Modeling Probability of Detection

'Aerial' Surveys

• Extension of Distance Estimation

• Probability of detection is related to more than just distance from observer

• Develop the model first under experimental conditions

• Apply the model to subsequent surveys to correct for missed animals

Sightability Model Development

1. Mark elk (deer, sheep, etc.) groups with radio-collars to establish a known population.

2. Fly aerial survey recording observations (of collared individuals) and conditions (group size, tree and shrub cover, snow cover, observers, type of helicopter, etc.) for each observation

3. Using a second survey (flown immediately after 1st with telemetry equipment to find all individuals), determine which groups were seen and which groups were missed; recording conditions

4. Develop a 'sightability' (i.e., probability of detection) model to relate conditions of each observation to the probability of detection

• Use logistic regression
• Prob. of detection = p_i

 

 

 

 

 

where, X₁, X₂, ..., X_i are the values for the conditions of each observation (e.g., group size, tree and shrub cover, snow cover, etc.)

Applying the Sightability Model

• Correct each group detected based on its probability of detection (visibility)

• Sum all corrected groups in a sample unit for an unbiased estimate of actual number of animals present.

Example: Lochsa Elk

Aerial Survey Program (see lab notes/exercise)

• All calculations easily performed.
• Variety of sightability models.
• see Unsworth, J. W., F. A. Leban, D. J. Leptich, E. O. Garton, and P. Zager. 1994. Aerial survey: user's manual, 2nd edition. Idaho Dep. Fish Game, Boise, Id. 84pp.

Revised: 25 October 2010