Scripts and data for: Integrating different facets of diversity into food web models: how adaptation among and within functional groups shape ecosystem functioning

Adaptation of communities to environmental fluctuations can emerge from different facets of biodiversity,  which may impact ecosystem functioning differently. Previous work examined how ecosystem functions can be influenced by two sources of adaptive potential: sorting (i.e., changes in community composition due to fitness differences) can occur when multiple species or groups are present (richness), and trait adaptability (i.e., trait adjustments within species or functional groups) can emerge from genetic or phenotypic diversity. However, their effect is typically studied separately, and often in the context of only one trophic level. Therefore, we used a bitrophic trait-based model varying in richness and in the presence of trait adaptability at each trophic level, to investigate how sorting and trait adaptability, at one or two trophic levels, separately or jointly shape ecosystem functions. We found that the adaptive potential emerging from any facet of diversity-induced changes in trophic interactions, in turn, affects biomass distributions within and across trophic levels, dynamical behaviour, and synchrony of biomass dynamics within a trophic level. Particularly, sorting and trait adaptability could contribute to a similar degree and at a similar time to temporal changes in ecosystem functions, but their respective contribution depended on the speed of trait adaptation, the trait range between similar functional groups, and trophic interactions. We thus suggest to consider multiple facets of diversity and their corresponding sources of adaptive potential to deepen our mechanistic understanding of ecosystem functioning, especially in a context of rapid biodiversity change.

Methods
The datasets were generated and not collected in the field and the laboratory. We briefly summarise the methods used, which are extensively explained in the associated Oikos article. We solved numerically the ordinary differential equations of an an extended Rosenzweig-MacArthur predator-prey model in C using the SUNDIALS CVODE solver 5.7.0}. Then, we used several packages in Python 3.10 among which NumPy, Pandas, and Matplotlib to analyse the biomass and trait dynamics, and to quantify ecosystem functions. We notably compared the temporal means and variation (coefficient of variation) of ecosystem functions and properties (e.g. total biomass, production, biomass-weighted mean trait, synchrony of prey and predators, and the ratio between prey losses due to predation and the sum of prey losses due to competition and predation) of food webs with different sources of adaptive potential.