Volume 8 Number 2 Spring 2005

Introduction to Salt Dilution Gauging for Streamflow Measurement Part III: cloud will cause EC to increase from its background value to a peak value, Slug Injection Using Salt in corresponding to the passage of the core of the cloud, followed by a decline to background EC as the Solution trailing edge of the cloud passes, resulting in a characteristic salt wave R.D. (Dan) Moore (Figure 2). Longitudinal dispersion reduces the peak EC of the salt wave as it travels downstream. The time Introduction Conceptual Basis required for the peak of the wave to revious Streamline articles In this approach, a volume of salt move past an observation point will Pintroduced the general principles solution, V (m³), is injected as a depend inversely on the mean of stream gauging by salt dilution near-instantaneous slug or gulp at one velocity of the streamflow, while the (Moore 2004a) and the procedure for location in the stream. Following duration of the salt wave will depend constant-rate injection (Moore injection, the salt solution mixes Continued on page 2 2004b). While constant-rate injection rapidly throughout the depth of the is best suited for use in small streams stream and less rapidly across the Inside this issue: at low flows (discharges stream width as it travels less than about 100 L/s downstream with the Introduction to Salt or 0.1 m³/s), slug Slug injection general flow of water. Dilution Gauging for injection can be used to works well in Because some portions of a Streamflow Measurement Part III: Slug Injection gauge flows up to 10 steep, highly stream flow faster than m³/s or greater, others (e.g., flow tends to Using Salt in Solution depending upon turbulent be faster in the centre than streams. An Inexpensive, channel characteristics. near the banks), the cloud Automatic Gravity-fed Slug injection works of salty water “stretches” Water Sampler for well in steep, highly downstream in a process Investigating Water turbulent streams, such as the called longitudinal dispersion. This bouldery mountain channel dispersion results in the cloud having Quality in Small Streams shown in Figure 1. This article a leading edge with relatively low Live Gravel Bar Staking introduces the conceptual concentrations of salt solution, a Channel Stabilization in basis and field procedures central zone of high concentrations, the Lower Elk River for slug injection using followed by a trailing edge of salt in solution. decreasing concentration. A Qualitative Hydro-Geomorphic Risk If the electrical conductivity (EC) is Analysis for British recorded at some point downstream, Columbia’s Interior where the tracer has been completely Watersheds: A Discussion mixed across the stream width, the Paper passage of the salt Re-creating Meandering Streams in the Central Oregon Coast Range, USA Results of Streamline Reader Survey 2004 Update Continued from page 1

Published by: FORREX–Forest Research Extension Partnership Suite 702, 235 1st Avenue Kamloops, BC V2C 3J4

Project Manager: Robin Pike Tel: (250) 387-5887

Distribution/Mailing List: Janet Jeffery Tel: (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. John Heinonen Jordan, M. Miles, R.D. Moore, R. Pike, D. Figure 1. Place Creek at high flow during summer glacier melt. Polster, P. Raymond, R. Scherer, M. Schnorbus, K. Swift, P. Teti, R. Winkler on the amount of longitudinal where T represents the salt wave Publication and Web Site Support: dispersion, which, in turn, depends on Jesse Piccin, Satnam Brar, Julie Schooling duration (s). Equation [2] can be how variable the stream velocities are rearranged to solve for Q: Graphic Layout: SigZag Design across the stream. The author has V found that the time required for the = Editing: Ros Penty Q [3] salt wave to pass typically varies from ò RC() t dt Cover Illustration: William McAusland a couple of minutes (e.g., Figure 2) to T McAusland Studios, Kamloops, B.C. over 20 minutes. Under low-flow Streamline is published twice a year by conditions with low velocities, the In practice, RC(t) is determined at the FORREX. All articles published in duration can be longer than desired downstream measurement point at a Streamline are reviewed to ensure reliable D and technically sound information is for accurate measurements (e.g., well discrete time interval t (e.g., 1 or 5 extended to our readers. Content over 30 minutes). s), and the integral is usually published in Streamline reflects the approximated as a summation: opinions and conclusions of the At any time (t) during the salt wave contributing author(s), not those of passage, the discharge of tracer ()@ ()D FORREX, our editorial staff, or our funding ò RC t dtå RC t t [4] partners. Please contact Robin Pike, solution q(t) (L/s or m³/s) past the Streamline Project Manager, for further point will be approximated by: Tn guidelines on article submission or with where n is the number of your comments and suggestions. qt()=× Q RCt () [1] measurements during the passage of the salt wave. The relative This publication is funded in part by the where Q is the stream discharge (L/s British Columbia Ministry of Forests concentration can be determined or m³/s) and RC(t) is the relative through the Forest Investment Account, from EC: Forest Science Program, and by the concentration of tracer solution (L/L) USDA Forest Service. in the flow at time (t). Equation [1] RC() t=- k [ EC () t ECbg ] [5] assumes that q(t) is much smaller than ISSN 1705-5989 Q, which should be true in virtually all where EC(t) is the electrical Printed in Canada cases. If the tracer discharge is conductivity measured at time t, ECbg © FORREX–Forest Research Extension integrated over the duration of the is the background electrical Partnership conductivity of the stream, and k is a Printed on recycled paper salt wave, and if the stream discharge is constant over that time, then the calibration constant. The calibration http://www.forrex.org/streamline following equation should hold for a constant, k, depends primarily on the conservative tracer (i.e., one that does salt concentration in the injection not react with other chemicals in the solution and secondarily on the water, bind to sediment, or otherwise chemical characteristics of the change as it flows downstream): streamwater. Combining Equations ==() () [3], [4], and [5], the following Vò qtdt Qò RCtdt[2] practical equation can be derived for TT computing discharge:

2 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 = V available. In addition, the salt solution is mixed in one container Q [6] concentrations and durations of then decanted into a second, ktD å[() ECt- ECbg ] exposure normally involved in pre-calibrated container (e.g., Østrem n discharge measurement are less than 1964). This procedure ensures that To apply Equation [6], we need to thresholds associated with deleterious the salt in the injection solution is know V, the volume of salt solution effects on organisms (Moore 2004a). completely dissolved, and allows injected; measure the resulting Wood and Dykes (2002) observed accurate measurement of the injection transient increases in invertebrate drift changes in EC at intervals of Dt until volume. during slug injection, but concluded EC returns to background levels; and that salt injection had a relatively Required Volumes of Injection determine the calibration constant, k. short-term effect and is unlikely to Solution Field Procedures have any long-term deleterious The accuracy of a measurement Choice of a Measurement Reach impacts on invertebrate communities depends on how much EC increases at most locations. above background during the salt Successful application of the slug wave passage, relative to the accuracy injection technique requires a stream The salt concentration in the injection of the conductivity probe. The change reach that generates complete lateral solution should be high enough to in EC during the salt wave passage mixing in a short distance. Selected increase EC reasonably when using depends, in turn, on the volume of reaches should have as little pool volumes of solution that can be easily salt solution and its concentration, as volume as possible because the slow handled, but it also needs to remain well as the mixing characteristics of exchange of tracer between the pool less than the solubility. Given the low the stream. Those streams with less volume and the flowing portion of the temperatures often associated with longitudinal dispersion will exhibit a stream will greatly increase the time field conditions, the maximum more peaked salt wave with higher required for the salt wave to pass. An concentration that will dissolve readily concentrations, and will require lower ideal reach begins with an injection is about 20%, or about 1 kg of salt in injection volumes. site upstream of a flow constriction 5 L of water (Østrem 1964; Kite (e.g., where the flow narrows around 1993). We have found that a mixture Kite (1993) suggested that peak EC a boulder, promoting rapid lateral of 1 kg of salt with 6 L of water should be 50% higher than (roughly a 17% solution) provides a background, while Hudson and Fraser 30 suitable (2002) suggested that peak EC should compromise be at least 5 times higher than 25 between strength background. Background EC in B.C. and ease of streams typically ranges from about 20 dilution. 10 mS/cm for stormflow conditions in

S/cm) 15 The injection streams draining catchments m solution does not underlain by granitic bedrock, to over EC ( 10 need to be mixed 400 mS/cm for low-flow conditions in from local streams sustained by groundwater 5 streamwater. discharge. The author suggests that Where access to increasing EC by 100–200% of 0 0 60 120 180 240 the stream does background should be adequate for Time (s) not involve a long streams with low background EC (less hike, it is often than about 50 mS/cm), while Kite’s Figure 2. Example of salt wave in Place Creek. convenient to (1993) guideline should be reasonable mixing) and contains no pools or pre-mix the for streams with background EC backwater areas below the injection solution to allow generous greater than about 100 mS/cm. constriction. A rough guideline is that time for dissolution and to minimize Table 1 summarizes the the mixing length should be at least time spent at the field site. masses/volumes of injected salt/salt 25 stream widths, but complete Note that the volume of the injection solution used by various authors. The mixing may require much longer or solution will be greater than the range reflects the diversity of channel shorter distances, depending on volume of water used to mix it. We morphologies and discharges stream morphology (Day 1976). have found that when a 1-kg box of encountered in the different studies. Mixing the Injection Solution salt is mixed with 6 L of water, the The author recommends starting with We use NaCl (table salt) as a tracer resulting solution has a volume of 1 L of 15–20% solution per m³/s. because it is inexpensive and readily 6.36 L (±0.01 L). Commonly, the salt Greater volumes of injected salt Continued on page 4

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 3 Continued from page 3 placed cobbles) so that it will not detailed description of the procedure solution may be appropriate for wider move during the measurement. To and the calculation of k. streams that require longer mixing position the probe in a strong current, Although ideally the calibration is reaches, while lower volumes may it may be useful to attach the probe performed in the field, particularly to work for narrower streams. To avoid to a rod weighted at the end that is maintain water temperature as close excessive salt concentrations in the placed in the water. to stream temperature as possible, it stream, one or more trial injections In some cases, the background EC can also be conducted in the should be conducted with low may vary. One possible cause is an laboratory. To perform the calibration volumes, working up to larger overly sensitive conductivity meter. off site, two 1-L samples of volumes as required. streamwater should be measured Table 1. Volumes/masses of injected salt used in different studies accurately into sample bottles using a Author Mass of salt Equivalent volume (L) volumetric flask. A sample of the injected per m³/s of 20% salt solution injection solution should also be taken streamflow (kg) (1 kg salt in 5 L water) in a small glass (not plastic) bottle to Østrem (1964) 0.5 2.5 avoid potential problems with salt Church and Kellerhals (1970) 0.2 1 sorbing onto the walls of a plastic Day (1976) 0.3 1.5 bottle. The calibration can then be Elder et al. (1990) 5 25 conducted following the procedure Hudson and Fraser (2002) 2 10 described by Moore (2004b).

Figure 2 illustrates a salt wave for Another cause of varying background Summary of Field Place Creek, where the author has EC is incomplete mixing of Procedures found the salt waves to be highly streamwater and groundwater (which Table 2 lists the equipment required. reproducible. Injecting 6.35 L of a typically has higher EC than the Suggested steps for conducting field roughly 17% solution into a flow of streamwater) within and immediately measurements are as follows: 2.66 m³/s produced a peak EC about downstream of groundwater 100% higher than background. discharge zones. Similar problems 1. Mix injection solution (either at with incomplete mixing can occur office or on site). Recording Electrical Conductivity downstream of tributaries. In these 2. Select measurement reach. Ideally, a data logger should be used latter cases, find an observation point 3. Use a pipette to extract a known to record the passage of the salt wave. where background EC is uniform volume of injection solution (e.g., 10 Some conductivity meters have across the channel and constant in mL) and add to the secondary built-in data logging, while others can time. output a signal that can be recorded solution bottle. Cap the bottle and using a separate data logger. If you do Determining k by Calibration store upright. not have data logging capacity, To determine k, a known volume of 4. Record background EC and water record EC manually at 5-s intervals. injection solution (typically 5 or 10 temperature at the downstream end Although this approach may not be as mL) is added to a known volume of of the measurement reach, and accurate as using a data logger and a streamwater (typically 1 L) to produce upstream of the injection point. 1-s recording interval, it can produce a secondary solution. Known 5. Set up the conductivity probe at satisfactory results. In most cases, two increments of this secondary solution the downstream end of the mixing people are required to conduct a salt are then added to a second known reach. Record the background EC and dilution measurement with manual volume of streamwater (typically 1 L), recording, while the use of a data water temperature. If you have a data to generate a set of EC values logger, start recording EC. logger allows a single person to make corresponding to different values of the measurement. relative concentration. The slope of 6. Inject a known volume of salt The conductivity probe should be the relation between relative solution at the upstream end of the placed within the main part of the concentration and EC provides the mixing reach. flow, not in a backwater. Avoid required value for k. This two-step 7. Record the passage of the salt locations with substantial aeration, as procedure dilutes the injection wave, continuing until EC returns to air bubbles passing through the probe solution to the relative concentrations background. If EC does not return to cause spurious drops in conductivity. observed during the salt wave without background, measure EC upstream of The probe should be firmly emplaced using large volumes of streamwater. the injection point again to determine (e.g., by wedging it between carefully See Moore (2004b) for a more whether the background changed.

