IMPACT OF PREY FEAR LEVEL IN A

PREY PREDATOR MODEL

A project report based on literature review

Submitted by -
Rupam ghosh
[2302020]
BARRACKPORE RASTRAGURU SURENDRANATH
COLLEGE

2025
IN

MATHEMATICS
 Abstract: This study investigates the impact of prey fear level on the stability of the prey-predator ecosystem.
Through the analysis of various equilibria's coexistence and stability conditions, we observe that once the
intrinsic growth rate of prey surpasses a critical threshold, the coexistence equilibrium becomes unstable and
undergoes Hopf bifurcation, leading to oscillations in all species. The direction of the Hopf bifurcation are
determined. Additionally, the study shows that bottom-up and top-down effects such as the intrinsic growth rate
of prey, level of fear of prey, and mortality rate of predator population play significant roles in facilitating
transitions between different steady states. Lastly, we perform numerical simulations to validate the analytical
findings. Our findings have important implications for understanding the complex dynamics of prey-predator
ecosystems and highlight the crucial role of prey fear level in shaping these dynamics.

INTRODUCTION : Traditional predator-prey models focus mainly on direct killing by predators as the primary
factor affecting prey populations. However, recent studies suggest that the fear of predators can also significantly
influence prey behavior, physiology, reproduction, and survival. Prey may alter their habitat use, reduce foraging,
or show stress responses even without actual predation, leading to population-level effects. Experimental
evidence, such as work on song sparrows, shows that perceived predation risk alone can reduce reproductive
success.

Despite these findings, most mathematical models do not account for the indirect effects of fear. This paper
introduces a new predator-prey model that incorporates the cost of fear to better understand its impact on
population dynamics. The model is analyzed under both linear and Holling type II functional responses. Results
show that fear can change the stability of equilibria and influence the occurrence and nature of Hopf bifurcations.
Numerical simulations further illustrate the potential role of fear in shaping predator-prey interactions.
THE MATHEMATICAL MODEL : The given problem deals with the dynamics of two populations, namely
prey, predator. These populations are characterized by their respective concentrations, denoted by x(t) and y(t) at
a given time t. The intrinsic growth rate and level of fear prey are represented by r and k respectively. The
maximal prey ingestion rate and conversion rate for growth are denoted by, and α₂( 𝛼₁,≤ α₂). respectively. The
mortality rates of prey and predator biomass are represented by d₁ and d₂ respectively. The grazing process is
assumed to follow the Holing type II functional form with half saturation constants k₁. The intra species
competition between prey population is 𝑟1 .

Based on the aforementioned biological assumptions, the model system is

𝑑𝑥 𝑟𝑥 𝛼1 (1−𝑚)𝑥𝑦
= − − 𝑑1𝑥 (2.1)
𝑑𝑡 1+𝑘𝑦 𝑘1 +(1−𝑚)𝑥

𝑑𝑦 𝛼₂ ( 1−𝑚)𝑥𝑦
= − 𝑑₂ 𝑦 − 𝜇𝑦 2 (2.2)
𝑑𝑡 𝑘1 +(1−𝑚)𝑥

We need the following condition to investigate the above system x(0)≥0, y(0)≥0

Some preliminary results:

Positive invariance Although the system (2.1) is not homogeneous, it has a property that
makes it biologically wellposed. Specifically, if the initial values of X(0) ∈ 𝑅2+ : = {(𝑠, 𝑡) ∈ 𝑅2 : 𝑠 ≥
0, 𝑡 ≥ 0} are chosen such that 𝑋𝑖 = 0 for i=1,2 then the corresponding Jacobian matrix 𝐽𝑖 (X) is
non-negative.

By this property, the system's solution remains non-negative. Therefore, the system exhibits a state
of biological well-posedness.

Boundedness of the system

Let f₁, f₂ be as in system (2.1). Then the functions t → f₁(x(t),y(t)), t → f₂(x(t), y(t)) are bounded on
[0, t₀] for t₀ > 0.

