Collision Risk Modelling (CRM)
Avian Collision Risk: from qualitative to quantitative assessments NZ Wind Energy Conference 11 – 13 April 2011
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Dr. Laurence Barea Senior Terrestrial Ecologist Golder Associates Ltd
E Waite
Introduction Planning goal
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
successful consent application.
•
Qualitative v quantitative assessments
•
Example = Collision Risk Modelling
•
Improving the current modelling process
•
Collision modelling can provide framework for stakeholder interaction that facilitates the RMA process
Birds and Wind Farms Birds often major focus when planning wind farm Concern Potential effects; • Displacement • Habitat loss • Collision risk NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Potential effects
Qualitative v Quantitative Globally collision effects are variable – often low, but not always. Little known in NZ context = uncertainty RMA process.
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Lack of knowledge and increasing number of applications = higher level of scrutiny in RMA process. Important to be prepared
Resource Management Act Section 17 – requires avoid, remedy, mitigate. Now - Biodiversity Offsets
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Need quantitative approach to mitigation planning • Too little = mitigation not scaled to effect • Too much = unnecessary cost! • Can effects be mitigated? • Are offsets really needed? • Monitoring targets?
Difficult to answer with a qualitative study
Qualitative v Quantitative Qualitative assessment • • • • •
Descriptive Species list Sometimes semi quantitative Risk = low, moderate, high Difficult to defend in court
Quantitative assessment
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
• • • •
Numerical data Statistical analysis Numerical results accounting for variation More defendable in court
Qualitative v Quantitative Problem - qualitative studies can lead to subjective interpretation. • • • • •
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Lack of objective basis for effects assessment Lack of transparency Open to values based interpretation Too much uncertainty for decision makers Expert opinion can be difficult to defend
Collision risk modelling is one way to quantify potential effects = reduce project risk.
Collision Risk Modelling (CRM) CRM is a Risk analysis tool • Objective use of quantitative data
Estimate of number of collisions • Quantitative assessment of potential effects • Mitigation/Offsets
Explicitly addresses uncertainty/variation
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
• • • • •
Is there an effect? Magnitude? Is Wind Farm appropriate? Framework for study design Data needs
Collision Risk Modelling (CRM) Collision Risk Modelling is quantitative not qualitative. It is not; Species present
Sometimes fly at rotor height
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
E Waite
Qualitative statement about risk – low, medium, high?
Not helpful – too much uncertainty!
Stage 1 – Estimate Collision Probability Turbine specifications
Bird metrics
Collision Probability
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Correct for Avoidance
Stage 2 - Bird Use and Wind Farm Size Flights through study area at rotor height
Proportion at risk
Wind farm shape / size
x Collision Probability NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Mortality Estimate
Band Model – Scottish Natural Heritage CALCULATION OF COLLISION RISK FOR BIRD PASSING THROUGH ROTOR AREA
K: [1D or [3D] (0 or 1 NoBlades MaxChord Pitch (degrees)
1 3 2.3 16
BirdLength Wingspan Flapping (0) or glide
0.5 1.2 0
Bird speed RotorDiam RotationPeriod
Bird aspect ratioo:
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
8 52 1.91
0.42
Calculation of alpha and p(collision) as a function of radius Upw ind: m
m m
r/R
c/C
collide
radius
chord
alpha
length
0.025
0.575
3.74
Dow nw ind: contribution collide
p(collision)
9.61
1.00
contribution
from radius r length
0.00125
p(collision)
8.88
1.00
from radius r
0.00125
0.075
0.575
1.25
3.45
0.68
0.00508
2.72
0.53
0.00400
0.125
0.702
0.75
2.50
0.49
0.00614
1.61
0.32
0.00396
0.175
0.860
0.53
2.20
0.43
0.00757
1.11
0.22
0.00382
m/sec
0.225
0.994
0.42
2.04
0.40
0.00903
0.78
0.15
0.00346
m
0.275
0.947
0.34
1.81