4 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 8. Measure a volume V0 (e.g., 1 L) of streamwater using the volumetric flask Worked Example and pour into the secondary solution Figure 2 shows a salt wave recorded during a slug-injection measurement bottle, which already contains the at Place Creek, located about 30 km northeast of Pemberton, B.C. Place sample of injection solution. Cap the Creek, a steep, bouldery mountain stream, would be impossible to gauge bottle and shake vigorously to mix the accurately using a current meter (Figure 1). The volume of injection streamwater and injection solution. solution was 6.35 L. This volume resulted from mixing 1 kg of salt with 6 L This mixture is the secondary solution. of 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. The 9. Measure a volume Vc (e.g., 1 L) of streamwater using the volumetric flask stream EC data were logged at 1-s intervals, and the calibration constant –6 and pour into the calibration tank. was 2.99·10 cm/mS. -33 Immerse the calibration tank in a V 635. × 10 m 3 Q = = = 266./ms shallow pool at the stream’s edge. ktDS[() ECtEC- ](.29910× -6 cm /m S )( 1 s)(797mScm / ) Keep the temperature in the tank as bg close to stream temperature as possible (Moore 2004b). To where the salt plasticware into the field as a backup help hold the calibration tank in wave is recorded. in case of breakage. place, position a “corral” of Under suitable In addition, cobbles around it. conditions, If it is raining during the discharge should measurement, ensure that the 10. Perform the calibration and streamflow not change calibration tank is sheltered. determine k using the measurements appreciably Otherwise, rain falling into the tank procedure described by Moore during the made by slug may dilute the concentrations below (2004b), then compute the injection trial. injection can be the calculated values, producing discharge using Equation [6]. precise to within Errors may arise biased calibrations. through about ±5%. The slug injection method may not be Errors and Limitations inaccuracies in appropriate when the channel Under suitable conditions, measuring the contains ice and (or) snow. In such streamflow measurements volumes of cases, low velocities may result in made by slug injection can be precise streamwater, injection solution, and poor lateral mixing and excessively to within about ±5% (Day 1976). secondary solution. These errors can long salt wave durations, particularly if Accurate measurements require that be effectively minimized if a salt solution flows into slush zones (1) the salt in the injection solution be volumetric flask is used to measure within the measurement reach. completely dissolved, and (2) the streamwater and glass pipettes used injection solution be fully mixed to measure the injection and The method will be subject to across the channel at the location secondary solutions. However, take substantial errors if the measurement Continued on page 6 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 passage 1Separate 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

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 5 Continued from page 5 reach is not sufficiently long to ensure higher flows (e.g., >5–10 m³/s), it is For further information, contact: complete lateral mixing. Unlike easier to inject dry salt than to mix Dan Moore, Ph.D., P.Geo. constant-rate injection, where lateral and inject adequate volumes of Associate Professor mixing can be verified once solution. However, a disadvantage of Departments of Geography and Forest steady-state conditions have been the dry salt method is that an Resources Management achieved, assessing mixing is more accurate scale to measure the mass of difficult with slug injection. If two salt or an adequate supply of 1984 West Mall University of British Columbia probes are available, then the salt pre-weighed salt in a range of Vancouver, BC V6T 1Z2 wave can be recorded at two quantities is needed. Where an downstream distances or on either accurate scale and pre-weighed Tel: (604) 822-3538 E-mail: [email protected] side of the stream. If mixing is quantities of salt are unavailable (e.g., complete, discharge calculated from at a remote site over an extended field References both probes should be in reasonable season), the slug injection method agreement. If this is not the case, a using salt solution would still be Church, M. and R. Kellerhals. 1970. Stream longer mixing reach is required. possible because the precise mass of gauging techniques for remote areas Alternatively, if only one probe is using portable equipment. Department salt in the injection solution does not of Energy, Mines and Resources Inland available, successive measurements need to be known, just the volume of Waters Branch, Ottawa, Canada. can be made during periods of steady the solution (Equation [6]). Technical Bulletin No. 25. flow using different distances. Day, T.J. 1976. On the precision of salt dilution gauging. Journal of Hydrology Problems can occur if the conductivity Summary 31:293–306. does not return to background. If the Streamflow measurement by slug Elder, K., R. Kattelmann, and R. Ferguson. measurements taken upstream show injection of salt solution has been 1990. Refinements in dilution gauging for mountain streams. In Hydrology in that the background has truly successfully applied in many locations Mountainous Regions. I - Hydrological changed, then an average of the around the world. It is particularly Measurements; the Water Cycle. original and final background values suitable for steep, bouldery mountain International Association for may be used in Equation [6]. It is streams, which are unsuitable for Hydrological Science (Proceedings of more problematic if EC has not gauging by conventional current two Lausanne symposia, August 1990). IAHS Publication No. 193, pp. returned to background due to a slow metering techniques. This article has 247–254. release of stored salt solution within described procedures that the author Hongve, D. 1987. A revised procedure for the mixing reach, as can occur in has found useful at sites throughout discharge measurements by means of reaches with pools, particularly at British Columbia. However, there is the salt dilution method. Hydrological Processes 1:267–270. lower flows. In such cases, one great scope to vary the details to suit Hudson, R. and J. Fraser. 2002. Alternative solution would be to extend the tail of individual circumstances and users are methods of flow rating in small coastal the salt wave by fitting an exponential encouraged to experiment with the streams. B.C. Ministry of Forests, decline to the values, although the outlined procedure. Vancouver Forest Region. Extension actual form of the decline will still be Note EN-014 Hydrology. 11 p. uncertain (Elder et al. 1990). Ideally, Kite, G. 1993. Computerized streamflow Acknowledgements measurement using slug injection. one should find a reach with minimal George Richards helped me Hydrological Processes 7:227–233. storage. experiment with variations on the slug Moore, R.D. 2004a. Introduction to salt dilution gauging for streamflow injection method while working at Injection of Salt in Solution measurement: Part 1. Streamline Place Creek. Tim Giles, Dave Watershed Management Bulletin Versus Injection of Dry Salt: Hutchinson, Scott Babakaiff, and John 7(4):20–23. A Comparison Heinonen helped refine the methods Moore, R.D. 2004b. Introduction to salt A number of authors have advocated by asking valuable questions and dilution gauging for streamflow measurement Part II: Constant-rate the use of dry salt injection as an bringing relevant articles to my injection. Streamline Watershed alternative to injection of salt in attention. Comments on earlier Management Bulletin 8(1):11–15. solution (Hongve 1987; Elder et al. versions of this article by John Østrem, G. 1964. A method of measuring 1990; Kite 1993; Hudson and Fraser Heinonen, Russell White, Michael water discharge in turbulent streams. 2002). A future Streamline article will Church, Robin Pike, and four Geographical Bulletin 21:21–43. anonymous reviewers helped improve Wood, P.J. and A.P. Dykes. 2002. The use focus on streamflow measurement by of salt dilution gauging techniques: dry salt injection. The key advantage its clarity. However, any errors remain ecological considerations and insights. of the method is that, for gauging my sole responsibility. Water Research 36:3054–3062.

6 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 represents one sample. A 30-day programmable clock allows for flexibility in sample scheduling (e.g., An Inexpensive, Automatic daily, weekly, or monthly). The master valve option (a valve that opens Gravity-fed Water Sampler whenever a sample valve is opened) is adaptable to control small pumps for Investigating Water where a gravity-fed approach is not feasible (e.g., lakes, ponds, large Quality in Small Streams rivers). We are presently developing an automated pump sampler for this purpose. Chad D. Luider, P. Jefferson Curtis, Rob A. Scherer, and David J. Arkinstall Installation The auto-sampler should be installed Introduction constructed from a PVC pipe with a outside of bankfull width to avoid ater samples are commonly piece of screen mesh secured on the damage during high flows. The intake W collected, either manually intake end to minimize large debris should be installed securely (e.g., (grab samples) or with automated from clogging the lines and valves. wedged between rocks, fastened to samplers, in many environmental The opposite end of the PVC intake rebar) within the streambed upslope monitoring and research programs pipe is connected to a 3–6 m length from the auto-sampler, with enough (e.g., Toews and Gluns 2003; Winkler of polyethylene pipe that forms the hydraulic head (e.g., 1–2 m of vertical et al. 2004). Manual sampling in main water supply line. The valve rise) for the gravity-fed intake system remote areas can be labour intensive manifold system distributes to function properly. For our and time consuming, whereas the streamwater from the intake pipe price of automated samplers (around through electronically controlled $4,000) may be prohibitive to many valves into individual sample monitoring programs. bottles (Figure 3). A plastic cargo box is used to contain the valves, This article describes a low-cost (< the sample containers, and the $600 per unit), gravity-fed, sealed battery-operated control automated water sampler timer, which is programmed to (auto-sampler) that can collect water control the electronic valves. samples from small streams less than Flexibility in sampling depends 5 m in bankfull width. The on the number of valves in the auto-sampler is best suited to collect manifold and the features samples for analyses of water quality associated with the control timer. measures in the dissolved phase, such Electronic control timers typically as pH, conductivity, carbon, control 4–12 valves, and can be phosphorus, and ammonia. Sediment connected in series to increase samples have not yet been collected the number of samples that the with the auto-sampler, and therefore auto-sampler can collect. For sediment sampling is not considered example, two timers controlling in this article. 12 valves each could be connected in series and Auto-Sampler Components programmed to collect a total of and Design 24 samples from the same Chad Ludier Figure 1. Auto-sampler setup adjacent to creek. A simple, lightweight gravity-fed sample site. auto-sampler can be constructed from Desirable features in a control commonly available irrigation timer include the ability to applications we have chosen ¼-inch supplies. The sampler consists of a independently program each valve, a valves although ½- or ¾-inch valves water intake system, a valve manifold 30-day programmable clock, and a can be used. Less hydraulic head is system, and a series of standard master valve option. Independent required for smaller valves compared sample bottles (Figures 1 and 2). The programming for each valve is with the larger valves, which require water intake system (Figure 3) is essential because each valve higher minimum operating pressures, Continued on page 8

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 7 Continued from page 7 and therefore more hydraulic head. the internal plumbing of the system bottles will depend on the flow rate Except for the valves and the solenoid with sample water. As the manual through the intake system. Similarly, plungers, the entire system could be flush valve is larger than the ¼-inch the automated flush valve should be made from stainless steel, sample valves, the flow rate is higher set to rinse the entire volume of the polyethylene, or Teflon where these and thus the intake system flushes system at least 5 times immediately materials are recommended for use in thoroughly. before sample collection. The amount sampling different water quality of time required to flush the intake Due to the design of the sampler, parameters. system can be calculated empirically clogged intake screens can by measuring flow rate through the Upon securing the intake within the significantly affect the operability of system and the length of the main stream, the following points should be the unit. Clogged intake screens can water supply line that is required for considered to minimize the potential be minimized by positioning the installation. For example, it would for air locks in the water supply line intake screen perpendicular to take 5 minutes to completely flush the volume of the intake line once given a flow rate of 1 L/min through 10 m of ½-inch intake line (volume of intake line is about 5 L; i.e., flushing time = volume of pipe/flow rate). The time required for sample collection can be calculated using the same approach (i.e., sample collection time = bottle volume/flow rate), but it is best to set the clock for more time than is required to account for decreases in flow rates. This approach ensures that sample bottles are completely filled and are flushed with sample water. Any excess water spilling from the sample bottles drains via holes in the bottom of the plastic cargo box. Sample Collection Protocol We designed and deployed the

