"

43 Recurrance Relations

As we’ll see in the next section, a differential equation looks like this: dPdt=0.03P. What I want to first talk about though are recurrence relations. Let me introduce these with a magic trick.

Magic Trick
Pick a number between 1 and 100, and I’m going to guess it. But not before we mix it up a bit.

  1. Take your number and divide by five, and round to the nearest whole number.
  2. Then add 36 to the result.
  3. Repeat steps (1) and (2) twice more, for a total of three iterations.

Done? I bet you ended with the number 45. Are you amazed?

This trick is based off the recurrence relation ft+1=15ft+36. Think of ft as the previous value, and ft+1 as the new value. How do you get from one to another? Well, ignoring the rounding, you divide by 5 and add 36, and that’s exactly what ft+1=15ft+36 is telling you to do. In order to use such a equation, we need an initial value or f0. In the trick, this was the original number you picked. Let’s create a graph with the initial value f0=100:

As you can see, this recurrence relation quickly converges to ft=45 by the time t=3. That’s why the trick works! If we started somewhere else, the graph looks much the same and it converges to 45 anyway.

However, recurrence relations are useful for more than just magic tricks.

Logging

Let ht be the biomass of a forest in year t. Suppose it expands by 1% each year, but also loses 2000 metric tonnes to logging. What might be a recurrence relation that explains this situation?

Well, since this is a recurrence relation, we want to relate the quantity under consideration, ht to its value the next year, which is ht+1. So it will look something like

ht+1=1.5ht+16

but those aren’t the right values yet — just want to have some idea of where this is going.

The first thing we need to encode is the expansion by 1%. We can take 1%, or 0.01 and multiply by ht like so: 0.01ht. But the old forest is still there (except for the logging, which we’ll worry about in a second), so let’s add ht as well: 0.01ht+ht. If we factor out ht, we get 0.01ht+ht=ht(0.01+1)=ht(1.01)=1.01ht
This is the growth by 1%. What about that logging? Well, that’s not a percent change, so we’ll just subtract the 2000 to represent the loss of biomass. So our final recurrence relation is–>

ht+1=1.01ht2000

What happens to the forest in the long run according to your recurrence relation?

Well, let’s play with it a bit and see what happens. But before we can do that, we need an initial value h0. Let’s guess something. Since we are losing 2000 a year, we’ll need a much bigger number than 2000. Let’s just guess that h0 is 50,000 metric tonnes.
Now we can compute several ht values:
h0=50000h1=1.01(50000)2000=48500h2=1.01(48500)2000=46985h3=1.01(46985)200045500
We can see that the biomass is going down — not a good sign for the forest. We can speed these calculations up quite a bit in excel. If you do that, you can see that the forest will be totally gone in by h29, in less than thirty years. However, that’s not a full answer, since it may depend on how much biomass we start with. Suppose it’s a larger forest with h0=300,000. Then we see
h0=300000h1=1.01(300000)2000=301000h2=1.01(301000)2000=302010h3=1.01(302010)2000303000
And the forest just grows from there.

Let’s see an even more complicated example.

Life Cycle of Cutthroat Trout

Recurrence relations can model the life of plant and animals species as they move from one stage of life to the next. For example, let ft0, ft1, ft2, ft3, ft4, ft5+ be the amount of cutthroat trout (oncorhynchus clarkii) in southwest Montana of age 0, 1, 2, 3, 4, and 5 or more respectively. Here, t is a measured in years starting from some t=0. Then according to https://compadre-db.org/Species/47501, these quantities follow the recurrence relation
ft+10=5.23ft3+18.0ft4+24.55ft5+ft+11=0.277ft0ft+12=0.3405ft1ft+13=0.4675ft2ft+14=0.4675ft3ft+15+=0.4675ft4+0.4675ft5+
Explain what each number in these recurrence relations mean.

My goodness, that’s a complicated mess of symbols. But with a little patience, we can figure it out I think.

Let’s start with the line ft+11=0.277ft0. We know from the problem statement that ft0 are the trout of age 0 at year t. The quantity ft+11 is the amount of one year old trout at year t+1. This equation is relating the number of 0 year olds with the number of 1 year olds a year later. What it is saying is 0.277 times the number of zero year olds gives you the number of one year olds a year later. In other words, this equation is giving a 27.7% survival rate from age zero to age one.
From here, we can now easily decode several other equations. ft+12=0.3405ft1 gives a 34.05% survival rate from age one to age two. The similar we find 46.75% survival rate from age two to age three, and the same rate from age three to age four. The equation ft+15+=0.4675ft4+0.4675ft5+ is a bit more complicated, since trout of age 5+ come from the age 4 trout, but also the age 5+ stay within that category, so there are two ways to get there. Both of these involve a 46.75% survival rate.
Notice these survival rates are pretty low by human standards. However, there is some good news for the species: look at the first equation ft+10=5.23ft3+18.0ft4+24.55ft5+. What do you make of this? That’s right — these are the new baby trout! As you can see, there are a lot of new babies that help balance the low survival rates we noticed before. In particular, each three year old produces roughly 5 new offspring, each four year old produces on average 18 new offspring, and an average 5+ year old produces almost 25 offspring.

Starting with f00=1500, f01=375, f02=125, f03=65, f04=30, and f05+=50, create a graph that shows how the population of the cutthroat trout changes over time. What does the graph show?

I created this graph in Excel (see the file “cutthroat-life-cycle.xlsx” for the formulas and data):

And here are just the last four stages to get a better look at these ones:

As we can see, the population seems to be fairly stable. One thing that stands out to me is how many age zero and age one fish there are compared to other groups.

License

Icon for the Creative Commons Attribution 4.0 International License

Informal Calculus Copyright © by Tyler Seacrest is licensed under a Creative Commons Attribution 4.0 International License, except where otherwise noted.