0.36
0.00978
0.61
0.12
0.00330
sec
0.325
0.899
0.29
1.64
0.32
0.01048
0.50
0.10
0.00320
0.375
0.851
0.25
1.51
0.30
0.01111
0.57
0.11
0.00420
0.425
0.804
0.22
1.40
0.27
0.01169
0.62
0.12
0.00516
0.475
0.756
0.20
1.31
0.26
0.01220
0.65
0.13
0.00606
0.525
0.708
0.18
1.23
0.24
0.01266
0.67
0.13
0.00691
0.575
0.660
0.16
1.16
0.23
0.01305
0.68
0.13
0.00769
0.625
0.613
0.15
1.09
0.21
0.01339
0.69
0.13
0.00841
0.675
0.565
0.14
1.03
0.20
0.01367
0.69
0.13
0.00908
0.725
0.517
0.13
0.98
0.19
0.01389
0.68
0.13
0.00969
0.775
0.470
0.12
0.92
0.18
0.01405
0.67
0.13
0.01023
0.825
0.422
0.11
0.87
0.17
0.01414
0.66
0.13
0.01072
0.875
0.374
0.11
0.83
0.16
0.01419
0.65
0.13
0.01115
0.925
0.327
0.10
0.78
0.15
0.01417
0.63
0.12
0.01151
0.975
0.279
0.10
0.74
0.14
0.01409
0.62
0.12
0.01182
Overall p(collision) =
Upw ind
22.2%
Average 17.9%
Dow nw ind
13.6%
Collision Risk Modelling (CRM) When should CRM be considered? Not all wind farms need CRM • Threatened species uses site • Ecologically important species • Migration corridor NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Collision Risk Modelling (CRM) CRM Worldwide UK – Band Model SNH Australia – Developed own model
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
USA – Bald and Golden Eagle Act (2011 USFWS guidelines) recommend CRM - incidental take permits, avoid federal litigation.
S Percival
Collision Risk Modelling (CRM) Taharoa 2008-2009 Shorebird migration – SIPO (28 collisions) 1.settled appeal 2.mitigation
HMR 2010 Shorebirds and bush birds
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
1.agreed conditions 2.mitigation and offsetting
Fuller et al. 2009
Collision Risk Modelling (CRM) Deterministic • Current model does not account for variation – Band model = single values in, single value out
• Problem for stakeholders - delays • Reduced confidence for decision makers = project risk NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
– further analysis required = costly delays – lost opportunity to agree conditions before EC/BOI – mitigation not scaled to potential effect
Deterministic Modelling Process
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Collision Risk Modelling (CRM) Improved Model – Probabilistic – Probability distributions in and out – Variation is quantified and analyzed – Provides for better defense of range of potential effects in court
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Probabilistic Modelling Process
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Creating a Better Model – Monte Carlo Input parameters
Sample variable parameters from probability distributions
Repeat
Dynamically simulate collision risk
Save results NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Assemble results into probability distributions
Avoidance
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.27
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.24
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.31
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.27
Collision Risk Modelling (CRM) Main benefits of CRM • Quantitative method accounting for variation – Stakeholders, Environment Court or BOI – Reduced project risk
• Framework for stakeholder interactions
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
– Magnitude of effect – objectively assessed – Site is appropriate for wind farm – Mitigation quanta objectively calculated & scaled appropriately – No net loss – targets defined – Agreed draft consent conditions
Facilitates a successful consent application
Golder New Zealand Take home messages
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
•
Quantitative studies provide more certainty in the RMA process = reduces project risk
•
Where avian issues are significant consider CRM
•
Use probabilistic models accounting for variation
•
Consider using a statistician
•
Involve stakeholders early and keep interacting
Thank You! Dr. Laurence Barea Senior Ecologist Golder Associates Ltd The Homestead AgResearch Campus East Street PO Box 19479 Hamilton, 3244 TEL +64 7 8592356 MOB + 64 021 956 395 NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Dr. Laurence Barea Senior Terrestrial Ecologist Golder Associates Ltd
E Waite
Introduction Planning goal
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
successful consent application.