David Arkinstall auto-sampler to collect specific water Figure 2. Auto-sampler components and parts. quality parameters. Therefore, this article will not detail sampling and clogging of the intake screen. Air streamflow and by designating one protocol, which varies with water locks tend to occur in high points of valve to flush the system immediately quality parameter of interest. For the main intake line, particularly in before activating the sample valves. further information regarding the highly aerated sections of stream. Flushing the intake system design of reliable monitoring Intake lines should therefore be immediately before sample collection programs using automated samplers, installed with a constant slope to the greatly reduces the risk of the intake refer to the Automated Water Quality valve manifold, thus avoiding loops in becoming plugged or blocked with Monitoring Field Manual (Resource debris by comparison to a continuous the line that trap air. In addition, the Inventory Committee 1999). intake screen should be submerged in flow setup. The system flush is non-turbulent, uniformly flowing programmed to be completed within Unit Performance less than 1 minute before activating water to minimize air bubbles We used two quality control (QC) the valve for sample collection. entering the system. A ball valve at measures to evaluate the precision the lower end of the valve manifold After the unit has been installed and and performance of the auto-sampler. should also be installed so that the the control timer set, sample bottles The first QC measure included the intake lines can be manually flushed are connected to the spouts of the direct collection of water samples (i.e., after installation and when changing valves designated for sample grab samples) at the auto-sampler the sample bottles (Figure 3). This collection with a piece of tubing and a intake in conjunction with samples flush allows water to flow through the two-holed stopper. The time required being collected with the auto-sampler. intake lines to remove air and rinse on the control timer to fill the sample The second QC measure was used to

8 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 evaluate whether leaching from and deionized water. Analyses of pH, investigation and possibly design adsorption to the water intake system conductivity, dissolved organic modifications: (1) high flows and was contaminating water quality carbon, and dissolved nutrients (PO4 freshet, (2) sediment sampling, (3) samples. We checked this by flushing and NH4) indicate no significant freezing conditions, and (4) streams deionized water through the difference between the grab samples, greater than 5 m bankfull width. auto-sampler at the end of the sample the deionized water, and samples In summary, the auto-sampler has season. We then compared these from the auto-sampler (p < 0.05). allowed us to collect water samples at samples with control samples of a higher frequency relative to manual Applications and sampling and at a reduced cost Constraints of the compared with commercially available Auto-Sampler auto-sampler units. Our auto-sampler In our study, we deployed performed very well during low and tested three summer flows and allowed us to auto-samplers for seven collect and analyze water samples for months (April to October several dissolved water quality 2004) in the Southern parameters. The auto-sampler was Interior of British reliable, cost effective, and easy to Columbia. Units were maintain. Future testing and installed in boulder–cobble improvements to the design of the streams with bankfull auto-sampler will likely increase its widths ranging between 1 suitability to more diverse sampling and 4 m and gradients environments and a greater variety of Figure 3. Schematic view of the auto-sampler. between 5 and 15%. water quality parameters. The labelled parts are listed in Table 1. Samples were collected from the units at a rate of For further information, contact: Table 1. 2–3 samples per week and Equipment list for a five-bottle auto-sampler during our field trials we P. Jeff Curtis Water intake system found that required Department of Earth and Environmental Sciences Item Quantity maintenance to the Okanagan University College A Screen mesh (10 x 10 cm) 1 auto-samplers was B 2” gear clamp 1 minimal. Only one repair 3333 University Way Kelowna, BC C 2” PVC pipe ~30 cm was required to a broken V1V 1V7 D 2” x 1/2” reducer bushing 1 fitting, which caused the loss of one sample. On Tel: (250) 762-5445, Ext 7521 E 1/2” threaded x 1.5 cm (1/2”) 1 E-mail: [email protected] barbed coupler average, the two 9V batteries in each control F 1/2” polyethylene pipe 3–6 m References timer were depleted by only 20% throughout the Valve manifold system Resource Inventory Committee. 1999. entire operation. All three Automated water quality monitoring, Item Quantity auto-samplers were field manual. Unpublished report G Plastic cargo container with lid 1 removed in late October prepared for B.C. Ministry of H 1/2” PVC threaded tee 6 Environment, Lands and Parks, Water due to freezing of the Management Branch for the Aquatic I 1/2” x 2” threaded coupler 6 valves and intake lines. Inventory Task Force. 61 p. Available J 1/2” x 1/8” reducer bushing 6 from: http://srmwww.gov.bc.ca/risc/ The auto-sampler unit is K 1/8” x 2” brass nipple 12 pubs/aquatic/index.htm aptly suited for our L 1/4” valve 6 Toews, D. and D. Gluns. 2003. Water monitoring purposes (i.e., quality sampling: an effective way to M Solenoid 6 chemical water quality monitor watershed condition? N 1/4” vinyl tubing 100–150 cm parameters, low summer Streamline Watershed Management O 1/4” thick-walled heat shrink 25–30 cm flows, ice free conditions). Bulletin 7(2):15–21. P Battery-operated control timer 1 However, the Winkler, R., D. Spittlehouse, T. Giles, B. Heise, G. Hope, and M. Schnorbus. auto-sampler has not been Q 1/2” threaded PVC manual 1 2004. Upper Penticton Creek: how ball valve tested or used under the forest harvesting affects water quantity R 500-mL polyethylene 1–6 following conditions that and quality. Streamline Watershed sample bottle would warrant further Management Bulletin 8(1):18–20.

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 9 mainstem channel. Polster (1999) discusses live gravel bar staking and other soil bioengineering techniques Live Gravel Bar Staking in detail. Channel Stabilization in Site Selection The Elk River, a tributary to Upper the Lower Elk River Campbell Lake, is located on northern Vancouver Island near the town of Gold River, B.C. The potential treatment sites were first selected by analyzing historical air photos from Iain D. Cuthbert and Ian D. Redden 1931 to 1995. The main site selection criteria were gravel bars (1) with easy equipment access, (2) in incipiently uring the past 70 years, the Elk Live Gravel Bar Staking: stable depositional areas, and (3) DRiver on northern Vancouver Background outside of the most active channel Island has evolved from a narrow, Live staking of gravel bars using sections that convey high flows. single-thread, stable channel to a willow (Salix spp.) and other plant Criteria 2 and 3 were extremely wide multi-thread, laterally unstable, species such as red-osier dogwood important as live gravel bar staking of aggraded channel. This change was (Cornus stolonifera) and black the more active channel sections in response to several factors cottonwood (Populus trichocarpa) can could reduce flood conveyance including: valley-bottom logging; be used to treat river channels that capacity and possibly accelerate bank channel relocation due to road have become aggraded and braided. erosion or channel shifting (M. Miles construction; a large landslide in the In live staking, cuttings (stakes) from and Associates 2004). river’s headwaters; and increased the selected pioneering species are flows resulting from the diversion of planted at high density into the gravel water into the Elk River from the bars. adjacent Heber River watershed. The net result: a 4–7 times increase in the During high flows, the treated areas unvegetated channel width in the are inundated; the friction caused by lower 13 km of river and degraded the protruding stakes traps very small fish habitat because pools infilled, woody debris and leads to local banks eroded, and cover was lost. deposition of sediment. Each winter, once enough sediment is deposited to Previous channel morphology studies cover the protruding stakes, Iain Cuthbert (e.g., M. Miles and Associates 1999) streamflow will top the bars without Figure 1. Function of live gravel bar staking. demonstrated the need to restore resistance. In the next growing channel processes in the lower Elk season, the cuttings will grow and A June reconnaissance trip finalized River to expedite the re-formation of a protrude above the gravel bar. This restoration site selection, determined stable, single channel. This project seasonal process of growth followed site access, and located suitable stock addresses this recommendation and by sediment and debris accumulation donor and soaking sites. During this does not incorporate any upland causes the gravel bars to progressively trip we discovered that the natural restoration activities that likely will be stabilize and elevate (Figure 1). At the recovery of many of the potential sites part of future restoration plans. Based same time, the accumulation of fines identified on the air photos was on successful treatments of rivers with and organics, such as small woody significant, and included deciduous similar conditions, such as the San debris, promotes the establishment of trees older than 5 years. We theorized Juan (Switzer 1999), we chose the soil additional riparian vegetation, further that this recovery, the greatest in the bioengineering technique of live stabilizing the bars. Over time the previous 45 years, was due to several gravel bar staking as the preferred gravel bars elevate, and become years with unusually wet summers restoration method to achieve our inundated less frequently. The and smaller than average flood flows. objective. This article describes the streamflow becomes increasingly The natural recovery observed was application of and lessons learned confined to the main channel, vigorous enough to eliminate several from live gravel bar staking in the redirecting the river’s energy to of the potential treatment sites. Thus, lower Elk River. scouring a narrower and deeper three additional sites not identified in

10 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 the main cutting site would not provide enough stock to plant the treatment areas, and two additional donor sites were located. Also, a limited quantity of stock at the main donor site met the diameter criteria. Due to the shortage of large stock, cuttings that were slightly smaller than 2 cm in diameter were also collected and were referred to as “undersize stock.” The cuttings need to be soaked in fresh water for 7–10 days to remove rooting inhibitors before planting (D. Polster, pers. comm., 2004). One challenge of this project was finding adequate soaking sites, as nearby

Ian Redden ponds were shallow and water levels Figure 2. Assembling limbed poles into bundles. dropped during the soaking period due to warm, dry weather. As a result, the office review were investigated in collected, topped, and limbed the cut bundles had to be repositioned the field. trees. Using high quality, relatively several times to avoid drying out. expensive pruning and lopping shears While many areas would have Beavers added another challenge: was invaluable, as smaller shears benefitted from live gravel bar they raided the soaking area; removed tended to break, disrupting staking, site access became the largest some of the largest cuttings, stripping production. The topped and limbed limiting factor. Although Highway 28 the bark and cambium layers from “poles,” which ranged from 2 to 4 m parallels much of the river, steep others; and sometimes took entire in length, were then placed on banks from the highway prevented bundles. sawhorses and tied with equipment from accessing the river. Once most of the donor stock had biodegradable sisal baling twine into The only other road in the area that been collected, the crew split up: one bundles of 7–10 stems (Figure 2). would have provided access to the crew continued cutting, while the Flagging tape was attached to each river had been deactivated for much second crew began planting stakes. bundle, with a different colour used of its length. As most of the lower The planting crew collected the for each day. When the weather river lies within Strathcona Provincial bundles from the soaking sites, taking conditions were cool and wet, Park, excavator access trail building the earliest cuttings first; and then bundles were loaded into trucks and needed to be minimized to preserve transported them to the treatment taken to the soaking site at the end of ecological values. In total, three sites sites where they were cut with each day. During warmer, sunny were selected for treatment. lopping shears into 1 m length stakes weather, bundles were taken to the in preparation for planting. Collecting and Preparing soaking sites throughout the day to the Stakes prevent desiccation (wilting) and Stake Planting death. The use of a Silva cool-tarp to The project began in September 2004 Due to the inherent difficulty of cover the bundles during collection with the collection of donor stock planting in gravel, an excavator with a would have been beneficial during from areas close to the restoration digging bucket was used to install the hot, dry weather. Production sites in Strathcona Park. Stock was cuttings in the coarse gravel bars. The averaged 2840 stakes per day for a collected by cutting down small use of the excavator minimized seven-person crew. deciduous trees close to the ground damage to the stock during planting, with chainsaws. The donors would The target size for stake collection was and ensured that the cuttings were coppice and regenerate in the 2 cm in diameter or larger. This size is planted deep enough to survive the following year. The stakes collected often referred to as the “rule of dry summer. The excavator did not were comprised of 85% willow thumb” as typically anything greater dig holes, but rather inserted the (Scouler’s and Sitka), 6% black in diameter than your thumb is the bucket into the gravel and pulled back cottonwood, and 9% red-osier desired size. After several days of the material, creating a 1 m wide gap dogwood. Crew members then harvesting it became apparent that into which the stakes were placed by Continued on page 12

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 11 Continued from page 11 an average density of 17 200 stems per hectare. Planting took an average of 4.5 days per hectare, with a crew of four people working with the excavator. With the live staking of gravel bars completed, the success of the project will depend on a number of factors, including the growth and survival of the stakes, mortality or stunting due to elk browse, and the response of the treated areas to peak flows.

