump Creek was eventually buriedor dislodged by another torrent in August 1983, approximately 110 m of good stream habitat with new LODremained more or less intact in this study.

Piper Creek and Tarundl Creek had no major debris torrents within or upstream of the study reach, andthey were therefore considered the least affected streams in the study. Schomar Creek was similarlyunaffected when the study began, but it had a major debris torrent jam immediately above the study reach,which broke open during the study and inundated the study reach with large quantities of gravel on at leasttwo occasions. A small tributary stream in the creek, where fish could escape conditions in the mainchannel, was also damaged during the study when a 5-m section of road 235 m upstream washed out andspilled downstream. Because none of the other streams had any tributaries or off-channel habitat that fishcould inhabit overwinter, survivals in all six streams are considered to be fairly representative of conditionsin the main channel.

SEQ 6553 JOB FISH-004-015 PAGE-0006 CHAP 3 REVISED 21SEP00 AT 08:17 BY BC DEPTH: 60 PICAS WIDTH 42.09 PICAS COLOR LEVEL 1

6

3 METHODS

3.1 Reach Selection and Habitat Measurements

Prospective study reaches in each stream in the synoptic survey were selected in the following manner.Since the entire length of each reach had to be accessible to anadromous salmonids, a preliminary surveyof each stream was first conducted to see how far juvenile coho salmon extended upstream. Streamgradients were then measured with hand-held levels in the accessible portion of each stream to locate allmain LOD steps, sediment wedges and basin gradient breaks. Bank-full channel widths were measuredevery 50 m. The stream was then divided into reaches that had a more or less constant gradient, the samelevel of access to anadromous fish, and a similar valley shape. Final selection of study reaches was basedon reaches that also had the same logging and mass wasting condition throughout.

Methods used to quantify juvenile fish rearing habitat in each study reach are described in detail byTripp and Poulin (1986a). Length, width and maximum depth were recorded for each pool, glide and riffle inthe study reach during the summer low flow period, along with the amount of rooted undercut bank andLOD cover present. Additional information collected from detailed study reaches included the specific typesof pools present as defined by Heifetz et al. (1986) and the agent responsible for the formation of each pool(e.g., stream banks, LOD, boulders, bedrock). To determine the proportion of each habitat type present inboth surveys, the area of each habitat type was summed and divided by total wetted area.

3.2 Fish Population Estimates

Population size in the synoptic survey portion of the study was estimated with a removal method. Typically,two pools, a glide and a riffle (or two pools and two riffles if glides were rare) were individually enclosed withfine mesh seine nets and fished intensively for two successive and equal passes, using both electro-shockers and pole seines in pools and glides. Electroshockers only were used in riffles. Numbers of fish ineach habitat unit were then estimated by species and age group (fry = age o+ fish; parr = age 1+ and olderfish) according to Seber and LeCren (1967), and averaged by habitat type. Densities in each habitat typewere in turn multiplied by the proportion of the study reach occupied by that habitat, and summed to obtainaverage density over the whole study reach. In this way, fish densities for each study reach were weightedaccording to the amount of different habitat present, to provide a better comparison between streams atdifferent stage heights and with widely varying portions of pool, glide and riffle habitat. Where glides werenot sampled, glide densities were assumed to be of the same magnitude, relative to pools and riffles, as inreaches where all three habitat types were sampled.

Fish numbers in the detailed survey portion of the study were estimated with a single census mark-recapture method. In fall, numbers were estimated for each habitat unit over the whole study reach in eachstream. The following summer, when fish movements downstream had ceased, the estimates wererepeated over the whole study reach in Macmillan, Saltspring and Southbay Dump creeks, and in threesubsections that together represented half of each study reach in Piper Creek, Schomar Creek and TarundlCreek. Throughout the estimates, stop nets were placed at the top and bottom ends of each reach or reachsubsection. Fish were therefore prevented from moving in or out of the enclosed areas during the estimate,although they were free to move between habitat units (pools, glides, riffles) within the enclosed areas. Fishwere collected with electroshockers and minnow seines, marked with an adipose fin clip, and redistributedaccording to the number initially captured in each habitat unit. Fish were resampled 24−36 hours later andthe number of fish present estimated according to Chapman (1951, in Ricker 1975). Where the number ofrecaptures in any one habitat unit was too small (i.e., <3) for an accurate estimate, results were pooled foran estimate of overall density in each habitat type. Where the number of recaptures in each habitat typewas too low for an estimate, results were pooled again for an estimate of overall density in the study reach.

In both the synoptic and detailed surveys, we regularly considered data only for coho salmon fry, cohosalmon parr, Dolly Varden parr and steelhead parr. Because of their small size and reclusive habits, reliableestimates were rarely obtained for sculpins or Dolly Varden fry and they had to be excluded from analysis.Steelhead fry were also difficult to capture in the early summer (June) of 1982, so they were onlyconsidered in the 1982 fall (September) and 1984 summer (July) estimates.


7

Fish captured during sampling were measured to the nearest millimetre fork length and designated asfry (o+ fish) or parr (1+ and older fish) on the basis of their length frequency distributions (all species) andscale age-length relationships (coho and steelhead) in each reach type. Mean weights were estimated fromlength-weight relationships derived separately from subsamples of fry and parr of each species that hadbeen preserved in dilute (1%) formalin and weighed within 24 hours to the nearest 0.1 g.

3.3 Spring Downstream Trapping

To estimate apparent overwinter survivals for each species, we compared the number of fish present ineach reach the previous fall (fry and parr) with the number of fish captured moving downstream in thespring, plus the number of fish (parr only) still left in the stream after all spring downstream movements hadceased. To enumerate smolts and fry moving downstream in the spring, temporary weir and trap facilitieswere installed at the lower end of each study reach. Separate 2.4 × 0.9 m panels covered with 6.4 mmgalvanized wire mesh were nailed and wired together into V-shaped patterns pointing downstream. Theseguided fish into baffled live-boxes via a 0.2 × 0.2 × 4.9 m long wooden trough. The trough was covered withblack plastic mesh and pinned to the weir with a 0.4 m long steel rod (1.25-cm rebar) passed horizontallythrough the trough and the end of each panel at the apex of the V. Wire mesh extending 25 cm past one sideof each panel overlapped the next panel and prevented fish from escaping through seams in the weir. Extramesh dug into the substrate on the bottom of each panel prevented fish from moving under the weir. Eachwing of the weir was also buried into a steep bank to keep fish from moving around the weir; a 1 m wide stripof heavy plastic film on the bottom of the weir prevented scouring downstream. The entire structure wassupported with steel rods driven into the substrate behind the panels, with ropes tied from the tops of eachsteel rod to trees upstream.

Weirs were operated for 40- to 80-day periods, between April 10−May 6 and June 11−29, 1984.Depending on the weather and flow conditions, each trap was checked daily for emigrant fish, and theessential data (species, age, fork lengths, marks) were recorded before fish were released downstream.Previous studies indicate that weirs of this type in small streams are very efficient at capturing downstreammigrating smolts over widely varying flow conditions (Tripp and McCart 1983; Tripp 1986). In this study,however, floods and excessive bedload movements prevented trapping for 3−10 days (avg. 6 days) in eachstudy stream during storms in late May. Since May was a critical period for emigrant coho smolts, theestimates obtained on smolt yield and overwinter survival in the detailed study streams are minimumestimates only.

3.4 Habitat Utilization

To examine habitat use by fish in the synoptic survey portion of the study, overall densities of each specieswere compared in pools, glides and riffles of each reach type. Habitat preferences by fish in the detailedstudy streams were quantified by relating the fraction of the population found in each particular type ofhabitat in fall (September 1983) and summer (July 1984) to the relative abundance of that habitat withineach study reach. The relationship is based on the electivity index of Ivlev (1961) as follows (from Bisson etal. 1982):

Ui = (Di - Dt)D2t

where: Ui = utilization of habitat type i;

Di = fish density in habitat type i;

Dt = overall fish density in the study, all habitat types combined.

Values of this coefficient (Ui) theoretically range from -1, indicating total non-use of a habitat type, topositive infinity as more and more fish in the population reside in the ‘‘preferred’’ habitat. A value of 0indicates that the habitat is neither ‘‘preferred’’ nor ‘‘avoided’’.


8

3.5 Data Analysis

For most of this report, juvenile densities are expressed as numbers of fish per metre of stream length, tofacilitate comparisons between the same study reaches sampled at different times and stage heights, andbetween different study reaches with different drainage areas or sediment budgets. The exception is wheredata on juvenile densities in pools, riffles and glides were compared, in which case densities are expressedas numbers of fish per square metre to compensate for differences in the shape of each habitat type. In bothcases, confidence limits about the arithmetic means are based on values transformed to log (x+1).

For the synoptic survey streams, differences in habitat and the number of each species of fish in thedifferent reach types were analysed with a full data set and a reduced data set. The full data set included allstudy reaches sampled, regardless of logging or mass wasting conditions upstream. Thus some unloggedstudy reaches had logging or mass wasting upstream, and some logged streams (no mass wasting) hadmass wasting upstream. To eliminate the possible downstream effects of logging or recent mass wastingupstream, a reduced data set was also analysed which included only those stream reaches with the sameconditions upstream as instream. Sample sizes for each data set in summer and fall are as follows:

Reach Type Summer Fall

Full data set (N)

Unlogged 11 7Logged 11 9Mass wasted 11 11

Reduced data set (N)

Unlogged 5 4Logged 7 7Mass wasted 8 8

Non-parametric statistics were used to analyse much of the data on fish abundance in the differentreach types, (1) because the error involved in estimating fish density for a whole study reach as describedabove was not determined, and (2) because not all the data could be normalized with standard transforma-tions. Streams where the total absence of a species was clearly unrelated to the quality of the rearinghabitat available were especially troublesome. The Mann-Whitney U-test was used to compare abundanceand percent survival overwinter in two different stream types, while a Sign test was used to comparepercent declines in abundance. Where abundance or preferences in more than two habitat or reach typeswere compared, a Kruskal-Wallis analysis of variance test was used to test for differences, with additionalU-tests whenever Kruskal-Wallis tests indicated a significant (P<0.05) difference.

Differences between the habitat characteristics of unlogged, logged and mass wasted reaches in thesynoptic survey were tested with analysis of variance (ANOVA) tests, followed by Tukey’s HSD tests onpairs of means where significant differences were indicated by analysis of variance. Other possiblerelationships between the habitat variables and fish abundance data were explored in a Pearson correlationmatrix, but then discontinued when probability values (Bonferroni adjusted) indicated no significant correla-tions between fish numbers and the habitat parameters measured.

Other analyses included a least squares regression analysis to compare the number of recentlyemerged coho fry (including downstream migrants), in each detailed study stream in spring 1983, with theamount of gravel scour recorded the same year in the same streams (data from Tripp and Poulin 1986b).Regression tests were also used to compare coho fry fork length and fish standing crop in the synopticsurvey, to see if coho fry growth may have been affected by density dependent factors. To reduce thevariability caused by differences in sampling times, the data were adjusted to August 30, which wasapproximately 1 week after the last stream was sampled in the first sampling period (June 15−August 23)and 1 day before the first stream was sampled in the second sampling period (August 31−September 30).To adjust fork lengths and fish densities to August 30, apparent instantaneous growth (G) and mortality (Z)


9

rates between the first and second sampling periods were calculated and the rates used to interpolatebetween the two sampling periods. The formulae used to calculate growth and mortality rates G and Z wereas follows:

G = lnL2 - lnL1

t

Z = - (ln N2 - ln N1)t

where: L1 and L2 = average fork length of coho fry in each study reach in the firstand second sampling periods, respectively;

N1 and N2 = fish densities in the first and second sampling periods;

t = proportion of the year sampled between the first and secondsampling periods.

4 RESULTS

4.1 Stream Habitats

Logged and mass wasted study reaches in the synoptic survey had a significantly (P<0.05, ANOVA)greater portion of their drainage basin logged (37 and 30%, respectively) than did unlogged reaches (7%).The reaches were otherwise similar in length, width, gradient and drainage area (Table 2). Distanceupstream from the mouth to each study reach also did not differ significantly, although it was highly variable.While most study reaches started within 275 m of the stream mouth, one logged reach (Riley Creek) andtwo mass wasted reaches (Bonanza and Riley creeks) were located 5.1−8.5 km upstream. When the latterthree reaches were deleted from the data, mean distance upstream from the mouth to the start of eachreach in each reach type was more similar — 210 m in unlogged reaches, 271 m in logged reaches, and163 m in mass wasted reaches.

