Potential vegetation estimations help to assess feasibility and expected effort needed in grassland restoration by shrub removal

Shrub removal is a commonly applied method to restore and preserve biodiverse open ecosystems. In order to maintain the optimal conservation state after clearing, costly intensive post-treatments and repeated shrub removal are often employed. Our hypothesis is that if the target vegetation is not self-sustainable, grassland restorations demand intensive post-treatments. Multiple potential natural vegetation (MPNV) estimates are indicative of the relative self-sustainability of various vegetation types. Therefore, MPNV estimations have the potential to predict the required post-treatment intensity. In the present study, our aim was to determine whether grassland and forest self-sustainability assessed using MPNV models could help in predicting the post-treatment intensity required to maintain grasslands restored by shrub removal. We collected data from grassland restoration projects in Hungary that employed shrub removal and that differed in the types and frequency of their post-treatments. We tested how grassland and forest self-sustainability estimated by MPNV models and project area size affected required post-treatment intensity using a cumulative link mixed model. Higher forest self-sustainability indicated a need for more intense post-treatment. If MPNV models estimate high self-sustainability for forests at restored sites, they are more likely to require intensive post-treatments after shrub removal, even beside considerable grassland self-sustainability. Based on our results, MPNV models can help predict the necessary post-treatment intensity in grassland restoration and be helpful in identifying sites that would require less intensive post-treatment, thus optimizing restoration costs. Finally, MPNV models can support successful restoration in maintenance-intensive sites by indicating the need for regular post-treatment measures.