Monitoring Long-term monitoring will allow us to assess the success of the live gravel bar staking in achieving the project goals. This information can also be

Ian Redden used to help direct future restoration Figure 3. Crew member instructing student on planting stakes. activities. The following measures

hand (Figure 3). The excavator then withdrew its bucket allowing the gravel to settle back in place. The stakes were planted with about three-quarters of their length in the gravel at a 45º or greater downstream angle (Figure 4). Four large stakes and three to five undersized stakes, if available, were placed into each opening, taking about one minute for each opening. While the undersized stakes may not flourish as well as the large stakes, they significantly increased the overall number of stakes planted, which should improve the chances of the project in overcoming Ian Redden mortality due to elk browse. Figure 4. Planted stakes. The excavator worked by backing upstream while planting in successive maximize the area covered with the were taken to assist in gauging project rows spaced 1.5–2 m apart and available stock, the stakes were success. staggered to prevent large open planted with tighter spacing and at The perimeters of the treated areas as patches within the planted areas. The higher densities on the first pass well as longitudinal and first pass of planting occurred nearest nearest the mainstem channel where cross-sectional profiles were surveyed the river channel with the excavator they would likely receive the greatest at each site using a total station positioned at the edge of the zone to flows. be planted. This ensured that the survey instrument. Benchmarks were edge of the row nearest the river was Live staking planting began on established at each site for future planted parallel to the flow. The September 29, 2004, and was reference during surveys and excavator then reached as far as completed on October 12. In total, monitoring. Fifty-one monitoring possible upland from the river. To 1.86 ha was planted at three sites at plots, including seven control plots,

12 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 were established at various locations O Have several stock donor sites Acknowledgements within the project area. Each research selected before beginning cutting. Mike Miles, a fluvial plot had a 3-m radius (28.3 m²). A A local Ministry of Forests office or geomorphologist, assisted the pro- subset of 16 research plots was forest company may offer some ject team by selecting suitable sites established to monitor changes in advice on donor sites. for live staking. David Polster, who substrate composition. Within each of largely pioneered this soil bioengi- O Use high-quality lopping shears these 16 plots the substrate in three neering technique in British Colum- and hand pruners. The loss in 0.5-m² squares was photographed bia, trained the crew to cut stakes to productivity due to the use of poor and documented. Vegetation surveys length. A core team of equipment will cost more than the at each of the 51 monitoring plots Mowachaht–Muchalaht First Nation initial expense of purchasing included recording the number, forestry workers, supported by volun- quality equipment. species, and size of stakes planted in teers from the Gold River each plot. Finally, three permanent Streamkeepers and Gold River Sec- O Use lopping shears rather than a photo points were established at each ondary School, completed this work. chainsaw to cut stakes to length. treatment site for future monitoring. The BC Hydro Bridge Coastal Fish The chainsaw tended to make In 2005, the areas treated in 2004 will and Wildlife Restoration Program rougher cuts and “shred” the bark be monitored and additional live funded this project. near the cut end. staking of gravel bars in the Elk River For further information, contact: will take place. O Avoid soaking sites near known beaver populations. The loss of Iain Cuthbert, M.Sc., R.P.Bio. bundled donor stock due to Streamline Environmental beavers was much higher than we Consulting Ltd. While securing had expected. Using beaver 786 Quilchena Crescent protection such as a rodent fence funding to study the Nanaimo, BC V9T 1P6 around the soaking bundles would Tel: (250) 758-7980 success of restoration have prevented some loss. Fax: (250) 758-8505 E-mail: [email protected] treatments is difficult, O Ensure soaking sites have stable monitoring of the live water levels. The sudden change References from wet weather to several days gravel bar staking M. Miles and Associates. 1999. Preliminary of dry weather caused one soaking assessment of the effects of the Crest project is needed to site to dry up, and the bundles Creek and Heber River diversions on continually improve required repositioning several channel morphology. Consultant’s times during the soaking period. report prepared for BC Hydro and Power Authority, Burnaby, B.C. the selection and Covering soaking bundles with Silva cool-tarps may be beneficial M. Miles and Associates. 2004. Selection of successful application sites for live gravel bar staking: Lower of restoration in hot, sunny weather. Elk River in Strathcona Park. Consultant’s report prepared for BC techniques in British O Flag bundles collected each day Hydro, Bridge Coastal Fish and Wildlife with a different colour flagging Restoration Program, Burnaby, B.C. Columbia. tape. This system allows for quick Polster, D. 1999. Soil bioengineering for and easy identification of the steep/unstable slopes and riparian restoration. Streamline Watershed bundles when collecting them Restoration Technical Bulletin Volume Summary of Lessons from the soaking site. Use bundles 4, Number 4, Winter 1999. Available Learned in the order that they were cut. from: http://www.forrex.org/streamline/ O While securing funding to study issue.asp?issue=14 O Often spring planting offers several the success of restoration Switzer, G. 1999. Observations of the benefits in soil bioengineering. Due success of the willow planting on the treatments is difficult, monitoring to funding timelines, we planted in sand, gravel and cobble bars of the of the live gravel bar staking the fall. lower San Juan River. Streamline project is needed to continually Watershed Restoration Technical Bulletin Volume 4, Number 4, Winter O Use the most recent air photos improve the selection and 1999. Available from: available for preliminary site successful application of restoration http://www.forrex.org/streamline/ selection and reconnaissance. techniques in British Columbia. issue.asp?issue=14

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 13 glacial history, and climate over the past 10 000 years (Clague [compiler] A Qualitative Hydro- 1989). The physical, chemical, and biological characteristics of low-order streams are closely linked to hillslope Geomorphic Risk Analysis (hydro-geomorphic) processes and riparian function (Montgomery and for British Columbia’s Buffington 1998; Gomi et al. 2002). Interior Watersheds: Natural disturbance events such as wildfire, pest epidemics, and floods A Discussion Paper routinely affect watersheds and constitute an intricate part of the dynamic and evolving landscape of Kim Green, M.Sc., P.Geo. the Southern Interior (Bragg 2000; Benda et al. 2003; Gayton 2003). Changes to streamflow, sediment Editor’s Notes: B.C. Ministry of Forests (Wise et al. delivery rates, and riparian function A preliminary version of this article was [editors] 2004) have developed provincial standards for landslide risk (collectively referred to as watershed published in ASPECT, May 2004. Since processes) following natural then, the article has been revised, based on analysis. These efforts have given disturbance define the natural numerous technical reviews. This article is professionals an understanding of the intended to stimulate discussion among terms and methods of risk analysis variability of a watershed over time forest hydrologists about the development needed for detailed terrain stability (Gomi et al. 2002; Miller et al. 2003). of a qualitative hydro-geomorphic risk mapping. Similar efforts to While natural disturbance events analysis for B.C. watersheds. As a discussion standardize methods typically have substantial, immediate paper, the author acknowledges some and terminology are impacts on channel limitations in the material presented below structure and aquatic in attempting to develop the framework. needed if hydrologic Under British risk analyses are to Columbia’s new values such as water Introduction become a widely quality and aquatic Forest and Range habitat (Rinne 1996), nder British Columbia’s new accepted and valued Practices Act, forest the influx of nutrients, Forest and Range Practices Act, component of forest U managers must sediment, and woody forest management is moving management in debris in the decades towards risk management and British Columbia. understand potential following the event can professional reliance. In this new risks to aquatic values The author play a vital role in regime, forest managers must associated with developed this maintaining the aquatic understand potential risks to aquatic hydro-geomorphic existing or proposed ecosystem of a values associated with existing or risk analysis for use in development in a watershed (Benda et al. proposed development in a Southern Interior watershed. 2003; Figure 1). watershed. B.C. watersheds and presents it to open A channel’s response to To date, industry and government discussion regarding disturbance events (i.e., professionals have had minimal a consistent methodology for risk the variability of channel morphology discussions about a standard analysis. Terminology used in this risk in time and space) depends on the approach to hydrological risk analysis. analysis is generally consistent with disturbance regime of a watershed, Inconsistencies in methods, definitions in Risk Management: which is a function of its geographic terminology, and elements being Guideline for Decision-Makers (CSA location (i.e., within British Columbia’s considered in hydrological risk 1997) and Landslide Risk Case Studies hydro-climatic and physiographic analyses are causing significant in Forest Development Planning and regions) and physical attributes differences in the way professionals Operations (Wise et al. [editors] 2004). including bedrock geology and estimate risk (e.g., Carver 2001; B.C. glacial/paraglacial history (B.C. Ministry of Forests 2001; Uunila Natural Variability and Ministry of Transportation and 2004). Channel Response Highways 1996; Montgomery and Recently, the Association of Low-order watersheds (<100 km²) in Buffington 1998; Hallett and Walker Professional Engineers and British Columbia’s Southern Interior 2000; Obedkoff 2002; Miller et al. Geoscientists of B.C. (2003) and the have been shaped by their geology, 2003).

14 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 AB Kim Green Figure 1. Woody debris recruitment and sediment influx following fire are key factors in maintaining aquatic ecosystems in many Southern Interior watersheds. Photos are from similar low-order streams in Caven Creek, southeastern British Columbia. Stream (A) burned during the 2003 Plumbob fire. Stream (B) experienced a similar fire about 70 years ago.

Streams draining steep mountain less natural variability in channel development considers two 1 slopes in the interior wet belt of British morphology. independent factors: the potential response of the channel to changes in Columbia have larger peak discharges Maintaining or improving aquatic watershed processes and the potential per unit area and experience a higher values of watersheds while impact of forest development on frequency of channel-forming events maximizing harvesting opportunities watershed processes. It is expressed as (e.g., debris flows, snow avalanches) is a primary management objective of the product of two components: than watersheds in arid, lowland forest development in British channel sensitivity (C) and hydrologic regions (Jakob and Jordan 2001; Columbia. Understanding watershed hazard (H). Obedkoff 2002). As a result, the processes and the natural variability in morphology of channels in the channel condition and aquatic values Risk=C×H interior wet belt typically have greater allows forest managers to apply Hydro-geomorphic risk is determined natural spatial and temporal variability management practices to reduce the for the main stem and significant than channels in arid and semi-arid risk of direct negative impacts to tributary channels upstream of a point regions where less frequent events low-order streams. In turn, this of interest (POI), such as a water such as wildfire and floods define the reduces the risk for cumulative intake structure or a specific fish disturbance regime. impacts in higher-order streams. habitat (elements at risk) for each Forest development in a watershed Hydro-geomorphic Risk watershed process or, where can cause changes to watershed Analysis appropriate, for each identified processes including increased hillslope aquatic value at the point of interest runoff and stream discharge (Troendle A qualitative risk analysis offers (1) a (e.g., water quality at the intake, et al. 2001; Wemple and Jones 2003; framework for forest hydrologists and channel stability on the fan). A simple Schnorbus and Alila 2004); increased geomorphologists to document matrix such as shown in Table 1 can rate of sediment delivery to streams critical watershed processes (i.e., be used to determine risk in this (Roberts and Church 1986; Gomi and stream discharge, rate of sediment qualitative analysis. delivery, and riparian function) that Sidle 2003); and reduced riparian Channel sensitivity, a measure of the are linked to aquatic values; and (2) function through removal of vulnerability (robustness or fragility) of recommendations for sustaining or streamside vegetation and direct the channel given changes to improving aquatic values within a impacts to channel bed and banks watershed processes, depends on the watershed. This reconnaissance-level (Bragg 2000; Faustini and Jones physical attributes of the channel. analysis is intended to help forest 2003). The potential for significant Channel sensitivity is equivalent to managers identify areas where a more (observable, long-term) change to consequence in the conventional detailed level of assessment is aquatic values in a watershed due to equation of Risk = Hazard × required. changes in watershed processes Consequence.2 The ratings of “low”, associated with forest development Simply stated, estimation of risk to “moderate” and “high” sensitivity will be greater in channels that have aquatic values from forest express the potential size of change to

