Our main objective is to determine the most efficient approaches to restore complex, multi-functional and resilient ecosystems, thus securing net environmental gain.
Because time is a major driver and impossible to manipulate within a four-year project, we will use a “natural experiment” approach. This involves overlaying a robust statistical design onto an existing network of sites that represent a range of potential drivers: time since restoration began, initial state and spatial proximity to existing woodlands or grasslands (Fig. 1).
The project has seven work packages, which are designed to elucidate the relationships between site factors and emergent properties (Fig. 1). After implementing a natural experiment (WP1), we will do a broad ecological survey in WP2 to link site factors to complexity, and targeted ecological surveys in WP3 to link complexity to function. Within a subset of sites we will do manipulative experiments to determine resilience to drought, bolstered by long-term data (WP4), and how management interventions can speed the development of complexity (WP5). Finally, in WP6 we will analyse and synthesise our findings, which will achieve impact through WP7.
WP1: Site selection (Lead – Prof. Kirsty Park)
We have selected existing restoration sites to represent specific variables for the natural experiment (Fig. 1): time since restoration began (10-50 years), initial state (contrasting intensive agriculture vs mining/quarrying) and spatial proximity to existing woodlands or grasslands. In total, we will survey 133 sites, including:
- 60 broadleaved and mixed woodlands – 30 located in Central Scotland and 30 in the Midlands of England
- 60 chalk or limestone grasslands – all located in Southern England
- 13 ‘Wildcard’ sites, located across the UK, which we can compare to the restoring sites. These will include ancient woodland, calcareous grassland and rewilding sites, which have been restored using novel or unconventional approaches.
“We will examine how the outcomes of restoration vary with proximity to other similar habitats, the initial state – in particular, agriculture and ex-quarry sites, and the methods used to restore them.
For our woodland sites, this will include planting versus natural colonisation, which is a very hot topic at the moment. Importantly, we will also look at how these outcomes vary over time, by including sites of varying restoration age.”Prof. Kirsty Park – University of Stirling
WP2: Ecosystem Complexity (Lead – Prof. James Bullock)
In each woodland, grassland and ‘Wildcard’ site (n=133) we will carry out multitrophic assessments, including soil microbiome, vegetation, and invertebrates (herbivores, predators, pollinators). These data will be used to calculate our measures of site-level ecological complexity. In our analyses, we will link site factors (i.e. age, initial state and proximity) to our measures of ecosystem complexity to assess ‘how site factors influence complexity measures’. These results will allow us to select suitable sites for WP3, and subset sites to represent gradients of complexity, from low to high.
WP3: Complexity effects on emergent properties (Lead – Prof. Jim Harris)
In a subset of sites (n=36) that represent a gradient of complexity, we will quantify if and how complexity supports net gains in the emergent functionality of ecosystems. This requires more detailed work on complexity measures (e.g. structural complexity, food web complexity, soil microbiome complexity and soundscape complexity), as well as emergent properties, including ecosystem function (e.g. litter decomposition rates, pollination services, and herbivory and predation rates) and soil thermodynamic efficiency. In our analyses, we will link ecological complexity to our measures of emergent properties to assess ‘how complexity influences ecosystem functioning’.
WP4: Ecosystem resilience and complexity (Lead – Prof. James Bullock)
Resilience is only rarely measured directly during restoration, and resilience along restoration gradients has not been quantified. We will do this in two complementary ways, by measuring ecosystem responses to experimental drought perturbations in a subset of our sites (n=10, representing a gradient in complexity) and using remote imagery over a 25-year archive to identify responses to drought.
We will impose droughts at our woodland and grassland sites, using passive rainout shelters and following well-defined protocols. The rainout shelters will be deployed for a period long enough to simulate a 1 in 100 year drought (i.e. 41 days). We will take measures of complexity and function in drought and control plots, just before the shelters go out, a week after we remove them and at the same date the following year.
We will measure resilience in terms of recovery to an average state following perturbation, and the ability to resist perturbations and maintain functions. In more complex ecosystems we expect functions to be resilient, even if complexity measures change with the perturbation.
“We cannot simply set in motion the restoration or rewilding of degraded places and hope for the best. The natural world and the benefits we get from nature, including carbon capture, clean water and beautiful landscapes, are threatened by climate change, pollution and mass extinction. The ecosystems we restore must be resilient to these threats, and we will investigate how to achieve this aim.”Prof. James Bullock – UK Centre for Ecology & Hydrology
WP5: Management interventions and ecosystem complexity (Lead – Prof. Kirsty Park)
In a subset of sites (n=20), we will test whether ecological complexity can be increased rapidly through targeted management, using a Before-After-Control-Intervention (BACI) approach. Sites will be chosen from WP2 to represent low to moderate complexity, and we will determine if interventions shift the pace of complexity development.
In grasslands, we will examine the effects of experimental additions of later successional plant species (e.g. Helianthemum, Campanula, Asperula spp.) on complexity, by comparing intervention plots (with plug plants added) to control plots. In woodlands, we will compare three plots: 1) with thinning to enhance structural diversity, 2) with thinning and the introduction of deadwood to provide additional resources and habitats, and 3) control plot with no interventions. We will carry out surveys to measure ecological complexity in each plot within our woodland and grassland sites, before interventions and 2 years later.