“by trial and error”
DESCRIPTION
“By Trial and Error”. Why do we say that?. FA in Business?. When Toyotas Fail……. Drive it back to root cause!. Sports?. Medicine?. We’re nominating you for the Darwin Award!. A 48” culvert plugs on an abandoned road. Why do we respond?. Did not pay your electric bill?. - PowerPoint PPT PresentationTRANSCRIPT
““By Trial and By Trial and Error”Error”
Why do we say Why do we say that? that?
FA in Business?FA in Business?
When Toyotas Fail……
Drive it back to root cause!
Sports? Sports?
Medicine?Medicine?
We’re nominating you for the
Darwin Award!
A 48” A 48” culvert culvert
plugs on an plugs on an abandoned abandoned
roadroad
Why do we Why do we respond?respond?
Did not pay your electric bill? Did not pay your electric bill?
A 48” culvert A 48” culvert plugs on an plugs on an abandoned abandoned
roadroad
StreaStream m
divertdiverts s
200 ft 200 ft down down
the the roadroad
……initiates a new initiates a new landslide.landslide.
……mobilized approximately 200,000 mobilized approximately 200,000 ydyd33 aand delivered 50% to Bluff nd delivered 50% to Bluff
Creek...Creek...
…a key watershe
d with Chinook
coho, and
steelhead.
The Cost of No The Cost of No MaintenanceMaintenance
This road This road was built in was built in
2001 for 2001 for approximatapproximat
ely ely $750,000 $750,000
The Future?The Future? Level 1 and 2 roads are
roughly 60% of our total road miles.
Level 1 and 2 roads are
roughly 60% of our total road miles.
Monitoring Road-Watershed Monitoring Road-Watershed PerformancePerformance
Suggested Initiative for Monitoring: Suggested Initiative for Monitoring:
Combine effort to complete DSRs Combine effort to complete DSRs and INFRA to achieve road and INFRA to achieve road performance monitoringperformance monitoring
Roads are a focus of Roads are a focus of watershed monitoringwatershed monitoring
But roads vary greatly in But roads vary greatly in performance performance
Most Most do notdo not fail fail Failures tend to cluster in areas of Failures tend to cluster in areas of
inherent instabilityinherent instability
Why?Why? Failure sites create a useful dataset for Failure sites create a useful dataset for
defining road performance through timedefining road performance through time
Failures define the limits of practice in Failures define the limits of practice in various landscape situationsvarious landscape situations
When experienced road managers retire, When experienced road managers retire, mission-critical knowledge could be mission-critical knowledge could be conserved conserved
Why?Why?
Little added effort for substantial Little added effort for substantial value returnedvalue returned INFRA in place and workingINFRA in place and working DSRs completed DSRs completed Related monitoringRelated monitoring
What you getWhat you get
Ability to determine thresholds of Ability to determine thresholds of performanceperformance
Ability to determine relative risk of Ability to determine relative risk of failurefailure
Quantitative description of risksQuantitative description of risks
Failure Rate vs Distance from Stream
Willamette NF Cumulative Road Failures by Distance to Stream
0
20
40
60
80
100
120
140
160
0.00
972
33.3
45405
88
66.6
81091
76
100.
01677
76
133.
35246
35
166.
68814
94
200.
02383
53
233.
35952
12
266.
69520
71
300.
03089
29
333.
36657
88
366.
70226
47
400.
03795
06
433.
37363
65
466.
70932
24
500.
04500
82
533.
38069
41
566.
71638
600.
05206
59
633.
38775
18
666.
72343
76
700.
05912
35
733.
39480
94
766.
73049
53
800.
06618
12
833.