1In general, channel types associated with transport-limited, alluvial valley segments (e.g., step–pool or riffle–pool channels) or those in supply-limited colluvial valleys dominated by forced alluvial reaches (LWD step–pool or step–bed channels) will have a greater potential for change than channel types in colluvial or bedrock valley segments (Montgomery and Buffington 1998). Continued on page 16

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 15 Continued from page 15

Table 1. Hydro-geomorphic risk matrix (B.C. Ministry of Forests 2002) past natural disturbance in a Hydro-geomorphic risk a Likelihood of a hydrologic hazard watershed forms a baseline for the assessment of hydrologic hazard.3 In Low Moderate High this assessment the likelihood of a Channel Low Very low Low Moderate sensitivity hydrological hazard is expressed Moderate Low Moderate High qualitatively as “low,” “moderate,” High Moderate High Very high and “high.” These ratings indicate aA rating of “negligible” can also be added to the matrix if channel condition is independent of a watershed process, or forest development does not affect watershed process. that the likelihood of a harmful or potentially harmful change to a the channel structure and associated present, and future dependence of watershed process occurring within aquatic values (collectively referred to channel condition on riparian the time span of the development is as channel condition) and are assessed vegetation (Montgomery 2003). A “negligible,” “not likely but possible,” for each of the watershed processes channel with deciduous riparian and “probable,” respectively. separately. What each sensitivity rating species such as alder and willow, implies in terms of probable level of which are indicative of frequent When detailed information such as impact to the channel/aquatic values flooding and snow avalanches, will be flood frequency, annual sediment is specific to a watershed and should less sensitive to disturbance of riparian budgets, and the frequency of be defined in the report. Example function than a channel with mature disturbance to riparian function is definitions are in the footnotes to coniferous riparian species supplying available, the risk analysis can be Table 2. large woody debris that contributes to adapted to be more quantitative. This Channel sensitivity to increases in channel bed and bank stability. Where is done by expressing and contrasting peak discharge considers the potential stream temperature is a concern, the the likelihood of a hydrologic hazard for increased bedload transport, dependence of a channel on riparian in the undeveloped (baseline) which is estimated by considering function considers channel condition and developed (disturbed) mean grain size, grain size orientation, hillslope gradient, and condition as the annual probability distribution, channel gradient, and riparian species (Brown 1980). (Pa) and the long-term probability (Px) hydraulic roughness (O’Connor and for the lifespan of the proposed Channel sensitivity is estimated for the development. Harr 1994; Buffington and main stem channel and larger Montgomery 1999; Church 2002). tributary channels through a For example, a stream that For example, a stream that has a combination of field assessment, experiences a major channel-forming cobble–boulder cascade morphology interpretation of current and historical flood event once every 50 years (1:50) will have a smaller change to channel air photos, and analysis of regional has an annual probability of 0.02 condition due to a given increase in hydrometric and climate information. (2%). If the development in question peak discharge than a low-gradient, Montgomery and MacDonald (2002) has a lifespan of 20 years the gravel, riffle–pool channel. describe in detail a similar approach to long-term probability (P20) of a Channel sensitivity to increases in assessment of channel condition and channel-forming flood event is x sediment delivery considers the sensitivity. Px = 1 – (1–(Pa)) capacity of the channel to transport so Key channel attributes that contribute 20 4 sediment as determined by channel P20 = 1 – (1–(1/50)) = 0.33 (33%). to the estimation of channel sensitivity gradient and sediment storage If development is estimated as opportunities. Due to increases in for the three watershed processes are summarized in Table 2. potentially increasing the annual sediment delivery in the headwater probability of a major reaches, a low-gradient (<5%) A hydrologic hazard is a harmful channel-forming flood from a 1:50 to meandering channel with intervening sustained change to a watershed 1:20 return period (e.g., Schnorbus wetland segments will have a smaller process. The hydrologic hazards and Alila 2004, scenario 2/3U, Table change to channel condition over the considered in this analysis are 3) the long-term probability (P20) of a length of the channel network than a increased peak discharge (Hp), major channel-forming flood is moderate gradient (5–15%) channel increased rate of sediment delivery increased to 0.64 (64%). In this case with limited sediment storage (Hs), and decreased riparian function the proposed development increases opportunities (Lisle 2000). (Hf) associated with proposed and the probability of a channel-forming Channel sensitivity to disturbances of existing development. The variability flood event from 33 to 64% (D31 riparian function considers the past, of watershed processes resulting from percentage points). Professionals 2In this case, channel sensitivity (consequence) equals vulnerability because the spatial and temporal probabilities of the elements being considered (channel structure 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 wasting or erosion events. 4See Wise et al. (editors, 2004), pp. 13–14, and Table, A4.2.

16 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 Table 2. Channel sensitivity

Watershed Channel b a Typical channel attributes that contribute to channel sensitivity process sensitivity • 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 watersheds with steep alpine headwaters Low • Coarse-textured bedload, not the result of a single anomalous flood event or an anthropogenic disturbance; numerous boulder cascade or bedrock reaches • Well-vegetated, overhanging banks (e.g., mature coniferous species with well-developed root system) and abundant 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 • 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 Increased • Some inherent capacity to withstand higher flows, such as overflow channels or an entrenched channel peak with resilient banks or non-alluvial segments Moderate discharge • Banks and riparian area vegetated with species that have well-developed root systems that protect the banks and forest floor from erosion • Often includes forced alluvial channels in colluvial or bedrock valley segments or transitional morphologies in alluvial valley segments • Does not experience frequent flood events; bed is dark and mossy, banks are overhanging, vegetated to bankfull, and show no or little evidence of old scour or overbank deposits • Contains fine-textured bedload that is susceptible to erosion High • 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 • Abundant locations for sediment storage, such as frequent functioning LWD jams or frequent low gradient 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 lacks Low the power to transport bedload (i.e., decoupled systems where source areas are isolated from downstream channels) • Headwaters are steep snow avalanche and (or) debris flow gullies that deliver large volumes of sediment Increased annually sedimentc • Colluvial valley segments with some storage capacity, such as some long (>100–200 m), low gradient delivery Moderate sections (<15%) that allow bedload sediment to settle out • Bordered by currently inactive, but relatively numerous natural landslide scars or debris flow gullies • 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 High 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 • Not dependent on LWD to control rate of sediment transport, such as a steep colluvial or bedrock Low channels or snow avalanche chutes • Low gradient, braided, or anastomosing channels, situated on a wide valley bottom vegetated with shrubs • 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 in colluvial valley segments) Decreased Moderate • Has tendency to migrate laterally across valley bottom and is unentrenched so that migration could be riparian accelerated if valley bottom is disturbed and banks destabilized (e.g., meandering step–pool to riffle–pool function channel in alluvial valley segment) • Some reaches are oriented such that the riparian canopy produces shade and moderates water temperatures • 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 High long-term stability (typically LWD forced alluvial step–pool to cascade pool channels in colluvial valley segments) • 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 a change in a watershed process. “High” implies extensive observable sustained negative impacts. “Moderate” implies local extensive or widespread moderate negative 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, and cumulative 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. Continued on page 18

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 17 Continued from page 16 undertaking the analysis must use with respect to unstable or potentially Summary unstable slopes, the connectivity of their judgment to define the hazard Under British Columbia’s new Forest hillslopes and channels, and the ratings in terms of change in and Range Practices Act, forest mechanism and frequency of natural probability (see Wise et al. [editors] management is moving towards risk sediment delivery events. Proposed 2004, Chapter 3, Table 2). analysis and professional reliance. In development on or above unstable or this new regime, forest managers The likelihood of a hydrologic hazard potentially unstable slopes adjacent to is estimated by considering the extent the channel in a watershed with few must thoroughly understand potential and location of existing or proposed natural sediment sources could have a risks to aquatic values associated with development in a watershed with high likelihood of increasing sediment existing or proposed development. respect to elevation, delivery if roads or Results of a hydro-geomorphic risk aspect, hillslope trails are proposed analysis can guide new forest gradient, and A qualitative (Jordan 2002). development, identify areas where hillslope–channel more detailed assessments are The likelihood of connectivity. The approach to required, or direct mitigative work. occurrence of biophysical conditions of hydro-geomorpho- The results can also be used to harmful changes to the watershed, including logic risk analysis is identify aquatic values and locations riparian function (H ) forest canopy and terrain an effective tool to f in the watershed that are suitable for considers the characteristics are also monitoring. identify the key location of existing considered. processes affecting or proposed The hydrologic risk analysis suggested The likelihood of aquatic values development with here is ideally suited for low-order occurrence of increased respect to the watersheds (<50 km²) but can be within a watershed adapted for use in smaller first-order peak discharge (Hp) and develop functioning riparian associated with existing area and the degree watersheds (<100 ha) as well as larger ³ or proposed practical of natural variability landscape-level watersheds ( 500 development depends recommendations (both spatial and km²). In a detailed analysis, watershed on the amount and to minimize risks to temporal) in riparian processes are adjusted to reflect distribution of the aquatic values from function through the hillslope processes and more detailed, site-specific information is required development; the forest development. watershed. A current and historical moderate amount of such as likelihood of landslides, terrain forest cover development in a and soil information, the nature of characteristics; and the riparian area where surface and subsurface runoff, slope extent that basin physiography, such immature coniferous and deciduous gradient and aspect, and forest as the amount of alpine area or the species offer limited riparian function canopy characteristics. The potential variation of elevations and aspects, will have a lower likelihood of for cumulative hydro-geomorphic allows for de-synchronization of development-related impacts than a impacts can be estimated in larger snowmelt runoff and controls similar level of development in a watersheds by dividing the landscape streamflow (Schnorbus and Alila riparian area with a climax stand of into smaller, hydrologically 2004). A low level of development mature coniferous species providing meaningful sub-basins and (<20%) that is distributed over a channel bed and bank stability. determining risk at each fan or range of different elevations and confluence along the main stem aspects in a forested watershed has a The likelihood of occurrence of a channel. Applying this risk analysis to low likelihood of increasing peak hydrological hazard is determined watersheds larger than about 50 km² discharge. A moderate level of through field assessment (focusing on could result in meaningless risk ratings development (20–40%) in a observations that give information on due to the increased variability in watershed that has an past disturbance history of the basin response at large scales (Bunte alpine-dominated peak discharge will watershed); observations of historical and MacDonald 1999; Miller et al. also have a low likelihood of and recent air photos; and 2003). increasing peak discharge (Schnorbus information from terrain stability, soil Risks to aquatic values exist regardless and Alila 2004). erosion, forest cover, and development maps. Examples of of forest development. Therefore, The likelihood of occurrence of watershed attributes and such development should not increased sediment delivery (Hs) in a development factors that contribute automatically be excluded from areas watershed associated with to the qualitative assessment of of higher risk. In these cases forest development considers the location of hydrologic hazard are presented in managers can adapt management existing or proposed development Table 3. practices to reduce the potential

