Slug Injection Using Salt in Solution - uvm.edu...time spent at the field site. Note that the volume...

Introduction to Salt Dilution Gauging for Streamflow Measurement Part III:

Slug Injection Using Salt inSolutionR.D. (Dan) Moore

Introduction

Previous Streamline articlesintroduced the general principles

of stream gauging by salt dilution(Moore 2004a) and the procedure forconstant-rate injection (Moore2004b). While constant-rate injectionis best suited for use in small streamsat low flows (dischargesless than about 100 L/sor 0.1 m³/s), sluginjection can be used togauge flows up to 10m³/s or greater,depending uponchannel characteristics.Slug injection workswell in steep, highlyturbulent streams, such as thebouldery mountain channel

shown in Figure 1. This articleintroduces the conceptual

basis and field proceduresfor slug injection using

salt in solution.

Conceptual BasisIn this approach, a volume of saltsolution, V (m³), is injected as anear-instantaneous slug or gulp at onelocation in the stream. Followinginjection, the salt solution mixesrapidly throughout the depth of thestream and less rapidly across the

stream width as it travelsdownstream with thegeneral flow of water.Because some portions of astream flow faster thanothers (e.g., flow tends tobe faster in the centre thannear the banks), the cloudof salty water “stretches”downstream in a process

called longitudinal dispersion. Thisdispersion results in the cloud havinga leading edge with relatively lowconcentrations of salt solution, acentral zone of high concentrations,followed by a trailing edge ofdecreasing concentration.

If the electrical conductivity (EC) isrecorded at some point downstream,where the tracer has been completely

mixed across the stream width, thepassage of the salt

cloud will cause EC to increase fromits background value to a peak value,corresponding to the passage of thecore of the cloud, followed by adecline to background EC as thetrailing edge of the cloud passes,resulting in a characteristic salt wave(Figure 2). Longitudinal dispersionreduces the peak EC of the salt waveas it travels downstream. The timerequired for the peak of the wave tomove past an observation point willdepend inversely on the meanvelocity of the streamflow, while theduration of the salt wave will depend

Inside this issue:Introduction to SaltDilution Gauging forStreamflow MeasurementPart III: Slug InjectionUsing Salt in Solution

An Inexpensive,Automatic Gravity-fedWater Sampler forInvestigating WaterQuality in Small Streams

Live Gravel Bar StakingChannel Stabilization inthe Lower Elk River

A QualitativeHydro-Geomorphic RiskAnalysis for BritishColumbia’s InteriorWatersheds: A DiscussionPaper

Re-creating MeanderingStreams in the CentralOregon Coast Range, USA

Results of StreamlineReader Survey 2004

Update

Slug injectionworks well insteep, highlyturbulentstreams.

Volume 8 Number 2 Spring 2005

Continued on page 2

on the amount of longitudinaldispersion, which, in turn, depends onhow variable the stream velocities areacross the stream. The author hasfound that the time required for thesalt wave to pass typically varies froma couple of minutes (e.g., Figure 2) toover 20 minutes. Under low-flowconditions with low velocities, theduration can be longer than desiredfor accurate measurements (e.g., wellover 30 minutes).

At any time (t) during the salt wavepassage, the discharge of tracersolution q(t) (L/s or m³/s) past thepoint will be approximated by:

q t Q RC t( ) ( )� � [1]

where Q is the stream discharge (L/sor m³/s) and RC(t) is the relativeconcentration of tracer solution (L/L)in the flow at time (t). Equation [1]assumes that q(t) is much smaller thanQ, which should be true in virtually allcases. If the tracer discharge isintegrated over the duration of thesalt wave, and if the stream dischargeis constant over that time, then thefollowing equation should hold for aconservative tracer (i.e., one that doesnot react with other chemicals in thewater, bind to sediment, or otherwisechange as it flows downstream):

V q t dt Q RC t dt� � �� ( ) ( ) [2]

T T

where T represents the salt waveduration (s). Equation [2] can berearranged to solve for Q:

QV

RC t dt�

� ( )[3]

T

In practice, RC(t) is determined at thedownstream measurement point at adiscrete time interval �t (e.g., 1 or 5s), and the integral is usuallyapproximated as a summation:

RC t dt RC t t( ) ( )� �� [4]

T nwhere n is the number ofmeasurements during the passage ofthe salt wave. The relativeconcentration can be determinedfrom EC:

RC t k EC t ECbg( ) [ ( ) ]� � [5]

where EC(t) is the electricalconductivity measured at time t, ECbg

is the background electricalconductivity of the stream, and k is acalibration constant. The calibrationconstant, k, depends primarily on thesalt concentration in the injectionsolution and secondarily on thechemical characteristics of thestreamwater. Combining Equations[3], [4], and [5], the followingpractical equation can be derived forcomputing discharge:

2 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005

Published by:

FORREX–Forest ResearchExtension Partnership

Suite 702, 235 1st AvenueKamloops, BC V2C 3J4

Project Manager:Robin PikeTel: (250) 387-5887

Distribution/Mailing List:Janet JefferyTel: (250) 371-3923

Technical review committee:R. Scherer, R.D. Moore, R. Pike

Technical reviewers this issue:S. Babakaiff, L. Barr, R. Doucet, B. Eaton,T. Giles, B. Grainger, D. Hutchinson, P.Jordan, M. Miles, R.D. Moore, R. Pike, D.Polster, P. Raymond, R. Scherer, M.Schnorbus, K. Swift, P. Teti, R. Winkler

Publication and Web Site Support:Jesse Piccin, Satnam Brar, Julie Schooling

Graphic Layout: SigZag Design

Editing: Ros Penty

Cover Illustration: William McAuslandMcAusland Studios, Kamloops, B.C.

Streamline is published twice a year byFORREX. All articles published inStreamline are reviewed to ensure reliableand technically sound information isextended to our readers. Contentpublished in Streamline reflects theopinions and conclusions of thecontributing author(s), not those ofFORREX, our editorial staff, or our fundingpartners. Please contact Robin Pike,Streamline Project Manager, for furtherguidelines on article submission or withyour comments and suggestions.

This publication is funded in part by theBritish Columbia Ministry of Foreststhrough the Forest Investment Account,Forest Science Program, and by theUSDA Forest Service.

ISSN 1705-5989Printed in Canada© FORREX–Forest Research ExtensionPartnershipPrinted on recycled paper

http://www.forrex.org/streamline

Figure 1. Place Creek at high flow during summer glacier melt.

Continued from page 1

John

Hei

none

n

QV

k t EC t ECbg

�� [ ( ) ]

[6]

nTo apply Equation [6], we need toknow V, the volume of salt solutioninjected; measure the resultingchanges in EC at intervals of �t untilEC returns to background levels; anddetermine the calibration constant, k.

Field ProceduresChoice of a Measurement ReachSuccessful application of the sluginjection technique requires a streamreach that generates complete lateralmixing in a short distance. Selectedreaches should have as little poolvolume as possible because the slowexchange of tracer between the poolvolume and the flowing portion of thestream will greatly increase the timerequired for the salt wave to pass. Anideal reach begins with an injectionsite upstream of a flow constriction(e.g., where the flow narrows arounda boulder, promoting rapid lateral

mixing) and contains no pools orbackwater areas below theconstriction. A rough guideline is thatthe mixing length should be at least25 stream widths, but completemixing may require much longer orshorter distances, depending onstream morphology (Day 1976).Mixing the Injection SolutionWe use NaCl (table salt) as a tracerbecause it is inexpensive and readily

available. In addition, the saltconcentrations and durations ofexposure normally involved indischarge measurement are less thanthresholds associated with deleteriouseffects on organisms (Moore 2004a).Wood and Dykes (2002) observedtransient increases in invertebrate driftduring slug injection, but concludedthat salt injection had a relativelyshort-term effect and is unlikely tohave any long-term deleteriousimpacts on invertebrate communitiesat most locations.

The salt concentration in the injectionsolution should be high enough toincrease EC reasonably when usingvolumes of solution that can be easilyhandled, but it also needs to remainless than the solubility. Given the lowtemperatures often associated withfield conditions, the maximumconcentration that will dissolve readilyis about 20%, or about 1 kg of salt in5 L of water (Østrem 1964; Kite1993). We have found that a mixtureof 1 kg of salt with 6 L of water(roughly a 17% solution) provides a

suitablecompromisebetween strengthand ease ofdilution.

The injectionsolution does notneed to be mixedfrom localstreamwater.Where access tothe stream doesnot involve a longhike, it is oftenconvenient topre-mix the

injection solution to allow generoustime for dissolution and to minimizetime spent at the field site.

Note that the volume of the injectionsolution will be greater than thevolume of water used to mix it. Wehave found that when a 1-kg box ofsalt is mixed with 6 L of water, theresulting solution has a volume of6.36 L (±0.01 L). Commonly, the salt

solution is mixed in one containerthen decanted into a second,pre-calibrated container (e.g., Østrem1964). This procedure ensures thatthe salt in the injection solution iscompletely dissolved, and allowsaccurate measurement of the injectionvolume.

Required Volumes of InjectionSolution

The accuracy of a measurementdepends on how much EC increasesabove background during the saltwave passage, relative to the accuracyof the conductivity probe. The changein EC during the salt wave passagedepends, in turn, on the volume ofsalt solution and its concentration, aswell as the mixing characteristics ofthe stream. Those streams with lesslongitudinal dispersion will exhibit amore peaked salt wave with higherconcentrations, and will require lowerinjection volumes.

Kite (1993) suggested that peak ECshould be 50% higher thanbackground, while Hudson and Fraser(2002) suggested that peak EC shouldbe at least 5 times higher thanbackground. Background EC in B.C.streams typically ranges from about10 �S/cm for stormflow conditions instreams draining catchmentsunderlain by granitic bedrock, to over400 �S/cm for low-flow conditions instreams sustained by groundwaterdischarge. The author suggests thatincreasing EC by 100–200% ofbackground should be adequate forstreams with low background EC (lessthan about 50 �S/cm), while Kite’s(1993) guideline should be reasonablefor streams with background ECgreater than about 100 �S/cm.

Table 1 summarizes themasses/volumes of injected salt/saltsolution used by various authors. Therange reflects the diversity of channelmorphologies and dischargesencountered in the different studies.The author recommends starting with1 L of 15–20% solution per m³/s.Greater volumes of injected salt

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 3

0

5

10

15

20

25

30

0 60 120 180 240

Time (s)

EC

( �S

/cm

)

Figure 2. Example of salt wave in Place Creek.

Continued on page 4

solution may be appropriate for widerstreams that require longer mixingreaches, while lower volumes maywork for narrower streams. To avoidexcessive salt concentrations in thestream, one or more trial injectionsshould be conducted with lowvolumes, working up to largervolumes as required.

Figure 2 illustrates a salt wave forPlace Creek, where the author hasfound the salt waves to be highlyreproducible. Injecting 6.35 L of aroughly 17% solution into a flow of2.66 m³/s produced a peak EC about100% higher than background.

Recording Electrical Conductivity

Ideally, a data logger should be usedto record the passage of the salt wave.Some conductivity meters havebuilt-in data logging, while others canoutput a signal that can be recordedusing a separate data logger. If you donot have data logging capacity,record EC manually at 5-s intervals.Although this approach may not be asaccurate as using a data logger and a1-s recording interval, it can producesatisfactory results. In most cases, twopeople are required to conduct a saltdilution measurement with manualrecording, while the use of a datalogger allows a single person to makethe measurement.

The conductivity probe should beplaced within the main part of theflow, not in a backwater. Avoidlocations with substantial aeration, asair bubbles passing through the probecause spurious drops in conductivity.The probe should be firmly emplaced(e.g., by wedging it between carefully

placed cobbles) so that it will notmove during the measurement. Toposition the probe in a strong current,it may be useful to attach the probeto a rod weighted at the end that isplaced in the water.

In some cases, the background ECmay vary. One possible cause is anoverly sensitive conductivity meter.

Another cause of varying backgroundEC is incomplete mixing ofstreamwater and groundwater (whichtypically has higher EC than thestreamwater) within and immediatelydownstream of groundwaterdischarge zones. Similar problemswith incomplete mixing can occurdownstream of tributaries. In theselatter cases, find an observation pointwhere background EC is uniformacross the channel and constant intime.

Determining k by Calibration

To determine k, a known volume ofinjection solution (typically 5 or 10mL) is added to a known volume ofstreamwater (typically 1 L) to producea secondary solution. Knownincrements of this secondary solutionare then added to a second knownvolume of streamwater (typically 1 L),to generate a set of EC valuescorresponding to different values ofrelative concentration. The slope ofthe relation between relativeconcentration and EC provides therequired value for k. This two-stepprocedure dilutes the injectionsolution to the relative concentrationsobserved during the salt wave withoutusing large volumes of streamwater.See Moore (2004b) for a more

detailed description of the procedureand the calculation of k.

Although ideally the calibration isperformed in the field, particularly tomaintain water temperature as closeto stream temperature as possible, itcan also be conducted in thelaboratory. To perform the calibrationoff site, two 1-L samples ofstreamwater should be measuredaccurately into sample bottles using avolumetric flask. A sample of theinjection solution should also be takenin a small glass (not plastic) bottle toavoid potential problems with saltsorbing onto the walls of a plasticbottle. The calibration can then beconducted following the proceduredescribed by Moore (2004b).

Summary of FieldProceduresTable 2 lists the equipment required.Suggested steps for conducting fieldmeasurements are as follows:

1. Mix injection solution (either atoffice or on site).

2. Select measurement reach.

3. Use a pipette to extract a knownvolume of injection solution (e.g., 10mL) and add to the secondarysolution bottle. Cap the bottle andstore upright.