Mass wasting substantially reduced the amount of summer rearing habitat available in directly affectedreaches (Table 2). Compared to unlogged stream reaches, mass wasted reaches had significantly lesspool/glide habitat, shallower pools, less LOD cover, and less stable (i.e., rooted) undercut bank cover(P<0.05). Compared to unlogged or logged reaches, mass wasted streams also had significantly smallerwetted stream widths relative to total channel width because of lateral channel movements, bank erosionand dewatering at major sediment deposits. Reducing the data to eliminate reaches with different logging ormass wasting conditions upstream did not change the overall pattern appreciably, although the smallersample sizes meant fewer significant differences. Mass wasted stream reaches in the reduced data set stillshowed significantly less surface flow, shallower pools and less LOD cover than in logged or unloggedreaches, but differences in pool/riffle area and stable undercut bank cover were no longer significant.

The habitat characteristics of logged reaches tended to be either very similar to those of unloggedreaches (e.g., pool/riffle areas, the ratio of wetted width to channel width) or intermediate between unloggedand mass wasted reaches (e.g. LOD cover, undercut bank area, net pool depth). Few of the differenceswere statistically significant (P<0.05). Logged reaches had less undercut bank area than did unloggedreaches (P<0.05, full data set only), but otherwise did not differ significantly from the latter in any otherstream habitat variable measured. Similarly, the only significant difference between logged and masswasted reaches was that wetted stream width relative to channel width in logged streams was twice thewetted width in mass wasted streams (full and reduced data sets).

SEQ 6574 JOB FISH-009-009 PAGE-0010 CHAP 4 (1) REVISED 21SEP00 AT 08:18 BY BC DEPTH: 60 PICAS WIDTH 42.09 PICAS COLOR LEVEL 1

10

TABLE 2. Stream habitat characteristics in the full and reduced sets of unlogged, logged and masswasted study reaches in the synoptic survey. Values are means ± 95% confidence limits.Significant differences among reach types (P<0.05, ANOVA) are denoted with an asterisk.

Habitat Reach typePvariable Unlogged Logged Mass wasted

Full data set

Sample size 11 11 11Stream length (km) 4.8 ± 2.1 6.0 ± 2.2 5.9 ± 2.5 0.618Reach basin area (km2) 8.0 ± 4.6 9.4 ± 3.6 7.7 ± 3.8 0.758Reach basin area logged (%) 6.7 ± 9.8 37.4 ± 14.2 29.9 ± 23.1 0.014*Reach distance upstream (m) 202.7 ± 214 715.0 ± 1071 1424.0 ± 1973 0.335 Reach length (m) 485.5 ± 83 573.0 ± 155 481.0 ± 110 0.397Gradient (%) 2.4 ± 0.8 2.1 ± 0.7 3.0 ± 1.0 0.212Wetted width (m) 8.2 ± 3.6 6.6 ± 2.8 5.0 ± 1.6 0.194Bankfull width (m) 24.1 ± 5.9 17.4 ± 5.9 25.0 ± 8.0 0.296Wetted width/bankfull width 0.4 ± 0.1 0.4 ± 0.1 0.2 ± 0.1 0.021*Pool, glide area (%) 65.8 ± 12.2 64.2 ± 9.1 43.5 ± 19.1 0.029*Riffle area (%) 34.2 ± 12.2 35.8 ± 9.1 56.5 ± 19.1 0.029*LOD cover (%) 11.8 ± 4.0 8.1 ± 2.7 5.4 ± 3.0 0.016*Undercut bank cover (%) 3.3 ± 1.4 1.3 ± 0.7 0.8 ± 0.6 0.001*Net pool depth (cm) 58.0 ± 16 52.0 ± 12 36.0 ± 14 0.042*

Reduced data set

Sample size 5 7 8Stream length (km) 5.6 ± 3.8 6.3 ± 3.1 5.5 ± 2.4 0.862Reach basin area (km2) 10.6 ± 9.5 10.5 ± 5.7 6.8 ± 3.4 0.342Reach basin area logged (%) 0.0 ± 0.0 36.4 ± 19.8 38.7 ± 29.5 0.000*Reach distance upstream (m) 50.0 ± 159 385.7 ± 655 845.0 ± 1674 0.559Reach length (m) 470.0 ± 172 648.6 ± 222 500.0 ± 144 0.196Gradient (%) 2.1 ± 1.6 2.1 ± 1.2 3.2 ± 1.3 0.224Wetted width (m) 11.6 ± 7.9 7.1 ± 4.6 4.8 ± 2.3 0.039*Bankfull width (m) 31.0 ± 23.8 16.9 ± 8.1 24.7 ± 9.8 0.189Wetted width/bankfull width 0.4 ± 0.3 0.4 ± 0.1 0.2 ± 0.1 0.004*Pool, glide area (%) 63.4 ± 31.3 58.7 ± 11.7 40.1 ± 23.7 0.140Riffle area (%) 36.6 ± 31.3 41.3 ± 11.7 59.9 ± 23.7 0.140LOD cover (%) 10.0 ± 6.5 7.3 ± 2.3 4.0 ± 3.3 0.029*Undercut bank cover (%) 1.9 ± 2.6 1.0 ± 0.6 0.5 ± 0.5 0.111Net pool depth (cm) 69.4 ± 40.7 55.0 ± 19.1 33.9 ± 16.4 0.033*

Results of the habitat surveys in the detailed stream reaches (Table 3) were similar to those of thesynoptic survey. As in the synoptic survey, mass wasted streams in the detailed survey had wider rootedwidths and narrower wetted widths (P<0.05) than did non-mass wasted streams. They also had signifi-cantly smaller and shallower pools, less total pool/glide area, and more riffle area. Total wetted area duringlow flows was therefore only 14% of total channel area in mass wasted streams, compared to 38% in non-mass wasted streams (P<0.05). The differences were not related to differences in drainage area or streamgradient. They were, however, consistent with the bank erosion, channel shifts and dewatering observed inmass wasted streams wherever large quantities of sediment had inundated the channel.

In the detailed stream survey, non-mass wasted streams had, on average, twice the number of LODpieces per metre of stream that mass wasted streams had (0.45 vs. 0.22 pieces per metre, P<0.05). Theyalso had twice the LOD coverage in planview, although in this case the differences were not signficantbecause the differences in the species composition of the riparian zones greatly increased the variability inthe length and diameter of the LOD present. The riparian zone along the first 500 m of the study reach inPiper Creek, for example, was A-frame logged in the 1950’s which effectively removed all of the instreamdebris. The latter was replaced by small, short pieces of debris from the red alder which now completelydominates the riparian zone. Conifer logs that were longer and thicker than the alder debris in Piper Creekdominated the LOD in the rest of the streams. Of these, Tarundl Creek had the largest average debris pieceswhich were derived mainly from the adjacent buffer strip. Debris in the other creeks came mainly from the


11

debris torrents or debris jams above each study reach. As described earlier, large conifer logs were alsoadded to Southbay Dump Creek.

Orientation of the debris relative to stream flows did not differ significantly between the mass and non-mass wasted detailed study streams (all logged). Average debris orientation was diagonal in all streamsexcept Southbay Dump Creek, where the extra debris added to the creek was placed at right angles to thestream flow. Position relative to the stream surface, on the other hand, did differ significantly (P<0.05) asshown by the differences in deflection values. In mass wasted streams, the debris tended to be perchedabove the water surface, on bouldery substrates in riffles or where new pools were scoured out underneathsmall debris jams.

Instream cover characteristics were highly variable. Mass wasted streams had about half the deepwater cover (5.3 vs. 10.7% of wetted stream area) and total overwinter cover (6.6 vs. 14.5%) present in non-mass wasted streams. The differences were not significant (P = 0.10), but they were suggestive. Therewere no significant differences between LOD cover, undercut bank cover or rock cover in the two streamtypes, nor were there any differences in the total amount of cover.

TABLE 3. Habitat characteristics of detailed study streams

Mass wasted streams Non-mass wasted streamsHabitat variable Macmillan Saltspring Southbay Schomar Piper Tarundl

Reach characteristicsDrainage area (km2) 6.2 6.3 4.0 6.6 4.2 9.6Length (m) 727.0 493.0 590.0 656.0 835.0 675.0Rooted width (m) 12.3 41.9 25.2 7.6 7.7 11.6Wetted width (m) 3.1 2.6 2.8 5.7 3.2 4.7Gradient (%) 3.1 3.5 3.9 2.1 2.2 4.0

Surface substrate characteristics% fines (<5 mm) 2 12 16 16 14 16% gravel (5−63 mm) 10 33 41 40 44 27% larges (>63 mm) 88 55 43 44 42 57D90 44 27 16 22 13 29

Pool characteristicsTotal pool area (%) 16.0 34.6 17.1 30.7 35.8 26.2Average pool area (m2) 10.2 12.1 10.2 42.2 13.7 23.1LOD pool area (%) 2.7 19.6 17.1 16.6 21.0 18.1No. pools / 100 m 4.8 7.3 4.7 4.1 7.9 5.6Net pool depth (cm) 14.0 23.6 18.3 30.5 25.8 39.0

LOD characteristicsNo. pieces/m 0.13 0.26 0.27 0.51 0.33 0.52Average length (m) 3.10 3.10 3.52 3.54 2.48 3.94Average diameter (m) 0.40 0.43 0.48 0.40 0.31 0.44Orientationa 1.98 2.16 2.50 2.09 2.06 2.09Deflectionb 1.33 1.58 1.74 2.07 1.30 1.84

Instream cover characteristicsLOD (%) 1.6 6.8 11.2 9.2 1.6 13.1Deep water (%) 3.7 9.6 2.5 16.2 4.7 11.1Undercut banks (%) 0.2 3.4 0.3 1.0 1.1 0.9Rock (%) 9.2 11.8 1.0 3.5 4.1 7.8Total (%) 14.7 31.6 15.0 29.9 11.5 32.9Total ‘‘winter’’ cover (%)c 1.0 12.0 6.9 18.9 7.8 16.9

a Average position of each LOD piece relative to stream flows: 1 = parallel; 2 = diagonal; 3 = perpendicular.b Average position of each LOD piece relative to the stream surface: 1 = over; 2 = in/out; 3 = submerged.c The sum of the instream LOD, deep water, rock and undercut bank cover in LOD controlled pools.


12

4.2 General Distribution and Abundance of Fish Species

Coho salmon were present in all but one of the synoptic survey reaches sampled in 1982. Only MacmillanCreek (mass wasted) lacked coho in 1982, although they were present in 1983 and 1984. Two otherreaches contained coho parr but no fry (Shelley Creek and Riley Creek tributary), while another one (upperSouthbay Dump Creek) had coho fry in fall but none in the summer. Dolly Varden, the next most widespreadspecies, were present in 82% (N = 27) of the study reaches; steelhead trout were present in 67% (N = 22) ofthe study reaches.

There were no significant differences in overall mean basin area, distance upstream or reach gradientbetween the three species (Table 4). Dolly Varden, nevertheless, appeared to use steeper reaches than dideither coho or steelhead. The average maximum stream gradient recorded for Dolly Varden in streamswithout any obvious barriers to upstream migrating fish was significantly higher (10.9%, P<0.05, Mann-Whitney U-tests) than the maximum gradient recorded for coho (7.0%) or steelhead (6.0%). No streamswith a basin area <5 km2 contained juvenile steelhead (unless it was part of a larger basin). It thus appearsthat coho and Dolly Varden are also capable of using smaller streams.

Coho salmon was the most common species in the study area. Overall, juvenile coho densities insummer (unlogged, logged and mass wasted reaches combined) averaged 7.04 fish per metre, whichrepresented 93.6% of all salmonids present (excluding Dolly Varden and steelhead fry), and 1.90 fish permetre in fall, or 57.4% of all salmonids present (including steelhead fry, Table 4). Dolly Varden densitiesrarely exceeded 0.2 fish per metre at any time and thus they were relatively uncommon in the streamsexamined, but widely distributed. Steelhead numbers were intermediate, averaging 0.33 fish per metre insummer when only parr were considered, increasing to 1.32 fish per metre after fry had completedemergence.

TABLE 4. General distribution characteristics, abundance and age composition of juvenile salmonidpopulations in 1982 synoptic survey reaches on the Queen Charlotte Islands. Numbers inbrackets are 95% confidence limits.