18 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 Table 3. Hydrological hazard a Watershed Likelihood b Watershed attributes and development factors contributing to hazard rating process • Watershed has significant alpine area and peak flows dominated by alpine snowmelt Low • 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 existing/proposed development in moderate to steep gradient, non-alpine watershed Increased Moderate • Moderate road density and ditches are concentrating and delivering runoff to stream network peak • Development is limited in distribution and (or) focused on 1 or 2 elevation/aspect zones that could discharge influence peak flows • Watershed is forested to the headwaters and has an upper broad basin or plateau where development is concentrated High • Limited elevation/aspect distribution and development are concentrated in one or two areas that likely control peak flows • Extensive existing/proposed development • High road density and ditches are carrying intercepted and concentrated runoff to stream network. • Low connectivity (coupling) between hillsides and valley bottom Low • 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 Increased • Some coupling between valley sides and stream channel with moderate density of roads/trails on or above sediment Moderate unstable or potentially unstable slopes adjacent to channel delivery • Moderate road density and number of stream crossing on steep slopes with erodible soils • Channels are directly coupled to valley sides with high road/trail density located on or above unstable terrain High • Watershed is small with no opportunity for sediment dilution • High road density with numerous stream crossings on moderate to steep slopes with erodible soils • No development in riparian zone Low • Appropriately sized riparian buffers in place • Few stream crossings by roads or trails • Significant amount of riparian area directly impacted by development Decreased Moderate • Undersized riparian buffers along some of the channel resulting in a reduction of LWD recruitment or riparian shade function of canopy function • High density of stream crossings by roads or trails • Large amount of development/disturbance in riparian area High • 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 the time 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. Different watersheds will respond differently to similar levels of road development and harvesting. hazards associated with development. all observations, interpretations, and minimize risks to aquatic values from Strategies to reduce the likelihood of assumptions should be appropriately forest development. occurrence of a hazard and thereby documented. Acknowledgements reduce development-related risk could Eventually, with continued research include undertaking detailed drainage initiatives directed at quantifying the The author is indebted to Doug plans to maintain natural drainage effects of timber harvest and road Vandine, P.Eng., P.Geo.; Rita Winkler, patterns, conducting riparian development on watershed processes RPF, Ph.D.; Peter Jordan, P.Geo., assessments to ensure block (e.g., Schnorbus and Alila 2004), the Ph.D.; Will Halleran, P.Geo.; and boundaries do not impinge on strength of risk analyses like the one Brett Eaton, Ph.D. Several anony- riparian function, or adjusting the size presented here will improve. Until mous reviewers offered valuable dis- or distribution of cutblocks to reduce then, a qualitative approach to cussions and comments to this the potential for increasing peak hydro-geomorphologic risk analysis is manuscript. Thanks also to Younes flows. an effective tool to identify the key Alila, Ph.D., P.Eng., for insightful dis- As with any analysis of qualitative risk, processes affecting aquatic values cussions regarding channel-forming this analysis is subject to professional within a watershed and develop flows and natural variability in water- experience and judgment. Therefore, practical recommendations to shed processes. Continued on page 20

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 19 Continued from page 19 Forests, Research Branch, Victoria, B.C. (editors). Springer-Verlag, New York, For further information, contact: Working Paper 57/2001. N.Y., pp. 13–42. Church, M. 2002. Geomorphic thresholds in Montgomery, D.R. and L.H. MacDonald. 2002. Diagnostic approach to stream Kim Green M.Sc., P.Geo. riverine landscapes. Freshwater Biology 47:541–557. channel assessment and monitoring. Apex Geoscience Consultants Ltd. Clague, J.J. (compiler). 1989. Quaternary Journal of the American Water 1220 Government Street geology of the Canadian Cordillera. Resources Association 38(1):1–16. Nelson, BC V1L 3K8 Chapter 1 in Quaternary Geology of Obedkoff, W. 2002. Streamflow in the E-mail: [email protected] Canada and Greenland. R.J. Fulton Kootenay Region. B.C. Ministry of (editor). Geological Survey of Canada, Sustainable Resource Management, Water Information Section, Aquatic References Ottawa, Ont. Geology of Canada, No. 1. Information Branch, Victoria, B.C. Association of Professional Engineers and Faustini, J.M. and J.A. Jones. 2003. O’Connor, M. and R.D. Harr. 1994. Geoscientists of B.C. (APEGBC). 2003. Influence of large woody debris on Bedload transport and large organic Overview of B.C. M.o.F. landslide channel morphology and dynamics in debris in steep mountain streams in hazard and risk case studies document. steep, boulder-rich mountain streams, forested watersheds on the Olympic Participant handouts for APEGBC western Cascades, Oregon. Peninsula, Washington. Final report. Annual Conference, Penticton, B.C., Geomorphology 51:187–205. Timber/Fish/Wildlife Sediment, October 23–25, 2003. Gayton, D.V. 2003. British Columbia Hydrology and Mass Wasting Steering Committee, and State of Washington B.C. Ministry of Forests. 2001. Watershed grasslands: monitoring vegetation Department of Natural Resources. assessment procedure guidebook. change. FORREX-Forest Research 122 p. Second edition. Version 2.1. Forest Extension Partnership, Kamloops, B.C. Practices Branch, Victoria, B.C. Forest FORREX Series 7. Rinne, J.H. 1996. Short-term effects of Practices Code Guidebook. Gomi, T. and R.C. Sidle. 2003. Bed load wildfire on fishes and aquatic macroinvertebrates in the southwestern B.C. Ministry of Forests. 2002. Forest road transport in managed steep-gradient United States. Journal of Fisheries engineering guidebook. Forest Practices headwater streams of southeastern Management 16:653–658. Branch, Victoria, B.C. Forest Practices Alaska. Water Resources Research Code of Guidebook. 39(12):1336. doi: Roberts, R.G. and M. Church. 1986. The sediment budget in severely disturbed B.C. Ministry of Transportation and 10.1029/2003WR002440. watersheds, Queen Charlotte Ranges, Highways. 1996. Natural hazards in Gomi, T., R.C. Sidle, and J.S. Richardson. British Columbia. Canadian Journal of British Columbia. Prepared by B.C. 2002. Understanding processes and Forest Research 16:1092–1106. Ministry of Transportation and downstream linkages of headwater Highways, Geotechnical and Materials systems. Bioscience 52(10):905–916. Schnorbus, M.A. and Y. Alila. 2004. Forest Engineering Branch in association with Hallett, D. and R. Walker. 2000. harvesting impacts on the peak flow VanDine Geological Engineering Ltd. Paleoecology and its application to fire regime in the Columbia Mountains of southeastern British Columbia: an Benda, L., D. Miller, P. Bigelow, and K. and vegetation management in investigation using long-term numerical Andras. 2003. Effects of post-wildfire Kootenay National Park, British modeling. Water Resources Research erosion on channel environments, Boise Columbia. Journal of Paleolimnology 40(5):W05205. River, Idaho. Forest Ecology and 24(4):401–414. doi:10.1029/2003WR002918. Management 178(1–2):105–119. Jakob, M. and P. Jordan. 2001. Design Troendle, C.A., M.S. Wilcox, G.S. Bevenger, Bragg, D.C. 2000. Simulating catastrophic flood estimates in mountain streams: and L.S. Porth. 2001. The Coon Creek and individualistic large woody debris the need for a geomorphic approach. Water Yield Augmentation Project: recruitment for a small riparian system. Canadian Journal of Civil Engineering implementation of timber harvesting Ecology 81(5):1383–1394. 28:425–439. technology to increase streamflow. Brown, G.W. 1980. Forestry and water Jordan, P. 2002. Landslide frequencies and Forest Ecology and Management quality. Oregon State University, terrain attributes in Arrow and 143:179–187. Corvallis, Oreg. 121 p. Kootenay Lakes Forest Districts. In Uunila, L. 2004. Final report, Crawford Buffington, J.M. and D.R. Montgomery. Terrain stability and forest Creek watershed hydrologic 1999. Effects of hydraulic roughness on management in the interior of B.C. P. assessment, Phases 1 and 2. Prepared surface textures of gravel-bed rivers. Jordan and J. Orban (editors). B.C. for B.C. Ministry of Forests, B.C. Timber Water Resources Research Ministry of Forests Technical Report Sales, Nelson, B.C. Summit 35(11):3507–3521. 003. Environmental Consultants Ltd. Bunte, K. and L.H. MacDonald. 1999. Scale Lisle, T.E. 2000. The fate of large sediment (Volumes 1 and 2) 114 p. + appendices considerations and the delectability of inputs in rivers: implications for and maps. sedimentary cumulative watershed watershed and waterway management. Wemple, B.C. and J.A. Jones. 2003. Runoff effects. U.S. National Council for Air AEG News 43(4):99. production on forest roads in a steep and Stream Improvement, January Miller, D., C. Luce, and L. Benda. 2003. mountain catchment. Water Resources 1999. Technical Bulletin No. 766. Time, space and the episodicity of Research 39(8):1220. doi: Canadian Standards Association (CSA). physical disturbance in streams. Forest 10.1029/2002WR001744. 1997. Risk management: guidelines for Ecology and Management Wise, M.P., G.D. Moore, and D.F. VanDine decision-makers, a national standard of 178:121–140. (editors). 2004. Landslide risk case Canada. CSA International, Montgomery, D.R. 2003. Wood in rivers: studies in forest development planning CAN/CSA-850-97, Etobicoke, Ont. interactions with channel morphology and operations. B.C. Ministry of Forests, Carver, M. 2001. Using indicators to assess and processes. Geomorphology 51:1–5. Research Branch, Victoria, B.C. Land hydrologic risk. In Watershed Montgomery, D.R. and J.M. Buffington. Management Handbook No. 56. assessment in the Southern Interior of 1998. Channel processes, classification, Available from: British Columbia. D.D.A. Toews and S. and response. In River ecology and www.for.gov.bc.ca/hfd/pubs/Docs/Lmh Chatwin (editors). B.C. Ministry of management. R. Naiman and R. Bilby 56.htm

20 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 valley floor was created. Other data collected included several Re-creating Meandering cross-sections of the existing ditch, pebble counts at the cross-section locations, a longitudinal profile of the Streams in the Central ditch, and cross-sections and pebble counts upstream from the Oregon Coast Range, USA channelized section. For reference on pre-channelized conditions, we used historical aerial photos (1955) that Barbara Ellis-Sugai and Johan Hogervorst showed the position, sinuosity, and meander geometry of the original Introduction pools? Our project goals included (1) stream channel in the valley above the or many years, the riparian and improving coho salmon rearing project area before channelization. Fstream functions of Bailey and habitat; (2) reconnecting channels to Determining the bankfull, or Karnowsky creeks on the Central floodplains; (3) restoring riparian “design,” flow was the most Oregon Coast Range have been vegetation; and (4) challenging aspect of impaired. The valleys were reducing data collection. We homesteaded in the 1800s; by the sedimentation. We Determining the determined this intended to create parameter via three early to mid-1900s, both streams bankfull, or were channelized into ditches to channels that are different sources of increase the amount of land available “stable”: in other “design,” flow information: (1) 16 years for pasture. This channelization words, they are able to was the most of correlated flow decreased sinuosity, resulting in transport the sediment challenging measurements to rainfall increased stream gradients and water load associated with aspect of data records (from Giese velocities. Both stream channels local deposition and collection. 1996); (2) measured subsequently incised into the easily scour while discharge in the field erodible valley floor. In Bailey Creek, maintaining a during winter flow the stream channel started to consistent channel size events; and (3) meander in the ditch, which and shape (Rosgen 1996). At the comparison of Bailey Creek to other increased bank erosion and sediment same time, we expected lateral nearby gauged watersheds. migration, via bank erosion and point deposition into Mercer Lake. In In our restoration plan, we wanted the bar deposition, over time. Karnowsky Creek, larger tributary new stream to flood frequently during streams with gradients under 5% This article describes the methods and the winter to re-establish seasonal were also channelized and downcut lessons learned in re-creating two wetland characteristics, and to to depths of greater than 3 m. These stream channels on the Central minimize the risk that the new channel conditions led to a loss of aquatic Oregon Coast Range. would readjust through bank erosion. habitat, disconnected floodplains, We considered designing either a lower groundwater tables, and Bailey Creek Restoration wide, shallow stream (a “C” channel increased bank erosion and Project type, width/depth [W/D] ratio > 12, sedimentation. In 1991, the U.S. Forest Service using Rosgen’s [1996] terminology), acquired Enchanted Valley in a large Since 1999, the Siuslaw National or a narrow, deep channel (an “E” land exchange with a timber Forest and partners have restored the channel type, W/D ratio < 12; Rosgen company; thus, the land changed stream channels and valley floor of [1996]). Based on the valley’s low from private to public ownership. Bailey and Karnowsky creeks. Lessons gradient, the geomorphic setting (an Bailey Creek flows through Enchanted learned in Bailey Creek were applied old lake bed), and reference stream Valley into Mercer Lake, near to Karnowsky Creek. Florence, Oregon. In 1995, we began Table 1. Enchanted Valley specifications For both streams, we had to answer the project by gathering data on Basin area 11.4 km² (4.4 mi.2) the questions: What type of stream existing stream conditions. We Valley slope 0.34% should we build that will fit the valley compared Bailey Creek with a similar type? What should the dimensions of coastal stream that had not been Valley length 954 m (3100 ft.) the new channel be, including width, cleared and channelized. A Streambank 60% silt/clay, depth, cross-sectional area, gradient, topographic map at a 0.3 m (1 ft.) substrate 40% sand sinuosity, and depth and length of contour interval of the Bailey Creek Continued on page 22