4. Record background EC and watertemperature at the downstream endof the measurement reach, andupstream of the injection point.

5. Set up the conductivity probe atthe downstream end of the mixingreach. Record the background EC andwater temperature. If you have a datalogger, start recording EC.

6. Inject a known volume of saltsolution at the upstream end of themixing reach.

7. Record the passage of the saltwave, continuing until EC returns tobackground. If EC does not return tobackground, measure EC upstream ofthe injection point again to determinewhether the background changed.



Table 1. Volumes/masses of injected salt used in different studies

Author Mass of saltinjected per m³/sstreamflow (kg)

Equivalent volume (L)of 20% salt solution

(1 kg salt in 5 L water)

Østrem (1964) 0.5 2.5

Church and Kellerhals (1970) 0.2 1

Day (1976) 0.3 1.5

Elder et al. (1990) 5 25

Hudson and Fraser (2002) 2 10

8. Measure a volume V0 (e.g., 1 L) ofstreamwater using the volumetric flaskand pour into the secondary solutionbottle, which already contains thesample of injection solution. Cap thebottle and shake vigorously to mix thestreamwater and injection solution.This mixture is the secondary solution.

9. Measure a volume Vc (e.g., 1 L) ofstreamwater using the volumetric flaskand pour into the calibration tank.Immerse the calibration tank in ashallow pool at the stream’s edge.Keep the temperature in the tank asclose to stream temperature aspossible (Moore 2004b). Tohelp hold the calibration tank inplace, position a “corral” ofcobbles around it.

10. Perform the calibration anddetermine k using theprocedure described by Moore(2004b), then compute thedischarge using Equation [6].

Errors and LimitationsUnder suitable conditions,streamflow measurementsmade by slug injection can be preciseto within about ±5% (Day 1976).Accurate measurements require that(1) the salt in the injection solution becompletely dissolved, and (2) theinjection solution be fully mixedacross the channel at the location

where the saltwave is recorded.In addition,discharge shouldnot changeappreciablyduring theinjection trial.

Errors may arisethroughinaccuracies inmeasuring thevolumes of

streamwater, injection solution, andsecondary solution. These errors canbe effectively minimized if avolumetric flask is used to measurestreamwater and glass pipettes usedto measure the injection andsecondary solutions. However, take

plasticware into the field as a backupin case of breakage.

If it is raining during themeasurement, ensure that thecalibration tank is sheltered.Otherwise, rain falling into the tankmay dilute the concentrations belowthe calculated values, producingbiased calibrations.

The slug injection method may not beappropriate when the channelcontains ice and (or) snow. In suchcases, low velocities may result inpoor lateral mixing and excessivelylong salt wave durations, particularly ifsalt solution flows into slush zoneswithin the measurement reach.

The method will be subject tosubstantial errors if the measurement


Worked ExampleFigure 2 shows a salt wave recorded during a slug-injection measurementat Place Creek, located about 30 km northeast of Pemberton, B.C. PlaceCreek, a steep, bouldery mountain stream, would be impossible to gaugeaccurately using a current meter (Figure 1). The volume of injectionsolution was 6.35 L. This volume resulted from mixing 1 kg of salt with 6 Lof water (to produce 6.36 L of solution), followed by extracting 10 mL(0.01 L) of injection solution for use in the calibration procedure. Thestream EC data were logged at 1-s intervals, and the calibration constantwas 2.99·10–6 cm/�S.

QV

k t EC t EC

m

cm S sbg

��

��

�

�

�� [ ( ) ]

.

( . / )(

635 10

2 99 10 1

3 3

6 � )( / ). /

7972 66

3

�S cmm s�

Table 2. Equipment list for field measurement of streamflow using slug injection of salt

Item Purpose

1-L volumetric flask Measuring streamwater

1-L plastic graduated cylinder Backup in case volumetric flask breaks

Plastic measuring cup with handle Pouring streamwater into volumetric flask

Squirt bottle Topping up streamwater in volumetric flask

5- and 10-mL pipettes1,2 Measuring injection solution to mix secondary solution

Pipette filler (rubber squeeze bulb) Drawing water into pipettes

1- or 2-L wide-mouth Nalgene water bottle Mixing the secondary solution

1- or 2-L Nalgene beaker or pail Calibration tank

2-, 5-, and 10-mL pipettes1,2 Measuring secondary solution

Plexiglas rod or tubing, 30 cm long Stir stick for calibration tank

Conductivity probe and meter Measuring EC during salt wave passage and for calibration

Data logger (desirable but optional) Recording EC during salt wave passage1Separate sets of pipettes need to be used for measuring the injection and secondary solutions.2Spare pipettes should be carried in case of breakage in the field. In addition, 10-mL plastic graduated cylinders or graduated pipettes could be carried as backups

Continued on page 6

Under suitableconditions,streamflowmeasurementsmade by sluginjection can beprecise to withinabout ±5%.

reach is not sufficiently long to ensurecomplete lateral mixing. Unlikeconstant-rate injection, where lateralmixing can be verified oncesteady-state conditions have beenachieved, assessing mixing is moredifficult with slug injection. If twoprobes are available, then the saltwave can be recorded at twodownstream distances or on eitherside of the stream. If mixing iscomplete, discharge calculated fromboth probes should be in reasonableagreement. If this is not the case, alonger mixing reach is required.Alternatively, if only one probe isavailable, successive measurementscan be made during periods of steadyflow using different distances.

Problems can occur if the conductivitydoes not return to background. If themeasurements taken upstream showthat the background has trulychanged, then an average of theoriginal and final background valuesmay be used in Equation [6]. It ismore problematic if EC has notreturned to background due to a slowrelease of stored salt solution withinthe mixing reach, as can occur inreaches with pools, particularly atlower flows. In such cases, onesolution would be to extend the tail ofthe salt wave by fitting an exponentialdecline to the values, although theactual form of the decline will still beuncertain (Elder et al. 1990). Ideally,one should find a reach with minimalstorage.

Injection of Salt in SolutionVersus Injection of Dry Salt:A ComparisonA number of authors have advocatedthe use of dry salt injection as analternative to injection of salt insolution (Hongve 1987; Elder et al.1990; Kite 1993; Hudson and Fraser2002). A future Streamline article willfocus on streamflow measurement bydry salt injection. The key advantageof the method is that, for gauging

higher flows (e.g., >5–10 m³/s), it iseasier to inject dry salt than to mixand inject adequate volumes ofsolution. However, a disadvantage ofthe dry salt method is that anaccurate scale to measure the mass ofsalt or an adequate supply ofpre-weighed salt in a range ofquantities is needed. Where anaccurate scale and pre-weighedquantities of salt are unavailable (e.g.,at a remote site over an extended fieldseason), the slug injection methodusing salt solution would still bepossible because the precise mass ofsalt in the injection solution does notneed to be known, just the volume ofthe solution (Equation [6]).

SummaryStreamflow measurement by sluginjection of salt solution has beensuccessfully applied in many locationsaround the world. It is particularlysuitable for steep, bouldery mountainstreams, which are unsuitable forgauging by conventional currentmetering techniques. This article hasdescribed procedures that the authorhas found useful at sites throughoutBritish Columbia. However, there isgreat scope to vary the details to suitindividual circumstances and users areencouraged to experiment with theoutlined procedure.

AcknowledgementsGeorge Richards helped meexperiment with variations on the sluginjection method while working atPlace Creek. Tim Giles, DaveHutchinson, Scott Babakaiff, and JohnHeinonen helped refine the methodsby asking valuable questions andbringing relevant articles to myattention. Comments on earlierversions of this article by JohnHeinonen, Russell White, MichaelChurch, Robin Pike, and fouranonymous reviewers helped improveits clarity. However, any errors remainmy sole responsibility.

For further information, contact:

Dan Moore, Ph.D., P.Geo.Associate ProfessorDepartments of Geography and ForestResources Management

1984 West MallUniversity of British ColumbiaVancouver, BC V6T 1Z2

Tel: (604) 822-3538E-mail: [email protected]

References

Church, M. and R. Kellerhals. 1970. Streamgauging techniques for remote areasusing portable equipment. Departmentof Energy, Mines and Resources InlandWaters Branch, Ottawa, Canada.Technical Bulletin No. 25.

Day, T.J. 1976. On the precision of saltdilution gauging. Journal of Hydrology31:293–306.

Elder, K., R. Kattelmann, and R. Ferguson.1990. Refinements in dilution gaugingfor mountain streams. In Hydrology inMountainous Regions. I - HydrologicalMeasurements; the Water Cycle.International Association forHydrological Science (Proceedings oftwo Lausanne symposia, August 1990).IAHS Publication No. 193, pp.247–254.

Hongve, D. 1987. A revised procedure fordischarge measurements by means ofthe salt dilution method. HydrologicalProcesses 1:267–270.

Hudson, R. and J. Fraser. 2002. Alternativemethods of flow rating in small coastalstreams. B.C. Ministry of Forests,Vancouver Forest Region. ExtensionNote EN-014 Hydrology. 11 p.

Kite, G. 1993. Computerized streamflowmeasurement using slug injection.Hydrological Processes 7:227–233.

Moore, R.D. 2004a. Introduction to saltdilution gauging for streamflowmeasurement: Part 1. StreamlineWatershed Management Bulletin7(4):20–23.

Moore, R.D. 2004b. Introduction to saltdilution gauging for streamflowmeasurement Part II: Constant-rateinjection. Streamline WatershedManagement Bulletin 8(1):11–15.

Østrem, G. 1964. A method of measuringwater discharge in turbulent streams.Geographical Bulletin 21:21–43.

Wood, P.J. and A.P. Dykes. 2002. The useof salt dilution gauging techniques:ecological considerations and insights.Water Research 36:3054–3062.



An Inexpensive, AutomaticGravity-fed Water Samplerfor Investigating WaterQuality in Small StreamsChad D. Luider, P. Jefferson Curtis, Rob A. Scherer,and David J. Arkinstall

Introduction

Water samples are commonlycollected, either manually

(grab samples) or with automatedsamplers, in many environmentalmonitoring and research programs(e.g., Toews and Gluns 2003; Winkleret al. 2004). Manual sampling inremote areas can be labour intensiveand time consuming, whereas theprice of automated samplers (around$4,000) may be prohibitive to manymonitoring programs.

This article describes a low-cost (<$600 per unit), gravity-fed,automated water sampler(auto-sampler) that can collect watersamples from small streams less than5 m in bankfull width. Theauto-sampler is best suited to collectsamples for analyses of water qualitymeasures in the dissolved phase, suchas pH, conductivity, carbon,phosphorus, and ammonia. Sedimentsamples have not yet been collectedwith the auto-sampler, and thereforesediment sampling is not consideredin this article.

Auto-Sampler Componentsand DesignA simple, lightweight gravity-fedauto-sampler can be constructed fromcommonly available irrigationsupplies. The sampler consists of awater intake system, a valve manifoldsystem, and a series of standardsample bottles (Figures 1 and 2). Thewater intake system (Figure 3) is

constructed from a PVC pipe with apiece of screen mesh secured on theintake end to minimize large debrisfrom clogging the lines and valves.The opposite end of the PVC intakepipe is connected to a 3–6 m lengthof polyethylene pipe that forms themain water supply line. The valvemanifold system distributesstreamwater from the intake pipethrough electronically controlledvalves into individual samplebottles (Figure 3). A plastic cargobox is used to contain the valves,the sample containers, and thesealed battery-operated controltimer, which is programmed tocontrol the electronic valves.Flexibility in sampling dependson the number of valves in themanifold and the featuresassociated with the control timer.Electronic control timers typicallycontrol 4–12 valves, and can beconnected in series to increasethe number of samples that theauto-sampler can collect. Forexample, two timers controlling12 valves each could beconnected in series andprogrammed to collect a total of24 samples from the samesample site.

Desirable features in a controltimer include the ability toindependently program each valve, a30-day programmable clock, and amaster valve option. Independentprogramming for each valve isessential because each valve

represents one sample. A 30-dayprogrammable clock allows forflexibility in sample scheduling (e.g.,daily, weekly, or monthly). The mastervalve option (a valve that openswhenever a sample valve is opened) isadaptable to control small pumpswhere a gravity-fed approach is notfeasible (e.g., lakes, ponds, largerivers). We are presently developingan automated pump sampler for thispurpose.

InstallationThe auto-sampler should be installedoutside of bankfull width to avoiddamage during high flows. The intakeshould be installed securely (e.g.,wedged between rocks, fastened torebar) within the streambed upslopefrom the auto-sampler, with enoughhydraulic head (e.g., 1–2 m of verticalrise) for the gravity-fed intake systemto function properly. For our

applications we have chosen ¼-inchvalves although ½- or ¾-inch valvescan be used. Less hydraulic head isrequired for smaller valves comparedwith the larger valves, which requirehigher minimum operating pressures,


Figure 1. Auto-sampler setup adjacent to creek.

Continued on page 8

Cha

dLu

dier

and therefore more hydraulic head.Except for the valves and the solenoidplungers, the entire system could bemade from stainless steel,polyethylene, or Teflon where thesematerials are recommended for use insampling different water qualityparameters.

