Basic Carcass Search Methodologies

1
Basic Carcass Search Methodologies

• Wally Erickson

2
Outline

• Objectives
• Precision, Uncertainty and Statistical Power
• Field Biases
• Summary/Future Research

3
Fatality Monitoring Objectives

• determine whether overall avian and bat fatality
  rates or raptor fatality rates are low, moderate,
  or high relative to other projects
• Determine whether predicted mortality is a
  reasonable estimate
• Determine whether a wind project has a fatality
  problem

4
Variance in Fatality Estimates

• Turbine to turbine variation
• Temporal variation
• Scavenging rates
• Observer detection

5
PRECISION EXAMPLES
6
Precision examples
3
2.5
2
ci - half width (fatalities/mw/yr)
1.5

• N25 turbines, vary search detection and
  scavenging

1
0.5
0
0
0.5
1
searcher detection

• Mean removal 3d, vary search detection and
  turbine sample size

7
Examples of Precision Estimates
Values in the table are approximate
cvse/mean
8
Approximate search hours
Values under migration and non-migration
represent the interval between searches in
days Assumes 2 hours per search
9
Other Objectives

• Compare fatality rates to exposure or relative
  abundance
• Estimate the influence of weather on fatality
  levels
• Determine effectiveness of mitigation measures
• Determine high risk turbines

10
Raptor Use Raptor Fatality
Uncertainty in the model
Facilities High Winds Diablo Winds Klondike State
line Nine Canyon Foote Creek Rim Vansycle Buffalo
Ridge Combine Hills
11
MEYERSDALE

• Negative correlation fatalities and wind speed

r-0.60
MOUNTAINEER
r-0.30
12
Statistical Power Case Study

• Calculated for 2000 turbines and assumes 2-year
  study
• 90 probability of showing effect of seasonal
  shutdown for all raptors, RTHA
• 45-65 for BUOW, AMKE, GOEA
• Higher with 3rd year of information

Note assumes turbines are Independent
experimental unit


13
Potential Biases in Field Methods

• Some casualties or injured specimens may land or
  move outside the search area
• Solution 1 ignore, if a small percentage
• Solution 2 estimate the percentage

14
(No Transcript)
15
Habitat Mapping

• Define search area
• Delineate habitat within search area

16
(No Transcript)
17
pictures
18
(No Transcript)
19
(No Transcript)
20
String A
Estimatelt10 missed based on this
distribution Caution Flat, mostly songbirds
String B
Estimatelt5 missed based on this
distribution Caution Flat
String C
21
Other Conditions
22
Searcher Efficiency

• Human Observers
• Randomly place bat fatalities (various stages of
  decomposition)
• Distance Sampling
• Dogs Humans
• Limited trials to determine efficiency of
  dogshumans

23
Searcher Efficiency
Searchers unaware of testing - Random dates -
Random turbines - Random /turbine
(0-4) Carcasses in various conditions - Fresh -
Desiccated - Scavenged Distributed randomly -
Throughout study period - Various weather
conditions - Various habitat types
24
Scavenging and Carcass Removal

• Follow fresh carcasses to determine removal rate
• Place additional carcasses in various habitats

25
(No Transcript)
26
Potential Biases

• Experimental carcasses/trials may not represent
  wind turbine casualties
• Possible Problem Feather spots may be more or
  less visible than intact carcasses used in
  trials.
• Possible Problem Placement, type and number of
  carcasses not representative of real mortality

27
Feather Spots
Feather Spots considered as project-related
fatalities
28
Reference or Background Mortality
29
Studies of Reference Mortality

• Buffalo Ridge, MN
• Estimates of fatality rate at plots without
  turbines 1/3 of estimate at turbines
• Note that without turbine, bird use is likely
  higher
• Other examples, Montana site (Harmata, NREL), San
  Gorgonio, Buffalo Mountain

30
Number and Timing of Trial Carcasses

• Factors
• How many to put out at one time?
• Species/surrogates
• Fresh versus frozen
• If variation by season, habitat and space, need
  adequate sample sizes by season and habitat
  (e.g., 20 small, 20 large per cell)

31
Summary/Future Research

• Objectives key to study design
• Currently recommend a mix of intensive searches
  at a sample of turbines with less intensive at
  others if broad objectives
• Research needs
• Model/study effective plot size especially for
  birds
• Fresh versus frozen birds, surrogates versus
  target species

32
(No Transcript)