Title: Population Dynamics Part III
1Population Dynamics Part III
2Deficiency of all models shown thus far
- Birth rate was considered to act instantly, based
on the present population size. - For many species there is a considerable time
delay (pregnancy, juveniles reaching maturity
etc.) - A more reasonable model should assume that
present birth rate (at time t) actually depends
on y(t-Td) and not on y(t), where Td is a known
pure time delay.
3Delay may cause an oscillatory behavior
Example (simple linear equation)
Above can be verified by direct substitution
4Linear Delay Model
We shall not dwell on the mathematics Is the
following true?
Well study this case later on in simulations only
5Logistic Delay Model
In this model the regulatory effect depends on
the population at an earlier time
6Logistic Delay Model Normalized
7Logistic Delay Model Simulation Results I
If DrTd lt p/2 there is no limit cycle (i.e. a
steady oscillation). Population y goes to K.
8Logistic Delay Model Simulation Results II
If DrTd gt p/2 there is a limit cycle about yK
with a period T which is in the order of
magnitude of T4Td
9Logistic Delay Model Simulation Results III
Example 1 DrTd 1.6 then Oscillation period
T4.03Td and ymax/ymin2.56
10Logistic Delay Model Simulation Results IV
Example 2 DrTd 2.1 then Oscillation period
T4.54Td and ymax/ymin42.3
11Logistic Delay Model Simulation Results V
Example 3 DrTd 2.5 then Oscillation period
T5.36Td and ymax/ymin2930
12Comparison of Nicholsons (1957) experimental
data for the population of the Australian
sheep-blowfly and a logistic delay model with
DrTd2.1
From Murrays book
- Observed oscillation period was 35-40 days.
- Model was fit using appropriate K,r,Td values.
13Comparison of Nicholsons (1957) experimental
data for the population of the Australian
sheep-blowfly and a logistic delay model with
DrTd2.1
- Model implied Td9 days (using the relationship
T4.54Td. Actual time for larva to mature into
adult is around 11 days. - Oscillation is actually independent of K (the
amount of food available to the insects)
14Can a First-order scalar equation without delay
have a periodic solution?
Here is a general first-order equation
Lets see if a periodic solution y(t)p(t)p(tT)
works
15First-order scalar equation without delay cannot
possibly have a periodic solution
Proof
But the left-hand side is positive. Therefore
periodic solution is impossible. The delay plays
a crucial role in the periodicity of the solution.