Understanding ecosystem recovery after perturbation is crucial for ecosystem conservation. Mutualisms contribute key functions for plants such as pollination and seed dispersal. We modelled the assembly of mutualistic networks based on trait matching between plants and their animal partners that have different degrees of specialization on plant traits. Additionally, we addressed the role of non-obligate animal mutualists, including facultative mutualists or non-resident species that have their main resources outside the target site. Our computer simulations show that non-obligate animals facilitate network assembly during the early stages, furthering colonization by an increase in niche space and reduced competition. While non-obligate and generalist animals provide most of the fitness benefits to plants in the early stages of the assembly, obligate and specialist animals dominate at the end of the assembly. Our results thus demonstrate the combined occurrence of shifts from diet, trait, and habitat generalists to more specialised animals.

Data was generated with computer simulations. The methods are described in the main text and the source code is provided here and described in the README.md file. The data produced with the simulations have been processed using Python scripts, which are also provided here and explained in the README.md file.