Towards a comprehensive evaluation and mechanistic perspective on grassland restoration

Schreiner, Falk-Rudhard; Hamer, Ute; Hölzel, Norbert; Neuenkamp, Lena

Zusammenfassung

Grassland degradation driven by land-use intensification has caused severe declines in biodiversity, ecosystem functioning and service provision worldwide. Effective restoration strategies are urgently needed to combat global grassland degradation. Such restoration strategies should achieve a comprehensive recovery of grasslands, i.e., reverse anthropogenic degradation with regard to a broad set of ecosystem attributes. Practical restoration ecology shows that the effectiveness of certain restoration activities usually varies with site-specific context, which complicates the prediction of restoration outcomes and the identification of optimal restoration strategies. We highlight that an ecological theory-based, comprehensive and mechanistic perspective enables a holistic understanding of restoration processes and ultimately the development of credible scenarios for expected restoration outcomes to inform policy and stakeholder decisions. In this review, we take initial steps towards a more comprehensive and mechanistic perspective on grassland restoration. We propose four overarching restoration goals aligned with international standards, each reflecting key ecosystem attributes. To operationalize these goals, we identify a set of quantitative indicators – including above- and belowground parameters - with a strong focus on indicators that may enhance mechanistic understanding of restoration processes. We argue that advancing mechanistic understanding will improve the reliability, transferability, and context-sensitivity of restoration planning. Moreover, it is essential for selecting indicators that allow a comprehensive evaluation of restoration success. To facilitate the development of such a mechanistic understanding we outline a conceptual model of potential causal links between ecosystem attributes, which may serve as blueprint for more sophisticated predictive models.