Including ecosystem functions into restoration ecology has been repeatedly suggested, yet there is limited evidence that this is taking place without bias to certain habitats, species, or functions. We reviewed the inclusion of ecosystem functions in restoration and potential relations to habitats and species by extracting 224 publications from the literature (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013). Most studies investigated forests, fewer grasslands or freshwaters, and fewest wetlands or marine habitats. Of all studies, 14% analyzed only ecosystem functions, 44% considered both biotic composition and functions, 42% exclusively studied the biotic component, mostly vascular plants, more rarely invertebrates or vertebrates, and least often microbes. Most studies investigating ecosystem functions focused on nutrient cycling (26%), whereas productivity (18%), water relations (16%), and geomorphological processes (14%) were less covered; carbon sequestration (10%), decomposition (6%), and trophic interactions (6%) were rarely studied. Monitoring of ecosystem functions was common in forests and grasslands, but the functions considered depended on the study organisms. These associations indicate research opportunities for certain habitats, species, and functions. Overall, the call to include ecosystem functions in restoration has been heard; however, a lack of clarity about the ecosystem functions to be included and deficits of feasible field methods are major obstacles for a functional approach. Restoration ecology should learn from recent advances in rapid assessment of ecosystem functions, and by a closer integration with biodiversity-ecosystem functioning research. Not all functions need to be measured in all ecosystems, but more functions than the few commonly addressed would improve the understanding of restored ecosystems.

Aim Species–area relationships (SARs) are fundamental scaling laws in ecology although their shape is still disputed. At larger areas, power laws best represent SARs. Yet, it remains unclear whether SARs follow other shapes at finer spatial grains in continuous vegetation. We asked which function describes SARs best at small grains and explored how sampling methodology or the environment influence SAR shape. Location Palaearctic grasslands and other non‐forested habitats. Taxa Vascular plants, bryophytes and lichens. Methods We used the GrassPlot database, containing standardized vegetation‐plot data from vascular plants, bryophytes and lichens spanning a wide range of grassland types throughout the Palaearctic and including 2,057 nested‐plot series with at least seven grain sizes ranging from 1 cm2 to 1,024 m2. Using nonlinear regression, we assessed the appropriateness of different SAR functions (power, power quadratic, power breakpoint, logarithmic, Michaelis–Menten). Based on AICc, we tested whether the ranking of functions differed among taxonomic groups, methodological settings, biomes or vegetation types. Results The power function was the most suitable function across the studied taxonomic groups. The superiority of this function increased from lichens to bryophytes to vascular plants to all three taxonomic groups together. The sampling method was highly influential as rooted presence sampling decreased the performance of the power function. By contrast, biome and vegetation type had practically no influence on the superiority of the power law. Main conclusions We conclude that SARs of sessile organisms at smaller spatial grains are best approximated by a power function. This coincides with several other comprehensive studies of SARs at different grain sizes and for different taxa, thus supporting the general appropriateness of the power function for modelling species diversity over a wide range of grain sizes. The poor performance of the Michaelis–Menten function demonstrates that richness within plant communities generally does not approach any saturation, thus calling into question the concept of minimal area.

Fig. 2. Spatial coverage of GrassPlot data from Morocco to Japan. Currently, the majority comes from sub-Mediterranean to hemiboreal Europe (black = multi-scale plots, grey = other plots). Current content v. 1.00 (January 2018) • 126 datasets • 198 data owners • 36 countries • 168,997 plots, among them 14,064 with data also for non-vascular plants • 66,000 0.01-m² plots, 17,206 1-m² plots, 5,520 10-(or 9-) m² plots, 2,545 100-m² plots • 2,797 nested-plot series (with at least 4 grain sizes)

Questions We investigated the effects of grassland management intensity and temporary conversion to other land uses on abiotic and biotic properties of subtropical grasslands. We used species‐rich permanent grasslands of medium management intensity (PG‐M) as a reference, and asked the following questions: (1) do permanent grasslands with low and high management intensities (PG‐L and PG‐H, respectively) have different plant diversity and species composition than reference grasslands; and (2) do secondary grasslands recovering from conversion to arable fields (SG‐A) or pine plantations (SG‐P) differ from permanent grasslands in their plant species composition and abiotic conditions? Location Highland grasslands, Campos de Cima da Serra, Rio Grande do Sul (RS), Brazil. Methods We analysed variation in plant species composition and diversity among 80 grassland sites, including three types of permanent grassland and two types of secondary grassland. An indicator species analysis was used to identify characteristic species for the different land‐use types. We used a linear discriminant analysis to investigate differences in soil conditions among land‐use types. Results Both PG‐L and PG‐H differed from PG‐M regarding plant species composition. Although PG‐L shared many typical grassland species with PG‐M, their communities were generally less diverse. PG‐H, on the other hand, not only had fewer species but also deviated from PG‐M in species composition. Secondary grasslands on former arable fields and plantations differed from PG‐M in species composition and showed lower diversity. Soil conditions of SG‐P were similar to those of PG‐L and PG‐M, but they were distinct from those of PG‐H and SG‐A. Conclusions All land‐use types showed deviations from reference grasslands (PG‐M). The decrease in the number of species in PG‐L may be reversed if traditional management intensity is re‐introduced, whereas strong compositional changes in SG‐P may require the re‐introduction of grassland species. This is also true for PG‐H and SG‐A: both showed marked deviations from reference grasslands in biotic and abiotic components. Overall, restoration of altered land‐use types to near‐natural subtropical grassland seems feasible, but suitable techniques have to be developed.

