Streamkeepers’ Physical Habitat Sampling Plan:  What Now?

 

Questions We Need to Answer SOON

 

Briefly, there is reason to think that Streamkeepers should perhaps discontinue its current set of physical-habitat monitoring (P-hab) parameters, and replace it with another set.  However, we haven’t had enough time to hold this discussion with our technical advisors in time to decide exactly what new set (if any) to put in place this summer.  So for now, our questions are more limited:

 

1.       Should we continue measuring our current set of P-hab parameters this summer, even if there’s a good chance we’ll be changing sets next year?  (Or should we measure them at new sites where haven’t measured P-hab before?)

 

2.       Should we explore alternative P-hab monitoring systems (between now and next summer), and if so, which ones?

 

3.       If we don’t monitor P-hab this summer, should we add an additional water-chem/flow monitoring session in July to the ones we already do in August and in mid-Sept—mid-Oct?  (The rationale would be that the volunteers would have less work to do, and the time saved by not monitoring P-hab could be spent doing additional water quality monitoring in July, which in some cases might be a critical water-quality month.)

 

We need answers to both questions by mid-June.

 

Genesis of Our Current P-hab Suite

 

Streamkeepers was created primarily to engage volunteers in gathering useful data about Clallam County’s streams.  We were spawned out of a predecessor program, the Eight Streams Project of WSU Cooperative Extension of Clallam County, which had created volunteer stream teams that were gathering data, but the data was not of a quality to be useful for most purposes.

 

When Streamkeepers inherited this program in 1999, we set about revising the monitoring program.  Our technical advisors (mostly local agency folks who were presumably the primary end-users of the data) wanted us to measure the chemical, biological, and physical health of streams at targeted sites, just as we had under the former project, only with more accurate and precise measurements.  For water-chemistry and flow measurements, this was easily accomplished by switching from educational test kits to electronic meters.  For biological monitoring, we adopted the genus-level ten-metric Benthic Index of Biological Integrity for the Puget Sound Lowlands, which is backed by extensive publications and has wide acceptance in scientific and regulatory circles.

 

P-hab monitoring was a more difficult matter.  We needed a program that would enable valid evaluation of the physical health of specific reaches of Clallam County’s streams.  There was a great variety of P-hab suites to choose from, but none seemed appropriate, for the following reasons:

 

1.       Many P-hab assessment systems involved numerous judgment-calls on the part of the evaluator, and even if adequate precision were possible with trained, professional technicians (a claim that has been disputed), that same precision did not seem possible with our volunteers.  (Example:  Barbour et al., Rapid Bioassessment Protocols, USEPA 1999; the 1989 Plafkin et al. version; and Hankin and Reeves’ visual habitat area estimation method.)

 

2.       Other systems were more quantitative and involved few or no judgment calls, but these were complex enough that they seemed beyond the scope of what volunteers could be expected to do, given the time-requirements for training and execution.  (Example:  Washington State’s TFW Ambient Monitoring Program.)

 

3.       Programs designed for volunteers yielded data lacking adequate precision, accuracy, and meaningfulness.  (Example:  Adopt-A-Stream Foundation’s Streamkeeper’s Field Guide.)

 

We decided to base our P-hab monitoring program on work done at the University of Washington[1] specifically addressing the question of which parameters offered the best combination of ease of measurement, precision, accuracy, and meaningfulness.  In consultation with Derek Booth, we decided to measure the following parameters:

 

 

It should be noted that at the recommendation of our technical advisors, our reaches were picked not randomly, but targeted in consideration of a number of factors, including intuitive representativeness of the suite of sites for the stream as a whole and of the particular site for its location in the watershed; adjacent land-use; access; and distance from channel alterations.  So our data can only properly be interpreted to represent their particular collection sites.