40186
71
Distance to Stream (m)
Fa
ilu
res
0%
20%
40%
60%
80%
100%
Cumulative %
SNF Mass Wasting Road Failure Rates by Stream Proximity
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
50 150 151
Distance to Stream
Fa
ilu
res
/mil
e o
f R
oa
d
ONF Mass Wasting Road Failure Rates by Distance to Stream
0.00
0.50
1.00
1.50
2.00
2.50
3.00
25 50 75 100 125 150 175 200 225 250 251
Distance from Stream (m)
Fa
ilure
s/m
ile o
f R
oa
d
ONF Mass Wasting Road Failure Rates by Distance to Stream
0.00
0.50
1.00
1.50
2.00
2.50
3.00
25 50 75 100 125 150 175 200 225 250 251
Distance from Stream (m)
Fa
ilure
s/m
ile o
f R
oa
dSNF Mass Wasting Road Failure Rates by
Stream Proximity
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
50 150 151
Distance to Stream
Fai
lure
s/m
ile
of
Ro
ad
Willamette NF Cumulative Road Failures by Distance to Stream
0
20
40
60
80
100
120
140
160
Distance to Stream (m)
Fa
ilu
res
0%
20%
40%
60%
80%
100%
Cumulative %
Mass wasting road failures per mile of road in slope classes(Multiple watersheds)
0.09 0.030.14 0.13
0.32
0.550.68
0.90
1.69
0.00.20.40.60.81.01.21.41.61.82.0
0-10 % 11-20% 21-30% 31-40% 41-50% 51-60% 61-70% 71-80% 81-90%
Slope class by percent
Ro
ad
fa
ilure
s/m
ile o
f ro
ad
Distribution by slope class of 229 mass wasting road failures sites
Slope Class Failures Road by Percent per mile Failures Mileage 0-10 % 0.09 4 45.91 11-20% 0.03 3 98.67 21-30% 0.14 21 147.86 31-40% 0.13 25 187.46 41-50% 0.32 53 165.52 51-60% 0.55 60 109.15 61-70% 0.68 36 53.33 71-80% 0.90 17 18.88 81-90% 1.69 8 4.73
SIUS NF Mass Wasting Road Failure Rates by Slope Position
0.0
0.5
1.0
1.5
2.0
2.5
3.0
30 70 71
Slope Class (%)
Fai
lure
s/m
ile
of
Ro
ad
Failure Rate vs Slope Class
<10%10-50%
>50%
<34
34to66
>66
0.00
0.05
0.10
0.15
0.20
0.25
Fai
lure
s/m
i
Slope Class
Slope Position
WNF: All Failures by Slope Postion and Slope Class
<34
34to66
>66
ONF Mass Wasting Road Failure Rates by Slope
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
5 15 25 35 45 55 65 75 85 95 500
Slope (%)
Fa
ilu
res
/mil
e o
f R
oa
dMass wasting road failures per mile of road in slope classes
(Multiple watersheds)
0.09 0.030.14 0.13
0.32
0.550.68
0.90
1.69
0.00.20.40.60.81.01.21.41.61.82.0
0-10 % 11-20% 21-30% 31-40% 41-50% 51-60% 61-70% 71-80% 81-90%
Slope class by percent
Ro
ad
fa
ilure
s/m
ile o
f ro
ad
Distribution by slope class of 229 mass wasting road failures sites
Slope Class Failures Road by Percent per mile Failures Mileage 0-10 % 0.09 4 45.91 11-20% 0.03 3 98.67 21-30% 0.14 21 147.86 31-40% 0.13 25 187.46 41-50% 0.32 53 165.52 51-60% 0.55 60 109.15 61-70% 0.68 36 53.33 71-80% 0.90 17 18.88 81-90% 1.69 8 4.73
<10%10-50%
>50%
<34
34to66
>66
0.00
0.05
0.10
0.15
0.20
0.25F
ailu
res/
mi
Slope Class
Slope Position
WNF: All Failures by Slope Postion and Slope Class
<34
34to66
>66
ONF Mass Wasting Road Failure Rates by Slope
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
5 15 25 35 45 55 65 75 85 95 500
Slope (%)
Fa
ilu
res
/mil
e o
f R
oa
d
SIUS NF Mass Wasting Road Failure Rates by Slope Steepness
0.0
0.5
1.0
1.5
2.0
2.5
3.0
30 70 71
Slope Class (%)
Fai
lure
s/m
ile
of
Ro
ad
Road failures per mile of road in slope positions (multiple watersheds)
0.14 0.17
0.078
0.38
0.50
0.14
0.51
0.67
0.06
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Upper Middle LowerSlope position
Roa
d fa
ilure
s/m
ile o
f roa
dSurface Erosion (90)
Mass Wasting (229)
All Failures (319)
<10%10-50%
>50%
<34
34to66
>66
0.00
0.05
0.10
0.15
0.20
0.25
Fai
lure
s/m
i
Slope Class
Slope Position
WNF: All Failures by Slope Postion and Slope Class
<34
34to66
>66
SNF Mass Wasting Road Failure Rates by Slope Position
0.00
0.05
0.10
0.15
0.20
0.25
33 67 68
Slope Position (%)
Fai
lure
s/m
ile o
f R
oad
ONF Mass Wasting Road Failure Rates by Slope Position
0.0
0.5
1.0
1.5
2.0
2.5
5 15 25 35 45 55 65 75 85 95
Slope Position (%)
Fai
lure
s/m
ile
of
Ro
ad
Slope Position vs Failure RateSlope Position vs Failure RateONF Mass Wasting Road Failure Rates by Slope
Position
0.0
0.5
1.0
1.5
2.0
2.5
5 15 25 35 45 55 65 75 85 95
Slope Position (%)
Fai
lure
s/m
ile
of
Ro
ad
<10%10-50%
>50%
<34
34to66
>66
0.000
0.050
0.100
0.150
0.200
0.250
Fai
lure
s/m
i
Slope Class
Slope Position
WNF: All Failures by Slope Postion and Slope Class
<34
34to66
>66
