Statistical Physics ↔ Epidemiology: Network percolation theory and epidemic threshold theory are the same mathematical object

🔭 Overview

In bond percolation on a network, a giant connected component emerges at a critical bond probability p_c — below p_c the outbreak is finite; above it a macroscopic fraction of nodes is infected. The epidemic threshold R_0=1 (basic reproduction number equals 1) is *exactly* this percolation threshold applied to a contact network where each edge is "bonded" with probability proportional to transmission probability. This mathematical identity has been known since Newman (2002) and Kenah-Robins (2007), but almost no working epidemiologists use percolation theory in practice. The result is that standard SIR/SEIR models treat populations as effectively infinite (mean-field), while condensed-matter physicists have 50 years of finite-size scaling (FSS) methods that correctly estimate critical behaviour in finite systems. The practical consequence: current outbreak risk models routinely over- or under-estimate the probability of a major epidemic in populations of 10,000- 100,000 people (the regime of healthcare facility outbreaks, island communities, schools) because they use mean-field R_0 estimates without finite-size corrections. Percolation FSS would give systematic, quantitative corrections to these estimates.

↔️ Translation Table

Domain A Term	Domain B Term	Note
bond percolation threshold p_c	epidemic threshold R_0 = 1 (herd immunity boundary)	—
giant connected component	epidemic — macroscopic fraction of population infected	—
finite cluster below p_c	self-limiting outbreak (index case dies out without major spread)	—
finite-size scaling (FSS) correction to p_c in N-node network	correction to R_0 threshold in finite population of size N	—
percolation correlation length xi	spatial/network scale over which infections correlate (cluster size distribution)	—
order parameter (giant component fraction P_inf)	final attack rate (fraction of population ultimately infected)	—
scaling function near p_c (P_inf ~ \|p-p_c\|^beta)	epidemic size as function of R_0 near threshold	—
universality class (mean-field percolation for random graphs, 2D percolation for spatial grids)	contact network topology class (random mixing vs spatially structured population)	—

🗺️ Why Hasn't This Been Unified?

The mathematical equivalence of SIR thresholds and percolation thresholds is established in the network science literature (Newman 2002, Kenah-Robins 2007, Miller 2010). However, the epidemiology literature overwhelmingly uses deterministic ODE models (SIR/SEIR) that are mean-field and infinite-population. The finite-size scaling tools are developed in condensed-matter physics and statistical mechanics journals that epidemiologists do not read. During COVID-19, finite-size effects in small institutional outbreaks (hospitals, nursing homes, schools) were handled with ad hoc simulation rather than systematic FSS — a direct consequence of this communication gap.

🌱 Cross-Pollination Opportunities

Finite-size scaling from condensed-matter physics gives a precise formula for how the effective epidemic threshold shifts in a population of N individuals: R_0^eff(N) = R_0^infinity + c * N^(-1/nu), where nu is the percolation correlation length exponent (nu=1 for random networks, nu=4/3 for 2D spatial). Epidemiologists could use this formula to report R_0 estimates with explicit finite-population uncertainty bands — a capability that does not currently exist.

The percolation cluster size distribution below threshold predicts the outbreak size distribution. Condensed-matter physicists know this distribution analytically (it is a power law with exponent tau=5/2 for mean-field percolation). This means the probability of an outbreak reaching any given size is derivable from the same framework, not just simulatable — giving public health a closed-form risk formula.

Vaccination strategies map directly to percolation bond removal. Targeted removal of high-degree nodes (hubs) in a network pushes p_c up nonlinearly — this is the mathematical basis of ring vaccination and contact tracing. Importing the full toolkit of percolation on heterogeneous networks (configuration model, degree sequence methods) would let epidemiologists design vaccination campaigns using rigorous phase-transition analysis rather than simulation.

The b-percolation-oncology bridge uses percolation to model tumour vascular networks. The same finite-size scaling methods apply — connecting the mathematical toolkits of oncology, epidemiology, and statistical physics through a single framework.

🧪 Crosscheck — Prove This Bridge

Runnable experiment protocols promoted from this bridge. USDR maps what connects; Crosscheck proves it.

Finite-size epidemic threshold — does bond percolation on a random graph show threshold shift consistent with R_0^eff(N) = R_0^∞ + c * N^(-1/nu) at nu = 1?

EXECUTED desktop

On Erdős–Rényi graphs G(N, p), the bond percolation threshold p_c(N) shifts systematically with N as p_c(N) = p_c(∞) + c * N^(-1/nu) with nu ≈ 1 (mean-field universality class for random graphs).

Protocol p-b-percolation-epidemiology-fss

Open in Colab · Protocol YAML

Crosscheck manifesto · Generate more drafts: python scripts/generate_crosscheck.py --bridge b-percolation-epidemiology --write

📚 References

DOI:10.1103/PhysRevE.66.016128 · Newman (2002) — proved SIR epidemic threshold = bond percolation threshold on random graphs

DOI:10.1093/aje/kwm120 · Kenah & Robins (2007) — epidemic as percolation on the transmission network

DOI:10.1103/PhysRevE.76.010101 · Miller (2007) — epidemic size and final size as percolation order parameter

arXiv:cond-mat/9212004 · Finite-size scaling in percolation — the FSS toolkit that epidemiology has not imported

DOI:10.1038/nature04153 · Newman (2006) — modularity and community structure in networks; relevant to spatial epidemic structure

Network percolation theory and epidemic threshold theory are the same mathematical object — the epidemic threshold R_0=1 is a percolation phase transition, and importing finite-size scaling from condensed-matter physics would transform how outbreak risk is estimated in finite populations.