ODE: Population Models

Ordinary differential equations (ODEs) are not just abstract mathematical constructs; they are powerful predictive tools. For engineers, scientists, and policymakers, translating biological assumptions into precise ODE models allows for the simulation, analysis, and management of dynamic systems, from managing fisheries to conserving endangered species. These models start from simple exponential growth and progress to systems that capture the complex interplay between competing or interdependent species.

1. The Foundation: Malthusian and Logistic Growth

The simplest model of population growth is the Malthusian exponential growth model. It assumes an unlimited environment: resources are infinite, and the population grows at a rate proportional to its current size. This leads to the familiar ODE:

$$\frac{dP}{dt}=kP$$

Here, $P(t)$ is the population at time $t$, and $k>0$ is the intrinsic growth rate constant. The solution is $P(t)=P_0{e}^{kt}$, where $P_0$ is the initial population. This model predicts unbounded growth, which is unrealistic for any sustained period but can be accurate for populations in early-stage colonization or bacterial growth in ideal lab conditions.

To introduce environmental limits, we use the logistic growth model. It modifies the growth rate by adding a term that slows growth as the population approaches a maximum sustainable size, called the carrying capacity, denoted by $K$. The logistic equation is:

$$\frac{dP}{dt}=rP\left(1-\frac{P}{K}\right)$$

The parameter $r>0$ is the maximum potential growth rate. The term $(1-P/K)$ acts as a brake. When $P$ is small, growth is nearly exponential ($dP/dt\approx rP$). As $P$ approaches $K$, the growth rate approaches zero. The solution curve is a sigmoid (S-shaped) function, modeling a smooth approach to the carrying capacity.

2. Models with External Intervention: Harvesting and Thresholds

Engineering applications often involve managing a population, such as harvesting fish or controlling pests. A harvesting model adds a constant or rate-dependent removal term to the logistic model. For a constant harvest rate $h>0$, the equation becomes:

$$\frac{dP}{dt}=rP\left(1-\frac{P}{K}\right)-h$$

This simple change dramatically alters the dynamics. The system can have two, one, or zero equilibrium populations. Analyzing this ODE helps determine a sustainable yield—a harvest rate $h$ that maintains a stable population—and identifies a critical over-harvesting threshold that leads to population collapse.

A related concept is the threshold population, often modeled with an Allee effect. Here, growth is negative at very low populations (e.g., due to difficulty finding mates) and becomes positive only after the population exceeds a certain threshold $T$. A simple form is:

$$\frac{dP}{dt}=rP\left(1-\frac{P}{K}\right)(P-T)$$

In this model, $0<T<K$. If the population falls below $T$, it will decline to extinction ($P=0$), making $T$ a critical threshold for survival. This is crucial for conservation engineering, where small populations require active intervention to boost them above this viability threshold.

3. Systems of Populations: Competition and Predation

Real ecosystems involve interactions. The competitive model describes two species competing for the same limited resources. An extension of the logistic model, the Lotka-Volterra competition equations are:

$$\begin{array}{rl}\frac{dx}{dt}& =r_{1}x\left(1-\frac{x+{\alpha }_{12}y}{K_1}\right)\\ \frac{dy}{dt}& =r_{2}y\left(1-\frac{y+{\alpha }_{21}x}{K_2}\right)\end{array}$$

Here, $x$ and $y$ are the two populations. The parameters ${\alpha }_{12}$ and ${\alpha }_{21}$ are competition coefficients: ${\alpha }_{12}$ measures the inhibitory effect of species $y$ on species $x$, in terms of $x$'s carrying capacity $K_1$. Outcomes include competitive exclusion (one species wins) or stable coexistence, determined by the relationship between these coefficients and the carrying capacities.

