ashander / landsim

efficient, forwards-time simulation of populations on raster-based landscapes

landsim : Efficient simulation of landscape-scale populations

This is an R package with tools to run efficient, forwards-time simulation of populations on habitats derived from GIS raster layers.

Install it by installing devtools and running

devtools::install_github("petrelharp/landsim")

Here are some vignettes showing what it can do:

1. Object structure : Describes the modeling strategy and how information is stored in population and demography objects.
2. Density dependence : Demonstrates a few different ways of implementing density-dependent population size regulation.
3. Types of selection : Demonstrates a few different ways of implementing genotype-dependent selection.
4. Spatial sweeps : Simulates and analyzes an expanding wave of an advantageous allele.
5. A selective sweep : Works through the modeling choices of a model of selection.
6. Steps in a generation : Works through the guts of the steps in a simulation.
7. Migration methods : Explains how to implement various types of migration: conservative or not, and different boundary types.
8. Migration method run times : Examines the details of two methods of migration, using precomputed sparse Matrix objects, and using raster methods.
9. Sierpinski habitat : Creates a random (Sierpinski gasket) habitat.

A more generally useful function is discretize_kernel(), which allows you to translate from continuous migration kernels to discrete migration rates between grid cells in a natural way that is consistent across different grid resolutions.