Burrowing seabirds that nest on islands transfer nutrients from the sea, disturb the soil through burrowing, damage tree foliage when landing, and thereby modify the surface litter. However, seabirds are in decline worldwide, as are their community- and ecosystem-level impacts, primarily due to invasive predatory mammals. The direct and indirect effects of seabird decline on communities and ecosystems are inherently complex. Here we employed network analysis, as a means of simplifying ecological complexity, to better understand the effects seabird loss may have on island invertebrate communities. Using data on leaf litter communities, we constructed invertebrate food webs for each of 18 offshore oceanic islands in northeastern New Zealand, nine of which have high seabird densities and nine of were invaded by rats. Ten network topological metrics (including entropy, generality, and vulnerability) were compared between rat-invaded and uninvaded (seabird-dominant) islands. We found that, on rat-invaded islands, the invertebrate food webs were smaller and less complex than on their seabird-dominated counterparts, which may be due to the suppression of seabird-derived nutrients and consequent effects on trophic cascades. This decreased complexity of food webs due to the presence of rats is indicative of lower ecosystem resistance via lower trophic redundancy. Our results show that rat effects on island ecosystems are manifested throughout entire food webs, and demonstrate how network analysis may be useful to assess ecosystem recovery status as these invaded islands are restored.

The progressive removal of invasive mammals from the Mercury Islands has led to over 25 years of field study designed to test the processes of restoration and natural recovery of these seabird-driven island ecosystems. Resulting from this work, four key restoration questions can now be identified as fundamental to designing island restoration programmes. The questions are: what is the regional context of the island (biogeography); how does each island ecosystem operate (ecosystem function); how have invasive species changed the ecosystem (response effects); and how can progress towards a restoration goal be defined (outcome measures)? Examples of how these questions influenced restoration in the Mercury Islands are provided with Korapuki Island as a case study. However, unpredicted and subtle responses can eventuate. In the Mercury Islands these included a hitherto unknown honeydew parasite-bird-gecko food web and subtle effects of rats on plant regeneration. Promising outcome measures of restoration progress are now being developed, including indices of marine influence using stable isotopes of nitrogen and the use of network analysis to analyse the composition of invertebrate food webs.

No abstract

No abstract

Responses of the terrestrial biosphere to rapidly changing environmental conditions are a major source of uncertainty in climate projections. In an effort to reduce this uncertainty, a wide range of global change experiments have been conducted that mimic future conditions in terrestrial ecosystems, manipulating CO2, temperature, and nutrient and water availability. Syntheses of results across experiments provide a more general sense of ecosystem responses to global change, and help to discern the influence of background conditions such as climate and vegetation type in determining global change responses. Several independent syntheses of published data have yielded distinct databases for specific objectives. Such parallel, uncoordinated initiatives carry the risk of producing redundant data collection efforts and have led to contrasting outcomes without clarifying the underlying reason for divergence. These problems could be avoided by creating a publicly available, updatable, curated database. Here, we report on a global effort to collect and curate 57,089 treatment responses across 3644 manipulation experiments at 1145 sites, simulating elevated CO2, warming, nutrient addition, and precipitation changes. In the resulting Manipulation Experiments Synthesis Initiative (MESI) database, effects of experimental global change drivers on carbon and nutrient cycles are included, as well as ancillary data such as background climate, vegetation type, treatment magnitude, duration, and, unique to our database, measured soil properties. Our analysis of the database indicates that most experiments are short term (one or few growing seasons), conducted in the USA, Europe, or China, and that the most abundantly reported variable is aboveground biomass. We provide the most comprehensive multifactor global change database to date, enabling the research community to tackle open research questions, vital to global policymaking. The MESI database, freely accessible at https://doi.org/10.5281/zenodo.7153253, opens new avenues for model evaluation and synthesis‐based understanding of how global change affects terrestrial biomes. We welcome contributions to the database on GitHub.