Fish speciesPopulation characteristics Coho salmon Dolly Varden Steelhead trout

Distribution% of reaches 97.0 81.8 66.6Reach gradient (%) 2.5 (2.0 - 2.9) 2.7 (2.2 - 3.2) 2.1 (1.7 - 2.6)Maximum gradient (%) 7.0 (6.5 - 7.5) 10.9 (7.6 - 14.3) 6.0 (3.9 - 8.1)Mean basin area (km2) 8.4 (6.4 - 10.5) 6.9 (5.1 - 8.8) 10.5 (7.9 - 13.1)Smallest basin (km2)a 2.3 2.3 5.0

AbundanceSummer density (n/m) 7.04 (5.13 - 9.66) 0.17 (0.15 - 0.17) 0.33 (0.30 - 0.37)Fall density (n/m) 1.90 (1.59 - 2.25) 0.10 (0.10 - 0.11) 1.32 (1.04 - 1.68)

Age structureSummer (% fry) 94.1 (91.4 - 96.7) NDb NDFall (% fry) 94.1 (91.2 - 97.0) ND 82.5 (75.5 - 89.5)

a Smallest basin that flows into the ocean (i.e., not part of a larger basin).b No data; fry estimates unreliable or emergence incomplete.

Most of the juvenile coho salmon and steelhead trout sampled were fry. Coho fry accounted for 94% ofthe coho present in both summer and fall; steelhead fry accounted for 83% of that species in fall samples.Coho parr were almost exclusively 1+ fish, while steelhead parr ranged from 1+ to 4+ fish. On the basis oftheir length-frequency distribution (Figure 2), at least four age classes (excluding fry) of Dolly Varden werealso present. The latter included both parr and adult stages, many of which (parr and adults) were sea runfish that had evidently moved into the streams in fall.

13

FIGURE 2. Length frequency distributions of coho salmon, steelhead trout and Dolly Varden char insynoptic survey streams on the Queen Charlotte Islands, 1982−1984.

4.3 Habitat Utilization

No significant differences occurred between the habitat preferences of juvenile coho salmon in loggedstreams with or without mass wasting. In both stream types, juvenile coho strongly preferred pools andavoided riffles (Kruskal-Wallis analyses, P<0.01−0.001). According to the combined data for the two streamtypes (Figure 3A), LOD pools formed as a result of scouring around LOD were further preferred over non-LOD pools formed from scouring around boulders, bedrock or along a bank (Mann-Whitney U-tests,P<0.05). Glides and non-LOD pools were in turn preferred over riffles (P<0.05), which coho avoided.

14

FIGURE 3. Habitat preferences and densities of juvenile salmonids in pools, glides and riffles. Habitatpreferences for each habitat type in 3A are the average of the fall and summer preferencesrecorded in all six detailed study streams sampled in 1983−1984. Densities in 3B are theaverage summer and fall densities recorded in pools, riffles and glides in the synoptic surveyin 1982.


15

Differences between the habitat preferences of Dolly Varden and steelhead in mass and non-masswasted streams could not be tested because the sample sizes for the mass wasted streams were too small.In the combined data (Figure 3A), Dolly Varden and steelhead trout parr both showed a significantpreference for LOD pools over glides and riffles, but only Dolly Varden appeared to avoid riffles (P<0.05).Steelhead fry tended to avoid non-LOD pools (P<0.05), but otherwise showed no preference for anyspecific habitat type; they were evenly distributed relative to the amount of LOD pool, glide and riffle habitatpresent.

With one exception, the apparent preferences of juvenile fish for pools, riffles or glides in the synopticsurvey portion of the study were similar to the distribution of the fish in the detailed study streams asdescribed above. Whereas coho fry in the detailed study streams strongly preferred pools over glides(Figure 3A), there were no significant differences in densities between the two habitats in summer or fall inthe synoptic survey (Figure 3B). Coho parr densities were also similar in pools and glides in summer, buthigher in pools than in glides during fall (P<0.05), and lowest in riffles. Densities of Dolly Varden parr,steelhead parr and steelhead fry were not significantly different between any habitats.

4.4 Effects of Logging and Mass Wasting on Fish Abundance

Logged stream reaches (no recent mass wasting) had significantly more coho salmon fry than unlogged ormass wasted (logged) reaches in both summer and fall (P<0.05, Table 5). They also had more fry thanunlogged or mass wasted reaches when the data were adjusted to a common date (August 30). This showsthat the differences were not just a function of different sample times. Summer coho fry densities averaged14.78 fish per metre in logged reaches, which was 4.1 and 6.7 times greater than densities in unlogged andmass wasted reaches, respectively. Fall densities averaged 2.61 fish per metre, which was 2.2 and 1.7times greater than unlogged or mass wasted reaches, respectively. Mean sample dates did not differsignificantly between reach types in summer or fall (Table 5).

Logged reaches, with the greatest initial densities in summer, showed the greatest decline in coho fryabundance from summer to fall (82.3%, P<0.05, Sign test), followed closely by unlogged reaches wheredensities declined 66.4% (P<0.05) from 3.60 to 1.21 fish per metre. Mass wasted streams, which had thelowest overall densities to begin with, showed the smallest overall decline (31.5%, P>0.05) from summer(2.19 fish per metre) to fall (1.50 fish per metre).

There were no significant differences between coho fry densities in unlogged and mass wasted streamreaches. Differences between the summer and fall carrying capacities of mass wasted and unloggedstreams, however, may have been obscured by poor egg-to-fry survivals or low escapements in masswasted streams. Four mass wasted stream reaches in 1982 had either no coho fry (Macmillan Creek, upperSouthbay Dump Creek, Riley Creek tributary) or very few coho fry (Shelley Creek, 0.20 fry per metre). Bycomparison, the next lowest summer coho fry densities recorded were 1.26 coho fry per metre (MountainCreek, mass wasted) and 1.42 coho fry per metre (Salmon River, unlogged). When streams without cohofry were deleted from the data, summer coho fry densities in mass wasted streams were essentially thesame as densities in unlogged streams (3.41 vs. 3.60 fry per metre), while fall densities were significantlyhigher (2.17 fry per metre in mass wasted streams vs. 1.21 fry per metre in unlogged streams, P<0.05,U-test).

Coho parr were equally abundant in each reach type in summer, but significantly lower in loggedreaches than in other reaches during the fall (P<0.05, Table 5). There were no significant differencesbetween reach types in the abundance patterns of the parr or fry of other species. As for coho fry, densitiesof Dolly Varden parr, steelhead parr and steelhead fry all tended to be highest in logged reaches. Unlike forcoho fry, however, the densities were also high in mass wasted reaches. Unlogged reaches in every casehad the lowest densities.

Of the seven logged reaches with steelhead in the synoptic survey, four showed a normal decline inabundance from summer to fall, two showed no change, and one increased. Mass wasted reaches showedthe opposite pattern. Of six reaches with steelhead, only one population (Riley Creek) declined fromsummer to fall. The remaining five all increased. Dolly Varden were more variable, partly because densitieswere usually very low to begin with, and partly because several streams had a mixture of resident (juvenile)


16

and sea run (adult and juvenile) fish. The latter were distinguished by their silvery coloration when theyreturned from the ocean in late summer. Of the seven logged reaches with Dolly Varden in the fall, twopopulations declined in abundance from summer to fall, one remained the same, and four increased. Ofseven mass wasted reaches with Dolly Varden, three declined, two remained the same, and two increased.

TABLE 5. Mean juvenile salmonid densities in the full set of unlogged, logged and mass wasted (logged)stream reaches sampled on the Queen Charlotte Islands in 1982. Means (number of fish permetre of stream) and 95% confidence limits (in brackets) as described in ‘‘Methods’’. Differ-ences between reach types were tested with a Kruskal-Wallis analysis of variance test (K).

Reach typeSpecies, age group Ka

Unlogged Logged Mass wasted

Summer (June 15−Aug. 23)

Sample size (N) 11 11 11Mean sample date July 26 (±12 days) July 14 (±16 days) July 22 (±18 days) 1.285

Coho fry 3.60 (2.76 - 4.69) 14.78 (8.02 - 27.25) 2.19 (1.30 - 3.70) 15.130***Coho parr 0.16 (0.15 - 0.17) 0.22 (0.20 - 0.24) 0.18 (0.16 - 0.20) 0.848Dolly Varden parr 0.20 (0.18 - 0.23) 0.11 (0.10 - 0.12) 0.17 (0.15 - 0.19) 1.000Steelhead parr 0.22 (0.19 - 0.25) 0.42 (0.37 - 0.48) 0.35 (0.28 - 0.44) 0.842

August 30

Sample size (N) 7 9 11Coho fry 1.70 (1.30 - 2.17) 3.64 (3.02 - 4.38) 1.50 (1.03 - 2.18) 11.041***Coho parr 0.12 (0.11 - 0.13) 0.06 (0.05 - 0.07) 0.15 (0.14 - 0.16) 3.077Dolly Varden parr 0.05 (0.04 - 0.06) 0.08 (0.07 - 0.09) 0.16 (0.14 - 0.18) 0.299Steelhead parr 0.07 (0.07 - 0.08) 0.32 (0.27 - 0.38) 0.33 (0.27 - 0.40) 1.983

Fall (Sept. 1−30)

Sample size (N) 7 9 11Mean sample date Sept. 20 (±3 days) Sept. 15 (±8 days) Sept. 22 (±6 days) 5.085

Coho fry 1.21 (0.91−1.60) 2.61 (2.27−3.00) 1.50 (1.02−2.20) 7.631**Coho parr 0.15 (0.13−0.18) 0.03 (0.03−0.03) 0.12 (0.11−0.13) 5.445*Dolly Varden parr 0.02 (0.02−0.02) 0.11 (0.10−0.12) 0.15 (0.13−0.17) 1.747Steelhead fry 0.49 (0.34−0.70) 1.41 (0.92−2.17) 1.14 (0.71−1.82) 2.236Steelhead parr 0.06 (0.06−0.07) 0.29 (0.25−0.34) 0.36 (0.29−0.45) 1.934

a * P<0.05, ** P<0.01, *** P<0.001.

Population densities in the detailed study streams over the 1983−1984 study period (Table 6) followedthe same general pattern as densities in the synoptic survey studies. Each species, regardless of age,tended to be most abundant in non-mass wasted streams. Coho densities in fall and summer (fry and parrcombined) were on average 3.4 and 7.3 times higher (P<0.05), respectively, in non-mass wasted streamsthan in mass wasted streams; Dolly Varden densities were 8.2 and 17.3 times higher (P<0.05). Steelheadwere present in only one mass wasted stream (Saltspring Creek) and occurred in about the same densityas in non-mass wasted streams in fall. However, they were 2.6 times more abundant in non-mass wastedstreams the following summer.

4.5 Upstream/Downstream Effects

Excluding from the data base the reaches where logging or mass wasting conditions upstream differed fromthose conditions in the study reach, we found differences in coho fry abundance between reach types to beexaggerated, although the overall patterns remained essentially unchanged (Table 7). As for the full dataset, coho salmon fry densities in the reduced data were significantly higher in logged reaches than inunlogged or mass wasted reaches during summer and fall. Coho parr and other species showed nosignificant differences between reach types.


17

TABLE 6. Juvenile coho salmon, Dolly Varden char and steelhead trout densities (number of fish permetre of stream) in the detailed study streams, fall (Sept. 12−Oct. 6, 1983) and summer (July13−23, 1984)

Mass wasted streams Non-mass wasted streamsSpecies, age Macmillan Saltspring Southbaya Schomarb Piperc Tarundl

Fall (Sept. 12−Oct. 6, 1983)

Coho fry 0.98 1.29 1.82 3.95 0.75 2.05Coho parr (1+) 0.00 0.00 0.04 0.49 0.14 0.21Dolly Varden fry 0.00 0.00 0.00 0.19 0.09 0.00Dolly Varden parr (1−3+) 0.10 0.17 0.23 0.69 0.77 1.34Steelhead fry 0.00 0.67 0.00 0.52 0.00 0.95Steelhead parr (1−3+) 0.00 0.51 0.00 0.60 0.00 1.01

Summer (July 13−23, 1984)

Coho fry 0.04 2.17 0.28 4.74 3.93 6.89Coho parr (1+) 0.00 0.00 0.04 0.02 0.23 0.18Dolly Varden fry 0.00 0.00 0.00 0.00 0.36 0.42Dolly Varden parr (1+) 0.00 0.00 0.00 0.07 0.33 0.04Dolly Varden parr (2+, 3+) 0.00 0.01 0.08 0.30 0.05 0.10Steelhead fry 0.00 2.80 0.00 2.11 0.00 0.66Steelhead parr (1+) 0.00 0.07 0.00 0.36 0.00 0.18 Steelhead parr (2+, 3+) 0.00 0.02 0.00 0.04 0.00 0.21

a LOD added to upper 315 m of study reach in 1982; upper 205 m re-torrented in August 1983, but the next 110 m of new habitatdownstream still provided good habitat over 1983−1984 winter.

b Not mass wasted at start of study, but both mainstem and tributary inundated with gravel when road crossing and debris torrent jamsupstream failed overwinter. Most gravel subsequently scoured out in later storms the same winter.

c Study reach was severely damaged by a major storm in August 1983.