 

Draft Physical Habitat Index (PHI)

 

After gathering five years’ worth of data, we drafted a 7-metric PHI, using the following parameters calculated using Streamkeeper data (to see our explanatory paper or scores, see http://www.clallam.net/streamkeepers/html/physical_habitat_index.htm), and comparing values for these parameters to established reference standards for healthy streams:

• LWD
• Pools
• Summer canopy closure
• Winter canopy closure
• Percent fines
• Bed substrate stability (a metric developed by the EPA’s EMAP program)
• Aggradation/degradation

 

We found our draft PHI scores to correlate with both B-IBI ratings and biologists’ intuitive P-hab ratings of sites, but the correlations were only strong at the extremes—cases of extremely good or bad habitat.  In the middle, the correlations were weak.  Some of the non-correlation relates to reality:  P-hab influences but does not in itself determine biological integrity.  Some of the correlation could be improved if our PHI were calibrated using standard statistical techniques.  And some further explanatory power might be gained by adding additional metrics for which we have data:  pebble embeddedness, invasive aquatic weeds, bank stability class, and cross-section instability.  But these tasks would require considerable time, which we currently don’t have.  (We do have some volunteers with sufficient scientific background to take on these tasks, and one of them might take on this task if we “pushed it.”)  It could be argued that even if we decide to abandon our current suite of P-hab protocols, we should complete the development of this PHI, in the attempt to synthesize five years’ worth of data.

 

Weaknesses of Our Current P-hab Monitoring Program

 

There are a number of weaknesses in our current suite of P-hab protocols, which might be grounds for abandoning or altering it.  The first weakness is our suite’s uniqueness.  After Booth published his 1999 suggestions, we assumed they would be followed by a spate of graduate-level research exploring Booth’s parameters and their usefulness in P-hab assessment, culminating in the development of a PHI.  That has turned out only partially to be the case, and though our PHI was significantly aided by Booth’s students’ work, none of them attempted to devise a PHI using Booth’s original set of parameters.  Since we aren’t using someone else’s standard technique, we don’t have other scores to compare ours to, and we would find greater acceptance of our work if we used a technique that had been developed, vetted, and published by some prominent arm of government or academia.  For example, the Washington State Comprehensive Monitoring Strategy for Watershed Health and Salmon Recovery (Dec. 2002) calls for monitoring of freshwater habitat status and trends using methods similar to the Environmental Monitoring and Assessment Program (EMAP) of the U.S. Environmental Protection Agency, which has been extensively applied in Pacific Northwest streams.

 

Our second great weakness is the length of our monitoring reaches.  When we began in 1999, we inherited reaches that had already been established by our predecessor program.  These were 100’ long.  Reaches established later tended to be 200’ or more, but even that length is considerably shorter than that recommended by scientists for a representative reach length—for example, EMAP recommends a minimum sample reach length of 150 m. or 40 times the baseflow wetted width, whichever is longer. 

 

One solution would be to just lengthen the reaches, but that would involve some complications:

1. There might not be enough uniform habitat or access in all of our reaches to lengthen them all coherently, and the process would require considerable staff time.
2. Several of our parameters are currently only measured at a single point or transect within a reach:  canopy closure, pebble count, and cross-section.  Protocols would need to be rewritten to gather samples more systematically along the reach.  At that point, they would begin to resemble other programs’ protocols, such as EPA’s EMAP, and it would be wise to consider whether to simply switch to one of those other programs.
3. Measuring these parameters along a significantly-lengthened reach would be time-consuming enough that we’d probably need to create special monitoring teams who would perform physical-habitat monitoring exclusively.  That would provide further incentive to consider a more drastic change in monitoring protocols, since a special team would need to be trained anyway.

 

Further weaknesses in our P-hab protocols are described in the table below:

 