•
Qualitative v quantitative assessments
•
Example = Collision Risk Modelling
•
Improving the current modelling process
•
Collision modelling can provide framework for stakeholder interaction that facilitates the RMA process
Birds and Wind Farms Birds often major focus when planning wind farm Concern Potential effects; • Displacement • Habitat loss • Collision risk NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Potential effects
Qualitative v Quantitative Globally collision effects are variable – often low, but not always. Little known in NZ context = uncertainty RMA process.
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Lack of knowledge and increasing number of applications = higher level of scrutiny in RMA process. Important to be prepared
Resource Management Act Section 17 – requires avoid, remedy, mitigate. Now - Biodiversity Offsets
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Need quantitative approach to mitigation planning • Too little = mitigation not scaled to effect • Too much = unnecessary cost! • Can effects be mitigated? • Are offsets really needed? • Monitoring targets?
Difficult to answer with a qualitative study
Qualitative v Quantitative Qualitative assessment • • • • •
Descriptive Species list Sometimes semi quantitative Risk = low, moderate, high Difficult to defend in court
Quantitative assessment
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
• • • •
Numerical data Statistical analysis Numerical results accounting for variation More defendable in court
Qualitative v Quantitative Problem - qualitative studies can lead to subjective interpretation. • • • • •
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Lack of objective basis for effects assessment Lack of transparency Open to values based interpretation Too much uncertainty for decision makers Expert opinion can be difficult to defend
Collision risk modelling is one way to quantify potential effects = reduce project risk.
Collision Risk Modelling (CRM) CRM is a Risk analysis tool • Objective use of quantitative data
Estimate of number of collisions • Quantitative assessment of potential effects • Mitigation/Offsets
Explicitly addresses uncertainty/variation
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
• • • • •
Is there an effect? Magnitude? Is Wind Farm appropriate? Framework for study design Data needs
Collision Risk Modelling (CRM) Collision Risk Modelling is quantitative not qualitative. It is not; Species present
Sometimes fly at rotor height
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
E Waite
Qualitative statement about risk – low, medium, high?
Not helpful – too much uncertainty!
Stage 1 – Estimate Collision Probability Turbine specifications
Bird metrics
Collision Probability
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Correct for Avoidance
Stage 2 - Bird Use and Wind Farm Size Flights through study area at rotor height
Proportion at risk
Wind farm shape / size
x Collision Probability NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Mortality Estimate
Band Model – Scottish Natural Heritage CALCULATION OF COLLISION RISK FOR BIRD PASSING THROUGH ROTOR AREA
K: [1D or [3D] (0 or 1 NoBlades MaxChord Pitch (degrees)
1 3 2.3 16
BirdLength Wingspan Flapping (0) or glide
0.5 1.2 0
Bird speed RotorDiam RotationPeriod
Bird aspect ratioo:
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
8 52 1.91
0.42
Calculation of alpha and p(collision) as a function of radius Upw ind: m
m m
r/R
c/C
collide
radius
chord
alpha
length
0.025
0.575
3.74
Dow nw ind: contribution collide
p(collision)
9.61
1.00
contribution
from radius r length
0.00125
p(collision)
8.88
1.00
from radius r
0.00125
0.075
0.575
1.25
3.45
0.68
0.00508
2.72
0.53
0.00400
0.125
0.702
0.75
2.50
0.49
0.00614
1.61
0.32
0.00396
0.175
0.860
0.53
2.20
0.43
0.00757
1.11
0.22
0.00382
m/sec
0.225
0.994
0.42
2.04
0.40
0.00903
0.78
0.15
0.00346
m
0.275
0.947
0.34
1.81
0.36
0.00978
0.61
0.12
0.00330
sec
0.325
0.899
0.29
1.64
0.32
0.01048
0.50
0.10
0.00320
0.375
0.851
0.25
1.51
0.30
0.01111
0.57
0.11