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 21 Continued from page 21 This method allowed us to been planted in the riparian zone easily adjust the proposed along the new channel. location of the stream. We subsequently designed a Karnowsky Creek Channel pool–riffle morphology for Restoration Project the stream bottom, with Karnowsky Creek, which was acquired the pools occupying the by the U.S. Forest Service in 1992, outside bends of flows into the Siuslaw River estuary meanders. The final between Florence and Mapleton, “string” map was then Oregon. In partnership with the digitized and put into a Siuslaw Watershed Council and the geographical information Siuslaw Soil and Water Conservation system. Stake co-ordinates District, we hired a student intern were calculated and team to develop a whole-watershed surveyed onto the ground. restoration plan during the summer of The new 1692 m (5500 ft.) 2001. This team researched watershed long channel was history, fish and wildlife habitat, and excavated in late summer plant communities, and subsequently 1999 (Figure 1). The drafted a restoration proposal. We outside bends of meanders used this proposal to apply for funds were revegetated with from the Oregon Watershed willow stakes in early Enhancement Board and the National spring 2000, and the new Forest Foundation.

USDA Forest Service channel was connected to The restoration plan for Karnowsky Figure 1. Low-elevation aerial photograph of the new the ditch in October 2000. Bailey Creek channel before connection to the old ditch, Creek was similar to Bailey Creek, September 2000. The abandoned ditch emphasizing the creation of summer downstream of the and winter rearing habitat for coho. cross-sections, we chose an “E” connection was then intermittently One heavily aggraded section of ditch channel. To ensure flooding during plugged with fill material that was that had suitable existing spawning the winter flows, we designed the new originally stockpiled during channel habitat was left to passively recover. In channel’s cross-sectional area to be construction, forming ponds in the contrast with the Bailey Creek channel 30% smaller than the existing ditch unfilled areas. The ponds were located design, we relied less on discharge (Tables 1 and 2). where small tributaries drained off the calculations and measurements, and The design parameters were then side slopes, and connected to the new more on bankfull cross-sections in the translated onto our base map, using a channel. Since then, wood has been existing ditch to determine the piece of string cut to the length of the added to the channel, and native cross-sectional area of the new new channel at the scale of our map. hardwoods, conifers, and shrubs have channel. 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-sections 4–4.5 m 6 m above channelized section) (13–15 ft.) (20 ft.) Bankfull depth No information 1.5–2 m 0.9 m (3 ft.) in riffles (5–7 ft.) 1.5 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) Not applicable 16 m (52 ft.) (average) 12–43 m (40–140 ft.) (range) Width/depth (W/D) ratio No information 2–3 in lower valley; 7 25 above channelized section 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.)

22 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 existing ditch’s estimated discharge and shear stress with that calculated To Siuslaw River for the new channel. We chose a W/D ratio of 9.3, a relatively narrow Segment E tidally influenced channel. The rationale was that the vegetation on the valley floor will support the higher shear stresses found in an “E” channel, and a Segment B narrower, deeper channel would have Segment D less direct solar heating. The new Segment A channel’s dimensions are shown in Table 3. Old ditch left as For Karnowsky Creek, the upper valley spawning channel Segment C is slightly steeper, while the lower, tidally influenced valley has a very low gradient. To fit the valley, we designed the new stream’s gradient to gradually decrease from 0.76% at the top to 0.11% in the tidally influenced zone. Likewise, sinuosity increases down valley, from 1.2 to 2.8. In the Upper Channel tidally influenced zone, where frequent winter flooding occurs, we wanted to create diverse fish habitat Barbara Ellis-Sugai

Figure 2. Map of the Karnowsky Creek valley floor and new channel. Key: black, bold solid As with Bailey Creek, once the slope lines-old ditches; dark blue, bold lines-new channel. and sinuosity were established, the new channel was laid out on the base had come into equilibrium with the stream should be a “C” or “E” map, and surveyed onto the ground bankfull flows, and would more channel (Rosgen 1996), we used the (Figure 2). The survey data for the accurately reflect the new channel’s W/D ratio from a nearby reference channel location and existing ground size requirement than flow equations. stream (9.5), and referred to Rosgen’s elevations were entered into a To cross-check, we calculated the (1996) classification system. Unlike spreadsheet. The expected bank bankfull discharge for the existing the Bailey Creek design, we allowed heights in the new channel, assuming ditch and compared it with the new more variation in the size and shape a constant stream gradient through a channel using a regional equation for of the meanders (Williams 1986). We reach, were then calculated. The small watersheds developed at ran several W/D combinations upstream and downstream locations Oregon State University (Adams et al. through Manning’s equation and for pools and riffles were added. 1986), the regional U.S. Geological shear stress equations to compare the Continued on page 24 Survey equations (Jennings et al. 1993), and Manning’s equation. As Table 3. New Karnowsky Creek channel dimensions with Bailey Creek, we wanted the Channel Width Riffle W/D Cross- Gradient Sinuosity New channel to frequently overtop its segment depth ratio section (%) channel banks. Therefore, the new channel’s area length cross-sectional area was designed to Upper 3.1 m 0.3 m 10 0.93 m² 0.76 1.2 684 m channel (10 ft.) (1 ft.) (10 ft.2) (2223 ft.) be 33% smaller than the existing A 4.3 m 0.46 m 9.3 2.0 m² 0.39 1.9 393 m ditch. 2 (14 ft.) (1.5 ft.) (21 ft. ) (1278 ft.) In our restoration plan, we explicitly B 4.3 m 0.46 m 9.3 2.0 m² 0.28 2.2 546 m 2 defined the desired width/depth (14 ft.) (1.5 ft.) (21 ft. ) (1773 ft.) (W/D) ratio, the slope, and the C 4.3 m 0.46 m 9.3 2.0 m² 0.38 1.6 356 m 2 sinuosity of the new channel. The (14 ft.) (1.5 ft.) (21 ft. ) (1157 ft.) W/D ratio is important because it is a D 4.3 m 0.77 m 5.6 3.3 m² 0.20 1.6 226 m (14 ft.) (2.5 ft.) (35 ft.2) (733 ft.) major control on shear stresses within E 4.3 m 0.77 m 5.6 3.3 m² 0.11 2.8 1356 m the channel. To determine whether (14 ft.) (2.5 ft.) (35 ft.2) (4406 ft.)

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 23 Continued from page 23 strategic locations, and water was habitat to complement rearing habitat diverted into the new channel. We created by the work discussed in this applied the lessons learned in Bailey paper. In 2004, the new upper Creek, where we had left an abrupt channel was connected to the existing vertical wall at the connection channel, and the ditch in the upper between the old ditch and new valley was filled in. channel. The old ditch was 1 m (3 ft.) below the new channel. We Monitoring erroneously assumed that sediment Both streams are being monitored would drop out at this point, as water with permanent cross-sections, slowed to enter the new channel at a low-elevation aerial photographs, J.B. Hogervorst on-the-ground photo points, Figure 3. Construction of the new Karnowsky lower gradient, and cause Creek channel, summer 2002. Stakes in aggradation in the old ditch. spawning surveys, juvenile fish counts, foreground identify the location of a However, a tail cut began to develop and collection of water-quality data. transition from a riffle to a pool. Note the downstream from this point as the In Karnowsky Creek, a network of excavated mounds of soil left on the groundwater wells is being measured floodplain. channel’s longitudinal profile came into equilibrium between the two monthly to track groundwater levels. The new Karnowsky stream channel elevations. In Karnowsky Creek, the Results of Restoration was built in late summer 2002 (Figure new channel was designed to 3). In the lower part of the valley, gradually slope up from the old Bailey Creek where wet soil conditions persist ditch’s bed elevation to the new The new channel increased channel throughout the year, the excavated channel, about a 0.3 m (1 ft.) length by 33% and doubled the pool material was piled in mounds and difference in elevation. A ramp of volume compared with the old ditch. shaped on the valley floor. This large logs was buried at grade in the The stream overflows its banks during method reduced both haul costs and new channel at the connection to winter, and the channel appears to be potential for soil compaction from prevent downcutting. relatively stable, although adjustments dump truck traffic. Mounds also In the fall of 2003, 130 large, whole are occurring. In some places, point offered high points in the floodplain trees were added to the new channel bars and mid-channel bars are being that provided good planting sites for and floodplain by helicopter to deposited, as expected (Figures 4a Sitka spruce and western redcedar. provide current and future cover for and 4b). Since the new Bailey Creek During the first winter after fish-rearing areas. Based on research was connected to the existing channel construction, willow stakes were by Roelof (2002), who completed the in 2000, U.S. Geological Survey river planted in the banks, and trees and planting plan for the project, we tried gauges on the Alsea River to the shrubs in the floodplain. At that time, to approach 10% coverage of the north, and the Siuslaw River to the water was not flowing in the main valley floor with this wood in 3–4 ha south have shown annual peak flows channel, which gave the willows a less than 2% of the valley floor. Work to be slightly below average. The chance to establish. in three steeper side tributaries and gauge record goes back to 1940 on During the second summer (2003), the upper main stem is ongoing, and the Alsea River. No gauges are on ditches were plugged in several may supply additional spawning nearby streams of comparable size. Jeff Schmalenberg J.B. Hogervorst Figure 4a. New Bailey Creek channel at photo point 7, summer 1999. Figure 4b. New Bailey Creek channel at photo point 7, summer 2003.

