Population Ecology: Population Dynamics

Global human population

United Nations
projections (2004)
(red, orange, green)

U. S. Census Bureau
modern (blue)
& historical (black)
estimates

Image from Wikimedia Commons

 Population Dynamics
The demographic processes that can change population size:
Birth, Immigration, Death, Emigration
B. I. D. E. (numbers of individuals in each category)

For an open population, observed at discrete time steps:

Nt+1 = Nt + B + I – D – E

For a closed population, observed through continuous time:

dN
= (b-d)N
dt
dN
= rN
dt
(b-d) can be considered a proxy for average per capita fitness
 Population Dynamics
5 main categories of population growth trajectories:

Exponential growth
Logistic growth
Population fluctuations
Regular population cycles
Chaos
 Population Dynamics
Deterministic logistic growth

Invariant density-dependent
vital rates

dN N
= rN 1 –
r dt K

Stable equilibrium
carrying capacity

Cain, Bowman & Hacker (2014), Fig. 11.5

 Population Dynamics
Deterministic vs. stochastic logistic growth

Invariant density-dependent “Fuzzy” density-dependent

vital rates vital rates

r ri

Stable equilibrium Fluctuating abundance within a

carrying capacity range of values for carrying
capacity
Cain, Bowman & Hacker (2014), Fig. 11.5
 Population Dynamics
Time lags can cause
delayed density dependence,
which can result in population cycles
If r is small,
logistic
Instead of growth tracking current
population size (as in logistic), growth
tracks density at  units back in time
If r is intermediate,
damped oscillations
dN N(t-)
= rN 1 –
dt K

If r is large,
stable limit cycle

Cain, Bowman & Hacker (2014), Fig. 11.10

 Population Dynamics
Time lags can cause delayed density dependence,
which can result in population cycles or chaos

Sir Robert May, Baron of Oxford

Photo from http://www.topbritishinnovations.org/PastInnovations/BiologicalChaos.aspx

 Population Dynamics
Population cycles & chaos

(scaled to max. size attainable)

Population size

Per capita rate of increase

 Variation in r and population growth

Is the long-term expected per capita growth rate (r) of a

population simply an average across years?

Consider this hypothetical example:

rgood = 0.5; rbad = -0.5

If the numbers of good & bad years are equal, is the following true?
rexpected = [rgood + rbad] / 2

At t0, N0=100
t1 is a bad year, so N1 = N0 + (rbad* N0) = 50
t2 is a good year, so N2 = N1 + (rgood*N1) = 75

The expected long-term r is clearly not 0 (the arithmetic mean of rgood & rbad)!
 Variation in  and population growth

Nt+1 = Nt

Nt+1
 =
Nt
A fluctuating
population

Arithmetic mean  = 1.02

Geometric mean  = 1.01

1.21

0.87

1.17

1.02

1.13

Cain, Bowman & Hacker (2014), Analyzing Data 11.1, pg. 258
 Variation in  and population growth

Nt+1 = Nt

Nt+1
 =
Nt
A steadily growing
population
1000 1.02
Arithmetic mean  = 1.02
1020 1.02
Geometric mean  = 1.02
1040 1.02

1061 1.02

1082 1.02

1104 1.02

1126 1.02

1148

Cain, Bowman & Hacker (2014), Analyzing Data 11.1, pg. 258
 Variation in  and population growth

Nt+1 = Nt

Nt+1
 =
Nt
A steadily growing
population
1000 1.01
Arithmetic mean  = 1.01
1010 1.01
Geometric mean  = 1.01
1020 1.01

1030 1.01

1040 1.01 Which mean (arithmetic

or geometric) best
1051 1.01
captures the trajectory of
1061 1.01 the fluctuating population
(the example given in the
1072
textbook)?

Cain, Bowman & Hacker (2014), Analyzing Data 11.1, pg. 258
 Population Size & Extinction Risk

Small populations are especially prone to extinction

from both deterministic and stochastic causes

Deterministic  r < 0

Genetic stochasticity & inbreeding

Demographic stochasticity  individual variability around r

(e.g., variance at any given time)

Environmental stochasticity  temporal fluctuations of r

(e.g., change in mean with time)

Catastrophes
 Population Size & Extinction Risk
Demographic stochasticity

Each student is a sexually reproducing, hermaphroditic,

out-crossing annual plant. Arrange the plants into small
sub-populations (2-3 plants/pop.).

In the first growing season (generation), each plant mates

(if there is at least 1 other individual in the population)
and produces 2 offspring.

Offspring have a 50% chance of surviving to the next season.

flip a coin for each offspring; “head” = lives, “tail” = dies.

Note that average r = 0; each parent adds 2 births to the population and on
average subtracts 2 deaths [self & 1 offspring – since 50% of offspring live and
50% die] prior to the next generation.
Population Size & Extinction Risk
Environmental stochasticity

How could the previous exercise be modified

to illustrate environmental stochasticity?
 Population Size & Extinction Risk
Natural catastrophes

What are the likely consequences to populations of sizes:

10; 100; 1000; 1,000,000
if 90% of individuals die in a flood?
 Population Size & Extinction Risk
Allee Effects occur when average per capita fitness
declines as a population becomes smaller

Birth (b)

Rate

Death (d)

K
Density (N)
Zone of
Allee Effects
 Spatially-Structured Populations
Patchy population
(High rates of inter-patch dispersal, i.e., patches are well-connected)
 Spatially-Structured Populations
Mainland-island model
(Unidirectional dispersal from mainland to islands)
 Spatially-Structured Populations
Classic Levins-type metapopulation (collection of populations) model
(Vacant patches are re-colonized from occupied patches
at low to intermediate rates of dispersal )

Assumptions of the basic model:

occupied
1. Infinite number of
identical habitat patches
occupied
unoccupied 2. Patches have identical
colonization probabilities
(spatial arrangement is
irrelevant)
unoccupied
occupied
3. Patches have identical local
extinction (extirpation)
probabilities

occupied 4. A colonized patch reaches K

instantaneously (within-patch
population dynamics are
ignored)
Original metapopulation idea from Levins (1969)
 Spatially-Structured Populations
Classic Levins-type metapopulation (collection of populations) model
(Vacant patches are re-colonized from occupied patches
at low to intermediate rates of dispersal )

dp
occupied = cp(1 - p) - ep
dt

occupied p = proportion of patches

unoccupied occupied

c = patch colonization rate

unoccupied e = patch extinction rate
occupied

Key results:
occupied
Metapopulation persistence
requires (e/c)<1

At equilibrium it is generally the case

Original metapopulation idea from Levins (1969) that not all patches are occupied
 Source-Sink Population Dynamics
Habitats vary in habitat quality;
occupied sink habitats broaden the realized niche

sink

source
source

source
sink

sink

Original source-sink idea from Pulliam (1988)