TABLE 7. Mean juvenile salmonid densities in the reduced set of unlogged, logged and mass wasted(logged) stream reaches sampled on the Queen Charlotte Islands in 1982. Means (number offish per metre of stream) and 95% confidence limits (in brackets) as described in ‘‘Methods’’.Significance K was determined with a Kruskal-Wallis analysis of variance test on densities ineach reach type.




Sample size (N) 5 7 8Mean sample date Aug. 4 (±15 days) July 1 (±14 days) July 13 (±22 days) 6.477*

Coho fry 2.65 (1.82−3.85) 20.78 (8.86−48.76) 2.35 (1.15−4.80) 9.635*Coho parr 0.19 (0.15−0.24) 0.07 (0.06−0.08) 0.22 (0.19−0.25) 2.337Dolly Varden parr 0.26 (0.19−0.36) 0.11 (0.09−0.13) 0.18 (0.15−0.22) 2.155Steelhead parr 0.28 (0.21−0.38) 0.50 (0.34−0.74) 0.38 (0.27−0.53) 0.825

August 30

Sample size (N) 4 7 8Coho fry 1.93 (1.29- 2.89) 3.72 (3.02−4.58) 1.39 (0.89−2.17) 10.695*Coho parr 0.15 (0.11−0.21) 0.06 (0.06−0.07) 0.15 (0.14−0.17) 2.743Dolly Varden parr 0.05 (0.04−0.06) 0.10 (0.09−0.11) 0.18 (0.15−0.21) 0.120Steelhead parr 0.09 (0.07−0.11) 0.29 (0.23−0.36) 0.33 (0.25−0.44) 1.135

Fall (Sept. 1−30)Sample size (N) 4 7 8Mean sample date Sept. 21 (±6 days) Sept. 12 (±10 days) Sept. 20 (±8 days) 2.600

Coho fry 1.40 (0.70−2.79) 2.88 (2.67−3.11) 1.37 (0.92−2.03) 6.870*Coho parr 0.19 (0.11−0.31) 0.04 (0.03−0.04) 0.12 (0.11−0.14) 2.976Dolly Varden parr 0.01 (0.01−0.02) 0.13 (0.11−0.15) 0.17 (0.15−0.20) 3.555Steelhead fry 0.62 (0.23−1.71) 1.53 (0.84−2.78) 0.60 (0.40−0.91) 2.326Steelhead parr 0.08 (0.06−0.10) 0.30 (0.24−0.38) 0.30 (0.23−0.39) 0.977

a * P<0.05.


18

Logged reaches with mass wasting upstream in the synoptic survey appeared to have fewer coho frythan did logged reaches with no mass wasting upstream, in both summer and fall samples (Table 8).Summer coho fry densities in logged reaches with no mass wasting upstream averaged 20.78 coho fry permetre (N = 7), while logged reaches with mass wasting upstream averaged 4.27 coho fry per metre (N = 4,P<0.05, Mann-Whitney U-test). Average distance upstream to the mass wasting was 550 m. In fall, thesame logged reaches with no mass wasting upstream averaged 2.88 fry per metre (N = 7), while loggedreaches with mass wasting upstream averaged 1.67 coho fry per metre (N = 2, P<0.05, one-tailed test).

TABLE 8. Coho fry densities (number of fish per metre of stream) and sample dates in logged streamreaches, with and without mass wasting upstream. Numbers in brackets are 95% confidencelimits. P values are for one-tailed Mann−Whitney U tests.

Logged reaches with Logged reaches withno mass wasting upstream mass wasting upstream P


N 7 4Density 20.78 (8.86−48.76) 4.27 (2.87−6.36) 0.012Sample date July 1 (±14 days) Aug. 5 (±25 days) 0.012

August 30

N 7 2Density 3.72 (3.02−4.58) 2.53 (0.38−16.93) 0.111

Fall (Sept. 1−30)

N 7 2Density 2.88 (2.67−3.11) 1.67 (1.65−1.68) 0.028Sample date Sept. 12 (±12 days) Sept. 24 (±51 days) 0.250

Differences in the mean sample date in the reduced data set were significant for the summer samplesonly, and only then between unlogged and logged reach types (Table 7). The mean sample date for loggedreaches was July 1, which was over a month earlier than the mean sample date for unlogged reaches(August 4, P<0.05). Logged reaches with no mass wasting upstream were also sampled, on average, overa month earlier (July 1) than logged reaches with mass wasting upstream (August 5, P<0.05, Table 8).Differences in sample times therefore probably account for a large portion of the variation among thesummer samples. Differences in sample times in fall did not appear to be as significant as in summer, butonly two logged streams with mass wasting upstream were sampled.

4.6 Effects on Coho Growth

Coho fork lengths at the end of the season in 1982 varied significantly within and between reach types(Figure 4). From means of 37−39 mm in early June, coho fry grew to 44−66 mm by the end of September,while parr grew from means of 71−76 to 83−100 mm. The largest mean fry (61−66 mm in September) wereall in mass wasted streams that maintained a continuous flow of water between pools in summer; thesmallest mean fry were located in logged reaches (42−55 mm) or in mass wasted reaches (47−54 mm)where fry were confined to isolated pools. Fry in unlogged streams had intermediate fork lengths (49−56mm) and showed the least variability between streams.

When the length data were adjusted to a common date (August 30), Kruskal-Wallis analysis ofvariance tests indicated growth in length was significantly related to reach type (P<0.05, Table 9). Furthercomparisons (Mann-Whitney U tests) indicated significantly better growth in mass wasted streams withcontinuous stream flow, but no differences in growth between other reach types.

Coho fry growth appeared to be density dependent inasmuch as logged reaches had not only some ofthe smallest fish and lowest growth rates overall, but the highest average mean densities as well (Table 9).When the data were adjusted to take into account differences in sampling time, however, density and forklength were not correlated. There were also no significant correlations between coho fry fork length andtotal salmonid density or total salmonid biomass.

19

FIGURE 4. Seasonal variation in mean fork length for coho salmon juveniles in unlogged, logged andmass wasted stream reaches on the Queen Charlotte Islands, June 9−September 30, 1982.


20

TABLE 9. Mean coho fry fork length, coho fry density, total salmonid density and total salmonid biomass inselected streams on the Queen Charlotte Islands, 1982. All values adjusted to August 30 basedon growth (G) and mortality (Z) rates for samples collected in summer and fall. G and Z belowfor coho fry only. Significance (K) tested with a Kruskal-Wallis analysis of variance test.

Total TotalCoho fry Coho fry salmonid salmonid

Stream fork length density density biomass G Z(mm) (n/m) (n/m) (g/m)

Unlogged streams

Cache 51.5 1.21 1.53 1.69 1.367 -0.014Hangover 51.9 2.66 3.34 3.81 1.064 -0.003Inskip 51.2 2.07 2.28 2.84 1.417 -0.020Matheson Head 52.1 1.39 1.43 2.01 1.406 -0.034Marshall Head 46.4 1.61 1.81 1.65 0.847 -0.018

Mean 50.6 1.79 2.08 2.40 1.220 -0.018

Logged streams

Landrick 51.7 3.59 3.65 5.08 0.942 -0.009Piper 47.2 5.96 6.01 6.41 1.281 -0.035Sachs 41.7 4.27 4.40 3.17 0.443 -0.033Schomar 51.7 2.01 2.27 2.84 0.907 -0.009Tarundl 44.4 2.94 3.29 2.63 0.832 -0.033Riley 44.0 3.06 3.65 2.67 1.330 -0.021

Mean 46.8 3.64 3.88 3.80 0.956 -0.023

Mass wasted streams (dewatered)

Mosquito Trib. 49.6 2.38 3.47 2.97 1.183 +0.008Mountain 48.1 0.98 1.96 1.12 0.947 -0.011Saltspring 46.7 3.46 3.64 3.61 0.712 +0.004

Mean 48.1 2.27 3.02 2.56 0.947 +0.000

Mass wasted streams (not dewatered)

Bonanza 55.6 1.88 2.71 3.31 2.120 -0.004Southbay Dump 62.6 2.34 2.37 5.87 2.140 -0.017Two Torrent 58.2 3.28 3.61 6.62 1.177 -0.001

Mean 58.8 2.50 2.90 5.27 1.812 -0.007

Ka 8.961** 6.056 7.325* 6.420* 5.542 6.907*

a * P<0.10, ** P<0.05.

Differences between coho fry growth in the detailed streams were more distinct than those in thesynoptic survey (Figure 5). In late May to early June 1983, coho fry in the two non-mass wasted streamssampled (Tarundl Creek and Piper Creek) averaged only 40.4−41.1 mm in fork length. By comparison, fryin mass wasted streams averaged 48.5 mm (Southbay Dump Creek) and 53.9 mm (Macmillan Creek) fornet differences of 8−13 mm (P<0.05). These differences persisted or increased throughout the ensuingyear, with the result that smolts from the 1983 brood year in mass wasted streams (96.0−107.6 mm) werealso significantly larger than the smolts in non-mass wasted streams (76.4−92.0 mm, P<0.05).

The age composition of juvenile coho in the detailed study streams varied according to growth theprevious year (Table 10). In September 1982, 99−100% of all coho present in each stream were 0+ fish.This suggests that growth the previous year was probably good, since few fish remained behind for asecond season. The poor growth recorded during the dry summer of 1982 in Piper, Saltspring and Tarundlcreeks, however, resulted in substantially more (9−17%) coho remaining behind for a second season inthese streams in 1983 than was the case in 1982. Growth during the wet summer of 1983 was significantly(P<0.05) better than in 1982 and the proportion of 1+ fish that remained behind for another year in 1984declined to 0−6%. In other streams (Macmillan, Schomar and Southbay Dump creeks), growth was rapidthroughout the study period and few coho spent more than 1 year in fresh water at any time.

21

FIGURE 5. Seasonal variation in juvenile coho salmon growth for the 1983 and 1984 year classes inmass and non-mass wasted streams.


22

TABLE 10. Annual variation in age composition and fry growth of coho salmon in detailed study streams.Values marked with an asterisk significantly different at P<0.05

Age composition (%0+) Mean fork length (mm)Stream Sept. 1982 Sept. 1983 July 1984 Sept. 1982 Sept. 1983

Mass wasted streamsMacmillan No coho 100 100 No coho 78.5Southbay 99 100 100 65.6 66.6Saltspring 100 83 100 52.2* 67.8*

Non-mass wasted streamsSchomar 100 100 100 54.5* 65.5*Tarundl 99 90 97 49.5* 58.3*Piper 100 82 94 50.1* 59.3*

The age composition of coho in mass wasted streams (e.g., Macmillan and Southbay Dump creeks)varied less than that in non-mass wasted streams, as long as there was a continuous flow of water and fishdid not end up concentrated in isolated pools (as was the case in Saltspring Creek). By comparison, growthand the age composition of coho in non-mass wasted streams varied considerably from year to year.Though Schomar Creek appeared to be an exception (Table 10), it was also the only non-mass wastedstudy reach to have a major debris torrent just upstream. Conditions affecting growth in Schomar Creekmay therefore have been more like conditions in mass wasted streams.

4.7 Effects on Biomass

Coho fry were the only juvenile fish to show significant differences in biomass between the three differentreach types (Table 11). In summer, coho fry biomass in mass wasted streams was 1.8−5.5 times lower thanin unlogged or logged reaches, respectively, but 2.5−1.3 times higher in fall. Only in summer were thedifferences statistically significant and only then between logged and mass wasted reaches (Mann-WhitneyU tests, P<0.05).

TABLE 11. Salmonid biomass (grams of fish per metre of stream) in unlogged, logged and mass wasted(logged) stream reaches on the Queen Charlotte Islands in 1982. Means and 95% confidencelimits were derived as described in ‘‘Methods’’. Significance K determined with a Kruskal-Wallis analysis of variance test.




