Line transect lecture
High seas salmon off BC’s Coast
Vegetation transects (Offwell, UK)
Duck transects along roads (N. Dakota)
Example 1: UK Butterfly monitoring scheme
Bald eagles
Red-tailed hawks
Short-eared owls
Example 2: Raptor Census - Kyle Elliott (2002) and the Vancouver Natural History Society
Q1. Why transects, not always quadrats?
Q2. What are potential biases in method?
Animals (in particular): detection bias
Animals (in particular): detection bias
Example: VNHS Raptor census (Elliott, 2002)
Two general methods (see Krebs)
1. Distance from random point to organism.
2. Distance from randomly selected organism to neighbouring organism.
12
Two general methods (see Krebs)
1. Distance from random point to organism.
r
Area of circle (π r 2) contains one individual
Inverse of: Density = individuals per unit area
nearest
Two general methods (see Krebs)
1. Distance from random point to organism.
r
r
r
All methods: calculate area per individual for each circle, calculate mean area per indiv., invert
= n π sum (r2)
byth-ripley
Two general methods (see Krebs)
1. Distance from random point to organism.
r
r
r
If look at third closest organism, we are calculating area per three organisms, or if divide by three, mean area per organism (n = 3).
= 3n - 1 π sum (r2)
ordered distance
Two general methods (see Krebs)
1. Distance from random point to organism.
2. Distance from randomly selected organism to neighbouring organism.
12
Two general methods (see Krebs)
2. Distance from randomly selected organism to neighbouring organism.
r
Area per two individuals, but two circles: cancels out to same π r 2 formula as before
Two general methods (see Krebs)
2. Distance from randomly selected organism to neighbouring organism.
r
Area per two individuals, but two circles: cancels out to same π r 2 formula as before
= n π sum (r2)
byth-ripley
Two general methods (see Krebs)
2. Distance from randomly selected organism to neighbouring organism.
Problem: how to randomly select first individual?Nearest organism to a random point: BIASED
Never selected Frequently
selected
WAYS TO RESOLVE PROBLEM:
1. Mark all organisms with a number, and then randomly select a few.
BUT if we could count all organisms, we wouldn’t need a census!
WAYS TO RESOLVE PROBLEM:
1. Mark all organisms with a number, and then randomly select a few.
2. Use a random subset of the area (mark organisms in random quadrats).
Byth and Ripley
WAYS TO RESOLVE PROBLEM:
1. Mark all organisms with a number, and then randomly select a few.
2. Use a random subset of the area (mark organisms in random quadrats).
3. Use a random point to locate organisms, but then ignore area between it and organism (biased to emptiness).
T-square
The 2 snipers
Excellent aim, crooked sights
Cross-eyed cat, Straight sights
Spatial pattern
More uniform More aggregated Random