The classic predator-prey model (Lotka-Volterra) captures a different interaction: $$\begin{array}{rlr}\frac{dx}{dt}& =\alpha x-\beta xy& \text{(Prey, e.g., rabbits)}\\ \frac{dy}{dt}& =\delta xy-\gamma y& \text{(Predators, e.g., foxes)}\end{array}$$

The prey ($x$) grow exponentially in the predator's absence ($\alpha x$). The interaction term $-\beta xy$ represents prey death due to predation. Predators ($y$) die off exponentially without food ($-\gamma y$), but grow via the interaction term $+\delta xy$, which converts consumed prey into new predators. This system produces neutrally stable cycles, where predator and prey populations oscillate indefinitely. While simplistic, it foundational for understanding cyclical dynamics in ecosystems.

4. Interpreting Equilibria and Stability

For any autonomous ODE model, finding equilibria (or fixed points) is the first step in long-term analysis. An equilibrium is a constant solution where the rate of change is zero. For a single equation $dP/dt=f(P)$, you find equilibria $P^{\ast }$ by solving $f(P^{\ast })=0$.

Finding equilibria is not enough; you must assess their stability. A stable equilibrium attracts nearby solutions, while an unstable one repels them. For a one-dimensional model, linear stability analysis is straightforward:

1. Find the equilibrium $P^{\ast }$.
2. Compute the derivative $f^{\prime }(P)=df/dP$.
3. Evaluate $f^{\prime }(P^{\ast })$.

• If $f^{\prime }(P^{\ast })<0$, the equilibrium is locally asymptotically stable.
• If $f^{\prime }(P^{\ast })>0$, it is unstable.
• If $f^{\prime }(P^{\ast })=0$, the test is inconclusive, and more advanced analysis is needed.

For systems (like the Lotka-Volterra models), you find equilibrium points $(x^{\ast },y^{\ast })$ by solving the simultaneous equations $dx/dt=0$ and $dy/dt=0$. Stability is determined by analyzing the eigenvalues of the Jacobian matrix evaluated at the equilibrium. In an ecological context, stability tells you whether a species mix will persist or if one will go extinct, guiding conservation and management decisions.

Common Pitfalls

1. Confusing Growth Rate Parameters: Misidentifying $k$ in the exponential model with $r$ in the logistic model is common. Remember, $r$ is the maximum potential growth rate, realized only when $P\approx 0$. The effective growth rate in the logistic model is $r(1-P/K)$.

1. Misapplying the Logistic Model: The standard logistic model assumes instantaneous and linear density dependence. It fails for populations with time lags (e.g., gestation periods) or where the carrying capacity itself changes. Using it for such systems without modification leads to inaccurate predictions.

1. Overlooking All Equilibria: When solving $f(P)=0$ for equilibria, it's easy to miss the trivial equilibrium, often $P=0$. For harvesting or threshold models, $P=0$ (extinction) is frequently an equilibrium whose stability must be checked, as it has dire practical implications.

1. Equating Existence with Stability: Finding a non-zero equilibrium for a harvesting model does not guarantee it's stable. You must perform the stability test. An unstable equilibrium is not a sustainable population level; any small perturbation will drive the system away from it, potentially to collapse.

Summary

• Malthusian growth ($dP/dt=kP$) models unbounded exponential increase, serving as a baseline but unrealistic for long-term modeling.
• The logistic growth model ($dP/dt=rP(1-P/K)$) introduces a carrying capacity $K$, producing S-shaped growth that approaches a stable equilibrium.
• Harvesting models add a removal term, crucial for determining sustainable yields and identifying collapse thresholds. Threshold population models incorporate an Allee effect, where populations below a critical level are driven to extinction.
• Competitive and predator-prey models (Lotka-Volterra) use systems of ODEs to model species interactions, predicting outcomes like exclusion, coexistence, or cyclic oscillations.
• The long-term behavior of any model is governed by its equilibria and their stability. Linear stability analysis (using $f^{\prime }(P^{\ast })$ or the Jacobian) is the essential tool for determining whether a population level or ecosystem state will persist.