Table 2. Model selection for ordinary differential equation models based on expected log predictive density for a new observation calculated using Pareto-smoothed importance sampling.

Each row shows the difference in expected log predictive density ( $\mathrm{\Delta }\mathrm{e}\mathrm{l}\mathrm{p}{\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{o}}$) between a given model and the best model in the top row, and the standard error (SE) of the difference.

Model	$\mathrm{\Delta }{\text{elpd}}_{\text{loo}}$	SE
Overgrowth	0	0
Protection	−32.0	18.1
Basic	−32.0	18.2
Settlement facilitation	−33.4	17.4
Growth facilitation	−34.3	16.1