Sample size (N) 11 11 11Coho fry 3.22 (2.12−4.10) 9.51 (3.83−15.19) 1.72 (0.50−2.94) 17.328*Coho parr 0.97 (0.33−2.62) 1.30 (0.00−2.62) 1.11 (0.36−1.86) 0.737Dolly Varden parr 4.37 (0.65−8.09) 2.29 (0.00−5.24) 3.30 (0.00−7.52) 1.054Steelhead parr 2.12 (0.34−3.91) 3.96 (0.00−8.03) 4.77 (0.38−9.16) 0.133All salmonids 10.58 (5.29−15.87) 17.06 (12.12−22.01) 10.90 (4.32−17.47) 5.606

Fall (Sept. 1−30)

Sample size (N) 7 9 11Coho fry 1.82 (1.38−2.26) 2.88 (2.55−3.20) 3.34 (2.30−4.39) 2.240Coho parr 1.34 (0.75−1.93) 0.27 (0.11−0.43) 1.11 (0.64−1.58) 5.177Dolly Varden parr 1.92 (0.85−2.99) 1.19 (0.82−1.56) 2.34 (1.25−3.44) 0.736Steelhead fry 0.29 (0.17−0.42) 0.59 (0.42−0.76) 1.12 (0.48−1.75) 0.893Steelhead parr 1.72 (1.02−2.43) 3.14 (2.19−4.10) 4.22 (2.68−5.77) 0.242All salmonids 7.10 (4.92−9.27) 8.08 (6.98−9.18) 12.14 (9.77−14.50) 3.430

a * P<0.05.


23

Because coho fry did not grow enough from summer to fall to compensate for declines in abundance,standing crop in unlogged and logged reaches declined from summer to fall by 43 and 70% (P<0.05),respectively. Because the decline in numbers from summer to fall was small, increases in fry size in masswasted streams resulted in a 94% increase in biomass (P<0.05). The relatively low numbers of fry presentin mass wasted streams in early summer were thus offset to a considerable extent by a combination oflower mortalities and significantly higher growth rates.

4.8 Smolt Yields and Overwinter Survivals

Non-mass wasted study reaches produced, on average, 1.5−3.3 times more smolts of each species of fishthan did mass wasted reaches (Table 12). For coho, average smolt production varied from 0.07 to 0.12 fishper metre of stream in mass and non-mass wasted streams, respectively; Dolly Varden and steelhead troutsmolt production varied from 0.03 to 0.09 and 0.02 to 0.04 fish per metre of stream, respectively. Averageapparent overwinter survivals were also 2.1−3.5 times higher in non-mass wasted streams, ranging from5.2 to 17.6% for coho, from 7.1 to 18.1% for steelhead, and from 32.4 to 68.3% for Dolly Varden.

TABLE 12. Densities, smolt yields and apparent overwinter survivals of juvenile coho salmon, DollyVarden char and steelhead trout in each detailed study stream, October 1983 to July 1984

Fall Spring Smolt Smolt ApparentStream density density yield yield overwinter

(no. fry, parr/m) (no. parr/m) (N) (n/m) survival (%)

Coho salmon

Mass wastedMacmillan 0.98 0.00 12 0.02 2.0Saltspring 1.29 0.00 2 0.02 1.6Southbay Dump 1.86 0.04 109 0.18 12.1

Non-mass wastedSchomar 4.44 0.02 92 0.14 3.6Tarundl 2.26 0.18 102 0.15 14.6Piper 0.89 0.23 66 0.08 34.7

Dolly Varden

Mass wastedMacmillan 0.10 0.00 0 0.00 0.0Saltspring 0.17 0.01 0 0.00 5.9Southbay Dump 0.23 0.08 48 0.08 71.3

Non-mass wastedSchomar 0.69 0.30 152 0.23 76.5Tarundl 1.34 0.10 6 0.01 8.2Piper 0.77 0.05 27 0.03 10.6

Steelhead troutMass wasted

Saltspring 0.51 0.02 8 0.02 7.1Non-mass wasted

Schomar 0.60 0.04 28 0.04 13.7Tarundl 1.01 0.21 11 0.02 22.4

Within each stream, coho salmon always had the lowest apparent overwinter survival of the threespecies, but because of their greater initial abundance the previous fall, they usually produced the mostsmolts. Dolly Varden, by comparison, always had the highest apparent overwinter survival (P<0.05, Mann-Whitney U-test), but — with one exception (Schomar Creek) — always produced fewer smolts than coho.This was partly because Dolly Varden densities were substantially lower in fall than coho, and partlybecause a large portion of the fish that survived to spring were juveniles too small to smolt. Steelhead hadon average the same survivals as coho in each reach type, but produced the fewest number of smolts foressentially the same reasons as for Dolly Varden.

The actual number of smolts captured in any one stream was never high. The highest number of cohosmolts was taken in Southbay Dump (N = 109), the lowest in Macmillan and Saltspring creeks (N = 12).


24

Numbers of coho smolts in the rest of the streams (all non-mass wasted) ranged from 66−102. Undoubtedlysome smolts were missed when floods prevented trapping for a 3- to 10-day period in each stream in lateMay. However, since the traps were inoperative in each stream at essentially the same time (Figure 6), thecomparison between mass and non-mass wasted streams should not have been seriously affected. Ifanything, since the daily catches of coho smolts were highest in Piper and Schomar creeks just before thefloods, but low or non-existent in the rest of the streams (including all the mass wasted streams), theflooding that occurred may just as easily have masked greater differences between smolt production andfish overwinter survivals in mass and non-mass wasted streams.

As noted earlier, some of the habitat in Southbay Dump Creek was improved when several large logswere added to the stream. As a result, parts of Southbay Dump Creek were more like unaffected streams,which may have accounted for the high coho survivals observed (15%) relative to other mass wastedstreams (1.8%). Conversely, overwinter conditions in Schomar Creek were probably more like mass wastedstreams because of the large jam above the study area that broke open during the winter and inundated theentire study area with gravel. Eliminating Southbay Dump and Schomar creeks from the data, averageapparent coho overwinter survival in the two remaining non-mass wasted study reaches (24.5%) was 12.9times greater than in the two remaining mass wasted reaches (1.8%), while the average number of smoltscaptured (0.12 fish per metre of stream) was 6 times greater than in mass wasted reaches (0.02 fish permetre of stream).

4.9 Coho Fry Recruitment

At the end of the study in July 1984, when fry movements downstream had ceased (Figure 6), the numberof coho fry left in Macmillan Creek and Southbay Dump Creek (0.16 fry per metre) was much lower than inthe previous fall (1.40 fry per metre, P<0.05). Reasons for the poor fry recruitment are unknown, but sincemass wasted streams also had significantly fewer fry displaced downstream during the spring outmigrationperiod (0.03 versus 5.34 fry per metre of stream P<0.05) than did non-mass wasted streams, excessive frydisplacement downstream was clearly not a factor. During the downstream migration period, coho fry wereconspicuously absent in mass wasted streams where a total of only 47 fry were captured compared to12240 fry (range 1071−9479, Figure 6) in non-mass wasted streams.

The extensive bedload movement recorded in mass wasted streams suggests that excessive egglosses due to scouring could be a critical factor in determining coho fry recruitment to mass wastedstreams. Data collected the same winter on mean scour depth (X) at pool/riffle breaks in each detailedstudy stream (data from Tripp and Poulin 1986b) showed a significant negative correlation with outmigrantfry density (Y, ln[Y + 1] = 10.048−2.911 ln X, r = -0.907, P<0.01). There was also a significant negativecorrelation between scour depth and the sum of the outmigrant plus July fry densities (r = -0.885, Figure 7).

25

FIGURE 6. Downstream movements of juvenile coho salmon in mass and non-mass wasted studystreams, April 10−June 29, 1984. The line graphs are numbers of fry, the vertical bars arenumbers of smolts. Bold horizontal bars on the axis are times that the weirs and traps werenot operating.

26

FIGURE 7. Correlation between mean scour depth and coho fry density (spring outmigrants plus Julyresident fish combined) in detailed study streams. Data on mean scour depths in each streamfrom Tripp and Poulin (1986b).


27

5 DISCUSSION

5.1 Effects of Logging and Mass Wasting on Rearing Habitat

The habitat characteristics of unlogged, logged and mass wasted stream reaches in this study wereessentially the same as those reported previously in the synoptic habitat survey (Tripp and Poulin 1986a).Since the latter survey included an additional seven reaches that were not readily accessible toanadromous fish, limiting the data to only accessible stream reaches did not affect the results very much.The results, if anything, were clearer, because there were no longer significant differences in gradient, basinarea or reach distance upstream to contend with between reach types. This does not mean that factorssuch as gradient had no effect on habitat characteristics; rather, the effects were homogeneous relative toreach type. Pool depth, for example, was significantly correlated with gradient, but since the slopes andintercepts of the relationships were the same for each reach type, there was no reason to include gradientas a covariate with other data when comparing reach types.

The only two meaningful changes observed in the rearing habitat of logged streams versus unloggedstreams were reductions in the amount of LOD (down 31%) and stable, rooted undercut bank habitat (down61%) present. While only the reduction in bank cover was statistically significant, both changes wereconsistent with the manner in which all but two of the reaches were logged (i.e., the total removal of alltimber in the riparian zone, including the stream banks). In several cases, many of the large logs originallypresent in the stream were also removed, though over-zealous clearing to improve fish passage may havebeen involved as well.

Since the riparian vegetation left along each logged study reach was composed almost entirely of redalder with relatively small stems (average age 24 years), there was less LOD left to enter the stream than inunlogged streams, and thus less LOD to compensate for LOD displaced downstream. There were alsosignificantly fewer large root masses to replace the deep undercut bank habitat present under theconiferous tree roots of unlogged streams. Young alder (<25 years old) may help to stabilize stream banks,but it rarely develops a root system large or rugged enough to provide fish more than the barest protectionfrom peak flows or predators. Where alder did form good undercut bank habitat, usually it was a group ofolder trees (>45 years old) clumped together on the remnants of old conifer logs or root wads embedded inthe bank.

Grette (1985) reported similar findings in old logged streams in Oregon, though the reductions ininstream LOD observed in Oregon were greater than in this study. As the present results show, the amountof debris in logged reaches may have declined more than that in unlogged streams, but the amount ofdebris left instream was still substantial. Assuming a mean diameter for LOD of 0.41 m (the average meandiameter recorded in the detailed study reaches) and a specific gravity of 0.6 (air-dried hemlock), thecoverage estimated in planview for logged streams is approximately equal to 27 kg/m2. This value iscomparable to the amounts of LOD recorded in unlogged, old-growth streams elsewhere in the PacificNorthwest (Harmon et al. 1986). The debris in the present study was possibly larger and thus less mobilethan in other regions. More of the debris in the present study may also be coming from upstream sources orthe adjacent hillslopes, and not just the riparian zone. Hogan (1986) also reported a decline in the amountof debris in more recently logged streams on the Queen Charlotte Islands, though changes in thedistribution and orientation of the debris appeared to be more significant than actual reductions in debrisvolumes. Single pieces tended to be reoriented parallel to the main axis of the stream, or clustered intolarger and more widely spaced log jams.

Mass wasting in logged streams further reduced the amount of LOD and undercut bank cover presentto approximately one-half and one-quarter, respectively, of the amount present in unlogged streams. Unlikelogged streams, where declines in LOD and undercut bank cover in the absence of mass wasting werelargely a function of changes in the riparian zone, changes in the habitat of mass wasted streams wererelated more to changes in sediment loads and the catastrophic redistribution of LOD. Of the two factors,the change in sediment loads appeared to be most important since its presence or absence was mainlyresponsible for the significant declines also observed in pool area (down 34%) and mean pool depth (down38%). Dewatering, as shown by the significant decline in wetted area (down 50%) relative to channel width,was also related primarily to the amount of sediment present in the channel.


28

The process by which mass wasting altered the stream habitat of logged streams depended stronglyon stream gradient. At gradients less than 3.5%, stream reaches affected by torrents tended to bedepositional zones, while higher gradients tended to be sediment transport zones (Tripp and Poulin 1986a).The impacts in transport zones are most commonly associated with debris torrents, though in terms ofoverall productivity it is probably the lower gradient depositional zones that show the greatest changes inproductivity. Rearing and spawning habitats are typically lost in the steep transport zones of small streamsbecause the sediments originally stored within the channel are scoured out and carried downstream, alongwith substantial portions of the stream banks and instream debris. Debris-controlled pools and spawninggravels are replaced by coarse, bouldery or bedrock substrates.

Habitat losses in depositional zones were mainly the result of sediments filling up the channel. In somecases, they actually spilled over the banks and spread over the floodplain. While abundant gravel waspresent to form new pools, the LOD needed to form new pools was buried, concentrated into large log jamsacross the stream, or windrowed along the valley walls. Further erosion and infilling often occurred as aresult of lateral channel movements or secondary torrents upstream. New channels are eroded in thedeposits until eventually the streambed becomes armoured with boulders, at which time further develop-ment slows down dramatically in the absence of large rough agents like LOD. Typical habitat in thedepositional zones of small debris-torrented streams on the Queen Charlotte Islands was a series of longcobble- or boulder-bedded riffles separated by a few small pools where LOD impinged on the stream.

