🌎 Spatial Epidemiology: Unveiling Disease Dynamics with Reaction–Diffusion Models

Partial Differential Equation (PDE) models, often expressed as Reaction–Diffusion systems, provide a mathematical framework for analyzing disease spread in continuous space and time. They extend traditional Ordinary Differential Equation (ODE) models by representing both the local spread of infection (reaction) and the geographic movement of hosts (diffusion). This approach is essential for understanding large-scale epidemic propagation and designing spatially targeted interventions.

⚙ Compartmental Structure and Flow Explanation

PDE-based epidemic models use standard epidemiological compartments such as Susceptible (S), Exposed (E), Infectious (I), and Recovered (R). Each compartment is represented as a spatial density function, uᵢ(x, t), defined over location x and time t.

The population dynamics combine two fundamental processes:

Reaction Term f(u)
Represents local disease processes at each point in space, including infection, recovery, births, and deaths.
Diffusion Term DΔu
Represents random host movement across space, typically modeled using Fick’s Law of diffusion. The Laplace operator Δ governs spatial dispersal.

Within a spatial domain Ω, individuals transition between compartments according to the reaction dynamics while simultaneously dispersing due to diffusion.

🧮 Mathematical Formulation (Reaction–Diffusion SIR System)

A spatial SIR model with constant demographic effects (Λ, μ) and diffusion coefficients (kₛ, kᵢ, kᵣ) is described by the system:

∂S(x,t)/∂t = kₛ ΔS(x,t) + Λ − β [S(x,t) I(x,t) / N] − μ S(x,t)
∂I(x,t)/∂t = kᵢ ΔI(x,t) + β [S(x,t) I(x,t) / N] − (γ + μ) I(x,t)
∂R(x,t)/∂t = kᵣ ΔR(x,t) + γ I(x,t) − μ R(x,t)

These PDEs are typically paired with Neumann boundary conditions (zero flux across the boundary ∂Ω), ensuring that population movement remains contained within the domain.

⚗ Parameter Definitions and Ranges

Parameter	Definition	Typical Range	Context
Λ, μ	Recruitment and natural death rates	Approximately 1/27,000	Often assumed spatially uniform.
β, γ	Transmission and recovery rates	0.2 – 1.0 per day	Represent infection risk and infectious period.
kₛ, kᵢ, kᵣ	Diffusion coefficients	Positive values	Quantify random spatial movement for each compartment.
N	Total population density	Constant	For cases where Λ = μN.

Applicability and Limitations

Applications

Spatial Spread Analysis
Reaction–diffusion models are essential for characterizing how epidemics propagate geographically, capturing wave fronts and spatial gradients of infection.
Wave Dynamics and Threshold Conditions
These models allow the calculation of minimal wave speeds and help determine whether an infection will successfully invade new territory.
Spatially Targeted Interventions
PDE frameworks can estimate required barrier sizes or identify optimal zones for intensified control measures to halt disease spread.

Assumptions and Weaknesses

Random Movement Assumption
Models typically assume undirected, random diffusion of hosts. Real movement patterns may be directional, constrained, or behaviorally influenced.
Local Homogeneity
Within any small spatial region, the model assumes homogeneous mixing, which may oversimplify complex behavioral or environmental heterogeneity.
Mathematical and Computational Complexity
Analytic solutions are rare. Numerical methods—finite differences, finite elements—are often required and can be computationally demanding.

📚 Selected References

Kermack, W. O., & McKendrick, A. G. A contribution to the mathematical theory of epidemics.
Murray, J. D., Stanley, E. A., & Brown, D. L. On the spatial spread of rabies among foxes.
Hethcote, H. W. The mathematics of infectious diseases.
Lotfi, E. M., Maziane, M., Hattaf, K., & Yousfi, N. Partial differential equations of an epidemic model with spatial diffusion.
Oshinubi, K., Chen, Y., Doerry, E., Gel, E. S., Hepp, C., Lant, T., Mehrotra, S., Sabo, S., & Mihaljevic, J. A systematic review of spatial epidemiological modeling approaches applied during the COVID-19 pandemic.