EcoSISTEM

Documentation	Build Status	DOI

EcoSISTEM (Ecosystem Simulation through Integrated Species-Trait Environment Modelling) is a Julia package for simulating species undergoing birth, death, competition and dispersal, under climate and habitat that change over time.

The package was primarily developed for global scale simulations of plant biodiversity. The underlying model is described in the arXiv paper arXiv:1911.12257 (q-bio.QM), Dynamic virtual ecosystems as a tool for detecting large-scale responses of biodiversity to environmental and land-use change.

A simulation is a spatial metacommunity: abundances per species per cell on a grid, rather than individuals. Each cell offers conditions a species is more or less suited to, and resources every species present competes for from a shared pool. Births and deaths respond to how the community's total demand compares with what a cell supplies, so a cell's carrying capacity is an outcome of who is living there rather than a number set in advance - conditions decide where a species can persist, and resources how many of it can.

The environment is built from layers, which may be uniform, drawn from real climate and land-cover data, or made to change over the course of a run. See How the model works for the mechanism in full.

using Pkg
Pkg.add("EcoSISTEM")

A first simulation needs a grid, an environment on it, a set of species, and a run:

using EcoSISTEM, EcoSISTEM.Units
using Unitful, Unitful.DefaultSymbols

# A synthetic grid, 50 km square, in cells of 10 km
area = StudyArea(extent = (50km, 50km), cellsize = 10km)

# One condition species are matched to, and one resource they compete for
env = GridHabitat(regime = UniformSpec(285.0K, axis = Temperature),
                  supply = UniformSpec(1.0e5kJ / (m^2 * day), axis = SolarRadiation),
                  area = area)

# Ten species suited to 285 K, sharing 10,000 individuals
species = build_species(10,
                        tolerance = (285.0K, 3.0K), toleranceaxis = Temperature,
                        demand = 1.0e9kJ / day, demandaxis = SolarRadiation,
                        dispersal = 5.0km,
                        abundance = 10_000,
                        seed = 1)

eco = build_ecosystem(species, env)
simulate!(eco, 10year, 1month_mean_duration)

sum(eco.abundances.matrix)      # individuals still alive

The documentation walks through this in more detail, and covers what an environmental layer means, where climate data comes from, how time works in a simulation, how to intervene in a running ecosystem, and how to measure the diversity of the result.

Two Pluto notebooks come with the source, in their own environment. With the repository cloned:

julia --project=notebooks -e 'import Pluto; Pluto.run()'

Then open notebooks/Introduction.jl for a guided tour, or notebooks/InteractiveAfrica.jl to watch an invasive species colonise Africa. Both may be slow to start, as the second downloads climate data.

This package is in beta now, so please raise an issue if you find any problems. For more information on how to contribute, please read our contributing guidelines. We are supported by NERC's Landscape Decisions small and large maths grants and an EPSRC studentship.