Proof. As the considered biological model deals with dynamics of three populations, the domain of
the functions f₁ and f₂ in (2.1) is ℝ2+ . Clearly, the functions f₁ and f₂ in (2.1) are well-defined and
continuous on ℝ2+ . Also, (x(t), y(t)) ∈ ℝ2+ for all t ≥ 0. Two populations in the considered model are
bounded, that is, t → (x(t), y(t)) is bounded in ℝ² on [0, 𝑡0 ] for 𝑡0 > 0. It is enough to show that the
functions f₁ and f₂ are bounded on any bounded set B ⊂ ℝ2+ . Let B ⊂ ℝ2+ be a bounded set in R².
Let 𝐵 − denotes the closure of B in 𝑅2 . Since B is bounded in ℝ², 𝐵− is bounded in R². Also
𝐵− ⊂ℝ2+ as ℝ2+ is closed set in ℝ². We know that a subset of ℝ² is compact iff it is closed and
 bounded. Therefore𝐵− is compact. Since f₁ and f₂ are continuous and the continuous image of a
compact set is compact, f₁(𝐵 − ) and f₂(𝐵− ) are compact. Hence f₁(𝐵− ) and f₂(𝐵− ) are bounded.

As B⊂ 𝐵− , f₁ and f₂ are bounded on any bounded set B⊂ ℝ2+ Hence the result follows.

Existence and stability of equilibria

Equilibrium points of the systems of equations (2.1) are obtained by solving the following algebraic equations:
f₁(x, y) = 0, f₂(x, y) = 0Let us denote E₀₀(0, 0), E₁(x₁, 0), E* (𝑥∗, y*) be the prey and predator free, predator free,
and coexistence equilibrium respectively .The purpose of this section is to analyze how system (2.1) behaves in
the local vicinity of its equilibrium points .To perform this analysis, we can evaluate the local stability of each
equilibrium point by computing the Jacobian matrix that corresponds to it. By examining the Jacobian matrix at
each equilibrium point, we can gain insights into the stability properties of the system and better understand its
dynamics. The Jacobian matrix at E = (x,y) is given by:

𝑟 𝛼 (1−𝑚)𝑦𝑘1 𝑟𝑘𝑥 𝛼1 (1−𝑚)𝑥

− (𝑘 1 2 − 𝑑1 − (1+𝑘𝑦)2 −
1+𝑘𝑦 1 + (1−𝑚)𝑥) 𝑘1 +(1−𝑚)𝑥
J(𝑥, 𝑦) = [ 𝛼 (1−𝑚)𝑦𝑘 𝛼2 (1−𝑚)𝑥
]
2 1
(𝑘1 +(1−𝑚)𝑥)2
− 𝑑2 − 2µ𝑦
𝑘1 +(1−𝑚)𝑥

(3.1)

Prey and predator free equilibrium

Prey & predator free equilibrium E₀₀(0, 0), is always feasible. As the Jacobian (3.1) evaluated at(0, 0) is

𝑟 − 𝑑1 0
J(0,0) = [ ]
0 − 𝑑2

has the eigen values are on the diagonal :

𝜆1 = 𝑟 − 𝑑1 ; 𝜆2 = −𝑑2 < 0

Stability Analysis

The stability of E₀ depends on the signs of these eigenvalues. We assume all parameters (r, k, r₁,
α₁, m, K₁, d₁, α₂, r, d₂, μ) are positive.

λ₂ = -d₂: Since d₂ > 0, λ₂ is always negative. This means that if the prey population x is zero, the
predator population y will decline to zero .

λ₁ = r - d₁: The sign of this eigenvalue depends on the relationship between the prey's intrinsic
growth rate r and its natural death rate d₁.

In this case, λ₁ = r - d₁ > 0.

 One eigenvalue is positive (λ₁ > 0) and one is negative (λ₂ < 0).

Therefore, E₀ = (0,0) is an unstable saddle point.

Biological Interpretation: If the prey's intrinsic growth rate is greater than its natural death rate, the
prey population can grow from small numbers in the absence of predators. The origin is unstable,
and trajectories will move away from it, at least initially in the prey direction.

 If r > d₁, E₀ is an unstable saddle point (prey can potentially grow).

Predator free equilibrium

At 𝐸1 (𝑥1 , 0), the Jacobian (3.1) evaluated at (𝑥1 , 0) is

α (1−𝑚)𝑥
𝑟 − 𝑑1 (−𝑟𝑘𝑥1 − 𝐾 1+(1−𝑚)𝑥1 )
1 1
J(𝑥1 , 0) = [ α2 (1−𝑟)𝑥1
]
0 − 𝑑2
𝐾1 +(1−𝑚)𝑥1

Eigenvalues:

λ₁ = d₁ - r

λ₂ = α₂(1-m)x₁ / (k₁+(1-m)x₁) - d₂

Stability Analysis
An equilibrium point is locally asymptotically stable if and only if all eigenvalues have negative real parts.

