Alcantara, J. M. | Search bioRxiv

bioRxiv2026

Assessing pollinator community recovery in restored agroecosystems using the recovery debt framework

Recovery debt (RD) quantifies the interim deficit of biodiversity and function during the recovery process after disturbance. Unlike typical recovery indices derived from data on experimental-control comparisons, RD further considers the target (reference) biodiversity level, modelling the rate at which it is approached over time. However, the application of the RD approach to active restoration has not been explicitly implemented to date. Here, we extend the RD framework to evaluate active ecological restoration in agricultural systems, defining the onset of recovery as the shift from intensive to wildlife-friendly management. We applied this approach to assess short-term pollinator recovery in 14 olive groves across a gradient of farming intensification and landscape complexity in southern Spain. Restoration actions included adopting low-intensity ground cover management and actively restoring field margins. At one, three, and five years post-restoration, we assessed community responses by quantifying bee abundance, species richness, plant-bee network properties, and flower visitation rates. Reference systems were defined by olive groves in complex landscapes with low-intensity herb cover management and organic farming practices. Following restoration, the RD of bee abundance decreased from 71% to 55%. We found no significant effects of pre-intervention agricultural management on RD. Instead, across sites, the reduction of the RD (i.e., recovery) of bee abundance, richness, network connectance and flower visitation rate was strongly mediated by the availability of high-quality semi-natural areas in the surrounding landscape and by the ecological contrast created by restoration interventions at both the farm and floral patch levels. RD for other network metrics showed no significant pattern of variation. Our study demonstrates that wildlife-friendly management and targeted habitat restoration can rapidly reduce recovery debt for bee abundance and function in permanent agroecosystems. However, the recovery of more complex interaction-network properties likely requires longer timescales.

ecology↗

bioRxiv2025

Maximum entropy networks show that plant-arbuscular mycorrhizal fungal associations are anti-nested and modular

Many applications of network theory to plant-mycorrhizal associations have used a bipartite description, in which one set of nodes is the plants, and the other set is the fungi. Most applications have relied on null models from algorithms that randomly rewire the observed connections to test for non-random patterns in the network. We used existing plant-arbuscular mycorrhizal (AM) fungal datasets to apply a new, well validated generation of network models relaxing the very limiting assumptions of traditional null models. We focused on nestedness and modularity, which have been related to the functioning and stability of communities. Given the existent literature, we expected nestedness and modularity to be prevalent. We modelled plant-AM fungal associations using maximum entropy networks with a degree sequence, soft constraint to generate null distributions for nestedness and modularity. Most plant-AM fungal associations were anti-nested and modular. This pattern was consistent across habitat types and multiple spatial scales. Anti-nestedness can easily emerge from modularity when network patterns are determined by the identity of the plant and AM fungal nodes. Future studies will have to test how the observed patterns determine the ability of the associations to adapt to environmental changes.

ecology↗