The flexible, customizable PoPS (Pest or Pathogen Spread) model simulates reproduction, dispersal, and establishment of pests (e.g., insects) and pathogens (i.e., viruses, bacteria, or other organisms that cause disease) through space and time. The PoPS model is the power-house behind the PoPS Forecasting Platform.

For every location in a landscape, at each time step, the PoPS model predicts the number of infested or infected hosts (Ψ). To better understand how the model works, let's take a detailed look at the equation governing the PoPS model.

To break down how the model works, let’s consider pests dispersing from a single location (**cell
i**) and arriving in another single location (**cell j**).

At time t, the number of infested/infected hosts in **cell j** as a result of
pests/pathogens in **cell i** is Ψijt, which can be described by the PoPS model equation:

**The PoPS model equation can be conceptualized in terms of:**

- Reproduction (
*How many pests leave*)**cell i**? - Dispersal (
*Where do the pests go?*) - Establishment (
*How many hosts in*)**cell j**become infected?

Beta (β), the number of pests or pathogens that disperse from a single host under optimal environmental conditions, is the starting point of the PoPS model.

Conditions are rarely optimal and locations contain multiple hosts, so β is modified by the number of currently infested or infected hosts (I) and environmental conditions in a location (i) at a particular time (t) to determine reproduction.

The reproduction component of the PoPS equation gives the number of pests/pathogens dispersing
from **cell i** to * any cell* at

Now that we know how many pests/pathogens are dispersing from **cell i**, we need to
determine where those pests/pathogens go.

This is where the dispersal component of the PoPS model comes in.

The dispersal kernel determines where the new dispersing propagules go; dispersal distance (d) is a function of gamma (γ), which indicates how much dispersal is due to natural processes (alpha-1, α1) or caused by human-mediated transport (alpha-2, α2). The distance of each propagule is determined by drawing from a distribution using either α1 or α2, and its direction is drawn from a distribution that accounts for predominant wind direction (ω) and wind strength (κ). If data are unavailable for these factors, then the distribution is a circle with equal probability in all directions.

The dispersal component of PoPS lets us know where the pests/pathogens from **cell
i** go. From that, we determine how many pests/pathogens arrive at **cell
j**.

Now that we know how many pests/pathogens arrive in **cell j** from **cell
i**, we need to calculate how many hosts in **cell j** become
infested/infected as a result.

Establishment depends on the environmental conditions in **cell j** and the
availability of suitable hosts, calculated as the number of susceptible hosts (S) divided by the
total number of potential hosts (N).

The establishment component of the PoPS equation predicts the number of infested or infected
hosts in **cell j** at **time t**.

Combining the reproduction, dispersal, and establishment components of PoPS, we can calculate
Ψijt (the number of infested/infected hosts in **cell j** at **time
t** as a result of pests/pathogens from **cell i**).

What we’ve described so far is a prediction of the number of infected hosts in a ** single
cell** as a result of pests/pathogens from another

The PoPS model performs these calculations for each cell in the landscape, for every time step in the simulation. The value of Ψ is predicted for each cell, forecasting the spread of a pest or pathogen from infested or infected hosts to susceptible hosts, among all cells, across the landscape.

Watch the video to see reproduction, dispersal and establishment across a landscape.

With considerable optimization and parallelization, PoPS runs quickly, even for landscapes with millions of cells.

Example PoPS simulation of a pest spreading across the United States

PoPS is a modular, spatially explicit, discrete-time model: various components (e.g., weather effects or long-range dispersal) can be included or excluded from the model as necessary (through intuitive on-off switches on the interface) depending on the drivers that influence the species of interest; the model accounts for spatial relationships and movements between grid cells in a landscape; and it forecasts across sequential time steps (which can be specified as either daily, weekly, monthly, or yearly).

Learn more about how PoPS handles:

PoPS is open source and freely available for anyone to use. There are two ways you can use PoPS:

Users may download the PoPS model code for:

- R
- GRASS GIS
- Python (
*coming soon*)

(*Available for public use soon.*)

No coding experience? We developed an easy-to-use interface to run the PoPS model in the cloud. The dashboard is currently being used internally and will be made available for public use soon.