Mass wasting simplified the rearing habitat of directly affected reaches, primarily by reducing pool areaand the number of pieces of LOD in the stream. The type of cover present also changed considerably, withmore small pools scoured out around boulders and fewer large pools scoured out around LOD. There wasless undercut bank in pools controlled by woody debris, but the pools themselves were comparatively smalland shallow in relation to the amount of debris present. Full-length logs with intact rootwads were rare.Critical winter cover in particular (deep, complex pools with abundant LOD and undercut banks) wasreduced in mass wasted streams.

5.2 Effects on Coho Salmon

5.2.1 Logging impacts

Most streams in this study tended to have relatively narrow floodplains, with very little tributary, pondor slough habitat available off-channel. This means that various factors such as annual ground-watertemperature regimes, sediment budgets and LOD inputs are likely quite different from those docu-mented in other streams with significantly wider alluvial plains (Hartman et al. 1984). It also meansthat the annual migration patterns of juveniles within each stream are probably simpler than those inmore complex drainages where the presence of off-channel habitats allow some fish to escape thewinter high-flow conditions present in mainstem habitats.

The results of this study indicate that logging with no mass wasting has a positive impact onsummer juvenile fish population levels on the Queen Charlotte Islands, particularly on coho salmon. Inthis regard, the results are consistent with numerous other studies in the Pacific Northwest that haveshown similar increases in summer fish numbers in response to clearcut logging (Murphy et al. 1981,1986; Hawkins et al. 1983; Bisson and Sedell 1984; Scrivener and Anderson 1984). What appears tobe unique about the increase in coho numbers in this study is that the increase bore no relationship tothe amount of light each reach was actually exposed to. Logging ended on average 24 years beforethe study, and virtually all of the reaches were heavily shaded during summer with a dense canopy ofalder. The apparent increases in fish production observed in the heavily shaded sections of olderlogged reaches may have been subsidized by increases in production in the upper reaches of thedrainage, which were logged much more recently and were therefore much more exposed to light.

5.2.2 Mass wasting impacts

Mass wasting effectively nullified the benefits that logging appeared to have on summerstanding crops. It achieved this in directly affected reaches in two ways:

1. by increasing the amount of gravel scouring in spawning areas and dramaticallyreducing egg-to-fry survivals overwinter — to a point where too few fry were availableto ‘‘seed’’ the stream and capitalize on logging-related increases in production; and


29

2. by eliminating overwinter habitats and reducing juvenile overwinter survivals and,ultimately, smolt production.

Poor egg-to-fry survivals due to scouring has been recognized for some time as apotentially important limiting factor in streams (Everest et al. 1987), but usually only duringcatastrophic (and therefore rare) events. During the present study, the average scour depthrecorded in late February to early March 1984 ranged from an estimated 27−35 cm in thethree debr is torrented streams examined, compared to 15−22 cm in the non-torrentedstreams (Tripp and Poulin 1986b). Based on the vertical distr ibution of coho salmon eggs innew redds, this represented a potential 58−80% egg loss in torrented streams, compared to6−19% in non-torrented streams. Because many scour monitors were in fact completelydisplaced in torrented streams, actual scour depths and egg losses in torrented streams werelikely underestimated.

The results indicate that scouring is a much more pervasive and chronic problem thanpreviously anticipated. It is particularly true of stream reaches that have been directly affectedby mass wasting or that appear to be aggrading gravel due to slides and torrents upstream. Indirectly affected streams, the problem appears to be serious enough that essentially wholeyear classes may be lost during even moderate floods. Though there are no adult countsavailable, we believe this happened in two of the three torrented streams examined in thisstudy (Macmillan and Southbay Dump creeks) in 1984. The close relationship between thedepth of scour recorded during the study in 1984 and the number of coho fry producedsupports this conclusion. We also believe it happened in 1982 when coho fry were absent in 4of the 11 mass wasted stream reaches surveyed (Macmillan Creek, upper Southbay DumpCreek, Shelley Creek and Riley Creek tr ibutary).

Obviously, poor adult returns could account for the exceedingly low fry densities observedin some mass wasted streams. However, given the small size of the affected streams, returnswould have to be very low indeed since as few as four females may be all that was needed toadequately seed each stream in the absence of scouring. This assumes an average fecundityof 2500 eggs/female, a 25% egg-to-fry survival rate overwinter, and fry densities averagingfive fry per square metre in pools and one fry per square metre in the rest of the stream. Evenone spawning pair would have produced substantially more fry than was actually present inthe more severely depressed streams.

Apart from underseeding, the only effects that mass wasting appeared to have on sum-mer-fall coho population characteristics was an increase in growth rates, though this wasnegated in some streams where the reach dewatered and fish became trapped in isolatedpools. Despite significant reductions in the amount of pool habitat available, fall coho densi-ties in mass wasted streams were slightly higher than those in unlogged streams, and almostas high as those in logged streams when streams that were clearly underseeded (no fry butparr present, or fry present in other years) were excluded from the analysis. Because of bettergrowth rates experienced by fish in mass wasted streams, these streams in fall also finishedthe growing season with higher standing crops of coho, compared to those crops in loggedand unlogged streams. Total salmonid standing crops were also highest in mass wastedstreams.

Reasons for the high growth rates in mass wasted streams that manage to maintain acontinuous flow of water all summer are unknown, but they did not appear to be stronglyrelated to either density dependent factors or differences in emergence times as noted inCarnation Creek (Hartman et al. 1984). They may have been related to differences in tem-perature or to the amount of light reaching the stream. Reductions in canopy cover due tologging have been correlated with increases in productivity at the lower trophic levels (Murphyet al. 1981, 1986; Hawkins et al. 1983). This would be consistent with the amount of canopycover that was present, since it was the mass wasted streams in this study that had little or nocanopy cover, while logged streams (average logging age 27 years, range 4−46 years)tended to have a dense canopy of red alder.


30

Differences in nutr ient loading rates may also account for the apparent differences inproductivity. Since debris torrents deposit large quantities of soil and decomposed organicmatter in streams, torrents may represent major infusions of inorganic nutr ients to streams.Although never chemically analysed, the substrates at the base of all debris torrent jams inthis study were invariably coated with iron precipitates. Isolated pools at the base of the jamstended to be covered with a film of nitrogenous-like compounds, similar to those found onpools at landfill sites and septic outfalls. Everest and Meehan (1981) found that large sedi-ment accumulations in torrented streams are highly active sites for the anaerobic processingof nutr ients. They also reported that coho salmon located downstream of such sites werelarger than those upstream.

The increase in growth rates and the large average size of coho fry in fall in mass wastedstreams were not enough to offset the impacts of mass wasting on overwinter survival. In thisregard, the results are similar to those of Mason (1976) who demonstrated how increases insummer/fall densities through supplemental feeding could be negated by a lack of overwinterhabitat and poor overwinter survivals. Since the streams with the largest fry had the lowestapparent survivals overwinter, the results also indicate that there is a limit to how muchincreases in size can compensate for declines in the quantity or quality of overwinter habitatpresent. Size in fall was not related to survival until spring, as was reported for coho salmon inCarnation Creek (Holtby 1987).

The differences observed in coho, steelhead and Dolly Varden overwinter survival ratesbetween mass and non-mass wasted logged streams in this study were consistent with theresults of the habitat surveys. Overwinter cover, as defined in this study, declined from 14.5%in logged streams to 6.6% in mass wasted streams, which was approximately the sameapparent decline observed in overwinter survivals of coho salmon (17.6 to 5.2%) and steel-head trout (18.1 to 7.1%). Overwinter survivals and the amount of winter rearing habitatpresent in the mainstem portions of mass wasted streams on the Queen Charlotte Islandswere therefore generally related, and more indicative of the overall direct impacts of masswasting on fish and fish habitats than were density or biomass during summer. The resultswere not unexpected. They agree with many other studies that have shown juvenile cohosalmon survivals overwinter are related primarily to the amount of quiet-water habitat avail-able during high-flow periods (Skeesick 1970; Bustard and Narver 1975; Tschaplinski andHartman 1983; Hartman and Brown 1987).

Because they were based only on fish actually captured moving downstream in spring,the overwinter survival estimates reported in this study are minimum estimates of the truesurvivals of coho salmon and steelhead trout in mass and non-mass wasted streams. They donot take into account fish which may have migrated downstream when the traps were outduring storms in May. As indicated earlier, this should not have seriously affected the compar-ison between mass and non-mass wasted streams, since the traps were inoperative in eachstream at essentially the same time. The estimates also do not account for fish that may havemigrated out of the study reaches after the fall estimates but before the traps were installed inthe spring. Again, this was probably not a serious error in the mass wasted study streams,where the entire stream population was being estimated and fish had to have smolted andgone to sea before trapping started. However, it may have been an important factor in two ofthe non-mass wasted study areas (e.g., Schomar and Tarundl creeks) where freshwaterhabitat was available downstream. If so, then the differences in apparent survivals betweenmass and non-mass wasted streams should have been greater than indicated. In Schomarand Tarundl creeks, fish migrating downstream overwinter were treated as total mortalities,although it seems unlikely that mortality would have been 100% had fish not been able tomove downstream.

5.2.3 Downstream effects

The comparatively low densities of coho salmon found in logged streams downstream ofmajor mass wasting events seem to suggest that mass wasting may not only nullify the


31

benefits of logging in directly impacted reaches, but may also nullify the benefits accrueddownstream of impacted reaches. If true, the phenomenon would have an important impact onfuture management strategies, since the incidence of major mass wasting upstream of fishbearing reaches is much more common than it is downstream. Unfortunately, the number oflogged streams sampled with mass wasting upstream was too small to be meaningful,particularly in fall when only two such reaches were sampled. Numbers of fry were alsosignificantly lower in the summer samples (N = 4), but the samples were taken, on average, amonth after the samples in logged streams without mass wasting upstream. The differences inabundance could easily be accounted for by the normal exponential declines in abundanceshown by most stream-dwelling salmonid fry in early summer.

If the decline in numbers below impacted stream reaches is real, one good candidate forthe cause of the decline is a decrease in egg-to-fry survivals due to increases in the amount offine and sand sized sediments present in spawning areas. Results of synoptic surveys on thegravel composition of streams affected by logging and mass wasting on the Queen CharlotteIslands (Tripp and Poulin 1986b) indicated that logging and mass wasting both significantlyincreased the amount of fine and sand sized sediment in spawning areas. Logging aloneincreased sediment levels as much as mass wasting upstream did in unlogged streams, butthe greatest declines in gravel quality were observed where logging and mass wastingupstream occurred together. When Fredle values (a measure of gravel quality based on meanparticle size and porosity) were converted to estimates of coho egg-to-fry survival relation-ships developed at Carnation Creek (Scrivener 1987), the survivals in logged streams withmass wasting upstream could be reduced to 5−15%, assuming natural egg-to-fry survivalsare 25%. This is approximately the same decrease in egg-to-fry survival shown for chum andcoho in Carnation Creek, for about the same increase in the amount of fines present(Scrivener and Brownlee 1989).

Juvenile populations downstream of directly affected reaches may also be reduced bysudden major downstream movements of gravel overwinter, as suggested in this study whenthe debris jam upstream of the study reach in Schomar Creek was breached during a storm inMay and large quantities of gravel completely inundated the stream. Pools over 40 m2 in areaand 2 m deep were completely filled in, while the area upstream of the fish weir at the mouth ofthe stream filled in to a depth of 1 m. Within a week, another storm washed all the gravel out ofthe stream, after which little visible evidence was left that the stream had ever been affectedin the first place. Physical injuries due to entombment by the gravel, followed by the rapidmovement of gravel downstream, presumably contr ibuted substantially to the low apparentoverwinter survivals of coho (3.6%) in Schomar Creek. Steelhead trout showed better sur-vivals (13.7%), though many of the fish captured after the above storms had large bruises.Extensive bruising was also noted on steelhead trout in Saltspring Creek, where apparentsurvival overwinter was comparatively low (7.1%). Few, if any, bruises were noted on fish inother streams, nor were they especially evident on smolts or juveniles captured in Schomar orSaltspring Creek before the storms. The bruises were confined almost exclusively to one orboth halves of the head, which was distinctly different from the bruises occasionally noted onthe backs or sides of fish that have been electroshocked.

5.3 Effects on Other Species

How logging and mass wasting affected the abundance and summer standing crop of steelheadtrout and Dolly Varden was not as evident as for coho. There are several possible reasons. First,because Dolly Varden fry were difficult to sample, they were excluded from analysis. Effects on thisspecies may therefore have been overlooked. Second, since Dolly Varden and steelhead trout parrwere not distinguished according to age class, poor survivals among the younger age classes (1-and 2-year-old fish) could have been masked by better survivals among older age classes (3- and4-year-old fish). Finally, since all the study streams contained resident Dolly Varden upstream, fishmoving downstream into the study reaches in spring likely inflated the estimates of overwintersurvival.


