The properties of a hypothetical environment that is completely homogeneous and well-mixed, and therefore lacking spatial structure (e.g. arrangement or configuration of habitat patches in a landscape), are usually unaffected by the movement of organisms and materials. Otherwise, spatial structure in the environment can be an important regulator of the dynamics of populations, communities, and ecosystems. Spatial structure has two main components: spatial hierarchy and spatial heterogeneity. Spatial hierarchy refers to the different levels of spatial organization of ecological systems. In the simplest view, this hierarchy involves a small scale (locality), where non-spatial processes may be particularly important (e.g. demography and species interactions), and a larger scale (region), where the dominant processes might differ (e.g. dispersal and colonization history). Spatial heterogeneity refers to the degree by which the spatial distribution of an ecological factor (such as temperature, moisture or nutrient availability) or process (such as predation pressure) varies over space (i.e. whether it is evenly or unevenly distributed). Heterogeneity and its spatial distribution can be important in affecting metapopulations (Hanski, 1994), metacommunities (Logue et al., 2011) and metaecosystems (Gounand et al., 2018). To date, metaecology has mostly addressed spatial structure by thinking of discrete habitat patches separated by an inhospitable matrix that limits dispersal, but ongoing work is likely to change this simplistic framework (e.g. Chase and Knight, 2013; Garzon-Lopez et al., 2014; Munkemuller et al., 2014; Leibold and Chase, 2017).

The movement of organisms and materials results from the interaction between their dispersal and flow rates, respectively, and the spatial organization of the environment. This interaction determines spatial connectedness, and it is this connectedness that drives important features of the overall system at both local and regional scales. For some processes, the dynamics can be thought to be ‘flow limited’ because colonists are not immediately available to establish populations, or because spatial flows of key materials are limiting to material cycles. For other processes the dynamics can be thought to be subject to ‘flow excess’ because the dispersal of colonists or the flow of materials from other locations can be so high as to overwhelm local dynamics. For example, flow excess can support source-sink population relations (i.e. populations in low quality environments – the sinks – are maintained through migration from high quality environments – the sources; Pulliam, 1988) or provide spatial subsidies of materials (Polis et al., 2004) in the extreme leading to spatial homogenization. Intermediate connectedness wherein colonization is (nearly) immediate and material flows are not limiting to material cycles nor are they so large as to lead to spatial homogenization have their own dynamics: one that leads to closer correspondence between local abiotic environmental conditions and biotic features of communities such as species composition and internally balanced material dynamics. Of course, this means that different processes can have different effective scales depending on the connectedness of the important agents involved.

Finally, metaecology recognizes that critical components of the dynamics of ecological systems are sensitive to, and occur in, spatially restricted locations. For individual species this corresponds to the classic meaning of a ‘local population’ and can be related to the dispersal of individuals. For communities and ecosystems, the spatial definition of ‘community’ or ‘ecosystem’ is less concrete, and indeed served as a long discussed topic (Wiens, 1989). Nevertheless, key elements of the dynamics of populations (e.g. demography and vital rates), communities (e.g. species interactions and responses to local environmental conditions), and ecosystems (e.g. element cycling and stoichiometric linkages among material fluxes) occur over strongly localized extents. The overall properties of the system emerge from the way these localized mechanisms interact with the spatial processes in metaecology.

Metapopulation ecology has been instrumental in developing optimization procedures to design protected area networks for the conservation of endangered species (Hanski, 2004). The false heath fritillary butterfly (Melitaea diamina) historically benefited from traditional agriculture in Finland because small-scale tilling and mowing favours its host plant, the pioneer valerian (Valeriana sambucifolia) (Fig. 2). However, due to natural succession and land use change, the species range plummeted to two small regions where it persisted as a metapopulation in a dynamic network of suitable habitat patches connected by dispersal (Cabeza, 2003). Metapopulation models and dispersal studies identified priority management sites, and motivated the development of applied conservation software for identifying optimal sets of conservation sites accounting for cost (Moilanen and Cabeza, 2002) and spatial dynamics (Cabeza, 2003). This successful application of metaecology highlights how the framework effectively confronts the challenges of developing optimal conservation strategies when suitable habitat patches shift in space and time, and eventually led to an evaluation of how well the spatial and temporal frame of actions of Finnish institutions matched the spatial and temporal dynamics of the species to be protected (Fabritius et al., 2017). Spatial structure, connectedness, turnover (i.e. replacement through time) and quality of local patches are now continuously monitored by local conservation authorities in Finland and used to revise alternative strategies for urban planning. These have been pivotal in guaranteeing the persistence of Melitaea diamina in the country.

Fig. 2.

(a–d) Selected case studies. (a,b) Metapopulation studies of the endangered false heath fritillary butterfly (Melitaea diamina) prompted the development of algorithms for identifying optimal sets of conservation sites accounting for both cost and spatiotemporal dynamics. Accounting for spatiotemporal dynamics was a necessity because the butterfly obligatory host is a pioneer plant (the valerian Valeriana sambucifolia), which distribution shifts as a function of forest succession and land use change. (c) Stream restoration historically focused on the manipulation of local habitat heterogeneity (e.g. reconfiguration of the channel, addition of logs, boulders and gravel), implicitly assuming that species would spontaneously recolonize. It was more recently found that, because contemporary dispersal among streams is very rare, community recovery requires consideration of proximity and active pathways to colonization sources, or deliberate stocking. (d) Pacific salmon (such as the chinook salmon Onchorhyncus tshawytscha) can accumulate in its biomass large quantities of biomagnifiying metals and persistent organic pollutants, and effectively transfer them to sometimes distant ecosystems via upstream migrations and death. Figures courtesy of Kale Meller (a, b), Herbythyme (c; GFDL/Creative Commons; image cropped) and Josh Larios (d; Creative Commons).