Road failures per mile of road in slope positions (multiple watersheds)
0.14 0.17
0.078
0.38
0.50
0.14
0.51
0.67
0.06
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Upper Middle LowerSlope position
Roa
d fa
ilure
s/m
ile o
f roa
d
Surface Erosion (90)
Mass Wasting (229)
All Failures (319)
SNF Mass Wasting Road Failure Rates by Slope Position
0.00
0.05
0.10
0.15
0.20
0.25
33 67 68
Slope Position (%)
Fai
lure
s/m
ile o
f R
oad
Mass wasting road failures per mile of road by bedrock geology(Bluff Creek Watershed)
0.87
0.710.61
0.27 0.240.12
0
0.2
0.4
0.6
0.8
1
1.2
Serpentinite Diorite Phyllite Metasediments Metavolcanics Schist
Bedrock geology units
Roa
d fa
ilure
s/m
ile o
f roa
d
Geology and Failure Rate
ONF Northwest DistrictONF Northwest District
Use Topograpy to Use Topograpy to Define Landscape Types Define Landscape Types
for for Chi-square AnalysisChi-square Analysis
Slope:<=15%, 15-30%, 30-45%,
>45%Slope Position:
<=20%, 20-55%, 55-85%, 85-100%
Distance to Stream:<34m, 34-74m, 74-135m,
<=135m
A Need for More Specific A Need for More Specific Risk InformationRisk Information
Combine 509 known failures with 1008 randomly selected locations.
Use slope, slope position, and stream proximity to estimate relative risk of road-related landslides.
Logistic Regression Modelling:
Logistic Regression Sample Logistic Regression Sample Units Units
Relative Odds of Road-Relative Odds of Road-Related Landslides Related Landslides
Reference Segment:Slope 3%Slpos 8%Distance 213m
Slope 7% Slopos 4% Distance 27mLandslide Odds 19XReference Segment95% CL: 7, 51
Slope 23% Slopos 19% Distance 27mLandslide Odds 39XReference Segment95% CL: 15, 100
Relative Odds Compared to 2% Relative Odds Compared to 2% Slope, 2% Slope Position, 200m to Slope, 2% Slope Position, 200m to
Stream Stream
73
167
127
72
50
53
65
17
Point swarms show problem areas clearly
How you get it…How you get it…
Add DSR points and attributes to Add DSR points and attributes to INFRAINFRA
Some work remaining on attributes Some work remaining on attributes to ensure they are optimalto ensure they are optimal
How you get itHow you get it
1.1. Modify description block in DSR to include:Modify description block in DSR to include:
Failure typeFailure type
CauseCause Volume Volume (quantity classes)(quantity classes)
TotalTotal To streamTo stream To riparian area To riparian area (within 50 m)(within 50 m)
Cause Attributes…Cause Attributes…QuestionsQuestions
Perpetrator or innocent bystanderPerpetrator or innocent bystander Context Context Impact Impact
Sometimes roads Sometimes roads catch and preventcatch and preventsediment deliverysediment delivery
CANL SystemsCANL Systems
ComplexComplex AdaptiveAdaptive Non-linearNon-linear Display emergent behaviors Display emergent behaviors Benefits emerge, destructive Benefits emerge, destructive
tendencies emergetendencies emerge
W W WW W WWhat Went Wrong?What Went Wrong?
CANL SystemsCANL Systems Context develops that controls Context develops that controls
perceptions and actions. perceptions and actions.
Activities that dampen disorder are Activities that dampen disorder are encouragedencouraged
Activities that create disorder are Activities that create disorder are discouraged. discouraged.
Disturbances create disorder and Disturbances create disorder and provide opportunity to rebalance the provide opportunity to rebalance the systemsystem
THATTO
THAT
SO THAT
SO THAT
Reinforce those
activities and
behaviors
Selectively develop
informationResrouces and
information are
distributed
Secure funding
Networks shift to
Activities shift to
Behaviors shift to
Arrangements shift so that
Our biggest challengeOur biggest challenge
If consequences are hidden, If consequences are hidden, slow, debatable, quiet….slow, debatable, quiet….
CANL systems will tend to CANL systems will tend to avoid detecting failures and avoid detecting failures and therefore errors will chronically therefore errors will chronically recur. recur.
S.M.O.O. S.M.O.O. (Same Mistakes, Over and Over)(Same Mistakes, Over and Over)
So…So…
Detect and analyze failuresDetect and analyze failures Our own failings are paydirt!Our own failings are paydirt! Without this:Without this:
S.M.O.O.S.M.O.O.Don’t let the system suppress your right to learn from your failures and teach others