Beyond the single index: Investigating ecological mechanisms underpinning ecosystem multifunctionality with network analysis.

Ecosystems simultaneously deliver multiple functions that relate to both the activities of resident species and environmental conditions. One of the biggest challenges in multifunctionality assessment is balancing analytical simplicity with ecosystem complexity. As an alternative to index-based approaches, we introduce a multivariate network analysis that uses network theory to assess multifunctionality in terms of the relationships between species' functional traits, environmental characteristics, and functions. We tested our approach in a complex and heterogeneous ecosystem, marine intertidal sandflats. We considered eight ecosystem function, five macrofaunal functional trait groups derived from 36 species, and four environmental characteristics. The indicators of ecosystem functions included the standing stock of primary producers, oxygen production, benthic oxygen consumption, DIN (ammonium and NOx efflux) and phosphate release from the sediments, denitrification, and organic matter degradation at the sediment surface. Trait clusters included functional groups of species that shared combinations of biological traits that affect ecosystem function: small mobile top 2 cm dwellers, suspension feeders, deep-dwelling worms, hard-bodied surface dwellers, and tube-forming worms. Environmental characteristics included sediment organic matter, %mud, %shell hash, and %sediment water content. Our results visualize and quantify how multiple ecosystem elements are connected and contribute to the provision of functions. Small mobile top 2 cm dwellers (among trait clusters) and %mud (among environmental characteristics) were the best predictor for multiple functions. Detailed knowledge of multifunctionality relationships can significantly increase our understanding of the real-world complexity of natural ecosystems. Multivariate network analysis, as a standalone method or applied alongside already existing single index multifunctionality methods, provides means to advance our understanding of how environmental change and biodiversity loss can influence ecosystem performance across multiple dimensions of functionality. Embedding such a detailed yet holistic multifunctionality assessment in environmental decision-making will support the assessment of multiple ecosystem services and social-ecological values.