Upon securing the intake within thestream, the following points should beconsidered to minimize the potentialfor air locks in the water supply line

and clogging of the intake screen. Airlocks tend to occur in high points ofthe main intake line, particularly inhighly aerated sections of stream.Intake lines should therefore beinstalled with a constant slope to thevalve manifold, thus avoiding loops inthe line that trap air. In addition, theintake screen should be submerged innon-turbulent, uniformly flowingwater to minimize air bubblesentering the system. A ball valve atthe lower end of the valve manifoldshould also be installed so that theintake lines can be manually flushedafter installation and when changingthe sample bottles (Figure 3). Thisflush allows water to flow through theintake lines to remove air and rinse

the internal plumbing of the systemwith sample water. As the manualflush valve is larger than the ¼-inchsample valves, the flow rate is higherand thus the intake system flushesthoroughly.

Due to the design of the sampler,clogged intake screens cansignificantly affect the operability ofthe unit. Clogged intake screens canbe minimized by positioning theintake screen perpendicular to

streamflow and by designating onevalve to flush the system immediatelybefore activating the sample valves.Flushing the intake systemimmediately before sample collectiongreatly reduces the risk of the intakebecoming plugged or blocked withdebris by comparison to a continuousflow setup. The system flush isprogrammed to be completed withinless than 1 minute before activatingthe valve for sample collection.

After the unit has been installed andthe control timer set, sample bottlesare connected to the spouts of thevalves designated for samplecollection with a piece of tubing and atwo-holed stopper. The time requiredon the control timer to fill the sample

bottles will depend on the flow ratethrough the intake system. Similarly,the automated flush valve should beset to rinse the entire volume of thesystem at least 5 times immediatelybefore sample collection. The amountof time required to flush the intakesystem can be calculated empiricallyby measuring flow rate through thesystem and the length of the mainwater supply line that is required forinstallation. For example, it wouldtake 5 minutes to completely flush thevolume of the intake line once given aflow rate of 1 L/min through 10 m of½-inch intake line (volume of intakeline is about 5 L; i.e., flushing time =volume of pipe/flow rate). The timerequired for sample collection can becalculated using the same approach(i.e., sample collection time = bottlevolume/flow rate), but it is best to setthe clock for more time than isrequired to account for decreases inflow rates. This approach ensures thatsample bottles are completely filledand are flushed with sample water.Any excess water spilling from thesample bottles drains via holes in thebottom of the plastic cargo box.

Sample Collection ProtocolWe designed and deployed theauto-sampler to collect specific waterquality parameters. Therefore, thisarticle will not detail samplingprotocol, which varies with waterquality parameter of interest. Forfurther information regarding thedesign of reliable monitoringprograms using automated samplers,refer to the Automated Water QualityMonitoring Field Manual (ResourceInventory Committee 1999).

Unit PerformanceWe used two quality control (QC)measures to evaluate the precisionand performance of the auto-sampler.The first QC measure included thedirect collection of water samples (i.e.,grab samples) at the auto-samplerintake in conjunction with samplesbeing collected with the auto-sampler.The second QC measure was used to


Figure 2. Auto-sampler components and parts.


Dav

idAr

kins

tall

evaluate whether leaching from andadsorption to the water intake systemwas contaminating water qualitysamples. We checked this by flushingdeionized water through theauto-sampler at the end of the sampleseason. We then compared thesesamples with control samples of

deionized water. Analyses of pH,conductivity, dissolved organiccarbon, and dissolved nutrients (PO4

and NH4) indicate no significantdifference between the grab samples,the deionized water, and samplesfrom the auto-sampler (p < 0.05).

Applications andConstraints of theAuto-SamplerIn our study, we deployedand tested threeauto-samplers for sevenmonths (April to October2004) in the SouthernInterior of BritishColumbia. Units wereinstalled in boulder–cobblestreams with bankfullwidths ranging between 1and 4 m and gradientsbetween 5 and 15%.Samples were collectedfrom the units at a rate of2–3 samples per week andduring our field trials wefound that requiredmaintenance to theauto-samplers wasminimal. Only one repairwas required to a brokenfitting, which caused theloss of one sample. Onaverage, the two 9Vbatteries in each controltimer were depleted byonly 20% throughout theentire operation. All threeauto-samplers wereremoved in late Octoberdue to freezing of thevalves and intake lines.

The auto-sampler unit isaptly suited for ourmonitoring purposes (i.e.,chemical water qualityparameters, low summerflows, ice free conditions).However, theauto-sampler has not beentested or used under thefollowing conditions thatwould warrant further

investigation and possibly designmodifications: (1) high flows andfreshet, (2) sediment sampling, (3)freezing conditions, and (4) streamsgreater than 5 m bankfull width.

In summary, the auto-sampler hasallowed us to collect water samples ata higher frequency relative to manualsampling and at a reduced costcompared with commercially availableauto-sampler units. Our auto-samplerperformed very well during lowsummer flows and allowed us tocollect and analyze water samples forseveral dissolved water qualityparameters. The auto-sampler wasreliable, cost effective, and easy tomaintain. Future testing andimprovements to the design of theauto-sampler will likely increase itssuitability to more diverse samplingenvironments and a greater variety ofwater quality parameters.


P. Jeff CurtisDepartment of Earth andEnvironmental SciencesOkanagan University College

3333 University WayKelowna, BCV1V 1V7Tel: (250) 762-5445, Ext 7521E-mail: [email protected]

References

Resource Inventory Committee. 1999.Automated water quality monitoring,field manual. Unpublished reportprepared for B.C. Ministry ofEnvironment, Lands and Parks, WaterManagement Branch for the AquaticInventory Task Force. 61 p. Availablefrom: http://srmwww.gov.bc.ca/risc/pubs/aquatic/index.htm

Toews, D. and D. Gluns. 2003. Waterquality sampling: an effective way tomonitor watershed condition?Streamline Watershed ManagementBulletin 7(2):15–21.

Winkler, R., D. Spittlehouse, T. Giles, B.Heise, G. Hope, and M. Schnorbus.2004. Upper Penticton Creek: howforest harvesting affects water quantityand quality. Streamline WatershedManagement Bulletin 8(1):18–20.


Table 1.Equipment list for a five-bottle auto-sampler

Water intake system

Item Quantity

A Screen mesh (10 x 10 cm) 1

B 2” gear clamp 1

C 2” PVC pipe ~30 cm

D 2” x 1/2” reducer bushing 1

E 1/2” threaded x 1.5 cm (1/2”)barbed coupler

1

F 1/2” polyethylene pipe 3–6 m

Valve manifold system

Item Quantity

G Plastic cargo container with lid 1

H 1/2” PVC threaded tee 6

I 1/2” x 2” threaded coupler 6

J 1/2” x 1/8” reducer bushing 6

K 1/8” x 2” brass nipple 12

L 1/4” valve 6

M Solenoid 6

N 1/4” vinyl tubing 100–150 cm

O 1/4” thick-walled heat shrink 25–30 cm

P Battery-operated control timer 1

Q 1/2” threaded PVC manualball valve

1

R 500-mL polyethylenesample bottle

1–6

Figure 3. Schematic view of the auto-sampler.The labelled parts are listed in Table 1.

Live Gravel Bar StakingChannel Stabilization inthe Lower Elk River

Iain D. Cuthbert and Ian D. Redden

During the past 70 years, the ElkRiver on northern Vancouver

Island has evolved from a narrow,single-thread, stable channel to awide multi-thread, laterally unstable,aggraded channel. This change wasin response to several factorsincluding: valley-bottom logging;channel relocation due to roadconstruction; a large landslide in theriver’s headwaters; and increasedflows resulting from the diversion ofwater into the Elk River from theadjacent Heber River watershed. Thenet result: a 4–7 times increase in theunvegetated channel width in thelower 13 km of river and degradedfish habitat because pools infilled,banks eroded, and cover was lost.

Previous channel morphology studies(e.g., M. Miles and Associates 1999)demonstrated the need to restorechannel processes in the lower ElkRiver to expedite the re-formation of astable, single channel. This projectaddresses this recommendation anddoes not incorporate any uplandrestoration activities that likely will bepart of future restoration plans. Basedon successful treatments of rivers withsimilar conditions, such as the SanJuan (Switzer 1999), we chose the soilbioengineering technique of livegravel bar staking as the preferredrestoration method to achieve ourobjective. This article describes theapplication of and lessons learnedfrom live gravel bar staking in thelower Elk River.

Live Gravel Bar Staking:BackgroundLive staking of gravel bars usingwillow (Salix spp.) and other plantspecies such as red-osier dogwood(Cornus stolonifera) and blackcottonwood (Populus trichocarpa) canbe used to treat river channels thathave become aggraded and braided.In live staking, cuttings (stakes) fromthe selected pioneering species areplanted at high density into the gravelbars.

During high flows, the treated areasare inundated; the friction caused bythe protruding stakes traps very smallwoody debris and leads to localdeposition of sediment. Each winter,once enough sediment is deposited tocover the protruding stakes,streamflow will top the bars withoutresistance. In the next growingseason, the cuttings will grow andprotrude above the gravel bar. Thisseasonal process of growth followedby sediment and debris accumulationcauses the gravel bars to progressivelystabilize and elevate (Figure 1). At thesame time, the accumulation of finesand organics, such as small woodydebris, promotes the establishment ofadditional riparian vegetation, furtherstabilizing the bars. Over time thegravel bars elevate, and becomeinundated less frequently. Thestreamflow becomes increasinglyconfined to the main channel,redirecting the river’s energy toscouring a narrower and deeper

mainstem channel. Polster (1999)discusses live gravel bar staking andother soil bioengineering techniquesin detail.

Site SelectionThe Elk River, a tributary to UpperCampbell Lake, is located on northernVancouver Island near the town ofGold River, B.C. The potentialtreatment sites were first selected byanalyzing historical air photos from1931 to 1995. The main site selectioncriteria were gravel bars (1) with easyequipment access, (2) in incipientlystable depositional areas, and (3)outside of the most active channelsections that convey high flows.Criteria 2 and 3 were extremelyimportant as live gravel bar staking ofthe more active channel sectionscould reduce flood conveyancecapacity and possibly accelerate bankerosion or channel shifting (M. Milesand Associates 2004).

A June reconnaissance trip finalizedrestoration site selection, determinedsite access, and located suitable stockdonor and soaking sites. During thistrip we discovered that the naturalrecovery of many of the potential sitesidentified on the air photos wassignificant, and included deciduoustrees older than 5 years. We theorizedthat this recovery, the greatest in theprevious 45 years, was due to severalyears with unusually wet summersand smaller than average flood flows.The natural recovery observed wasvigorous enough to eliminate severalof the potential treatment sites. Thus,three additional sites not identified in


Figure 1. Function of live gravel bar staking.

Iain

Cut

hber

t

the office review were investigated inthe field.

While many areas would havebenefitted from live gravel barstaking, site access became the largestlimiting factor. Although Highway 28parallels much of the river, steepbanks from the highway preventedequipment from accessing the river.The only other road in the area thatwould have provided access to theriver had been deactivated for muchof its length. As most of the lowerriver lies within Strathcona ProvincialPark, excavator access trail buildingneeded to be minimized to preserveecological values. In total, three siteswere selected for treatment.

Collecting and Preparingthe StakesThe project began in September 2004with the collection of donor stockfrom areas close to the restorationsites in Strathcona Park. Stock wascollected by cutting down smalldeciduous trees close to the groundwith chainsaws. The donors wouldcoppice and regenerate in thefollowing year. The stakes collectedwere comprised of 85% willow(Scouler’s and Sitka), 6% blackcottonwood, and 9% red-osierdogwood. Crew members then

collected, topped, and limbed the cuttrees. Using high quality, relativelyexpensive pruning and lopping shearswas invaluable, as smaller shearstended to break, disruptingproduction. The topped and limbed“poles,” which ranged from 2 to 4 min length, were then placed onsawhorses and tied withbiodegradable sisal baling twine intobundles of 7–10 stems (Figure 2).Flagging tape was attached to eachbundle, with a different colour usedfor each day. When the weatherconditions were cool and wet,bundles were loaded into trucks andtaken to the soaking site at the end ofeach day. During warmer, sunnyweather, bundles were taken to thesoaking sites throughout the day toprevent desiccation (wilting) anddeath. The use of a Silva cool-tarp tocover the bundles during collectionwould have been beneficial duringhot, dry weather. Productionaveraged 2840 stakes per day for aseven-person crew.

The target size for stake collection was2 cm in diameter or larger. This size isoften referred to as the “rule ofthumb” as typically anything greaterin diameter than your thumb is thedesired size. After several days ofharvesting it became apparent that

the main cutting site would notprovide enough stock to plant thetreatment areas, and two additionaldonor sites were located. Also, alimited quantity of stock at the maindonor site met the diameter criteria.Due to the shortage of large stock,cuttings that were slightly smallerthan 2 cm in diameter were alsocollected and were referred to as“undersize stock.”

The cuttings need to be soaked infresh water for 7–10 days to removerooting inhibitors before planting (D.Polster, pers. comm., 2004). Onechallenge of this project was findingadequate soaking sites, as nearbyponds were shallow and water levelsdropped during the soaking perioddue to warm, dry weather. As a result,bundles had to be repositionedseveral times to avoid drying out.Beavers added another challenge:they raided the soaking area; removedsome of the largest cuttings, strippingthe bark and cambium layers fromothers; and sometimes took entirebundles.

Once most of the donor stock hadbeen collected, the crew split up: onecrew continued cutting, while thesecond crew began planting stakes.The planting crew collected thebundles from the soaking sites, takingthe earliest cuttings first; and thentransported them to the treatmentsites where they were cut withlopping shears into 1 m length stakesin preparation for planting.

