Indices and Detection Probability p

1
Indices and Detection Probability (p)

• Next Time Predation

2
(No Transcript)
3
Census

• Assume p 1.00 or N (parameter) C (statistic)
• Application largely limited to small sampling
  areas
• US census is congressional mandated
• Statistical sampling is illegal!

4
Index

• Measure of variable related to parameter of
  interest or incomplete enumeration
• Generally used as measure of abundance
• Shortcut in statistical inference
• Easy to measure
• Assumption C (statistic) is a constant
  proportion of N (parameter) or p is constant
• Standardize, standardize, standardize, and
  interpret with caution

5
Index
6
Indices

• Birds?
• Mammals?
• Fish?
• Amphibians and Reptiles?

7
Example BBS
8
Constant Proportion Of Truth?
9
Meeting the Assumption

• Standardize, Standardize, Standardize
• http//www.pwrc.usgs.gov/BBS/
• Measure Related Variables
• E.g., BBS

10
Estimate

• Use statistics and estimators to obtain estimates
• Estimate detection probability and adjust
  statistics accordingly

11
Detection Probability (p)

• Probability that an individual is detected at a
  specific place and time given that an individual
  is present
• Usually between 0 1.0
• Can exceed 1.0 in some sampling scenarios (aerial
  surveys)
• Visibility correction factor (VCF)
• Roll up

12
(No Transcript)
13
(No Transcript)
14
How can we use estimates of detection probability
to obtain estimates of abundance and other
parameters?
How do we estimate detection probability?
15
Several Approaches

• Double Sampling
• Gassaway surveys
• Videography surveys
• Distance Sampling
• Line transect
• Variable circular plot
• Capture-Mark-Recapture

16
Previously - Census w/i plots

• Reasonable number of plots can be sampled
• Complete census in each plot
• Moose in Fairbanks City Limits

17
Double Sampling

• When?
• A large areas can be surveyed inexpensively and
  incompletely
• A subset of area can be censused or used to
  estimate p

• How?
• The extensive (incomplete) count is adjusted for
  detection probability estimated from intensive
  counts to obtain an estimate of abundance for the
  area that is sampled extensively

18
Assumptions

• Detection probability for the area sampled
  intensively is 1.0 or known
• Detection probability estimated from the
  intensive areas can be applied to area only
  sampled extensively

19
Extensive Count (Ce) Gassaway Survey
20
Intensive Count (Ci)
21
How do you estimate p for the entire area that
was sampled extensively?

• p Ce/Ci for the units that were sampled both
  extensively and intensively (CeltCi)
• p may be spatially (habitat) specific

22
How do you estimate abundance for the entire area
that was sampled extensively?
23
Abundance of Sea Otters

• Double Survey
• Coast of California
• Aerial Survey for Extensive Count
• Ground Survey for Intensive Count

24
(No Transcript)
25
(No Transcript)
26
How does this work for Distance Sampling?
27
What do we do with this?