John M. Buffington

13
John M. Buffington Research Geomorphologist US Forest Service Rocky Mountain Research Station Boise, Idaho, USA PNAMP Protocol Comparison Meeting February 1, 2006, Portland, OR

description

PNAMP Protocol Comparison Meeting February 1, 2006, Portland, OR. John M. Buffington. Research Geomorphologist US Forest Service Rocky Mountain Research Station Boise, Idaho, USA. The truth …. Overview of study design & “truth” protocol. Treatments: 3 channel types - PowerPoint PPT Presentation

Transcript of John M. Buffington

Page 1: John M. Buffington

John M. Buffington

Research GeomorphologistUS Forest Service Rocky Mountain Research Station Boise, Idaho, USA

PNAMP Protocol Comparison MeetingFebruary 1, 2006, Portland, OR

Page 2: John M. Buffington

The truth…

Page 3: John M. Buffington

Overview of study design & “truth” protocol

Treatments:3 channel types• Plane-bed (Tinker, Bridge, Camas, Potamus)• Pool-riffle (WF Lick, Crane, Trail, Big)• Step-pool (Whiskey, Myrtle, Indian, Crawfish)

plane-bed pool-riffle step-pool

Page 4: John M. Buffington

Protocol overview

Treatments:4 reaches of each channel type, representing a range of

topographic complexity• e.g., simple, self-formed pool-riffle (WF Lick Ck) vs. complex,

wood-forced pool-riffle (Big Ck)

WF Lick, simple pool-riffle Big, complex pool-riffle

Page 5: John M. Buffington

Protocol overview

Treatments:Reach length• 3 long reaches (80 bankfull channel widths): Bridge (plane-

bed), Crane (pool-riffle), Myrtle (step-pool)• 9 short reaches (40 bankfull channel widths)• bankfull width initially estimated by making measurements

every 20 m over first 200 m of reach

Page 6: John M. Buffington

Protocol overviewMeasurements:• Cross sections (surveyed with total station):

• 1 every 0.5 bankfull widths for first 40 channel widths, and every 1 bankfull width for second 40 channel widths (long reaches only)

• 5 PNAMP cross sections (starting point, and every 10 channel widths)

• cross sections extend out onto floodplain, or confining hillslopes; i.e., well beyond top of bank

• Longitudinal profile of bed along center-line of bankfull width (total station)

• Pebble counts: 10 grains per cross section, evenly spaced across bed• Pools:

• residual depth• type of pool: self-formed vs. forced (wood, bedrock, etc.)• surface area

Page 7: John M. Buffington

Protocol overviewMeasurements:• Wood:

• count of all pieces > 1 m in length & > 10 cm in diameter • geomorphic function:

• forming a pool• assisting in pool scour, but not dominant cause• not causing scour

• location: within bankfull channel, suspended above, or both• inventory divided into 4 subsections between each PNAMP

cross section• no length or volume measurements

• Undercut banks:• depth of undercut if > 10 cm• left & right banks at each cross section• not done at all sites

Page 8: John M. Buffington

9650 9700 9750 9800 9850 9900 9950 10000

9850

9900

9950

10000 Crane Ck(reach length = 80 channel widths)

survey points

pool

bar

contour interval = 10 cm

riffle

Page 9: John M. Buffington

Protocol overview

Derived data:• channel dimensions: width and depth for:

• bankfull flow• any given stage (for protocols that measure wetted values)

• reach slope• number and size of channel units: pools vs. shallows

• shallows could be divided into different types (glide, riffle, rapid, etc.) based on slope and flow depth

• grain size: reach-average, by cross section, or by channel unit• wood: number, geomorphic function, spatial distribution • undercut banks: spatial distribution and depth of undercut

Page 10: John M. Buffington

The “truth”There are many shades of the “truth”, depending on 1. Flow stage (bankfull vs. wetted measurements on a given day)

• Working on bankfull analysis. • Should we also examine solutions for other stages for

protocols that use wetted parameters?2. Method of analysis3. Spatial scale of analysis

• Results of each protocol are likely sensitive to sample domain (spatial extent) and sampling density within that domain, both of which vary between protocols.

• Correct that we will compare protocols “as is” (i.e., mixed-scale analysis)?

• Will reach length and sample size be factors of analysis?• Could attempt to match scales (i.e., clip protocols with

long reaches to extent of “truth” data, and clip “truth” data to extent of protocols with short reaches).

Page 11: John M. Buffington

“truth” is sensitive to methodology

Example: average bankfull width at Trail Ck, derived from “truth” data set

Method 1: average of 5 PNAMP cross sections (8.48 m)

Method 2: average of all 75 cross sections (8.79 m)

Method 3: total bankfull volume of the channel divided by total bankfull surface area, both determined from topographic map constructed from all total station data (7.81 m)

Page 12: John M. Buffington

10000

10000.5

10001

10001.5

10002

10002.5

10003

10003.5

10004

0 5 10 15 20 25 30 35 40

bankfull elevation (m)

bank

full

ele

vatio

n (m

)

distance (channel widths)

Longitudinal profile of bankfull elevation• Noise implies high potential error for small sample size

(e.g., 5 PNAMP cross sections)

Crane Ck

Page 13: John M. Buffington

Partial truth

Were not able to survey all sites• Plane-bed:

• Bridge (long reach), Tinker, Camas, Potamus • Pool-riffle:

• Crane (long reach), WF Lick, Trail, Big• Step-pool:

• Myrtle (long reach), Whiskey, Crawfish, Indian

Black = completedBlue = 14 cross sections + 5 PNAMP cross sectionsRed = only 5 PNAMP cross sections, with 10 grain-size measurements

each