Parameter/Protocol	Problems Noted	Proposed Changes
Pebble count:  Surface sediment sampled annually in half-phi size classes	Measures surface armoring only, not substrate, which is critical for spawning/rearing; and cannot accurately measure fines < 8 mm	Create special team to perform core-sampling in selected cases; also, sample pebbles at regular intervals along a longer reach, as in EMAP
Large woody debris:  Tallied by minimum-size threshold and channel intrusion zone	LWD volume data is needed to accurately calculate Bed Substrate Stability for our PHI.  (We currently estimate this factor.)	Measure LWD volume within the bankfull channel (as in EMAP)
Pools:  Tallied along with residual depth, by minimum-depth threshold according to stream size	Residual-pool volume data is needed to accurately calculate Relative Bed Stability for our PHI.  (We currently estimate this factor.)	Eliminate pools tally and add a thalweg-profile component in order to calculate residual-pool area (as in EMAP)
Cross section:  Measured annually at one spot per site, between permanent monuments	Micro-location may impact the stability or instability of the particular spot; monuments can go missing	Eliminate permanent monuments and measure multiple cross sections at regular intervals along the monitoring reach (as in EMAP)
Gradient:  Measured annually along a 50-100’ line within the monitoring reach	·         The site selected may poorly reflect overall gradient within the larger reach ·         Basic measurement problems	·         Measure gradient (& thalweg profile) along entire length of longer reaches (as in EMAP) ·         Use surveying techniques
Canopy closure:  Measured biannually at a single point within the reach	Single point may bias the data	Measure canopy closure at regular intervals along the monitoring reach (as in EMAP)
Conifer stem count:  All conifer stems of any size are counted within site length & width	Sites are too short for adequate reach characterization (see above).  Stems aren’t measured, so resulting riparian habitat info is coarse & this count isn’t used in our draft PHI. Also, this is the only current parameter that requires a reach map, a protocol with its own problems.	Discontinue; implement a random-plot procedure assessing species & size by category (as in EMAP); or implement remote-sensing analysis

 

 

Broader Questions to Consider

 

In order to decide on what suite of P-hab parameters to measure, we need to pose the following questions:

 

1.      Should Streamkeepers do P-hab monitoring at all?  If so, why?

The stated goals of our monitoring include status, trends, red-flags, problem investigation, and restoration planning and effectiveness monitoring.  Are all of those goals still valid, and should P-hab monitoring be conducted to help support those goals?  Do the goals need to be prioritized?

 

2.      Where and when should we monitor P-hab, and with what parameters?

These questions address overall sampling design.  Our technical advisors have consistently asked us to do targeted rather than randomized sampling, although some have lately begun to argue for the latter.  For example, Bruce Crawford of the Statewide Monitoring Oversight Committee is involved in an effort to devise a statewide WRIA-level randomized monitoring framework that would designate about 20 randomized sites per WRIA, whereby various entities (including volunteer groups such as Streamkeepers) could partner to monitor a full suite of parameters (physical, chemical, and biological) at the framework sites.  It may be that eventually Streamkeepers could perform both targeted and randomized sampling, perhaps with our entire suite of parameters—biological, chemical, and physical—in order to meet the valid goals of both approaches.

 

3.      How should we monitor those parameters?

In order to simplify the myriad of sampling protocols and suites of protocols available, we’ll break the choices down to a few basic options, with pros and cons listed for each approach:

 

A.     Streamkeepers’ current P-hab suite

We could continue using our present suite of P-hab protocols, possibly lengthening reaches and making some/all of the modifications suggested in the table above.

·         Pros:

·         Would require the least retooling of our program.

·         New data could be correlated with old data, thus salvaging a five-year data set.

·         PHI development could continue, possibly resulting in elimination of parameters with redundancy or high signal/noise ratio.

·         Cons:  Listed above under “Weaknesses.”  Also, PHI calibration could lead up a blind alley if our suite of parameters measured turns out to be inadequate.

 

B.     Refined University of Washington P-hab suite

Some of Booth’s students at UW continued working along the lines of his 1999-2000 suggestions.  Sossa and Booth (2004) produced a Physical In-Stream Condition Index (PSCI) that ignores the riparian area and includes just 4 parameters:  bank stability and sediment cementation (categorized qualitatively), and numbers of LWD and pools.  This index provides a coarse discrimination of “low,” “intermediate,” or “high” quality physical in-stream condition.  McBride and Booth (2003) produced an index that included bank stability, cementation, and LWD, plus embeddedness and channel complexity (categorized qualitatively), and a comparison of channel cross-section with watershed size.  Each of the 6 metrics uses a 4-point scale.  This index was found to correlate well with measures of biological integrity and landscape cover.

