Table 1.

List of variables to describe communities

Metric		Definition
Diversity	Number of species: with at least one individual present in the landscape.	S
	Number of individuals:	$N={\sum }_{i=1}^S{n}_i$, where ni is the number of individuals belonging to species i.
	Shannon index: measures evenness in distribution of species abundances.	$D_{\mathrm{Shannon}}=-{\sum }_{i=1}^S{r}_i\log \left(r_i\right)$, where ri is the relative abundance of species i.
	Shannon equitability: is the Shannon index, normalised to control for number of species present.	$E_{\mathrm{Shannon}}=\frac{D_{\mathrm{Shannon}}}{\log \left(S\right)}$, where DShannon is the Shannon index defined above.
	RATP: relative abundance of top predator species.	$\mathrm{RATP}=\left({\sum }_{j=1}^P{n}_j\right)∕\left({\sum }_{i=1}^S{n}_j\right)$, where P is the number of top predator species, i indexes all species, and j indexes top predators only.
Network	Number of links: the number of links present in the realised interaction network. (Presence defined as at least one interaction event in the landscape during 200 time steps.)	L
	Compartmentalisation: the degree to which species share common neighbours across the network51.	$C=\frac{1}{S\left(S-1\right)}{\sum }_{j\ne i}{\sum }_{i=1}^S{c}_{ij}$, where cij is the number of species with which both i and j interact, divided by the number of species with which neither i or j interact.
	Nestedness: the extent to which specialist species interact with subsets of the species with whom generalists interact40.	Calculated for the mutualistic sub-network only, using the NODF algorithm52.
	Generality: weighted quantitative generality53.	$G_q={\sum }_{k=1}^S\frac{b_{.k}}{b_{..}}{n}_{N,k}$, where $b_{.k}$ is the total amount of biomass going into species k, and $b_{..}$ is the total amount of biomass flowing through the entire ecological network. $n_{N,k}$ is the number of species predated on by species k. Here, the biomass flowing from one species to another was calculated as the number of individuals of a given prey species eaten by individuals of predator species k.
	Vulnerability: weighted quantitative vulnerability53.	$V_q={\sum }_{k=1}^S\frac{b_{k.}}{b_{..}}{n}_{P,k}$, where $b_{k.}$ is the total biomass emanating from species k. $b_{..}$ is the total amount of biomass flowing through the entire ecological network. $n_{P,k}$ is the number of predator species that feed upon prey species k. Here, the biomass flowing from one species to another was calculated as the number of individuals of prey species k eaten by a given predator species.
	Mean interaction strength (IS): average inter-specific interaction strength (averaged over all interactions in realized network)	${\sum }_{i\ne j}\frac{b_{ij}}{2n_i{n}_j}$, where $b_{ij}$ is the total biomass flowing from prey species i to predator species j—quantified here as the total number of individuals (or fractions of it, in the case of plants) from species i eaten by individuals of species j. This way of calculating interaction strengths quantifies the per-capita effect of a predator species over its prey, and it is thus analogous to Paine’s index and Lotka–Volterra interaction coefficients9,54. Hence, these values allow to assess and understand community stability based on the strengths of ecological interactions.
Stability	CV population: mean coefficient of temporal variation in species population abundances.	$\frac{1}{S}{\sum }_{i=1}^S\frac{\sigma \left(n_i\right)}{\mu \left(n_i\right)}$, where $\mu \left(n_i\right)$ and $\sigma \left(n_i\right)$ are the mean and standard deviation in the abundance $n_i$ of species i over 200 simulation time steps.
	CV range: mean coefficient of temporal variation in species range area.	$\frac{1}{S}{\sum }_{i=1}^S\frac{\sigma \left(a_i\right)}{\mu \left(a_i\right)}$, where $\mu \left(a_i\right)$ and $\sigma \left(a_i\right)$ are the mean and standard deviation in the range area $a_i$ of species i over 200 simulation time steps. The range area of a species is defined as the area of the circle, centred on the centre of mass of the species spatial distribution, that contains 95% of the individuals belonging to that species.