Condition for Stability:

λ₁ < 0:

d₁ - r < 0 => d₁ < r

This is the same condition required for the equilibrium x₁ to be positive. It means the prey's intrinsic growth rate r
must be greater than its natural death rate d₁ for the population to sustain itself.

λ₂ < 0:

α₂(1-m)x₁ / (k₁+(1-m)x₁) - d₂ < 0

α₂(1-m)x₁ / (k₁+(1-m)x₁) < d₂

This condition has a clear biological meaning. The term α₂(1-m)x₁ / (k₁+(1-m)x₁) represents the per-capita growth
rate of the predator population when the prey is at its carrying capacity x₁. The condition states that this growth
rate must be less than the predator's death rate d₂. In other words, even with the maximum available prey (in the
absence of predators), the predator population cannot grow and will die out.
• Co-existence equilibrium
A coexistence equilibrium occurs when both populations exist at a stable positive level. This is the
point

(x∗, y∗) where x∗>0, y∗>0 and both derivatives are zero.
Then we find

1 𝛼 (1−𝑚)𝑥 ∗
𝑦 ∗ = 𝜇 (𝑘 2+(1−𝑚)𝑥 ∗ − 𝑑2 )
1

For a biologically meaningful coexistence (y∗>0), we require the condition:

𝛼2 (1−𝑚)𝑥 ∗
< 𝑑2
𝑘1 +(1−𝑚)𝑥 ∗

Here ,
𝑎2 (1 − 𝑚)𝑥 − 𝑑2 (𝑘1 + (1 − 𝑚)𝑥)
𝑟 𝑎1 (1 − 𝑚) ( )
𝜇(𝑘1 + (1 − 𝑚)𝑥)
− 𝑑1 − =0
𝑎 (1 − 𝑚)𝑥 − 𝑑2 (𝑘1 + (1 − 𝑚)𝑥) 𝑘1 + (1 − 𝑚)𝑥
1+𝑘( 2 )
𝜇(𝑘1 + (1 − 𝑚)𝑥)

𝑥 ∗ is the root of the equation.

𝑎11 𝑎12
Then , J* = [𝑎 𝑎22 ]
21

𝛼1 (1 − 𝑚)2 𝑥 ∗ 𝑦 ∗
𝑎11 =
(𝑘1 + (1 − 𝑚)𝑥 ∗ )2
−𝑟𝑘𝑥 ∗ 𝛼 (1−𝑚)𝑥 ∗
𝑎12 = (1+𝑘𝑦 ∗ )2 − 𝑘 1+(1−𝑚)𝑥 ∗
1

𝛼2 (1 − 𝑚)𝑘1 𝑦 ∗
𝑎21 =
(𝑘1 + (1 − 𝑚)𝑥 ∗ )2

𝑎22 = (𝜇𝑦 ∗) − 2𝜇𝑦 ∗ = - 𝜇𝑦 ∗

Obviously, det(J*) = a11 a22 − a12 a21 > 0 and then the stability of E* is determined by the
α (1−𝑚)2𝑥 ∗ 𝑦 ∗
Trace(𝐽∗ ) = (𝑘1 +(1−𝑚)𝑥 ∗)2 − µ𝑦 ∗
1

So , E* is asymptotically stable iff

Trace(𝐽∗ ) < 0
3.4. Hopf-Bifurcation

Theorem 3. The necessary and sufficient conditions for Hopf bifurcation of the system around 𝐸 ∗
at 𝑟 = 𝑟 𝑐
𝑑
are [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 = 0, [𝑑𝑒𝑡(𝑗 ∗ )]𝑟=𝑟 𝑐 > 0 and 𝑑𝑟 [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 ≠ 0

Proof. The condition [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 = 0 gives ,

𝑟 α (1−𝑚)𝑦 ∗
(1+𝑘𝑦 ∗)
+ 2𝑟1 𝑥 ∗ − (1+α
1
∗ )2 = 0, in which [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 = 0
1 𝑟𝑥

