Estimating Population Size: Mark-Recapture Method
Credits: This lab exercise is based on the following:
([Link]
[Link])
([Link]
k-recapture)
Introduction
One of the goals of population ecologists is to explain patterns of species distribution
and abundance. In today’s lab we will learn some methods for estimating population
size and for determining the distribution of organisms. Measuring Abundance:
Mark-Recapture Mobile animals are usually simpler to define as individuals, but harder
to count, because they tend to move around, mix together, and hide from ecologists.
Quadrats are not a good approach with mobile animals because immigration and
emigration in and out of the study site make it hard to know what area the entire
population occupies. For largemouth bass in a farm pond, you could easily draw a line
around a map of the population, but how would you define the edges of a population of
house sparrows in your community? Although house sparrows tend to be more
concentrated in towns and urban areas, they do not stop and turn back at the city limit
sign. For zoologists, a fuzzy definition of the space occupied by the population often
forces an arbitrary designation of the survey group, such as the "population" of robins
nesting on your campus in the spring. Knowing the number of animals in a designated
study area is interesting, but we must bear in mind that the ecological population is
defined in terms of interactions among organisms of the same species, and not by the
ecologist's convenience. After defining the individual and establishing the limits of the
population you wish to count, your next task is to choose a counting method. Arctic and
prairie habitats lend themselves to accurate survey by aerial reconnaissance. This
approach works poorly in forests, at night, underwater, or in soil habitats. If animals can
be collected or observed in a standard time or collecting effort, you can get an idea of
relative abundance, but not absolute numbers. For example, the number of
grasshoppers collected in 50 swings with an insect net through an old field community
produces data that could be used to compare relative abundance in different fields, but
would not tell you how many grasshoppers were in the population. For estimates of
absolute numbers, mark-recapture methods can be very effective. The first step is to
capture and mark a sample of individuals. Marking methods depend on the species:
birds can be banded with a small aluminum ankle bracelet, snails can be marked with
waterproof paint on their shells, butterflies can have labels taped to their wings, large
mammals can be fitted with collars, fish fins can be notched, and amphibians can have
nontoxic dyes injected under the skin. Marked animals are immediately released as
1
close as possible to the collection site. After giving the animals time to recover and to
mix randomly with the whole population, the ecologist goes out on a second collecting
trip and gathers a second sample of the organisms. The size of the population can then
be estimated from the number of marked individuals recaptured on the second day. The
assumption behind mark-recapture methods is that the proportion of marked individuals
recaptured in the second sample represents the proportion of marked individuals in the
population as a whole. In algebraic terms,
Example:
This method is called the Lincoln-Peterson Index of population size.
2
The Lincoln-Peterson method is fairly simple, and its calculations are straightforward,
but it does depend on several assumptions. Violating the conditions of the
Lincoln-Peterson model can seriously affect the accuracy of your estimate, so it is very
important to bear these assumptions in mind as you interpret your results:
1. Individuals with marks have the same probability of survival as other members of the
population. It is important to choose a marking method that does not harm your animal.
If a predator used your paint marks to locate and capture marked turtles at a higher rate
than other turtles, your number of recaptures would be lower, and the estimate would
therefore be too high.
2. Births and deaths do not occur in significant numbers between the time of release
and the time of recapture. If marked individuals die and are replaced with newborns,
then you will recapture few or no marked individuals, and your estimate will be too high.
This is not a large concern in studies of box turtles, but can significantly affect estimates
for rapidly breeding organisms.
3. Immigration and emigration do not occur in significant numbers between the time of
release and the time of recapture. If marked individuals leave the study area and new
unmarked individuals come in to replace them, you will get fewer recaptures than the
equilibrium population size would lead you to expect. To think about this another way,
the real population covers a much larger area than the habitat you thought you were
studying.
4. Marked individuals mix randomly with the population at large. If your marked turtles
do not move among unmarked turtles, and you recapture them near the place you
released them, then recaptured turtles may be overrepresented in your second sample,
driving down your population estimate.
5. Marked animals are neither easier, nor harder, to capture a second time. If marking
an animal frightens it so that it hides from you a second time, then recaptures will be
underrepresented in a second sample. If animals become tame and are easier to
recapture, then the opposite error is introduced.
6. Marks do not come off of your marked organisms. Invertebrates molt and shed
marks, mammals can wriggle out of their collars, and many things can happen to
obscure your marks. If this happens, recaptures will be undercounted, and your
estimate will be too high.
7. Recapture rates are high enough to support an accurate estimate. The
Lincoln-Peterson calculation tends to overestimate the population size, especially if the
number of recaptures is small.
3
Assumption 7 is often violated, because it is difficult to generate sufficient recaptures in
large populations. To provide a better estimate, ecologists sometimes use multiple
marks and recaptures and the Schnabel index to estimate population size:
where Mi = the total number of previously marked animals at time i, Ci = the number
caught at time i, and Ri = the number of marked animals caught at time i.
As an example, suppose we caught and marked 100 animals in our first sample,
captured 85 animals (15 marked and 70 unmarked) in a second sample, and then
captured 105 animals (25 marked and 80 unmarked) in a third sample. We would then
have the following:
Table: Example of Schnabel index calculation.
Note that Mi at time 1 is zero; this happens because we start this sampling with no
marked animals. For each subsequent time period, Mi is simply the sum of all previous
values in the “New marked” column. From the above, we can estimate population size
as:
Note that had we sampled only two times, our population size estimate would be 567,
so this value is dependent on the number of samples taken.
4
Methods
Overview
In this experiment you’ll estimate frog population size using a mark-recapture study.