24 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 Fish numbers, both returning one of the few areas where spawning the spring migration, but frequent spawning adults and rearing juvenile counts increased in 2003. The control floodplain inundation, along with lack coho, have increased. Spawning data indicate that from 2001 to 2003, of funding, labour resources, and surveys over the last 4 years indicate a the number of juvenile coho was research groups prohibits this level of definite increase in numbers, 1.5–2.0 times higher compared with monitoring. averaging over 322 fish per kilometre the two previous years’ samples. At Spawning surveys are also ongoing, (200 fish per mile) compared with an the same time, there was roughly a particularly in a 0.8-km (0.5-mi.) average of 113 fish per kilometre (70 10-fold increase in numbers of section of upper main stem that was fish per mile) annually during the juvenile coho in the project area in reconstructed and connected to water 4-year period before the new channel 2001–2003. For 2003, the control in 2004. In December 2004, coho was built (1996–1999). The increase estimate was 0.5 coho per square salmon were observed spawning at from 2002 to 2003 alone was 130 fish metre while the project area estimate the top of the upper main-stem per kilometre (81 fish per mile). The was 1.1 coho per square metre. channel, just 2 months after that 2003 spawning adults were the first Karnowsky Creek section of the new channel was juveniles reared in the project area to connected to the existing stream Although it is too early to have have returned. The assumption is that channel. juveniles of this year’s class took significant monitoring results, we are advantage of both favourable already seeing abundant coho smolts Summary of Lessons conditions in the new channel and and fry in the new Karnowsky Creek Learned channel. The channel functioned well the ocean to produce the 2003 O Cross-sections of the existing ditch through the first two winters, with spawning numbers. Bailey Creek was are probably more reliable than frequent floodplain inundation. regional flow equations or Willows and other riparian vegetation discharge measurements when are growing well, and point bars are determining the size of the new being deposited on the inside of channel. meander bends in the lower channel. O Creating hummocks in the Little, if any, bank erosion is evident floodplain aids in re-establishing (Figures 5 and 6). The mounds built vegetation in areas infested with into the floodplain of the lower valley reed canary grass, and provides are successful nurseries for young micro-topographic sites. It also conifers and shrubs. saves hauling of the excavated Future monitoring of fish populations

Barbara Ellis-Sugai material. will include summer snorkel counts in Figure 5. The new Karnowsky Creek channel O Grade control and a smooth in the lower valley during high winter flows. pools of the new channel and spot transition from the existing ditch to Conifer seedlings in plastic tubes are planted checks of ponds created from the old the new channel will prevent on the mounds left on the floodplain. ditches that were plugged. We downcutting in the new channel. considered running a smolt trap for Barbara Ellis-Sugai Figure 6. The new Karnowsky Creek channel with large wood added, winter 2003. New conifer seedlings are planted along the bank and protected by plastic tubing. Continued on page 26

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 25 Continued from page 25

O These restoration projects require intensive data-gathering and planning by an interdisciplinary Results of Streamline team of hydrologists, geomorphologists, fisheries Reader Survey 2004 biologists, botanists, and surveyors, and benefit from review by other technical experts. Robin Pike

For further information, contact: HANK YOU to everyone who Streamline is relevant, Barbara Ellis-Sugai Tparticipated in our reader survey scientifically sound, user Forest Hydrologist this past fall. The survey was designed friendly, and easy to access. Siuslaw National Forest to help us assess our performance Most readers prefer the 4077 Research Way and, most importantly, solicit your current format. Corvallis, OR 97333 feedback on areas for improvement Of those completing the survey, 90% Tel: (541) 750-7056 and suggestions for future articles. indicated that they prefer the current E-mail: [email protected] Overall, respondents told us that we mix of short newsletter-style and Johan Hogervorst are on track in presenting objective longer technical articles. Preference South Zone Hydrologist and reliable watershed management for print versus online versions of the Siuslaw National Forest information. Here are the highlights publication was more evenly split 4480 Hwy 101, Bldg G of the survey. among respondents, with 51% Florence, OR 97439 favouring online access, 41% print, Tel: (541) 902-6956 Streamline articles are and 8% both publication formats. We E-mail: [email protected] technically reviewed and those readers polled trust will seriously consider these data if References limited funding in the future does not information presented in allow us to produce print versions. Adams, P.W., A.J. Campbell, R.C. Sidle, R.L. Streamline. Beschta, and H.A. Froelich. 1986. Regarding our publication format, Of those polled, 42% were unaware Estimating stream flows on small most respondents agreed or strongly forested watersheds for culvert and that all articles published in Streamline agreed that articles in Streamline are bridge design in Oregon. Oregon State undergo a technical peer review. As a relevant and applicable (92%), University, Forest Research Laboratory, result, we will better communicate the College of Forestry, Corvallis, Oreg. scientifically sound (82%), readable in measures we use to ensure that Research Bulletin 55. 8 p. style (94%), well laid out (84%), easy Giese, T.P. 1996. Phosphorus export from reliable and sound information is the Clear and Mercer Lake watersheds, extended. Despite this finding, 92% Oregon State University, Corvallis, of those surveyed indicated that they Oreg. M.S. thesis, 136 p. Jack Minard of the Tsolum either have a lot of trust (25%) or a Jennings, M.E., W.O. Thomas Jr., and H.C. River Restoration Society, fair amount of trust (67%) in Riggs. 1993. U.S. Geological Survey Courtenay, B.C., won our Water-Resources Investigations Report Streamline. Only 5% of the 94-4002: nationwide summary of U.S. respondents indicated that they had survey draw-prize—a $75 Geological Survey regional regression little trust in Streamline as an gift certificate to Chapters. equations for estimating magnitude and frequency of floods for ungaged information source. sites. 196 p. Roelof, S. 2002. Of fire and fog: Table 1. Client satisfaction in Streamline’s format representing landscape processes and catalyzing nonlinear transformation in Question: In your opinion, Strongly Agree Neutral Somewhat Strongly a Coast Range riparian zone. M.S. are Streamline articles… agree (%) (%) disagree disagree thesis draft.University of Oregon, (%) (%) (%) Department of Landscape Architecture, 1. Relevant and applicable 26 66 7 1 0 Eugene, Oreg. Rosgen, D. 1996. Applied river 2. Scientifically sound 19 63 17 1 0 morphology. Wildland Hydrology 3. Readable in style 27 67 4 2 0 [consultant], Pagosa Springs, Colo. 363 p. 4. Well laid out 31 53 15 0 1 Williams, G.P. 1986. River meanders and 5. Easy to access 36 56 7 1 0 channel size. Journal of Hydrology 88:147–164. 6. Innovative in content 14 52 30 4 0

26 Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 Table 2. Reader suggestions for future content future issues. We grouped the 105 Proposed article topic / content direction Ranka suggestions into categories (Table 2). More of the same mix 1 Notably, most of these suggestions Watershed/stream restoration 2 support the current direction of Monitoring/assessment 3 Streamline’s expanded watershed More forestry 4 management mandate. The most Fish habitat/fisheries 4 common response was for a Methods and techniques 5 continuation in the diversity of topics Riparian issues and science 5 covered. Water conservation/social issues 5 Stewardship 6 Streamline is providing Water chemistry, quality, and health 6 timely access to Hydrology/Watershed processes 6 information, thus increasing Non-forested watersheds/urban watershed management 6 Biology/Biodiversity 6 knowledge of watershed Broaden scope to include international case studies 6 management research, Standards/BMPs/Policy 6 hydrologic processes and Groundwater 7 the effects of watershed Less forestry 8 management. Wildfire effects 8 Finally, we asked readers if we were Opinion pieces 8 achieving our objectives (Table 3). Current updates from around B.C. 8 a Reader suggestions were grouped into the above categories. The number of suggestions per While these results are qualitative and category was tallied and rank was assigned (1 to 9). Tied rank values represent categories have their limitations, 74% agreed with an equal tally. that we’re giving them timely access to access (92%), and innovative in assisted them in their work activities to information, 93% indicated that content (66%). As a result we are not or decisions. Most commonly, Streamline increases their knowledge planning to change the format. respondents stated they use of current B.C watershed Complete results, including Streamline to increase their awareness management research and expertise, percentage of neutral responses and and (or) background knowledge of and 76% indicated that Streamline those in disagreement, are presented watershed management issues and increases their knowledge of natural in Table 1. science. A few readers said they use hydrologic processes and the effects Streamline articles for teaching and of watershed management. Overall, Streamline articles elevate training. the survey results will greatly assist us readers’ knowledge of in managing Streamline, leveraging watershed management Readers were also asked if they could financial support, and ultimately science and issues. Those identify a favourite article over the last offering a reader-focused and relevant readers polled indicated 2 years. Many suggestions were put publication. that they are happy with the forward, the most common being current mix of topics and “Shade and Stream Temperature” by If you would like to provide further feedback on Streamline or suggest an recommended that P. Teti, followed by “Wildfires and Watershed Effects in the Southern B.C. idea for an article, please contact Streamline continue to Robin Pike at feature diverse topics. Interior” by D. Scott and R. Pike. Respondents were asked to comment Readers also offered feedback on the (250) 387-5887, or by e-mail at on how Streamline has improved or types of articles they want to see in [email protected].

Table 3. Client satisfaction in Streamline’s performance against objectives Question: Strongly Agree Neutral Somewhat Strongly Does Streamline… agree (%) (%) disagree disagree (%) (%) (%) 1) Provide timely access to watershed management 21 53 25 1 0 information? 2) Increase your knowledge of current B.C. watershed 38 55 6 1 0 management research and expertise? 3) Increase your knowledge of natural hydrologic 19 57 23 1 0 processes and the effects of watershed management?

Streamline Watershed Management Bulletin Vol. 8/No. 2 Spring 2005 27 UPDATE Upcoming Events

April 16–20, 2005 May 25–27, 2005 August 16–19, 2005 33rd Annual BCWWA Conference and North American Association of Fisheries Second North American Lake Trout Trade Show. Economists (NAAFE) 3rd Biennial Forum. Symposium. Penticton, BC Vancouver, BC Yellowknife, NWT http://www.bcwwa.org/newsdocs/mailer/ http://www.feru.org/events/naafe.htm http://www.laketroutsymposium2005.ca/ BCWWA-AGM-Exhibition-Brochure-v1.4.pdf August 17–19, 2005 May 31–June 4, 2005 Hydrotechnical Engineering: Cornerstone April 26–27, 2005 2005 Joint International Conference on of a Sustainable Environment. Implications of Climate Change in BC’s Landslide Risk Management and 18th Edmonton, AB Interior Forests. Annual Vancouver Geotechnical Society www.hydrotechnics.ca/hydro2005 Revelstoke, BC Symposium. http://www.cmiae.org/ Vancouver, BC September 13–15, 2005 http://cgs.ca/2005ICLRM/ 10th International Specialist Conference May 4–7, 2005 on Watershed and River Basin Manage- 49th Annual BCWF Convention: Navi- June 14–17, 2005 ment 2005. gating the Turbulent Waters of Conserva- 58th Annual CWRA National Conference, Calgary, AB tion in British Columbia. Reflections on our Future...a New Century http://content.calgary.ca/CCA/City+Hall/ Nanaimo, BC of Water Stewardship. Business+Units/Waterworks/Events/I http://www.bcwf.bc.ca/s=142/ Banff, AB WA+Watershed+Conference+2005.htm bcw1089397688895/ http://www.reflectionsonourfuture.ca/ September 29 to October 1, 2005 May 8–11, 2005 August 8–11, 2005 The Coastal Cutthroat Symposium: Biol- ogy, Status, Management, and Earth System Processes II. Geological Canadian Geophysical Union Annual Sci- Conservation. entific Meeting. Society of America. Fort Worden State Park Banff, AB Calgary, AB. (near Port Townsend, Washington) http://www.ucalgary.ca/%7Ecguconf/ http://www.geosociety.org/meetings/esp2/ http://www.orafs.org/cutthroat.html

Recent Publications

Ecosystems and Management 5(2). http://www.forrex.org/jem/2004/ Available from: vol5/no2/art2.pdf http://www.forrex.org/jem/2004/vol5/ no2/art3.pdf Morford, S., D. Robinson, F. Mazzoni, C. Corbett, and H. Schaiberger. 2005. Gayton, D. 2004. Nature conservation Participatory research in rural in an era of indifference. BC Journal of communities in transition: A case study Ecosystems and Management 5(2). of the Malaspina–Ucluelet Research Available from: Alliance. BC Journal of Ecosystems and http://www.forrex.org/jem/2004/vol5/ Management 5(2). Available from: no2/art1.pdf www.forrex.org/jem/2004/vol5/no2/art5.pdf Krzic, M., H. Page, R.F. Newman, and Prescott, C., L. Blevins, and C. Staley. K. Broersma. 2004. Aspen 2005. Litter decomposition in British regeneration, forage production, and Columbia forests: Controlling factors soil compaction on harvested and and influences of forestry activities. BC grazed boreal aspen stands. BC Journal Journal of Ecosystems and of Ecosystems and Management 5(2). Management 5(2). Available from: Available from: http://www.forrex.org/jem/2004/vol5/ http://www.forrex.org/jem/2004/vol5/ no2/art6.pdf no2/art4.pdf

BC Journal of Ecosystems and Lewis, J.L., S.R.J. Sheppard, and K. New text book Management Volume 5, Number 2 Sutherland. 2004. Computer-based Northcote, T.G. and G.F. Hartman Densmore, N., J. Parminter, and V. visualization of forest management: A (editors). 2004. Fishes and Forestry: Stevens. 2004. Coarse woody debris: primer for resource managers, worldwide watershed interactions Inventory, decay modelling, and communities, and educators. BC and management. Blackwell Science, management implications in three Journal of Ecosystems and London, U.K. 789 p. biogeoclimatic zones. BC Journal of Management 5(2). Available from:

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