Now, [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 > 0 which is equivalent to the characteristic equation λ2 + [det(𝐽∗ )] 𝑟=𝑟 𝑐 = 0
whose roots are purely imaginary, for 𝑟 = 𝑟 𝑐 , the characteristic can be written as

𝜒2 + 𝜔=0
(3.4.1)

Where 𝜔 = [𝑑𝑒𝑡(𝑗 ∗)]𝑟=𝑟 𝑐 > 0 . Therefore, the above equation has two roots of the form 𝜒₁ =
+𝑖√𝜔 and 𝜒₁ = −𝑖√𝜔. Let at any neighboring point r of 𝑟 𝑐 , we can express the above roots in
general form like

𝑡𝑟(𝑗 ∗ ) 𝑡𝑟(𝑗 ∗)
χ₁,₂ = θ₁(r) ± 𝑖θ₂(r) , where θ₁(r) = 2
and 𝜃2 (𝑟) = √𝑑𝑒𝑡(𝑗 ∗ ) − 2
. Now it is to be verified the
𝑑
transversality condition 𝑑𝑟 (𝑅𝑒(𝜒𝑖 (𝑟))𝑟=𝑟 𝑐 ≠ 0 𝑓𝑜𝑟 j=1,2.

Substituting χ₁ = θ₁(r) + 𝑖θ₂(r) in (3.4.1) and calculate the derivative, we have

2θ₁(r)θ₁'(r) - 2θ₂(r)θ₂'(r) + ω' = 0
2θ₂(r)θ₁'(r) + 2θ₁(r)θ₂'(r) = 0 (3.4.2)
𝑑 2𝜃 𝜔′ 𝑑
Solving (3.4.2), we get 𝑑𝑟 (𝑅𝑒(𝜒𝑖 (𝑟))𝑟=𝑟 𝑐 = − 2(𝜃21+𝜃2 ) ≠ 0 i.e., [ [𝑡𝑟(𝑗 ∗ )]𝑟=𝑟 𝑐 ≠ 0
1 2 𝑑𝑟

which satisfy the transversality condition. This implies that the system undergoes a Hopf-
bifurcation at

𝑟 = 𝑟𝑐 .
4. Numerical simulations

In order to visualize the analytical finding, we perform the numerical simulations over the set of
parametric values

Table 1

Parameter Definition Values

𝒓 1
𝑘 Level of fear in prey population 0.4
𝛼1 Injection rate of prey 1
𝛼2 Conversion rate of prey 0.8
m Refuse rate of prey population 0.72
𝐾1 Half saturation constant (Holling type 2 3
constant)
𝑑1 Natural death rate of prey population 0.6
𝑑2 Natural death rate of predator population 0.2
𝜇 Intra-species competition of predator 0.035
population

(a) (b)

Fig 1. For the parameter values specified in the table above, the equilibrium point E* is asymptotically
stable.

(a) Phase series., (b)Time series

4.3 Effects of k:
When all other parameters are held constant, the mathematical model (2.1) exhibits oscillatory patterns
around E* as the value of k is increased from 0.4 to 4 (refer to Fig. 2). The behavior of a system can be
described by equation (2. 1) which is studied with varying values of parameter k.

(a)
Fig 2. As the parameter k increases from 0.4 to 4, the figure demonstrates oscillatory behavior around at E*
of system (2.1). (a) Phase series; (b) Time series.

4.2 Effects of m:
When all other parameters are held constant, the mathematical model (2.1) exhibits oscillatory patterns
around E* as the value of m is decreased from 0.72 to 0.52 (refer to Fig. 3). The behavior of a system can
be described by equation (2. 1) which is studied with varying values of parameter m.

(a) (b)
Fig 3. As the parameter, decreases from 0.72 to 0.52. the figure demonstrates oscillatory behavior around at
E* of system (2.1). (a) Phase series (b) Time series.
4.2 Effects of 𝝁:
When all other parameters are held constant, the mathematical model (2.1) exhibits oscillatory patterns
around E* as the value of 𝝁 is decreased from 0.035 to 0.02 (refer to Fig. 4). The behavior of a system can
be described by equation (2.1) which is studied with varying values of parameter 𝝁.

(a) (b)

Fig 4. As the parameter, decreases from 0.035 to 0.02. the figure demonstrates oscillatory behavior around
at E* of system (2.1). (a) Phase series (b) Time series.