Stake PlantingDue to the inherent difficulty ofplanting in gravel, an excavator with adigging bucket was used to install thecuttings in the coarse gravel bars. Theuse of the excavator minimizeddamage to the stock during planting,and ensured that the cuttings wereplanted deep enough to survive thedry summer. The excavator did notdig holes, but rather inserted thebucket into the gravel and pulled backthe material, creating a 1 m wide gapinto which the stakes were placed by


Figure 2. Assembling limbed poles into bundles.

Continued on page 12

Ian

Redd

en

hand (Figure 3). The excavator thenwithdrew its bucket allowing thegravel to settle back in place. Thestakes were planted with aboutthree-quarters of their length in thegravel at a 45º or greater downstreamangle (Figure 4). Four large stakes andthree to five undersized stakes, ifavailable, were placed into eachopening, taking about one minute foreach opening. While the undersizedstakes may not flourish as well as thelarge stakes, they significantlyincreased the overall number of stakesplanted, which should improve thechances of the project in overcomingmortality due to elk browse.

The excavator worked by backingupstream while planting in successiverows spaced 1.5–2 m apart andstaggered to prevent large openpatches within the planted areas. Thefirst pass of planting occurred nearestthe river channel with the excavatorpositioned at the edge of the zone tobe planted. This ensured that theedge of the row nearest the river wasplanted parallel to the flow. Theexcavator then reached as far aspossible upland from the river. To

maximize the area covered with theavailable stock, the stakes wereplanted with tighter spacing and athigher densities on the first passnearest the mainstem channel wherethey would likely receive the greatestflows.

Live staking planting began onSeptember 29, 2004, and wascompleted on October 12. In total,1.86 ha was planted at three sites at

an average density of 17 200 stemsper hectare. Planting took an averageof 4.5 days per hectare, with a crew offour people working with theexcavator.

With the live staking of gravel barscompleted, the success of the projectwill depend on a number of factors,including the growth and survival ofthe stakes, mortality or stunting dueto elk browse, and the response of thetreated areas to peak flows.

MonitoringLong-term monitoring will allow us toassess the success of the live gravelbar staking in achieving the projectgoals. This information can also beused to help direct future restorationactivities. The following measures

were taken to assist in gauging projectsuccess.

The perimeters of the treated areas aswell as longitudinal andcross-sectional profiles were surveyedat each site using a total stationsurvey instrument. Benchmarks wereestablished at each site for futurereference during surveys andmonitoring. Fifty-one monitoringplots, including seven control plots,


Figure 4. Planted stakes.

Ian

Redd

en

Ian

Redd

en

Figure 3. Crew member instructing student on planting stakes.


were established at various locationswithin the project area. Each researchplot had a 3-m radius (28.3 m²). Asubset of 16 research plots wasestablished to monitor changes insubstrate composition. Within each ofthese 16 plots the substrate in three0.5-m² squares was photographedand documented. Vegetation surveysat each of the 51 monitoring plotsincluded recording the number,species, and size of stakes planted ineach plot. Finally, three permanentphoto points were established at eachtreatment site for future monitoring.In 2005, the areas treated in 2004 willbe monitored and additional livestaking of gravel bars in the Elk Riverwill take place.

Summary of LessonsLearned

Often spring planting offers severalbenefits in soil bioengineering. Dueto funding timelines, we planted inthe fall.

Use the most recent air photosavailable for preliminary siteselection and reconnaissance.

Have several stock donor sitesselected before beginning cutting.A local Ministry of Forests office orforest company may offer someadvice on donor sites.

Use high-quality lopping shearsand hand pruners. The loss inproductivity due to the use of poorequipment will cost more than theinitial expense of purchasingquality equipment.

Use lopping shears rather than achainsaw to cut stakes to length.The chainsaw tended to makerougher cuts and “shred” the barknear the cut end.

Avoid soaking sites near knownbeaver populations. The loss ofbundled donor stock due tobeavers was much higher than wehad expected. Using beaverprotection such as a rodent fencearound the soaking bundles wouldhave prevented some loss.

Ensure soaking sites have stablewater levels. The sudden changefrom wet weather to several daysof dry weather caused one soakingsite to dry up, and the bundlesrequired repositioning severaltimes during the soaking period.Covering soaking bundles withSilva cool-tarps may be beneficialin hot, sunny weather.

Flag bundles collected each daywith a different colour flaggingtape. This system allows for quickand easy identification of thebundles when collecting themfrom the soaking site. Use bundlesin the order that they were cut.

While securing funding to studythe success of restorationtreatments is difficult, monitoringof the live gravel bar stakingproject is needed to continuallyimprove the selection andsuccessful application of restorationtechniques in British Columbia.

Acknowledgements

Mike Miles, a fluvialgeomorphologist, assisted the pro-ject team by selecting suitable sitesfor live staking. David Polster, wholargely pioneered this soil bioengi-neering technique in British Colum-bia, trained the crew to cut stakes tolength. A core team ofMowachaht–Muchalaht First Nationforestry workers, supported by volun-teers from the Gold RiverStreamkeepers and Gold River Sec-ondary School, completed this work.The BC Hydro Bridge Coastal Fishand Wildlife Restoration Programfunded this project.


Iain Cuthbert, M.Sc., R.P.Bio.Streamline EnvironmentalConsulting Ltd.

786 Quilchena CrescentNanaimo, BC V9T 1P6

Tel: (250) 758-7980Fax: (250) 758-8505E-mail: [email protected]

References

M. Miles and Associates. 1999. Preliminaryassessment of the effects of the CrestCreek and Heber River diversions onchannel morphology. Consultant’sreport prepared for BC Hydro andPower Authority, Burnaby, B.C.

M. Miles and Associates. 2004. Selection ofsites for live gravel bar staking: LowerElk River in Strathcona Park.Consultant’s report prepared for BCHydro, Bridge Coastal Fish and WildlifeRestoration Program, Burnaby, B.C.

Polster, D. 1999. Soil bioengineering forsteep/unstable slopes and riparianrestoration. Streamline WatershedRestoration Technical Bulletin Volume4, Number 4, Winter 1999. Availablefrom:http://www.forrex.org/streamline/issue.asp?issue=14

Switzer, G. 1999. Observations of thesuccess of the willow planting on thesand, gravel and cobble bars of thelower San Juan River. StreamlineWatershed Restoration TechnicalBulletin Volume 4, Number 4, Winter1999. Available from:http://www.forrex.org/streamline/issue.asp?issue=14


While securingfunding to study thesuccess of restorationtreatments is difficult,monitoring of the livegravel bar stakingproject is needed tocontinually improvethe selection andsuccessful applicationof restorationtechniques in BritishColumbia.

A Qualitative Hydro-Geomorphic Risk Analysisfor British Columbia’sInterior Watersheds:A Discussion PaperKim Green, M.Sc., P.Geo.

Editor’s Notes:

A preliminary version of this article waspublished in ASPECT, May 2004. Sincethen, the article has been revised, based onnumerous technical reviews. This article isintended to stimulate discussion amongforest hydrologists about the developmentof a qualitative hydro-geomorphic riskanalysis for B.C. watersheds. As a discussionpaper, the author acknowledges somelimitations in the material presented belowin attempting to develop the framework.

Introduction

Under British Columbia’s newForest and Range Practices Act,

forest management is movingtowards risk management andprofessional reliance. In this newregime, forest managers mustunderstand potential risks to aquaticvalues associated with existing orproposed development in awatershed.

To date, industry and governmentprofessionals have had minimaldiscussions about a standardapproach to hydrological risk analysis.Inconsistencies in methods,terminology, and elements beingconsidered in hydrological riskanalyses are causing significantdifferences in the way professionalsestimate risk (e.g., Carver 2001; B.C.Ministry of Forests 2001; Uunila2004).

Recently, the Association ofProfessional Engineers andGeoscientists of B.C. (2003) and the

B.C. Ministry of Forests (Wise et al.[editors] 2004) have developedprovincial standards for landslide riskanalysis. These efforts have givenprofessionals an understanding of theterms and methods of risk analysisneeded for detailed terrain stabilitymapping. Similar efforts tostandardize methodsand terminology areneeded if hydrologicrisk analyses are tobecome a widelyaccepted and valuedcomponent of forestmanagement inBritish Columbia.

The authordeveloped thishydro-geomorphicrisk analysis for use inSouthern InteriorB.C. watersheds andpresents it to opendiscussion regardinga consistent methodology for riskanalysis. Terminology used in this riskanalysis is generally consistent withdefinitions in Risk Management:Guideline for Decision-Makers (CSA1997) and Landslide Risk Case Studiesin Forest Development Planning andOperations (Wise et al. [editors] 2004).

Natural Variability andChannel ResponseLow-order watersheds (<100 km²) inBritish Columbia’s Southern Interiorhave been shaped by their geology,

glacial history, and climate over thepast 10 000 years (Clague [compiler]1989). The physical, chemical, andbiological characteristics of low-orderstreams are closely linked to hillslope(hydro-geomorphic) processes andriparian function (Montgomery andBuffington 1998; Gomi et al. 2002).

Natural disturbance events such aswildfire, pest epidemics, and floodsroutinely affect watersheds andconstitute an intricate part of thedynamic and evolving landscape ofthe Southern Interior (Bragg 2000;Benda et al. 2003; Gayton 2003).Changes to streamflow, sedimentdelivery rates, and riparian function(collectively referred to as watershedprocesses) following naturaldisturbance define the naturalvariability of a watershed over time(Gomi et al. 2002; Miller et al. 2003).While natural disturbance eventstypically have substantial, immediate

impacts on channelstructure and aquaticvalues such as waterquality and aquatichabitat (Rinne 1996),the influx of nutrients,sediment, and woodydebris in the decadesfollowing the event canplay a vital role inmaintaining the aquaticecosystem of awatershed (Benda et al.2003; Figure 1).

A channel’s response todisturbance events (i.e.,

the variability of channel morphologyin time and space) depends on thedisturbance regime of a watershed,which is a function of its geographiclocation (i.e., within British Columbia’shydro-climatic and physiographicregions) and physical attributesincluding bedrock geology andglacial/paraglacial history (B.C.Ministry of Transportation andHighways 1996; Montgomery andBuffington 1998; Hallett and Walker2000; Obedkoff 2002; Miller et al.2003).


Under BritishColumbia’s newForest and RangePractices Act, forestmanagers mustunderstand potentialrisks to aquatic valuesassociated withexisting or proposeddevelopment in awatershed.

Streams draining steep mountainslopes in the interior wet belt of BritishColumbia have larger peak dischargesper unit area and experience a higherfrequency of channel-forming events(e.g., debris flows, snow avalanches)than watersheds in arid, lowlandregions (Jakob and Jordan 2001;Obedkoff 2002). As a result, themorphology of channels in theinterior wet belt typically have greaternatural spatial and temporal variabilitythan channels in arid and semi-aridregions where less frequent eventssuch as wildfire and floods define thedisturbance regime.

Forest development in a watershedcan cause changes to watershedprocesses including increased hillsloperunoff and stream discharge (Troendleet al. 2001; Wemple and Jones 2003;Schnorbus and Alila 2004); increasedrate of sediment delivery to streams(Roberts and Church 1986; Gomi andSidle 2003); and reduced riparianfunction through removal ofstreamside vegetation and directimpacts to channel bed and banks(Bragg 2000; Faustini and Jones2003). The potential for significant(observable, long-term) change toaquatic values in a watershed due tochanges in watershed processesassociated with forest developmentwill be greater in channels that have

less natural variability in channelmorphology.1

Maintaining or improving aquaticvalues of watersheds whilemaximizing harvesting opportunitiesis a primary management objective offorest development in BritishColumbia. Understanding watershedprocesses and the natural variability inchannel condition and aquatic valuesallows forest managers to applymanagement practices to reduce therisk of direct negative impacts tolow-order streams. In turn, thisreduces the risk for cumulativeimpacts in higher-order streams.

Hydro-geomorphic RiskAnalysisA qualitative risk analysis offers (1) aframework for forest hydrologists andgeomorphologists to documentcritical watershed processes (i.e.,stream discharge, rate of sedimentdelivery, and riparian function) thatare linked to aquatic values; and (2)recommendations for sustaining orimproving aquatic values within awatershed. This reconnaissance-levelanalysis is intended to help forestmanagers identify areas where a moredetailed level of assessment isrequired.

Simply stated, estimation of risk toaquatic values from forest

development considers twoindependent factors: the potentialresponse of the channel to changes inwatershed processes and the potentialimpact of forest development onwatershed processes. It is expressed asthe product of two components:channel sensitivity (C) and hydrologichazard (H).

Risk = C × H

Hydro-geomorphic risk is determinedfor the main stem and significanttributary channels upstream of a pointof interest (POI), such as a waterintake structure or a specific fishhabitat (elements at risk) for eachwatershed process or, whereappropriate, for each identifiedaquatic value at the point of interest(e.g., water quality at the intake,channel stability on the fan). A simplematrix such as shown in Table 1 canbe used to determine risk in thisqualitative analysis.

Channel sensitivity, a measure of thevulnerability (robustness or fragility) ofthe channel given changes towatershed processes, depends on thephysical attributes of the channel.Channel sensitivity is equivalent toconsequence in the conventionalequation of Risk = Hazard ×Consequence.2 The ratings of “low”,“moderate” and “high” sensitivityexpress the potential size of change to


Figure 1. Woody debris recruitment and sediment influx following fire are key factors in maintaining aquatic ecosystems in many SouthernInterior watersheds. Photos are from similar low-order streams in Caven Creek, southeastern British Columbia. Stream (A) burned during the2003 Plumbob fire. Stream (B) experienced a similar fire about 70 years ago.