Initial set-up involves sampling the frogs, marking individuals with a red tag, and
releasing them back into the pond. You’ll then collect samples where you count how
many marked and unmarked tadpoles you capture, from which you estimate population
size. You will run two variations of this experiment. In part 1, you will calculate
Lincoln-Peterson Index based estimates of population size. In part 2, you will calculate
the Schnabel Index based estimates of population size.
Part 1: Simulation based test of Lincoln-Peterson Index.
Initial set-up
1. Open the Virtual Biology Lab animated frog sampling tool and orient yourself to the
controls. If the website doesn't function properly, try using your browser's incognito
mode (if it has one), and/or clear your browser cache, and/or try using a different
browser. We have tested it on Chrome/Firefox/Safari, Windows/Mac and Android. We
can confirm that it works on all these systems.
On the sampling tool, in the green boxes on the left, select these settings:
● Pond size: medium
● Population size: large
● Net size: medium
2. Click on “Setup” to complete the setup process. Then click “Go” and you’ll see
tadpoles begin to swim. Allow them to mix for a few seconds. Note that there’s no time
limit for the simulation, so there’s no need to rush through steps.
3. Click “Dip Net” (bottom right corner). You should see tadpoles appear in the Sample
Bucket.
4. With that collection of tadpoles, click “Mark” 20 times. Note the count appear below in
the yellow boxes. Then click “Release” to see the red-marked tadpoles go back into the
pond with the others.
5. You’re now ready to begin data collection using the prepared spreadsheet. Use the
spreadsheet provided to record data from your samples.
5
Sampling & Data Collection
Note: Enter data in only the blue boxes. The green boxes have formulae and will
auto populate. Do not enter anything in the green boxes.
Graphs are auto generated as you enter the data. Save these graphs. You will
need to submit the graph.
6. With 20 marked individuals in the pond, and with the medium net still selected, click
on Dip Net but do not mark more individuals. On your spreadsheet, under Round 1
where Number Marked = 20, do the following:
● For “Marked in Bucket(Recaptured, R) ”, record the number from the yellow
box “Marked in Bucket”.
6
● For “Total in Bucket (Second Sample, S) ", record the number from the yellow
box “Total in Bucket”.
● Notice that the population estimate (L-P Index) column should auto-fill. Make
sure you can explain how the Population estimate was made - it may help to
write the equation out on a piece of paper.
7. To consider how multiple samples (or a larger sample) would impact our estimate,
click on Hold. This moves the tadpoles from the Sample Bucket to the Holding Pen.
8. Repeat steps 6 - 7 two more times, then press Empty Pen to release all the tadpoles.
The Total row should auto-fill as you enter data for your three samples. Make sure that
you understand how the numbers and terms of the equation map on to each other in
Sample 2 and 3.
Note: Between marking and sampling, allow for a little time so that the marked
tadpoles are well mixed in the pond.
You’ll repeat this process in three separate rounds where you vary the number of
marked individuals:
● Round 1: Initial set up with 20 marked individuals x 3 samples
● Round 2: Initial set up with 50 marked individuals x 3 samples
● Round 3: Initial set up with 100 marked individuals x 3 samples
Note: Between each round, make sure that you click “Unmark All” to remove all
tags from all the tadpoles.
Part 2: Simulation based test of Schnabel Index.
Initial set-up
1. Open the Virtual Biology Lab animated frog sampling tool and orient yourself to the
controls. If the website doesn't function properly, try using your browser's incognito
mode (if it has one), and/or clear your browser cache, and/or try using a different
browser. We have tested it on Chrome/Firefox/Safari, Windows/Mac and Android. We
can confirm that it works on all these systems.
On the sampling tool, in the green boxes on the left, select these settings:
● Pond size: medium
● Population size: large
● Net size: medium
7
2. Click on “Setup” to complete the setup process. Then click “Go” and you’ll see
tadpoles begin to swim. Allow them to mix for a few seconds. Note that there’s no time
limit for the simulation, so there’s no need to rush through steps.
3. Click “Dip Net” (bottom right corner). You should see tadpoles appear in the Sample
Bucket.
4. With that collection of tadpoles, click “Mark” repeatedly to mark ALL the collected
tadpoles. Note the count appear below in the yellow boxes.
Sampling & Data Collection
Note: Enter data in only the blue boxes. The green boxes have formulae and will
auto populate. Do not enter anything in the green boxes.
8
Graphs are auto generated as you enter the data. Save these graphs. You will
need to submit the graph.
6. On your spreadsheet (Sheet 2: Schnabel Index), enter the data for Number Captured
in this trial Ci and Number of NEW individuals marked in this trial. Note that both will be
the same number for Trial 1.
7. Back to the simulation. Click “Release” to see the red-marked tadpoles go back into
the pond with the others.
8. With marked individuals in the pond, and with the medium net still selected, click on
Dip Net. A sample of tadpoles will appear in the sample bucket. On your spreadsheet,
record the data for Number Captured in this trial Ci and Number of already marked
individuals recaptured in this trial Ri.
9. Now mark all the unmarked individuals in the sample bucket. Enter this data (Number
of NEW individuals marked in this trial) in the spreadsheet. For example, if your sample
bucket had a total of 23 tadpoles, of which 4 were already marked, Your data for that
Trial would be Ci = 23, Ri = 4 and Number of new individuals marked in this trial =
(23-4) = 19.
10. Click “Release” to see the red-marked tadpoles go back into the pond with the
others.
Repeat steps 8-10 till you fill the data sheet (10 Trials).
As you fill up the data, make sure that you understand how the Schnabel index is being
calculated.
Note: Allow for a little time so that the marked tadpoles are well mixed in the
pond.
Submission:
Answer all the questions in this Google form. Upload a single PDF containing (a) filled
data table for part 1 (b) Graph of part 1 (c) filled data table for part 2 and (d) Graph of
part 2.
Link: [Link]