0.00420
0.425
0.804
0.22
1.40
0.27
0.01169
0.62
0.12
0.00516
0.475
0.756
0.20
1.31
0.26
0.01220
0.65
0.13
0.00606
0.525
0.708
0.18
1.23
0.24
0.01266
0.67
0.13
0.00691
0.575
0.660
0.16
1.16
0.23
0.01305
0.68
0.13
0.00769
0.625
0.613
0.15
1.09
0.21
0.01339
0.69
0.13
0.00841
0.675
0.565
0.14
1.03
0.20
0.01367
0.69
0.13
0.00908
0.725
0.517
0.13
0.98
0.19
0.01389
0.68
0.13
0.00969
0.775
0.470
0.12
0.92
0.18
0.01405
0.67
0.13
0.01023
0.825
0.422
0.11
0.87
0.17
0.01414
0.66
0.13
0.01072
0.875
0.374
0.11
0.83
0.16
0.01419
0.65
0.13
0.01115
0.925
0.327
0.10
0.78
0.15
0.01417
0.63
0.12
0.01151
0.975
0.279
0.10
0.74
0.14
0.01409
0.62
0.12
0.01182
Overall p(collision) =
Upw ind
22.2%
Average 17.9%
Dow nw ind
13.6%
Collision Risk Modelling (CRM) When should CRM be considered? Not all wind farms need CRM • Threatened species uses site • Ecologically important species • Migration corridor NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Collision Risk Modelling (CRM) CRM Worldwide UK – Band Model SNH Australia – Developed own model
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
USA – Bald and Golden Eagle Act (2011 USFWS guidelines) recommend CRM - incidental take permits, avoid federal litigation.
S Percival
Collision Risk Modelling (CRM) Taharoa 2008-2009 Shorebird migration – SIPO (28 collisions) 1.settled appeal 2.mitigation
HMR 2010 Shorebirds and bush birds
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
1.agreed conditions 2.mitigation and offsetting
Fuller et al. 2009
Collision Risk Modelling (CRM) Deterministic • Current model does not account for variation – Band model = single values in, single value out
• Problem for stakeholders - delays • Reduced confidence for decision makers = project risk NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
– further analysis required = costly delays – lost opportunity to agree conditions before EC/BOI – mitigation not scaled to potential effect
Deterministic Modelling Process
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Collision Risk Modelling (CRM) Improved Model – Probabilistic – Probability distributions in and out – Variation is quantified and analyzed – Provides for better defense of range of potential effects in court
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Probabilistic Modelling Process
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Creating a Better Model – Monte Carlo Input parameters
Sample variable parameters from probability distributions
Repeat
Dynamically simulate collision risk
Save results NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Assemble results into probability distributions
Avoidance
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.27
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.24
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.31
Collision Risk Modelling (CRM)
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
Deterministic estimate ~0.27
Collision Risk Modelling (CRM) Main benefits of CRM • Quantitative method accounting for variation – Stakeholders, Environment Court or BOI – Reduced project risk
• Framework for stakeholder interactions
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
– Magnitude of effect – objectively assessed – Site is appropriate for wind farm – Mitigation quanta objectively calculated & scaled appropriately – No net loss – targets defined – Agreed draft consent conditions
Facilitates a successful consent application
Golder New Zealand Take home messages
NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
•
Quantitative studies provide more certainty in the RMA process = reduces project risk
•
Where avian issues are significant consider CRM
•
Use probabilistic models accounting for variation
•
Consider using a statistician
•
Involve stakeholders early and keep interacting
Thank You! Dr. Laurence Barea Senior Ecologist Golder Associates Ltd The Homestead AgResearch Campus East Street PO Box 19479 Hamilton, 3244 TEL +64 7 8592356 MOB + 64 021 956 395 NEW ZEALAND: tel +64 9 486 8068 fax. +64 9 486 8072 www.golder.co.nz