A B

1In general, channel types associated with transport-limited, alluvial valley segments (e.g., step–pool or riffle–pool channels) or those in supply-limited colluvialvalleys dominated by forced alluvial reaches (LWD step–pool or step–bed channels) will have a greater potential for change than channel types in colluvial orbedrock valley segments (Montgomery and Buffington 1998).

Kim

Gre

en


the channel structure and associatedaquatic values (collectively referred toas channel condition) and are assessedfor each of the watershed processesseparately. What each sensitivity ratingimplies in terms of probable level ofimpact to the channel/aquatic valuesis specific to a watershed and shouldbe defined in the report. Exampledefinitions are in the footnotes toTable 2.

Channel sensitivity to increases inpeak discharge considers the potentialfor increased bedload transport,which is estimated by consideringmean grain size, grain sizedistribution, channel gradient, andhydraulic roughness (O’Connor andHarr 1994; Buffington andMontgomery 1999; Church 2002).For example, a stream that has acobble–boulder cascade morphologywill have a smaller change to channelcondition due to a given increase inpeak discharge than a low-gradient,gravel, riffle–pool channel.

Channel sensitivity to increases insediment delivery considers thecapacity of the channel to transportsediment as determined by channelgradient and sediment storageopportunities. Due to increases insediment delivery in the headwaterreaches, a low-gradient (<5%)meandering channel with interveningwetland segments will have a smallerchange to channel condition over thelength of the channel network than amoderate gradient (5–15%) channelwith limited sediment storageopportunities (Lisle 2000).

Channel sensitivity to disturbances ofriparian function considers the past,

present, and future dependence ofchannel condition on riparianvegetation (Montgomery 2003). Achannel with deciduous riparianspecies such as alder and willow,which are indicative of frequentflooding and snow avalanches, will beless sensitive to disturbance of riparianfunction than a channel with matureconiferous riparian species supplyinglarge woody debris that contributes tochannel bed and bank stability. Wherestream temperature is a concern, thedependence of a channel on riparianfunction considers channelorientation, hillslope gradient, andriparian species (Brown 1980).

Channel sensitivity is estimated for themain stem channel and largertributary channels through acombination of field assessment,interpretation of current and historicalair photos, and analysis of regionalhydrometric and climate information.Montgomery and MacDonald (2002)describe in detail a similar approach toassessment of channel condition andsensitivity.

Key channel attributes that contributeto the estimation of channel sensitivityfor the three watershed processes aresummarized in Table 2.

A hydrologic hazard is a harmfulsustained change to a watershedprocess. The hydrologic hazardsconsidered in this analysis areincreased peak discharge (Hp),increased rate of sediment delivery(Hs), and decreased riparian function(Hf) associated with proposed andexisting development. The variabilityof watershed processes resulting from

past natural disturbance in awatershed forms a baseline for theassessment of hydrologic hazard.3 Inthis assessment the likelihood of ahydrological hazard is expressedqualitatively as “low,” “moderate,”and “high.” These ratings indicatethat the likelihood of a harmful orpotentially harmful change to awatershed process occurring withinthe time span of the development is“negligible,” “not likely but possible,”and “probable,” respectively.

When detailed information such asflood frequency, annual sedimentbudgets, and the frequency ofdisturbance to riparian function isavailable, the risk analysis can beadapted to be more quantitative. Thisis done by expressing and contrastingthe likelihood of a hydrologic hazardin the undeveloped (baseline)condition and developed (disturbed)condition as the annual probability(Pa) and the long-term probability (Px)for the lifespan of the proposeddevelopment.

For example, a stream thatexperiences a major channel-formingflood event once every 50 years (1:50)has an annual probability of 0.02(2%). If the development in questionhas a lifespan of 20 years thelong-term probability (P20) of achannel-forming flood event is

Px = 1 – (1–(Pa))x

soP20 = 1 – (1–(1/50))

20= 0.33 (33%).

4

If development is estimated aspotentially increasing the annualprobability of a majorchannel-forming flood from a 1:50 to1:20 return period (e.g., Schnorbusand Alila 2004, scenario 2/3U, Table3) the long-term probability (P20) of amajor channel-forming flood isincreased to 0.64 (64%). In this casethe proposed development increasesthe probability of a channel-formingflood event from 33 to 64% (�31percentage points). Professionals


2In this case, channel sensitivity (consequence) equals vulnerability because the spatial and temporal probabilities of the elements being considered (channelstructure and aquatic values) are both equal to 1 (e.g., Wise et al. [editors] 2004, p. 16).3For example, the frequency of channel-forming floods, return period of fire or forest health epidemics, distribution and frequency of occurrence of mass wastingor erosion events.4See Wise et al. (editors, 2004), pp. 13–14, and Table, A4.2.


Table 1. Hydro-geomorphic risk matrix (B.C. Ministry of Forests 2002)

Hydro-geomorphic risk a Likelihood of a hydrologic hazard

Low Moderate High

Channelsensitivity

Low Very low Low Moderate

Moderate Low Moderate High

High Moderate High Very highaA rating of “negligible” can also be added to the matrix if channel condition is independent of a watershedprocess, or forest development does not affect watershed process.


Table 2. Channel sensitivityWatershed

processChannel

sensitivitya Typical channel attributes that contribute to channel sensitivityb

Increasedpeak

discharge

Low

• Experiences frequent large, rapid peak flows: banks and floodplain vegetated with alder and willow,bright, scoured channel bed and banks, historically active fan; typical of channels draining watershedswith steep alpine headwaters

• Coarse-textured bedload, not the result of a single anomalous flood event or an anthropogenicdisturbance; numerous boulder cascade or bedrock reaches

• Well-vegetated, overhanging banks (e.g., mature coniferous species with well-developed root system) andabundant functioning large woody debris (LWD) and debris jams that provide channel and bank stability

• Often includes channels in supply-limited, colluvial, or bedrock valley segments

Moderate

• Experienced larger flood events in the past, indicated by numerous, multi-aged vegetated bank sloughs,levees, or old woody debris jams at obstructions with minimal long-term changes to channel stability

• Some inherent capacity to withstand higher flows, such as overflow channels or an entrenched channelwith resilient banks or non-alluvial segments

• Banks and riparian area vegetated with species that have well-developed root systems that protect thebanks and forest floor from erosion

• Often includes forced alluvial channels in colluvial or bedrock valley segments or transitionalmorphologies in alluvial valley segments

High

• Does not experience frequent flood events; bed is dark and mossy, banks are overhanging, vegetated tobankfull, and show no or little evidence of old scour or overbank deposits

• Contains fine-textured bedload that is susceptible to erosion• Partially or entirely confined and lacks structures, such as overflow channels, low gradient marshy reaches,

and abundant functioning LWD that help reduce flow velocity• Generally includes fine-textured, transport-limited plane–bed to riffle–pool channels or forced alluvial

channels

Increasedsedimentc

delivery

Low

• Abundant locations for sediment storage, such as frequent functioning LWD jams or frequent lowgradient unconfined sections (e.g., alluvial valley segments with riffle–pool channels)

• Contains slow-flowing, meandering stream (e.g., flows through marsh or wetland segments) and lacksthe power to transport bedload (i.e., decoupled systems where source areas are isolated fromdownstream channels)

• Headwaters are steep snow avalanche and (or) debris flow gullies that deliver large volumes of sedimentannually

Moderate• Colluvial valley segments with some storage capacity, such as some long (>100–200 m), low gradient

sections (<15%) that allow bedload sediment to settle out• Bordered by currently inactive, but relatively numerous natural landslide scars or debris flow gullies

High

• Laterally confined, forced alluvial and riffle–pool to cascade–pool systems that will become aggraded• Channel has little or no storage capacity so that increases in sediment delivery are likely to cause lateral

avulsion or channel aggradation• Additional sediment input will be rapidly transported through system to P.O.I. due to steep headwater

tributaries and ephemeral channels (>10%) with minimal opportunity for storage of sediment

Decreasedriparianfunction

Low

• Not dependent on LWD to control rate of sediment transport, such as a steep colluvial or bedrockchannels or snow avalanche chutes

• Low gradient, braided, or anastomosing channels, situated on a wide valley bottom vegetated withshrubs

Moderate

• Requires some LWD in a number of reaches to offer long-term storage, moderate bedload transport rate,or shade and cover for aquatic habitat (e.g., forced alluvial, LWD step–pool, or step–bed channels incolluvial valley segments)

• Has tendency to migrate laterally across valley bottom and is unentrenched so that migration could beaccelerated if valley bottom is disturbed and banks destabilized (e.g., meandering step–pool to riffle–poolchannel in alluvial valley segment)

• Some reaches are oriented such that the riparian canopy produces shade and moderates watertemperatures

High

• Entirely dependent on LWD to control bedload transport rates and maintain bank integrity• Appears to migrate over floodplain/valley bottom frequently and requires a wide effective riparian area for

long-term stability (typically LWD forced alluvial step–pool to cascade pool channels in colluvial valleysegments)

• Dependent on riparian canopy to maintain water temperature and habitat values

Notes:a “High,” “moderate,” or “low” channel sensitivity is a measure of the size of observable, sustained impacts to channel morphology/aquatic values in response to achange in a watershed process. “High” implies extensive observable sustained negative impacts. “Moderate” implies local extensive or widespread moderatenegative impacts. “Low” implies local moderate to no observable negative impacts.b The list of channel attributes here is incomplete and is only for illustration. The attributes must be considered and interpreted in a temporal, spatial, andcumulative context, not in isolation.c The sensitivity of the channel to increases in bedload sediment and increases in suspended sediment should be considered separately. In small headwater streams,suspended sediment is typically transported through the system rapidly, resulting in short-term negative changes to water quality.


undertaking the analysis must usetheir judgment to define the hazardratings in terms of change inprobability (see Wise et al. [editors]2004, Chapter 3, Table 2).

The likelihood of a hydrologic hazardis estimated by considering the extentand location of existing or proposeddevelopment in a watershed withrespect to elevation,aspect, hillslopegradient, andhillslope–channelconnectivity. Thebiophysical conditions ofthe watershed, includingforest canopy and terraincharacteristics are alsoconsidered.

The likelihood ofoccurrence of increasedpeak discharge (Hp)associated with existingor proposeddevelopment dependson the amount anddistribution of thedevelopment; thecurrent and historicalforest covercharacteristics; and theextent that basin physiography, suchas the amount of alpine area or thevariation of elevations and aspects,allows for de-synchronization ofsnowmelt runoff and controlsstreamflow (Schnorbus and Alila2004). A low level of development(<20%) that is distributed over arange of different elevations andaspects in a forested watershed has alow likelihood of increasing peakdischarge. A moderate level ofdevelopment (20–40%) in awatershed that has analpine-dominated peak discharge willalso have a low likelihood ofincreasing peak discharge (Schnorbusand Alila 2004).

The likelihood of occurrence ofincreased sediment delivery (Hs) in awatershed associated withdevelopment considers the location ofexisting or proposed development

with respect to unstable or potentiallyunstable slopes, the connectivity ofhillslopes and channels, and themechanism and frequency of naturalsediment delivery events. Proposeddevelopment on or above unstable orpotentially unstable slopes adjacent tothe channel in a watershed with fewnatural sediment sources could have ahigh likelihood of increasing sediment

delivery if roads ortrails are proposed(Jordan 2002).

The likelihood ofoccurrence ofharmful changes toriparian function (Hf)considers thelocation of existingor proposeddevelopment withrespect to thefunctioning riparianarea and the degreeof natural variability(both spatial andtemporal) in riparianfunction through thewatershed. Amoderate amount ofdevelopment in ariparian area where

immature coniferous and deciduousspecies offer limited riparian functionwill have a lower likelihood ofdevelopment-related impacts than asimilar level of development in ariparian area with a climax stand ofmature coniferous species providingchannel bed and bank stability.

The likelihood of occurrence of ahydrological hazard is determinedthrough field assessment (focusing onobservations that give information onpast disturbance history of thewatershed); observations of historicaland recent air photos; andinformation from terrain stability, soilerosion, forest cover, anddevelopment maps. Examples ofwatershed attributes anddevelopment factors that contributeto the qualitative assessment ofhydrologic hazard are presented inTable 3.

SummaryUnder British Columbia’s new Forestand Range Practices Act, forestmanagement is moving towards riskanalysis and professional reliance. Inthis new regime, forest managersmust thoroughly understand potentialrisks to aquatic values associated withexisting or proposed development.Results of a hydro-geomorphic riskanalysis can guide new forestdevelopment, identify areas wheremore detailed assessments arerequired, or direct mitigative work.The results can also be used toidentify aquatic values and locationsin the watershed that are suitable formonitoring.

The hydrologic risk analysis suggestedhere is ideally suited for low-orderwatersheds (<50 km²) but can beadapted for use in smaller first-orderwatersheds (<100 ha) as well as largerlandscape-level watersheds (500km²). In a detailed analysis, watershedprocesses are adjusted to reflecthillslope processes and more detailed,site-specific information is requiredsuch as likelihood of landslides, terrainand soil information, the nature ofsurface and subsurface runoff, slopegradient and aspect, and forestcanopy characteristics. The potentialfor cumulative hydro-geomorphicimpacts can be estimated in largerwatersheds by dividing the landscapeinto smaller, hydrologicallymeaningful sub-basins anddetermining risk at each fan orconfluence along the main stemchannel. Applying this risk analysis towatersheds larger than about 50 km²could result in meaningless risk ratingsdue to the increased variability inbasin response at large scales (Bunteand MacDonald 1999; Miller et al.2003).