Biomes are constructs for organising knowledge on the structure and functioning of the world's ecosystems, and serve as useful units for monitoring how the biosphere responds to anthropogenic drivers, including climate change. The current practice of delimiting biomes relies on expert knowledge. Recent studies have questioned the value of such biome maps for comparative ecology and globalchange research, partly due to their subjective origin. Here we propose a flexible method for developing biome maps objectively. The method uses range modelling of several thousands of plant species to reveal spatial attractors for different growth-form assemblages that define biomes. The workflow is illustrated using distribution data from 23 500 African plant species. In an example application, we create a biome map for Africa and use the fitted species models to project biome shifts. In a second example, we map gradients of growth-form suitability that can be used to identify sites for comparative ecology. This method provides a flexible framework that (1) allows a range of biome types to be defined according to user needs and (2) enables projections of biome changes that emerge purely from the individualistic responses of plant species to environmental changes.

Niche‐based selection and stochastic processes can operate simultaneously to generate spatial and temporal variation in species composition. Yet, the conditions under which ecological dynamics are dominated by niche‐based versus stochastic processes are poorly understood. Using a field experiment in early‐successional temperate grassland and null models of beta diversity, this study investigates the effects of soil nutrient supply on the relative importance of niche‐based selection versus stochastic dynamics for variation in species composition among sites. Nutrient availability was manipulated experimentally, individual seed mixtures with 25 species were sown in each experimental plot, and then stochastic and deterministic niche‐based assembly processes were allowed to happen. We found that compositional variation among grassland plots with low nutrient supply was driven by stochastic immigration and extinctions. In contrast, nutrient enrichment reduced the importance of stochasticity and imposed a deterministic environmental filter that homogenized communities through the selection of few species with greater competitive ability for light. This demonstrates that soil nutrient availability is a critical environmental feature that dictates the degree to which terrestrial plant communities are controlled by niche‐based selection versus stochastic assembly processes. Our study shows further that alternative states of eutrophic grasslands emerge from initial stochastic variation in the composition of a particular functional group of species that can become dominant at high nutrient supply. We discuss potential mechanisms underlying the shift from stochastic to niche‐driven dynamics along soil nutrient gradients.

Species distribution modeling is a widely used tool in many branches of ecology and evolution. Evaluations of the transferability of species distribution models—their ability to predict the distribution of species in independent data domains—are, however, rare. In this study, we contrast the transferability of a process‐based and a correlative species distribution model. Our case study uses 664 Australian eucalypt and acacia species. We estimate models for these species using data from their native Australia and then assess whether these models can predict the adventive range of these species. We find that the correlative model—MaxEnt—has a superior ability to describe the data in the training data domain (Australia) and that the process‐based model—TTR‐SDM—has a superior ability to predict the distribution of the study species outside of Australia. The implication of this analysis, that process‐based models may be more appropriate than correlative models when making projections outside of the domain of the training data, needs to be tested in other case studies.

Increasing human pressure on strongly defaunated ecosystems is characteristic of the Anthropocene and calls for proactive restoration approaches that promote self-sustaining, functioning ecosystems. However, the suitability of novel restoration concepts such as trophic rewilding is still under discussion given fragmentary empirical data and limited theory development. Here, we develop a theoretical framework that integrates the concept of 'ecological memory' into trophic rewilding. The ecological memory of an ecosystem is defined as an ecosystem's accumulated abiotic and biotic material and information legacies from past dynamics. By summarising existing knowledge about the ecological effects of megafauna extinction and rewilding across a large range of spatial and temporal scales, we identify two key drivers of ecosystem responses to trophic rewilding: (i) impact potential of (re)introduced megafauna, and (ii) ecological memory characterising the focal ecosystem. The impact potential of (re)introduced megafauna species can be estimated from species properties such as lifetime per capita engineering capacity, population density, home range size and niche overlap with resident species. The importance of ecological memory characterising the focal ecosystem depends on (i) the absolute time since megafauna loss, (ii) the speed of abiotic and biotic turnover, (iii) the strength of species interactions characterising the focal ecosystem, and (iv) the compensatory capacity of surrounding source ecosystems. These properties related to the focal and surrounding ecosystems mediate material and information legacies (its ecological memory) and modulate the net ecosystem impact of (re)introduced megafauna species. We provide practical advice about how to quantify all these properties while highlighting the strong link between ecological memory and historically contingent ecosystem trajectories. With this newly established ecological memory-rewilding framework, we hope to guide future empirical studies that investigate the ecological effects of trophic rewilding and other ecosystem-restoration approaches. The proposed integrated conceptual framework should also assist managers and decision makers to anticipate the possible trajectories of ecosystem dynamics after restoration actions and to weigh plausible alternatives. This will help practitioners to develop adaptive management strategies for trophic rewilding that could facilitate sustainable management of functioning ecosystems in an increasingly human-dominated world.