Line transect lecture

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