Risks to aquatic values exist regardlessof forest development. Therefore,such development should notautomatically be excluded from areasof higher risk. In these cases forestmanagers can adapt managementpractices to reduce the potential



A qualitativeapproach tohydro-geomorpho-logic risk analysis isan effective tool toidentify the keyprocesses affectingaquatic valueswithin a watershedand developpracticalrecommendationsto minimize risks toaquatic values fromforest development.

hazards associated with development.Strategies to reduce the likelihood ofoccurrence of a hazard and therebyreduce development-related risk couldinclude undertaking detailed drainageplans to maintain natural drainagepatterns, conducting riparianassessments to ensure blockboundaries do not impinge onriparian function, or adjusting the sizeor distribution of cutblocks to reducethe potential for increasing peakflows.

As with any analysis of qualitative risk,this analysis is subject to professionalexperience and judgment. Therefore,

all observations, interpretations, andassumptions should be appropriatelydocumented.

Eventually, with continued researchinitiatives directed at quantifying theeffects of timber harvest and roaddevelopment on watershed processes(e.g., Schnorbus and Alila 2004), thestrength of risk analyses like the onepresented here will improve. Untilthen, a qualitative approach tohydro-geomorphologic risk analysis isan effective tool to identify the keyprocesses affecting aquatic valueswithin a watershed and developpractical recommendations to

minimize risks to aquatic values fromforest development.

Acknowledgements

The author is indebted to DougVandine, P.Eng., P.Geo.; Rita Winkler,RPF, Ph.D.; Peter Jordan, P.Geo.,Ph.D.; Will Halleran, P.Geo.; andBrett Eaton, Ph.D. Several anony-mous reviewers offered valuable dis-cussions and comments to thismanuscript. Thanks also to YounesAlila, Ph.D., P.Eng., for insightful dis-cussions regarding channel-formingflows and natural variability in water-shed processes.


Table 3. Hydrological hazard

Watershedprocess

Likelihooda

Watershed attributes and development factors contributing to hazard rating b

Increasedpeak

discharge

Low

• Watershed has significant alpine area and peak flows dominated by alpine snowmelt• Watershed has wide elevation/aspect component and openings are appropriately distributed• Minimal existing/proposed development• Minimal road density and ditches are not concentrating runoff

Moderate

• Moderate existing/proposed development in moderate to steep gradient, non-alpine watershed• Moderate road density and ditches are concentrating and delivering runoff to stream network• Development is limited in distribution and (or) focused on 1 or 2 elevation/aspect zones that could

influence peak flows

High

• Watershed is forested to the headwaters and has an upper broad basin or plateau where development isconcentrated

• Limited elevation/aspect distribution and development are concentrated in one or two areas that likelycontrol peak flows

• Extensive existing/proposed development• High road density and ditches are carrying intercepted and concentrated runoff to stream network.

Increasedsedimentdelivery

Low

• Low connectivity (coupling) between hillsides and valley bottom• Large watershed with the capacity to dilute local forestry related sedimentation events• Stable and non-erodible terrain is adjacent to channel• Low road density and few stream crossings

Moderate• Some coupling between valley sides and stream channel with moderate density of roads/trails on or above

unstable or potentially unstable slopes adjacent to channel• Moderate road density and number of stream crossing on steep slopes with erodible soils

High

• Channels are directly coupled to valley sides with high road/trail density located on or above unstableterrain

• Watershed is small with no opportunity for sediment dilution• High road density with numerous stream crossings on moderate to steep slopes with erodible soils

Decreasedriparianfunction

Low• No development in riparian zone• Appropriately sized riparian buffers in place• Few stream crossings by roads or trails

Moderate

• Significant amount of riparian area directly impacted by development• Undersized riparian buffers along some of the channel resulting in a reduction of LWD recruitment or

shade function of canopy• High density of stream crossings by roads or trails

High

• Large amount of development/disturbance in riparian area• Undersized or no riparian buffers along more than half of channel• Main stem channel oriented east–west with moderate or low gradient hillsides. Development/disturbance

has removed significant amount of riparian vegetation on south side of channel.Notes:a The ratings of “low,” “moderate,” and “high” indicate that the likelihood of a harmful or potentially harmful change to a watershed process occurring within thetime span of the development is negligible, not likely but possible, and probable, respectively.b The typical watershed attributes and development factors given for the “low,” “moderate,” and “high” hazard ratings are for discussion purposes only. Differentwatersheds will respond differently to similar levels of road development and harvesting.



Kim Green M.Sc., P.Geo.Apex Geoscience Consultants Ltd.1220 Government StreetNelson, BC V1L 3K8E-mail: [email protected]

References

Association of Professional Engineers andGeoscientists of B.C. (APEGBC). 2003.Overview of B.C. M.o.F. landslidehazard and risk case studies document.Participant handouts for APEGBCAnnual Conference, Penticton, B.C.,October 23–25, 2003.

B.C. Ministry of Forests. 2001. Watershedassessment procedure guidebook.Second edition. Version 2.1. ForestPractices Branch, Victoria, B.C. ForestPractices Code Guidebook.

B.C. Ministry of Forests. 2002. Forest roadengineering guidebook. Forest PracticesBranch, Victoria, B.C. Forest PracticesCode of Guidebook.

B.C. Ministry of Transportation andHighways. 1996. Natural hazards inBritish Columbia. Prepared by B.C.Ministry of Transportation andHighways, Geotechnical and MaterialsEngineering Branch in association withVanDine Geological Engineering Ltd.

Benda, L., D. Miller, P. Bigelow, and K.Andras. 2003. Effects of post-wildfireerosion on channel environments, BoiseRiver, Idaho. Forest Ecology andManagement 178(1–2):105–119.

Bragg, D.C. 2000. Simulating catastrophicand individualistic large woody debrisrecruitment for a small riparian system.Ecology 81(5):1383–1394.

Brown, G.W. 1980. Forestry and waterquality. Oregon State University,Corvallis, Oreg. 121 p.

Buffington, J.M. and D.R. Montgomery.1999. Effects of hydraulic roughness onsurface textures of gravel-bed rivers.Water Resources Research35(11):3507–3521.

Bunte, K. and L.H. MacDonald. 1999. Scaleconsiderations and the delectability ofsedimentary cumulative watershedeffects. U.S. National Council for Airand Stream Improvement, January1999. Technical Bulletin No. 766.

Canadian Standards Association (CSA).1997. Risk management: guidelines fordecision-makers, a national standard ofCanada. CSA International,CAN/CSA-850-97, Etobicoke, Ont.

Carver, M. 2001. Using indicators to assesshydrologic risk. In Watershedassessment in the Southern Interior ofBritish Columbia. D.D.A. Toews and S.Chatwin (editors). B.C. Ministry of

Forests, Research Branch, Victoria, B.C.Working Paper 57/2001.

Church, M. 2002. Geomorphic thresholds inriverine landscapes. Freshwater Biology47:541–557.

Clague, J.J. (compiler). 1989. Quaternarygeology of the Canadian Cordillera.Chapter 1 in Quaternary Geology ofCanada and Greenland. R.J. Fulton(editor). Geological Survey of Canada,Ottawa, Ont. Geology of Canada, No.1.

Faustini, J.M. and J.A. Jones. 2003.Influence of large woody debris onchannel morphology and dynamics insteep, boulder-rich mountain streams,western Cascades, Oregon.Geomorphology 51:187–205.

Gayton, D.V. 2003. British Columbiagrasslands: monitoring vegetationchange. FORREX-Forest ResearchExtension Partnership, Kamloops, B.C.FORREX Series 7.

Gomi, T. and R.C. Sidle. 2003. Bed loadtransport in managed steep-gradientheadwater streams of southeasternAlaska. Water Resources Research39(12):1336. doi:10.1029/2003WR002440.

Gomi, T., R.C. Sidle, and J.S. Richardson.2002. Understanding processes anddownstream linkages of headwatersystems. Bioscience 52(10):905–916.

Hallett, D. and R. Walker. 2000.Paleoecology and its application to fireand vegetation management inKootenay National Park, BritishColumbia. Journal of Paleolimnology24(4):401–414.

Jakob, M. and P. Jordan. 2001. Designflood estimates in mountain streams:the need for a geomorphic approach.Canadian Journal of Civil Engineering28:425–439.

Jordan, P. 2002. Landslide frequencies andterrain attributes in Arrow andKootenay Lakes Forest Districts. InTerrain stability and forestmanagement in the interior of B.C. P.Jordan and J. Orban (editors). B.C.Ministry of Forests Technical Report003.

Lisle, T.E. 2000. The fate of large sedimentinputs in rivers: implications forwatershed and waterway management.AEG News 43(4):99.

Miller, D., C. Luce, and L. Benda. 2003.Time, space and the episodicity ofphysical disturbance in streams. ForestEcology and Management178:121–140.

Montgomery, D.R. 2003. Wood in rivers:interactions with channel morphologyand processes. Geomorphology 51:1–5.

Montgomery, D.R. and J.M. Buffington.1998. Channel processes, classification,and response. In River ecology andmanagement. R. Naiman and R. Bilby

(editors). Springer-Verlag, New York,N.Y., pp. 13–42.

Montgomery, D.R. and L.H. MacDonald.2002. Diagnostic approach to streamchannel assessment and monitoring.Journal of the American WaterResources Association 38(1):1–16.

Obedkoff, W. 2002. Streamflow in theKootenay Region. B.C. Ministry ofSustainable Resource Management,Water Information Section, AquaticInformation Branch, Victoria, B.C.

O’Connor, M. and R.D. Harr. 1994.Bedload transport and large organicdebris in steep mountain streams inforested watersheds on the OlympicPeninsula, Washington. Final report.Timber/Fish/Wildlife Sediment,Hydrology and Mass Wasting SteeringCommittee, and State of WashingtonDepartment of Natural Resources.122 p.

Rinne, J.H. 1996. Short-term effects ofwildfire on fishes and aquaticmacroinvertebrates in the southwesternUnited States. Journal of FisheriesManagement 16:653–658.

Roberts, R.G. and M. Church. 1986. Thesediment budget in severely disturbedwatersheds, Queen Charlotte Ranges,British Columbia. Canadian Journal ofForest Research 16:1092–1106.

Schnorbus, M.A. and Y. Alila. 2004. Forestharvesting impacts on the peak flowregime in the Columbia Mountains ofsoutheastern British Columbia: aninvestigation using long-term numericalmodeling. Water Resources Research40(5):W05205.doi:10.1029/2003WR002918.

Troendle, C.A., M.S. Wilcox, G.S. Bevenger,and L.S. Porth. 2001. The Coon CreekWater Yield Augmentation Project:implementation of timber harvestingtechnology to increase streamflow.Forest Ecology and Management143:179–187.

Uunila, L. 2004. Final report, CrawfordCreek watershed hydrologicassessment, Phases 1 and 2. Preparedfor B.C. Ministry of Forests, B.C. TimberSales, Nelson, B.C. SummitEnvironmental Consultants Ltd.(Volumes 1 and 2) 114 p. + appendicesand maps.

Wemple, B.C. and J.A. Jones. 2003. Runoffproduction on forest roads in a steepmountain catchment. Water ResourcesResearch 39(8):1220. doi:10.1029/2002WR001744.

Wise, M.P., G.D. Moore, and D.F. VanDine(editors). 2004. Landslide risk casestudies in forest development planningand operations. B.C. Ministry of Forests,Research Branch, Victoria, B.C. LandManagement Handbook No. 56.Available from:www.for.gov.bc.ca/hfd/pubs/Docs/Lmh56.htm



Re-creating MeanderingStreams in the CentralOregon Coast Range, USABarbara Ellis-Sugai and Johan Hogervorst

Introduction

For many years, the riparian andstream functions of Bailey and

Karnowsky creeks on the CentralOregon Coast Range have beenimpaired. The valleys werehomesteaded in the 1800s; by theearly to mid-1900s, both streamswere channelized into ditches toincrease the amount of land availablefor pasture. This channelizationdecreased sinuosity, resulting inincreased stream gradients and watervelocities. Both stream channelssubsequently incised into the easilyerodible valley floor. In Bailey Creek,the stream channel started tomeander in the ditch, whichincreased bank erosion and sedimentdeposition into Mercer Lake. InKarnowsky Creek, larger tributarystreams with gradients under 5%were also channelized and downcutto depths of greater than 3 m. Theseconditions led to a loss of aquatichabitat, disconnected floodplains,lower groundwater tables, andincreased bank erosion andsedimentation.

Since 1999, the Siuslaw NationalForest and partners have restored thestream channels and valley floor ofBailey and Karnowsky creeks. Lessonslearned in Bailey Creek were appliedto Karnowsky Creek.

For both streams, we had to answerthe questions: What type of streamshould we build that will fit the valleytype? What should the dimensions ofthe new channel be, including width,depth, cross-sectional area, gradient,sinuosity, and depth and length of

pools? Our project goals included (1)improving coho salmon rearinghabitat; (2) reconnecting channels tofloodplains; (3) restoring riparianvegetation; and (4)reducingsedimentation. Weintended to createchannels that are“stable”: in otherwords, they are able totransport the sedimentload associated withlocal deposition andscour whilemaintaining aconsistent channel sizeand shape (Rosgen 1996). At thesame time, we expected lateralmigration, via bank erosion and pointbar deposition, over time.

This article describes the methods andlessons learned in re-creating twostream channels on the CentralOregon Coast Range.