32

Differences between the life history strategies of coho salmon and steelhead trout couldaccount for some of the differences in the response of these two species to mass wasting orlogging. In contrast to coho or Dolly Varden, steelhead trout are spring spawners. Thus, as long asthere are enough adults there should be little danger of streams being underseeded because ofexcessive scour in spawning areas. As shown in this study, major storms can occur at almost anytime of the year on the Queen Charlotte Islands. Most, however, occur in fall and winter, whichmeans that coho eggs run a proportionately greater r isk of being scoured out than do steelheadeggs. Other studies have documented similar changes in species composition when the eggs ofone species were exposed to more scouring than the eggs of another species (Kondolf et al. 1991).

Differences in habitat use could also be important. The results of this and other studies(Hartman 1965; Glova 1978; Tschaplinski and Hartman 1983) have shown that coho are stronglydependent on stable pool habitats in both summer and winter, particularly those with abundantLOD. Rainbow trout also prefer pools, but they sometimes show a greater use than coho of otherhabitats as well, including the lateral margins and boulder shadows of r iffles (Murphy et al. 1984).Winter habitat use by steelhead trout is evidently also more variable than is coho’s betweenstreams. In Carnation Creek, Hartman and Brown (1987) found that trout abundance in winter wasnot correlated with the amount of cover (primarily LOD) as was the case for coho (Tschaplinski andHartman 1983). Rather, numbers were highest in open pool and r iffle locations with the leastamount of woody cover. In SE Alaska streams, by contrast, Heifetz et al. (1986) found very fewcoho, rainbow trout, or Dolly Varden in open pools, r iffles or glides in winter. Almost all fish werelocated in deep pools with cover.

The generally high rainbow trout densities and biomass found in logged and mass wastedstreams, relative to unlogged streams in the synoptic survey portion of this study, agree with similarfindings in California (Burns 1972) and western Washington (Bisson and Sedell 1984). They differmarkedly from Carnation Creek, however, where rainbow trout numbers declined dramatically afterlogging (Hartman and Holtby 1982). Hartman (1987) suggests that the decrease in rainbow troutnumbers is largely the result of declines in the amount of suitable overwinter habitat present in themainstem portion of the stream. Since rainbow trout are less likely than coho or cutthroat trout touse the small tr ibutary or slough habitats along Carnation Creek (Hartman and Brown 1987),Hartman (1987) further suggests that rainbow trout may be more vulnerable than coho or cutthroattrout to physical changes in the mainstem after logging.

The above hypothesis does not consider the fact that while off-channel habitats are valuable,more than 75% of the post-logging smolt production in Carnation Creek still came from themainstem portion of the stream (Brown and McMahon 1987). If the mainstem was capable ofproducing more coho, despite a decline in the amount of suitable overwinter habitat available, itshould have supported more trout as well. Better coho survivals in Carnation Creek overwinterafter logging were attr ibuted primarily to the increase in the average size of fish entering thewinter. Given the larger size of rainbow trout parr, they too should have fared well.

Abundant good habitat in the form of deep pools with complex LOD and stable undercut banksdoes not appear to be a prerequisite for high fish densities, or even good overwinter juvenilesurvivals. Torrented r ight down to the ocean, Shelley Creek on the west coast of the QueenCharlotte Islands was a small, severely affected stream with very limited habitat. It neverthelessmanaged to support the largest concentrations of steelhead parr found anywhere in the synopticsurvey in both summer and fall (1.36−1.55 parr per metre of stream). Coho numbers, by compari-son, were among the lowest recorded (0.46−0.49 fish per metre, fry and parr combined).

The apparent differences between rainbow trout survivals in this and the Carnation Creekstudy may be related to other factors, such as the degree of competition that occurs between cohosalmon and rainbow trout in early spring before smolting. In a study on the effects of cohoheadwater stocking programs on cutthroat trout in two streams on southern Vancouver Island, theperiod between early March and the smolt outmigration period was found to be a period of veryrapid growth for coho salmon and trout in streams (Tripp and McCart 1983). Competition between


33

coho pre-smolts and trout may therefore be more intense at this time than at any other time of theyear. In the above study, coho salmon reduced trout production by as much as 40% in 1 year. Thegreatest reduction in trout production occurred in the stream that had the largest concentration ofcoho at the beginning of the spring growth period.

The variations in survival shown between rainbow trout and coho salmon in response tologging and mass wasting require further study. If competition between coho salmon and rainbowtrout is a factor in other streams during the spring pre-smolt period, it could help explain whyrainbow trout numbers dropped so precipitously in Carnation Creek after logging in response to asudden increase in the number of coho pre-smolts present. It could also help account for thecomparatively high numbers of rainbow trout present in those logged or mass wasted streams onthe Queen Charlotte Islands where coho numbers are low.

5.4 Management Implications

For coho salmon, and to a lesser extent steelhead trout, the declines in smolt yields attr ibutable tomass wasting appear to be substantial in directly affected reaches. In small streams, where theentire channel has been torrented and no habitat exists other than that in the mainstem, theimpacts may in fact be too great to ensure an adequate return of adults or, therefore, self-sustaining populations. Spawning and overwinter habitats must first improve. As shown inSchomar Creek, even juvenile populations in logged reaches with mass wasting upstream may beat r isk if they are subjected to regular major sediment releases from debris jams that are breakingdown upstream. While the data are somewhat equivocal on this point, spawning habitat, egg-to-frysurvivals and fry recruitment may also be adversely affected in logged areas located downstreamof major mass wasting events.

Reductions in overwinter juvenile survivals and smolt yields as a result of mass wasting havean obvious impact on the harvestable surplus of coho salmon and steelhead trout, which needs tobe taken into account when the optimum fishing level for these species are assessed. What is notclear is that poor egg-to-fry survivals and fry recruitment may also reduce the available catch forthese species further. Considerably more spawners than normal might be required to ensure thatall of the available rearing habitat in affected streams is fully used, and that all of the potentialsmolt yield is realized. Typically, poor egg-to-fry survivals are considered to have a direct impactonly on the harvestable surplus of species such as chum or pink salmon, which go to sea shortlyafter emergence.

Despite the best planning, even a well-executed forest harvesting operation is vulnerable tolandslides in steep coastal forests where mass wasting is the dominant geomorphic process. Inthese areas it is not possible to identify all unstable sites within generally acceptable cut-blocks.Nevertheless, prevention of upslope mass wasting and chronic sources of surface soil erosionmust remain a priority in landslide-prone areas. In particular, improvements and standardization ofterrain stability mapping and engineering layout are critical for managing both fisheries andforestry resources. Standardization would lead to a greater uniformity in the interpretation ofcutblock submissions and the application of terrain analyses, and hence a greater likelihood thatspecific problem areas would be identified. More onsite inspections during and after logging arealso needed in sensitive areas (e.g., gullies, steep slopes, and sensitive road locations) to assessthe effectiveness of the various prescriptions recommended to prevent mass wasting.

The importance of providing forest streams with long-term sources of LOD is well acceptednow (Bisson et al. 1987). Blowdown, however, is sufficiently acute on the Queen Charlotte Islandsto require logging in many r iparian areas to the stream edge. This practice minimizes bank damagecaused by blowdown, but reduces the management options available to fishery managers formaintaining long-term debris supplies to streams. Even current practices of leaving one side of awatershed forested until clearcut blocks green-up may significantly limit natural debris fall intostreams, if the logged sides of the stream exhibit greater rates of bank erosion than the unloggedsides. With rotational periods of 100 years or less, it is uncertain whether enough large debris


34

would ever be available to stream channels to maintain the LOD that currently helps stabilizegravel and create the preferred pool habitat typical of streams on the Queen Charlotte Islands.

Land use managers must recognize that good streamside logging practices and debris man-agement in the r iparian zone are not the only solutions to problems of mass wasting. For somestreams, irrespective of the amount of debris present or available to the stream, if the sedimentsupply entering the stream from mass wasting or erosion upstream exceeds the capacity of thecurrent channel configuration to transport the material, then the channel will adjust its form inresponse to the increase in sediment loads. In most cases this will be to the detr iment of theexisting fish habitat. Land use managers must also recognize that not all debris comes from theriparian zone, and that the relative importance of other sources of LOD to the channel must beconsidered. In some streams, the bulk of the debris in fish bearing sections of the stream maycome from the adjacent hillslopes in the form of torrents, in which case it may be best to leave thetimber along the gully systems where mass wasting is most likely to occur. Leave str ips might thenbe confined along the stream to only those stream sections where bank erosion and lateral channelmovements are the principal agents responsible for bringing extra debris into the channel. In thisway, leave str ips could be tailored by size and location to better reflect the importance of eachstand of timber as an LOD source for the stream.

There is an extensive backlog of streams on the Queen Charlotte Islands that have beenseverely damaged by mass wasting in logged areas. These streams represent a legacy of pastpractices that are becoming increasingly difficult problems to solve. Declining access due to roaddeactivation following logging, combined with a lack of cost-recovery programs for streamimprovement, underline the need for preventing new slope failures that could lead to streamdamage. Current logging practices have not been studied long enough to determine the rate atwhich fish populations or stream channels recover from the effects of mass wasting, but it seemsreasonable to assume that the process will be long term unless some alternative source of LOD isprovided and sediment supplies are controlled.

Stream rehabilitation options on the Queen Charlotte Islands are limited. Because streamssubjected to mass wasting are typically incised channels with narrow floodplains and limited off-channel areas suitable for rearing pools or spawning channels, rehabilitation techniques mustfocus on instream techniques that establish juvenile overwinter habitat and stable spawning areasin the mainstem portions of the stream. Ongoing experimentation with LOD structures in streamson the Queen Charlotte Islands suggests that log placements in channels can improve pool to r iffleratios, increase width and depth variability and initiate the development of diverse gravel storageareas. Regular, large scale bedload movements and secondary torrents, however, continue toaffect the experimental work and reduce the effectiveness of the LOD emplacements.

To be fully effective, stream rehabilitation efforts must be combined with a management planthat also minimizes upslope hazards. However, given that regular large scale movements of debrisand sediments downstream (i.e., secondary torrents) are characteristic of most severely masswasted streams, rehabilitation efforts must also focus on structures that are capable of maintainingor re-establishing critical habitat when affected again, or that take advantage of the large scalemovements of debris and sediment downstream to enhance the existing habitat. Rehabilitationmust also take into account the species to be benefited, since increases in survival and productionof one species may be gained at the expense of other species.

5.5 Recommendations for Further Study

Recommendations for further fisheries studies, based on the results of the present study, are asfollows:

1. Further study is needed to determine if increases in production of recently logged headwa-ter areas are transferred downstream to fish-bearing reaches of older logged areas.Further study is also needed to determine how such a ‘‘subsidy,’’ if it exists at all, istransferred downstream (i.e., how much of the subsidy is due to increased temperatures


35

and inorganic nutr ient concentrations, and how much is due to higher drift rates ofparticulate organic matter, periphyton, or benthic invertebrates).

2. The studies to date on the the effects of mass wasting on streams on the Queen CharlotteIslands have focused mainly on smaller, moderate gradient streams where the distancestravelled downstream by debris torrents can be substantial. Further study is required onthe effects of mass wasting on fish and fish habitat in larger streams, where the distancescovered by mass wasting events may be relatively short, but the number of eventsimpinging on the channel may be great; and where channel-scale effects related to stemlength, root strength and bank erosion characteristics are likely to be quite different fromthose of smaller streams. The origin, transport rates and storage characteristics of LODand sediment in larger streams are also likely to be quite different from those in smallstreams, as are the responses exhibited by juvenile fish populations.

3. The present study focused primarily on the effects that mass wasting had on streams 4−6years after they were affected. Further study is required on the longer term effects of masswasting on egg-to-fry survivals and smolt yields in logged and unlogged streams.

4. Egg-to-fry survivals, juvenile overwinter survivals and smolt yields need to be betterquantified in logged and unlogged streams, with and without mass wasting. The presentstudy assessed these variables in logged streams only.

5. The possibility that the adverse effects of mass wasting may extend downstream beyonddirectly affected reaches has major implications for the management of fish stocks in lowerreaches of streams. Further study is recommended so that the cause, extent and durationof these downstream impacts can be more clearly defined.

6. Coho salmon and rainbow and steelhead trout may exhibit different responses to loggingand mass wasting, possibly because of fundamental differences in their spawning timesand thus egg-to-fry survivals, and possibly because of their competitive interactions forfood and cover after emergence. Further study is recommended to define how logging andmass wasting affect these interactions, and what the net effects are on smolt yields andadult returns.