·         Pros:

·         These are truly rapid assessment protocols, designed for assessment at a broad geographic scale.

·         Multimetric indexes have been developed, at least in draft form.

·         Cons:

·         Their use requires a good deal of subjective assessment, which would run into precision problems with our large volunteer group.

·         Assessments are at a fairly crude level, and trend analysis would not be very discriminating.

·         There is no single set of protocols broadly accepted by a variety of agencies.

 

C.     Nationally- or regionally-developed P-hab suite, such as EMAP

Bruce Crawford, the statewide Monitoring Oversight Committee coordinator, is working with state agencies on modifications to EMAP to improve efficiency and precision, most notably the use of remote-sensing images to do riparian analysis.  We could coordinate with Bruce’s effort, get trained, and then train a sub-group of our volunteers in EMAP protocols.  Phil Kaufman of EPA says that we could train volunteers in 2 days, and that a team of 2 can monitor a reach in about 3.5 hours.  If we could train 8-10 volunteers willing to work 8 days apiece over the summer and fall, we could assess 30-50 reaches per year.  These could be at Streamkeepers’ already-targeted sites, new randomized sites, or a combination of both.

·         Pros:

·          Data would be consistent with federal and state databases, and therefore more likely to be used at those levels.

·          Qualitative assessments are minimized, and a subset of our volunteers could be trained to gather high-quality data.

·          Protocols and data sheets are well developed, and Oregon DEQ has programmed data-entry into hand-held computers to incorporate data-validation and reduce data-entry time and error.

·          EPA is working on calibrated indexes of P-hab quality based on EMAP data.

·         Cons:

·         Local agencies have little experience with EMAP.

·         Our five-year data set might end up going for naught.

·         Implementing EMAP would be a major initiative requiring a good deal of staff time.  We would hopefully be able to get trained and do pilot studies with our volunteers in 2005, and then begin monitoring in 2006.

 

D.     Locally-developed P-hab suite

Local natural-resource professionals have developed their own P-hab monitoring systems.  For example, the watershed assessment done on Siebert Creek in 2004 used a modified TFW protocol developed by Derek Booth and Chris May, combining qualitative and quantitative approaches, and including the following parameters:

·         Categorization of habitat units (pool, riffle, cascade, glide)

·         Qualitative assessment of dominant spawning gravel particle size, embeddedness, and bank stability

·         LWD count, key pieces, and sizes

·         Pool depths plus qualitative assessment

·         Fish passage blockage assessment

·         Photo documentation

·         Riparian assessment using aerial photos

·         Other qualitative observations relevant to habitat quality

When assessing restoration project effectiveness, local project managers will often add stream-profile surveys using standard surveying techniques.

·         Pros:

·         Many local professionals are comfortable with this approach.

·         Cons:

·         A good deal of qualitative assessment is required, and precision would be a problem.

·         This assessment system is not standardized and accepted across a broad variety of agencies, so data comparability and communication may be a problem.

·         No attempt has been made to develop a multimetric index using such a system.

·         Formal surveying would be difficult and time-consuming for our volunteers.

 

Here’s our opinion; what’s yours?

 

Given our research to this point, we’re leaning toward EMAP as modified by Washington’s Statewide Monitoring Oversight Committee.  This approach is robust, within the ability of our volunteers to perform well, broadly accepted and applied, supported by the EPA and DOE in terms of data management and interpretation, and most likely to become “standard currency” in years to come.  Any major transition for a volunteer program such as ours requires a vast outlay of effort for both staff and volunteers, so we don’t take these decisions lightly.  Transition to EMAP or any other new system would not be easy, but continuing our current protocols out of inertia is not a good option.  If we are going to make a transition, it makes sense to switch to something that promises to be useful for the most purposes for the longest time, and EMAP seems to be that choice.  But we want to hear the opinions of our technical advisors and volunteers before making a decision as involved as this one.  So please let us know your thoughts!

Thanks, Ed and Hannah