Bailey Creek RestorationProjectIn 1991, the U.S. Forest Serviceacquired Enchanted Valley in a largeland exchange with a timbercompany; thus, the land changedfrom private to public ownership.Bailey Creek flows through EnchantedValley into Mercer Lake, nearFlorence, Oregon. In 1995, we beganthe project by gathering data onexisting stream conditions. Wecompared Bailey Creek with a similarcoastal stream that had not beencleared and channelized. Atopographic map at a 0.3 m (1 ft.)contour interval of the Bailey Creek

valley floor was created. Other datacollected included severalcross-sections of the existing ditch,pebble counts at the cross-sectionlocations, a longitudinal profile of theditch, and cross-sections and pebblecounts upstream from thechannelized section. For reference onpre-channelized conditions, we usedhistorical aerial photos (1955) thatshowed the position, sinuosity, andmeander geometry of the originalstream channel in the valley above theproject area before channelization.Determining the bankfull, or“design,” flow was the most

challenging aspect ofdata collection. Wedetermined thisparameter via threedifferent sources ofinformation: (1) 16 yearsof correlated flowmeasurements to rainfallrecords (from Giese1996); (2) measureddischarge in the fieldduring winter flowevents; and (3)

comparison of Bailey Creek to othernearby gauged watersheds.

In our restoration plan, we wanted thenew stream to flood frequently duringthe winter to re-establish seasonalwetland characteristics, and tominimize the risk that the new channelwould readjust through bank erosion.We considered designing either awide, shallow stream (a “C” channeltype, width/depth [W/D] ratio > 12,using Rosgen’s [1996] terminology),or a narrow, deep channel (an “E”channel type, W/D ratio < 12; Rosgen[1996]). Based on the valley’s lowgradient, the geomorphic setting (anold lake bed), and reference stream


Determining thebankfull, or“design,” flowwas the mostchallengingaspect of datacollection.

Table 1. Enchanted Valley specifications

Basin area 11.4 km² (4.4 mi.2)

Valley slope 0.34%

Valley length 954 m (3100 ft.)

Streambanksubstrate

60% silt/clay,40% sand


cross-sections, we chose an “E”channel. To ensure flooding duringthe winter flows, we designed the newchannel’s cross-sectional area to be30% smaller than the existing ditch(Tables 1 and 2).

The design parameters were thentranslated onto our base map, using apiece of string cut to the length of thenew channel at the scale of our map.

This method allowed us toeasily adjust the proposedlocation of the stream. Wesubsequently designed apool–riffle morphology forthe stream bottom, withthe pools occupying theoutside bends ofmeanders. The final“string” map was thendigitized and put into ageographical informationsystem. Stake co-ordinateswere calculated andsurveyed onto the ground.

The new 1692 m (5500 ft.)long channel wasexcavated in late summer1999 (Figure 1). Theoutside bends of meanderswere revegetated withwillow stakes in earlyspring 2000, and the newchannel was connected tothe ditch in October 2000.The abandoned ditchdownstream of the

connection was then intermittentlyplugged with fill material that wasoriginally stockpiled during channelconstruction, forming ponds in theunfilled areas. The ponds were locatedwhere small tributaries drained off theside slopes, and connected to the newchannel. Since then, wood has beenadded to the channel, and nativehardwoods, conifers, and shrubs have

been planted in the riparian zonealong the new channel.

Karnowsky Creek ChannelRestoration ProjectKarnowsky Creek, which was acquiredby the U.S. Forest Service in 1992,flows into the Siuslaw River estuarybetween Florence and Mapleton,Oregon. In partnership with theSiuslaw Watershed Council and theSiuslaw Soil and Water ConservationDistrict, we hired a student internteam to develop a whole-watershedrestoration plan during the summer of2001. This team researched watershedhistory, fish and wildlife habitat, andplant communities, and subsequentlydrafted a restoration proposal. Weused this proposal to apply for fundsfrom the Oregon WatershedEnhancement Board and the NationalForest Foundation.

The restoration plan for KarnowskyCreek was similar to Bailey Creek,emphasizing the creation of summerand winter rearing habitat for coho.One heavily aggraded section of ditchthat had suitable existing spawninghabitat was left to passively recover. Incontrast with the Bailey Creek channeldesign, we relied less on dischargecalculations and measurements, andmore on bankfull cross-sections in theexisting ditch to determine thecross-sectional area of the newchannel. We assumed that the ditch


Table 2. Bailey Creek measurements

Parameter Historic channel Ditch New channel

Sinuosity 1.7 1.1 1.8

Gradient 0.22% 0.32% 0.19%

Bankfull width 8.5–12 m (28.5–40 ft.) (based on cross-sectionsabove channelized section)

4–4.5 m(13–15 ft.)

6 m(20 ft.)

Bankfull depth No information 1.5–2 m(5–7 ft.)

0.9 m (3 ft.) in riffles1.5 m (5 ft.) in pools

Belt width 68–85 m (220–275 ft.) Not applicable 46–77 m (150–250 ft.)

Meander length 77 m (250 ft.) (average) Not applicable 73 m (239 ft.) (average)

Radius of curvature 25 m (82 ft.) (average)12–43 m (40–140 ft.) (range)

Not applicable 16 m (52 ft.) (average)

Width/depth (W/D) ratio No information 2–3 in lower valley;25 above channelized section

7

Cross-section area No information 8.36 m² (90 ft.2) 6.27 m² (60 ft.2)

Total channel length No information 1105 m (3627 ft.) 1692 m (5500 ft.)

Figure 1. Low-elevation aerial photograph of the newBailey Creek channel before connection to the old ditch,September 2000.


USD

AFo

rest

Serv

ice

had come into equilibrium withbankfull flows, and would moreaccurately reflect the new channel’ssize requirement than flow equations.To cross-check, we calculated thebankfull discharge for the existingditch and compared it with the newchannel using a regional equation forsmall watersheds developed atOregon State University (Adams et al.1986), the regional U.S. GeologicalSurvey equations (Jennings et al.1993), and Manning’s equation. Aswith Bailey Creek, we wanted thechannel to frequently overtop itsbanks. Therefore, the new channel’scross-sectional area was designed tobe 33% smaller than the existingditch.

In our restoration plan, we explicitlydefined the desired width/depth(W/D) ratio, the slope, and thesinuosity of the new channel. TheW/D ratio is important because it is amajor control on shear stresses withinthe channel. To determine whether

the stream should be a “C” or “E”channel (Rosgen 1996), we used theW/D ratio from a nearby referencestream (9.5), and referred to Rosgen’s(1996) classification system. Unlikethe Bailey Creek design, we allowedmore variation in the size and shapeof the meanders (Williams 1986). Weran several W/D combinationsthrough Manning’s equation andshear stress equations to compare the

existing ditch’s estimated dischargeand shear stress with that calculatedfor the new channel. We chose a W/Dratio of 9.3, a relatively narrowchannel. The rationale was that thevegetation on the valley floor willsupport the higher shear stressesfound in an “E” channel, and anarrower, deeper channel would haveless direct solar heating. The newchannel’s dimensions are shown inTable 3.

For Karnowsky Creek, the upper valleyis slightly steeper, while the lower,tidally influenced valley has a very lowgradient. To fit the valley, wedesigned the new stream’s gradient togradually decrease from 0.76% at thetop to 0.11% in the tidally influencedzone. Likewise, sinuosity increasesdown valley, from 1.2 to 2.8. In thetidally influenced zone, wherefrequent winter flooding occurs, wewanted to create diverse fish habitat

As with Bailey Creek, once the slopeand sinuosity were established, thenew channel was laid out on the basemap, and surveyed onto the ground(Figure 2). The survey data for thechannel location and existing groundelevations were entered into aspreadsheet. The expected bankheights in the new channel, assuminga constant stream gradient through areach, were then calculated. Theupstream and downstream locationsfor pools and riffles were added.


Table 3. New Karnowsky Creek channel dimensions

Channelsegment

Width Riffledepth

W/Dratio

Cross-section

area

Gradient(%)

Sinuosity Newchannellength

Upperchannel

3.1 m(10 ft.)

0.3 m(1 ft.)

10 0.93 m²(10 ft.2)

0.76 1.2 684 m(2223 ft.)

A 4.3 m(14 ft.)

0.46 m(1.5 ft.)

9.3 2.0 m²(21 ft.2)

0.39 1.9 393 m(1278 ft.)

B 4.3 m(14 ft.)

0.46 m(1.5 ft.)

9.3 2.0 m²(21 ft.2)

0.28 2.2 546 m(1773 ft.)

C 4.3 m(14 ft.)

0.46 m(1.5 ft.)

9.3 2.0 m²(21 ft.2)

0.38 1.6 356 m(1157 ft.)

D 4.3 m(14 ft.)

0.77 m(2.5 ft.)

5.6 3.3 m²(35 ft.2)

0.20 1.6 226 m(733 ft.)

E 4.3 m(14 ft.)

0.77 m(2.5 ft.)

5.6 3.3 m²(35 ft.2)

0.11 2.8 1356 m(4406 ft.)

To Siuslaw River

Segment D

Segment C

Segment B

Segment A

Upper Channel

Segment E tidally influenced

Old ditch left asspawning channel

Figure 2. Map of the Karnowsky Creek valley floor and new channel. Key: black, bold solidlines-old ditches; dark blue, bold lines-new channel.


Barb

ara

Ellis

-Sug

ai

The new Karnowsky stream channelwas built in late summer 2002 (Figure3). In the lower part of the valley,where wet soil conditions persistthroughout the year, the excavatedmaterial was piled in mounds andshaped on the valley floor. Thismethod reduced both haul costs andpotential for soil compaction fromdump truck traffic. Mounds alsooffered high points in the floodplainthat provided good planting sites forSitka spruce and western redcedar.During the first winter afterconstruction, willow stakes wereplanted in the banks, and trees andshrubs in the floodplain. At that time,water was not flowing in the mainchannel, which gave the willows achance to establish.

During the second summer (2003),ditches were plugged in several

strategic locations, and water wasdiverted into the new channel. Weapplied the lessons learned in BaileyCreek, where we had left an abruptvertical wall at the connectionbetween the old ditch and newchannel. The old ditch was 1 m (3 ft.)below the new channel. Weerroneously assumed that sedimentwould drop out at this point, as waterslowed to enter the new channel at alower gradient, and causeaggradation in the old ditch.However, a tail cut began to developdownstream from this point as thechannel’s longitudinal profile cameinto equilibrium between the twoelevations. In Karnowsky Creek, thenew channel was designed togradually slope up from the oldditch’s bed elevation to the newchannel, about a 0.3 m (1 ft.)difference in elevation. A ramp oflarge logs was buried at grade in thenew channel at the connection toprevent downcutting.

In the fall of 2003, 130 large, wholetrees were added to the new channeland floodplain by helicopter toprovide current and future cover forfish-rearing areas. Based on researchby Roelof (2002), who completed theplanting plan for the project, we triedto approach 10% coverage of thevalley floor with this wood in 3–4 haless than 2% of the valley floor. Workin three steeper side tributaries andthe upper main stem is ongoing, andmay supply additional spawning

habitat to complement rearing habitatcreated by the work discussed in thispaper. In 2004, the new upperchannel was connected to the existingchannel, and the ditch in the uppervalley was filled in.

MonitoringBoth streams are being monitoredwith permanent cross-sections,low-elevation aerial photographs,on-the-ground photo points,spawning surveys, juvenile fish counts,and collection of water-quality data.In Karnowsky Creek, a network ofgroundwater wells is being measuredmonthly to track groundwater levels.

Results of Restoration

Bailey Creek

The new channel increased channellength by 33% and doubled the poolvolume compared with the old ditch.The stream overflows its banks duringwinter, and the channel appears to berelatively stable, although adjustmentsare occurring. In some places, pointbars and mid-channel bars are beingdeposited, as expected (Figures 4aand 4b). Since the new Bailey Creekwas connected to the existing channelin 2000, U.S. Geological Survey rivergauges on the Alsea River to thenorth, and the Siuslaw River to thesouth have shown annual peak flowsto be slightly below average. Thegauge record goes back to 1940 onthe Alsea River. No gauges are onnearby streams of comparable size.


Figure 3. Construction of the new KarnowskyCreek channel, summer 2002. Stakes inforeground identify the location of atransition from a riffle to a pool. Note theexcavated mounds of soil left on thefloodplain.


J.B.H

oger

vors

t

Figure 4b. New Bailey Creek channel at photo point 7, summer 2003.Figure 4a. New Bailey Creek channel at photo point 7, summer 1999.

J.B.H

oger

vors

t

Jeff

Schm

alen

berg

Fish numbers, both returningspawning adults and rearing juvenilecoho, have increased. Spawningsurveys over the last 4 years indicate adefinite increase in numbers,averaging over 322 fish per kilometre(200 fish per mile) compared with anaverage of 113 fish per kilometre (70fish per mile) annually during the4-year period before the new channelwas built (1996–1999). The increasefrom 2002 to 2003 alone was 130 fishper kilometre (81 fish per mile). The2003 spawning adults were the firstjuveniles reared in the project area tohave returned. The assumption is thatjuveniles of this year’s class tookadvantage of both favourableconditions in the new channel andthe ocean to produce the 2003spawning numbers. Bailey Creek was

one of the few areas where spawningcounts increased in 2003. The controldata indicate that from 2001 to 2003,the number of juvenile coho was1.5–2.0 times higher compared withthe two previous years’ samples. Atthe same time, there was roughly a10-fold increase in numbers ofjuvenile coho in the project area in2001–2003. For 2003, the controlestimate was 0.5 coho per squaremetre while the project area estimatewas 1.1 coho per square metre.