SEQ 6570 JOB FISH-007-005 PAGE-0036 LIT CIT REVISED 21SEP00 AT 08:18 BY BC DEPTH: 60 PICAS WIDTH 42.09 PICAS COLOR LEVEL 1

36

LITERATURE CITED

Bisson, P. A., J. L. Nielsen, R. A. Palmason, and L. E. Grove. 1982. Stream habitat definition 1. Summer lowflow habitats and their utilization by salmonid fishes. In Proc. Symp. on Acquisition and Utilization ofAquatic Habitat Inventory Information. West. Div. Am. Fish. Soc., Portland, Oreg.

Bisson, P. A. and J. R. Sedell. 1984. Salmonid populations in streams in clearcut vs. old-growth forests ofwestern Washington. In Fish and wildlife relationships in old-growth forests. Proc. Symp., Juneau,Alaska, 12−15 April 1982. W. R. Meehan, T. R. Merrell Jr., and T. A. Hanley (editors). Am. Inst. Fish.Res. Biol., Juneau, Alaska, pp. 121−129.

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 in the Pacific northwest: past, presentand future. In Streamside management: an interdisciplinary symposium on forestry and fisheryinteractions. E. O. Salo and T. Cundy (editors). Fish. Res. Inst. and Coll. For. Resources, Univ. Wash.,Seattle, Wash. pp. 143−190.

Brown, T. G. and T. McMahon. 1987. Winter ecology of juvenile coho salmon in Carnation Creek: summaryof findings and management implications. In Proc. Workshop on Applying 15 years of CarnationCreek Results. T. W. Chamberlin (editor). Carnation Creek Steering Committee, Pac. Biolog. Sta.,Nanaimo, B. C., pp. 108−117.

Burns, J. W. 1972. Some effects of logging and associated road construction on Northern Californiastreams. Trans. Am. Fish. Soc. 101:1−16.

Bustard, D. R. and D. W. Narver. 1975. Aspects of the winter ecology of juvenile coho salmon(Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Board Can.32(5):667−680.

Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoologicalsample censuses. Univ. Calif. Publ. Stat. 1:131−160.

Everest, F.H., R.L. Beschta, J.C. Scrivener, K V. Koski, J.R. Sedell, and C.J. Cederholm. 1987. Finesediment and salmonid production: a paradox. In Streamside management: an interdisciplinarysymposium on forestry and fishery interactions. E. O. Salo and T. Cundy (editors). Fish. Res. Inst. andColl. For. Resources, Univ. Wash., Seattle, Wash. pp. 98−142.

Everest, F.H. and W.R. Meehan. 1981. Forest management and anadromous fish habitat productivity. In46th North Am. Wildl. and Nat. Resource Conf. Trans. Wildl. Manage. Inst., Wash., D.C., pp.521−530.

Gimbarzevsky, P. 1988. Mass wasting on the Queen Charlotte Islands: a regional overview. B. C. Min. For.Lands, Victoria, B. C. Land Manage. Rep. 29. 95 p.

Glova, G.J. 1978. Pattern and mechanism of resource partitioning between stream populations of juvenilecoho salmon (Oncorhynchus kisutch) and coastal cutthroat trout (Salmo clarki clarki). PhD thesis.Univ. B.C. Vancouver, B.C. 170 p.

Grette, G. 1985. The role of large organic debris in juvenile salmonid rearing habitat in small streams. M.S.thesis. Univ. Wash., Seattle, Wash. 105 p.

Harmon, M. E., J. F. Franklin, F. J. Swanson. 1986. Ecology of coarse woody debris in temperateecosystems. Adv. Eco. Res. 15:133−302.

Hartman, G.F. 1965. The role of behavior in the ecology and interaction of underyearling coho salmon(Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Board of Canada22:1031−1081.


37

. 1987. Some preliminary comments on results of studies of trout biology and logging impacts inCarnation Creek. In Proc. Workshop on Applying 15 years of Carnation Creek Results. T. W.Chamberlin (editor). Carnation Creek Steering Committee, Pac. Biolog. Sta., Nanaimo, B. C., pp.175−180.

Hartman, G.F., B.C. Andersen, and J.C. Scrivener. 1982. Seaward movement of coho salmon(Oncorhynchus kisutch) fry in Carnation Creek, an unstable coastal stream in British Columbia. Can.J. Fish. Aquat. Sci. 39:588−597.

Hartman, G. F. and T. G. Brown. 1987. Use of small, temporary, flood plain tributaries by juvenile salmonidsin a west coast rain-forest drainage basin, Carnation Creek, British Columbia. Can. J. Fish. Aquat.Sci. 44:262−270.

Hartman, G.F. and L.B. Holtby. 1982. An overview of some biophysical determinants of fish production andfish population responses to logging in Carnation Creek, British Columbia. In Proc. Carnation CreekWorkshop: A 10-year review. G.F. Hartman (editor). Pac. Biolog. Sta., Nanaimo, B. C. pp. 348−374.

Hartman, G.F., L.B. Holtby, and J.C. Scrivener. 1984. Some effects of natural and logging-related winterstream temperature changes on the early life history of coho salmon (Oncorhynchus kisutch) inCarnation Creek, British Columbia. In Fish and wildlife relationships in old-growth forests. Proc.Symp., Juneau, Alaska, 12−15 April 1982. W. R. Meehan, T. R. Merrell Jr., and T. A. Hanley (editors).Am. Inst. Fish. Res. Biol., Juneau, Alaska, pp. 141−149.

Hartman, G. F., J. C. Scrivener, L. B. Holtby, and L. Powell. 1987. Some effects of different streamsidetreatments on physical conditions and fish population processes in Carnation Creek, a coastal rain-forest stream in British Columbia. In Streamside management: an interdisciplinary symposium onforestry and fishery interactions. E. O. Salo and T. Cundy (editors). Fish. Res. Inst. and Coll. For.Resources, Univ. Wash., Seattle, Wash. pp. 330−372.

Hawkins, C. P., M. L. Murphy, N. H. Anderson, and M. A. Wilzbach. 1983. Density of fish and salamandersin relation to riparian canopy and physical habitat in streams of the northwestern United States. Can.J. Fish. Aquat. Sci. 40:1173−1185.

Heifetz, J., M.L. Murphy, and K V. Koski. 1986. Effects of logging on winter habitat of juvenile salmonids inAlaskan streams. N. Am. J. Fish. Manage. 6:52−58.

Hogan, D. L. 1986. Channel morphology of unlogged, logged, and debris torrented streams in the QueenCharlotte Islands. B.C. Min. For. Lands, Victoria, B. C. Land Manage. Rep. 49. 85 p. + appendices.

Holtby, L. B. 1987. The effects of logging on the coho salmon of Carnation Creek. In Proc. Workshop onApplying 15 years of Carnation Creek Results. T. W. Chamberlin (editor). Carnation Creek SteeringCommittee, Pac. Biolog. Sta., Nanaimo, B. C., pp. 159−174.

.1988. Some effects of logging-related stream temperature changes on a population of cohosalmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 45:502−515.

Holtby, L. B. and G. F. Hartman. 1982. The population dynamics of coho salmon (Oncorhynchus kisutch) ina west coast rain forest stream subjected to logging. In Proc. Workshop on Applying 15 years ofCarnation Creek Results. T. W. Chamberlin (editor). Carnation Creek Steering Committee, Pac.Biolog. Sta., Nanaimo, B. C., pp. 308−347.

Ivlev, V. S. 1961. Experimental ecology of the feeding of fishes. (Translated from the Russian by DouglasScott), Yale Univ. Press, New Haven, Conn.

Kondolf, G. M., G. F. Cada, M. J. Sale, and T. Felando. 1991. Distribution and stability of potential salmonidspawning gravel in steep boulder-bed streams of the eastern Sierra Nevada. Trans. Am. Fish. Soc.120:177−186.

Mason, J. C. 1976. Response of underyearling coho salmon to supplemental feeding in a natural stream. J.Wildl. Manage. 40:775−788.

Murphy, M. L., C. P. Hawkins, and N. H. Anderson. 1981. Effects of canopy modification and accumulatedsediment on stream communities. Trans. Am. Fish. Soc. 110:469−478.


38

Murphy, M. L., J. Heifetz, S. W. Johnson, K V. Koski, and J. F. Thedinga. 1986. Effects of clear-cut loggingwith and without buffer strips on juvenile salmonids in Alaskan streams. Can. J. Fish. Aquat. Sci.43:1521−1533.

Murphy, M. L., J. F. Thedinga, K V. Koski, and G. E. Grette. 1984. A stream ecosystem in an old-growthforest in Southeastern Alaska. Part V: Seasonal changes in habitat utilization by juvenile salmonids. InFish and wildlife relationships in old-growth forests. Proc. Symp., Juneau, Alaska, 12−15 April 1982.W. R. Meehan, T. R. Merrell Jr., and T. A. Hanley (editors). Am. Inst. Fish. Res. Biol., Juneau,Alaska,.pp. 89−98.

Poulin, V. A. 1984. A research approach to solving fish/forestry interactions in relation to mass wasting onthe Queen Charlotte Islands. B.C. Min. For., Victoria, B.C. Land Manage. Rep. No. 27.

Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish andMarine Serv. Bull. No. 191. 382 p.

Roberts, R.G. and M. Church. 1986. The sediment budget in severely disturbed watersheds, QueenCharlotte Ranges, British Columbia. Can. J. For. 16:1092−1106.

Rood, K. M. 1984. An aerial photography inventory of the frequency and yield of mass wasting on theQueen Charlotte Islands. B.C. Min. For., Victoria, B.C. Land Manage. Rep. No. 34.

Scrivener, J. C. 1987. Changes in composition of the streambed between 1973 and 1985 and the impactson salmonids in Carnation Creek. In Proc. Workshop on Applying 15 years of Carnation CreekResults. T. W. Chamberlin (editor). Carnation Creek Steering Committee, Pac. Biolog. Sta., Nanaimo,B. C., pp. 59−65.

Scrivener, J. C. and B. C. Anderson. 1984. Logging impacts and some mechanisms that determine the sizeof spring and summer populations of coho salmon fry (Oncorhynchus kisutch) in Carnation Creek,British Columbia. Can. J. Fish. Aquat. Sci. 41:1097−1105.

Scrivener, J. C. and M. C. Brownlee. 1989. Effects of forest harvesting on gravel quality and incubationsurvival of chum (Oncorhynchus keta) and coho salmon (Oncorhynchus kisutch) in Carnation Creek,British Columbia. Can. J. Fish. Aquat. Sci. 46:681−696.

Seber, G. A. F. and E. D. LeCren. 1967. Estimating population parameters from catches large relative to thepopulation. J. Anim. Ecol. 36(3):631−643.

Skeesick, D.G. 1970. The fall immigration of juvenile coho salmon into a small tributary. Res. Rep. FishComm. Oreg. 2:90−95.

Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. Trans. Am. Geophys. Union38:913−922.

Tripp, D. 1986. Using large organic debris to restore fish habitat in debris torrented streams. B.C. Min. For.Lands, Victoria, B.C. Land Manage. Rep. No. 47. 25 p. + appendix.

Tripp, D. and P. McCart. 1983. Effects of different coho stocking strategies on coho and cutthroat troutproduction in isolated headwater streams. Can. Tech. Rep. Fish. Aquat. Sci. 1212:xi + 176 p.

Tripp, D.B. and V.A. Poulin. 1986a. The effects of mass wasting on juvenile fish habitats in streams on theQueen Charlotte Islands. B.C. Min. For. Lands, Victoria, B.C. Land Manage. Rep. No. 45. 36 p. +appendix.

. 1986b. The effects of logging and mass wasting on salmonid spawning habitat in streams on theQueen Charlotte Islands. B.C. Min. For. Lands, Victoria, B.C. Land Manage. Rep. No. 50. 29 p.

Tschaplinski, P.J. and G.F. Hartman. 1983. Winter distribution of juvenile coho salmon (Oncorhynchuskisutch) before and after logging in Carnation Creek, British Columbia, and some implications foroverwinter survival. Can. J. Fish. Aquat. Sci. 40:452−461.

Varnes, D.J. 1978. Slope movement types and processes. In Landslides: analysis and control. Transport.Res. Board, Nat. Acad. Sci., Nat. Res. Council, Washington, D.C. Spec. Rep. 176, pp. 11−13.

The Effects of Logging and Mass Wasting on Juvenile ... · The Effects of Logging and Mass Wasting on Juvenile Salmonid Populations in Streams on the Queen Charlotte Islands Land

Documents