Karnowsky CreekAlthough it is too early to havesignificant monitoring results, we arealready seeing abundant coho smoltsand fry in the new Karnowsky Creekchannel. The channel functioned wellthrough the first two winters, withfrequent floodplain inundation.Willows and other riparian vegetationare growing well, and point bars arebeing deposited on the inside ofmeander bends in the lower channel.Little, if any, bank erosion is evident(Figures 5 and 6). The mounds builtinto the floodplain of the lower valleyare successful nurseries for youngconifers and shrubs.

Future monitoring of fish populationswill include summer snorkel counts inpools of the new channel and spotchecks of ponds created from the oldditches that were plugged. Weconsidered running a smolt trap for

the spring migration, but frequentfloodplain inundation, along with lackof funding, labour resources, andresearch groups prohibits this level ofmonitoring.

Spawning surveys are also ongoing,particularly in a 0.8-km (0.5-mi.)section of upper main stem that wasreconstructed and connected to waterin 2004. In December 2004, cohosalmon were observed spawning atthe top of the upper main-stemchannel, just 2 months after thatsection of the new channel wasconnected to the existing streamchannel.

Summary of LessonsLearned

Cross-sections of the existing ditchare probably more reliable thanregional flow equations ordischarge measurements whendetermining the size of the newchannel.Creating hummocks in thefloodplain aids in re-establishingvegetation in areas infested withreed canary grass, and providesmicro-topographic sites. It alsosaves hauling of the excavatedmaterial.

Grade control and a smoothtransition from the existing ditch tothe new channel will preventdowncutting in the new channel.


Figure 5. The new Karnowsky Creek channelin the lower valley during high winter flows.Conifer seedlings in plastic tubes are plantedon the mounds left on the floodplain.

Barb

ara

Ellis

-Sug

ai

Figure 6. The new Karnowsky Creek channel with large wood added, winter 2003. New conifer seedlings are planted along the bank andprotected by plastic tubing.

Barb

ara

Ellis

-Sug

ai


These restoration projects requireintensive data-gathering andplanning by an interdisciplinaryteam of hydrologists,geomorphologists, fisheriesbiologists, botanists, and surveyors,and benefit from review by othertechnical experts.


Barbara Ellis-SugaiForest HydrologistSiuslaw National Forest4077 Research WayCorvallis, OR 97333Tel: (541) 750-7056E-mail: [email protected]

Johan HogervorstSouth Zone HydrologistSiuslaw National Forest

4480 Hwy 101, Bldg GFlorence, OR 97439Tel: (541) 902-6956E-mail: [email protected]

References

Adams, P.W., A.J. Campbell, R.C. Sidle, R.L.Beschta, and H.A. Froelich. 1986.Estimating stream flows on smallforested watersheds for culvert andbridge design in Oregon. Oregon StateUniversity, Forest Research Laboratory,College of Forestry, Corvallis, Oreg.Research Bulletin 55. 8 p.

Giese, T.P. 1996. Phosphorus export fromthe Clear and Mercer Lake watersheds,Oregon State University, Corvallis,Oreg. M.S. thesis, 136 p.

Jennings, M.E., W.O. Thomas Jr., and H.C.Riggs. 1993. U.S. Geological SurveyWater-Resources Investigations Report94-4002: nationwide summary of U.S.Geological Survey regional regressionequations for estimating magnitudeand frequency of floods for ungagedsites. 196 p.

Roelof, S. 2002. Of fire and fog:representing landscape processes andcatalyzing nonlinear transformation ina Coast Range riparian zone. M.S.thesis draft.University of Oregon,Department of Landscape Architecture,Eugene, Oreg.

Rosgen, D. 1996. Applied rivermorphology. Wildland Hydrology[consultant], Pagosa Springs, Colo.363 p.

Williams, G.P. 1986. River meanders andchannel size. Journal of Hydrology88:147–164.

Results of StreamlineReader Survey 2004Robin Pike

THANK YOU to everyone whoparticipated in our reader survey

this past fall. The survey was designedto help us assess our performanceand, most importantly, solicit yourfeedback on areas for improvementand suggestions for future articles.Overall, respondents told us that weare on track in presenting objectiveand reliable watershed managementinformation. Here are the highlightsof the survey.

Streamline articles aretechnically reviewed andthose readers polled trustinformation presented inStreamline.Of those polled, 42% were unawarethat all articles published in Streamlineundergo a technical peer review. As aresult, we will better communicate themeasures we use to ensure thatreliable and sound information isextended. Despite this finding, 92%of those surveyed indicated that theyeither have a lot of trust (25%) or afair amount of trust (67%) inStreamline. Only 5% of therespondents indicated that they hadlittle trust in Streamline as aninformation source.

Streamline is relevant,scientifically sound, userfriendly, and easy to access.Most readers prefer thecurrent format.Of those completing the survey, 90%indicated that they prefer the currentmix of short newsletter-style andlonger technical articles. Preferencefor print versus online versions of thepublication was more evenly splitamong respondents, with 51%favouring online access, 41% print,and 8% both publication formats. Wewill seriously consider these data iflimited funding in the future does notallow us to produce print versions.

Regarding our publication format,most respondents agreed or stronglyagreed that articles in Streamline arerelevant and applicable (92%),scientifically sound (82%), readable instyle (94%), well laid out (84%), easy


Table 1. Client satisfaction in Streamline’s format

Question: In your opinion,are Streamline articles…

Stronglyagree(%)

Agree(%)

Neutral(%)

Somewhatdisagree

(%)

Stronglydisagree

(%)

1. Relevant and applicable 26 66 7 1 0

2. Scientifically sound 19 63 17 1 0

3. Readable in style 27 67 4 2 0

4. Well laid out 31 53 15 0 1

5. Easy to access 36 56 7 1 0

6. Innovative in content 14 52 30 4 0

Jack Minard of the TsolumRiver Restoration Society,Courtenay, B.C., won oursurvey draw-prize—a $75gift certificate to Chapters.


to access (92%), and innovative incontent (66%). As a result we are notplanning to change the format.Complete results, includingpercentage of neutral responses andthose in disagreement, are presentedin Table 1.

Streamline articles elevatereaders’ knowledge ofwatershed managementscience and issues. Thosereaders polled indicatedthat they are happy with thecurrent mix of topics andrecommended thatStreamline continue tofeature diverse topics.Respondents were asked to commenton how Streamline has improved or

assisted them in their work activitiesor decisions. Most commonly,respondents stated they useStreamline to increase their awarenessand (or) background knowledge ofwatershed management issues andscience. A few readers said they useStreamline articles for teaching andtraining.

Readers were also asked if they couldidentify a favourite article over the last2 years. Many suggestions were putforward, the most common being“Shade and Stream Temperature” byP. Teti, followed by “Wildfires andWatershed Effects in the Southern B.C.Interior” by D. Scott and R. Pike.

Readers also offered feedback on thetypes of articles they want to see in

future issues. We grouped the 105suggestions into categories (Table 2).Notably, most of these suggestionssupport the current direction ofStreamline’s expanded watershedmanagement mandate. The mostcommon response was for acontinuation in the diversity of topicscovered.

Streamline is providingtimely access toinformation, thus increasingknowledge of watershedmanagement research,hydrologic processes andthe effects of watershedmanagement.Finally, we asked readers if we wereachieving our objectives (Table 3).While these results are qualitative andhave their limitations, 74% agreedthat we’re giving them timely accessto information, 93% indicated thatStreamline increases their knowledgeof current B.C watershedmanagement research and expertise,and 76% indicated that Streamlineincreases their knowledge of naturalhydrologic processes and the effectsof watershed management. Overall,the survey results will greatly assist usin managing Streamline, leveragingfinancial support, and ultimatelyoffering a reader-focused and relevantpublication.

If you would like to provide furtherfeedback on Streamline or suggest anidea for an article, please contactRobin Pike at

(250) 387-5887, or by e-mail [email protected].


Table 2. Reader suggestions for future contentProposed article topic / content direction Ranka

More of the same mix 1Watershed/stream restoration 2Monitoring/assessment 3More forestry 4Fish habitat/fisheries 4Methods and techniques 5Riparian issues and science 5Water conservation/social issues 5Stewardship 6Water chemistry, quality, and health 6Hydrology/Watershed processes 6Non-forested watersheds/urban watershed management 6Biology/Biodiversity 6Broaden scope to include international case studies 6Standards/BMPs/Policy 6Groundwater 7Less forestry 8Wildfire effects 8Opinion pieces 8Current updates from around B.C. 8

a Reader suggestions were grouped into the above categories. The number of suggestions percategory was tallied and rank was assigned (1 to 9). Tied rank values represent categorieswith an equal tally.

Table 3. Client satisfaction in Streamline’s performance against objectives

Question:Does Streamline…

Stronglyagree(%)

Agree(%)

Neutral(%)

Somewhatdisagree

(%)

Stronglydisagree

(%)

1) Provide timely access to watershed managementinformation?

21 53 25 1 0

2) Increase your knowledge of current B.C. watershedmanagement research and expertise?

38 55 6 1 0

3) Increase your knowledge of natural hydrologicprocesses and the effects of watershed management?

19 57 23 1 0

UPDATEUpcoming Events

April 16–20, 200533rd Annual BCWWA Conference andTrade Show.Penticton, BC

http://www.bcwwa.org/newsdocs/mailer/BCWWA-AGM-Exhibition-Brochure-v1.4.pdf

April 26–27, 2005Implications of Climate Change in BC’sInterior Forests.Revelstoke, BC

http://www.cmiae.org/

May 4–7, 200549th Annual BCWF Convention: Navi-gating the Turbulent Waters of Conserva-tion in British Columbia.Nanaimo, BC

http://www.bcwf.bc.ca/s=142/bcw1089397688895/

May 8–11, 2005Canadian Geophysical Union Annual Sci-entific Meeting.Banff, AB

http://www.ucalgary.ca/%7Ecguconf/

May 25–27, 2005North American Association of FisheriesEconomists (NAAFE) 3rd Biennial Forum.Vancouver, BC

http://www.feru.org/events/naafe.htm

May 31–June 4, 20052005 Joint International Conference onLandslide Risk Management and 18thAnnual Vancouver Geotechnical SocietySymposium.Vancouver, BC

http://cgs.ca/2005ICLRM/

June 14–17, 200558th Annual CWRA National Conference,Reflections on our Future...a New Centuryof Water Stewardship.Banff, AB

http://www.reflectionsonourfuture.ca/

August 8–11, 2005Earth System Processes II. GeologicalSociety of America.Calgary, AB.

http://www.geosociety.org/meetings/esp2/

August 16–19, 2005Second North American Lake TroutSymposium.Yellowknife, NWThttp://www.laketroutsymposium2005.ca/

August 17–19, 2005Hydrotechnical Engineering: Cornerstoneof a Sustainable Environment.Edmonton, ABwww.hydrotechnics.ca/hydro2005

September 13–15, 200510th International Specialist Conferenceon Watershed and River Basin Manage-ment 2005.Calgary, ABhttp://content.calgary.ca/CCA/City+Hall/Business+Units/Waterworks/Events/IWA+Watershed+Conference+2005.htm

September 29 to October 1, 2005The Coastal Cutthroat Symposium: Biol-ogy, Status, Management, andConservation.Fort Worden State Park(near Port Townsend, Washington)http://www.orafs.org/cutthroat.html


Recent Publications

BC Journal of Ecosystems andManagement Volume 5, Number 2Densmore, N., J. Parminter, and V.Stevens. 2004. Coarse woody debris:Inventory, decay modelling, andmanagement implications in threebiogeoclimatic zones. BC Journal of

Ecosystems and Management 5(2).Available from:http://www.forrex.org/jem/2004/vol5/no2/art3.pdf

Gayton, D. 2004. Nature conservationin an era of indifference. BC Journal ofEcosystems and Management 5(2).Available from:http://www.forrex.org/jem/2004/vol5/no2/art1.pdf

Krzic, M., H. Page, R.F. Newman, andK. Broersma. 2004. Aspenregeneration, forage production, andsoil compaction on harvested andgrazed boreal aspen stands. BC Journalof Ecosystems and Management 5(2).Available from:http://www.forrex.org/jem/2004/vol5/no2/art4.pdf

Lewis, J.L., S.R.J. Sheppard, and K.Sutherland. 2004. Computer-basedvisualization of forest management: Aprimer for resource managers,communities, and educators. BCJournal of Ecosystems andManagement 5(2). Available from:

http://www.forrex.org/jem/2004/vol5/no2/art2.pdf

Morford, S., D. Robinson, F. Mazzoni,C. Corbett, and H. Schaiberger. 2005.Participatory research in ruralcommunities in transition: A case studyof the Malaspina–Ucluelet ResearchAlliance. BC Journal of Ecosystems andManagement 5(2). Available from:www.forrex.org/jem/2004/vol5/no2/art5.pdf

Prescott, C., L. Blevins, and C. Staley.2005. Litter decomposition in BritishColumbia forests: Controlling factorsand influences of forestry activities. BCJournal of Ecosystems andManagement 5(2). Available from:http://www.forrex.org/jem/2004/vol5/no2/art6.pdf

New text bookNorthcote, T.G. and G.F. Hartman(editors). 2004. Fishes and Forestry:worldwide watershed interactionsand management. Blackwell Science,London, U.K. 789 p.

Slug Injection Using Salt in Solution - uvm.edu...time spent at the field site. Note that the volume...

Documents

Transcript of Slug Injection Using Salt in Solution - uvm.edu...time spent at the field site. Note that the volume...