(0.9MB).

Beyond a single endangered, umbrella or flagship species, protected area networks should satisfy the requirements of multiple species with contrasting life histories and movement ecologies. Albert et al. (2017) applied the framework of metaecology to prioritize forest remnants for the long-term conservation of fourteen vertebrate species ranging from salamanders to bears in the St. Lawrence lowlands surrounding Montreal, Canada. For conservation to be effective in the long term, the authors accounted for future land use and climate change, and their uncertainties. The solution comprised a combination of well-connected, large forest patches favoring the short-range connectedness that is required for the persistence of metapopulations within the network, with corridors of smaller stepping-stone patches promoting long-range connectivity that facilitates climate-driven range shifts across adjacent networks. The design of such habitat network was a direct request from the Ministry of Sustainable Development, Environment and Parks of the Quebec government and a strong relationship with local stakeholders ensured that these results are now being concretely used in the design of Montreal’s greenbelt.

While the previous case study considered the individual responses of multiple, non-interacting species (i.e., a collection of metapopulations), ecologists and practitioners are sometimes interested in managing particular species interactions or groups of interactions (i.e. interactions networks) that are associated to a particular ecosystem service. This is the case of plant-pollinator mutualisms. Because many crop species depend heavily on animal pollination to produce fruits and seeds (Garibaldi et al., 2013), the increased loss of remnants of natural vegetation in intensively managed agricultural landscapes poses a serious threat not only to biodiversity conservation but also to agricultural yields (Bengtsson, 2009). Metaecological approaches have provided explicit recommendations as to how to manage habitat structure, restore non-agricultural habitat, and/or enhance connectedness between agricultural fields and surrounding natural or seminatural habitats. For example, restoring hedgerows in agricultural landscapes can improve the richness and abundance of pollinators which, by flowing into interconnected fields, may maintain and enhance the yield, quality and diversity of agricultural products (Morandin and Kremer, 2013). Furthermore, recent developments have shown how these interaction networks might also be interconnected across space forming metanetworks (i.e. a set of local interaction networks that are linked by dispersal of multiple interacting species; Emer et al., 2018). Reducing the impacts of agricultural intensification reqs taking into account how such changes might modify interactions at local scales but also how different metanetworks might be interconnected and distributed across space.

Stream restoration has historically focused on the manipulation of local habitat heterogeneity (e.g. reconfiguration of the channel, addition of logs, boulders and gravel; Fig. 2), implicitly assuming that individuals of different species would spontaneously recolonize restored reaches. However, many of these restoration projects failed because target species were either absent in the regional species pool or unable to disperse to restored sites (Palmer et al., 2014). Considering dispersal and metacommunity dynamics is critical to sustain biodiversity in river networks since isolated tributaries receive few immigrants from the main river corridor and most freshwater organisms rarely disperse overland among streams or catchments (Hughes, 2007). Thus, communities in isolated headwaters are predicted to be more vulnerable to environmental change than communities inhabiting well-connected downstream branches. Swan and Brown (2017) tested this idea using metacommunity ecology to study how restoration efforts fared in headwater versus mainstem sites. Their study clearly demonstrates that restoration efforts are context-dependent, enhancing biodiversity and stability of ecological communities in more isolated headwater sites, where local processes play a major role, while similar efforts in well-connected mainstems are largely ineffective because regional processes dominate. Even so, successful restoration of headwaters requires recolonization and maintenance of viable populations in restored stream sites through proximity to colonization sources (Parkyn and Smith, 2011) or deliberate introductions (Stranko et al., 2012). This case study demonstrates that building management actions considering spatial structure and dispersal among target streams can increase the success of restoration activities.

Spatially separated ecosystems may be connected to each other by the flow of energy and materials – including chemical contaminants. Throughout the 20th century, the Laurentian Great Lakes and their basins were subject to heavy discharges of persistent organic pollutants (POPs) including PCBs (polychlorinated biphenyls), PBDEs (polybrominated diphenyl ethers) and DDE (dichlorodiphenyldichloroethylene), which were used in household and industrial products such as electrical transformers, flame retardants, and pesticides. Known primary routes of contaminant dispersal from release sites included volatilization and atmospheric deposition, as well as runoff and downstream transport of water and eroded sediment. Not considered was the spatial redistribution of these POPs by the regular stocking of salmonids for recreational and commercial fisheries (745 million fish stocked in the Great Lakes since 1967; Crawford, 2001). PCBs, PBDEs and DDE biomagnify (i.e., increase in concentration) along food chains and predatory pacific salmon such as Oncorhynchus tshawytscha (Fig. 2) can accumulate high POP concentrations in their biomass and transfer them to tributaries via upstream spawning migrations and death (Janetski et al., 2012). Decomposition and consumption of salmon carcasses facilitate POP flow through local food webs and contaminate resident brook trout, a species targeted by recreational anglers and not regularly assessed for contaminant levels nor included in health consumption advisories (Janetski et al., 2012). Interestingly, whereas an increase in connectedness was a target for environmental managers in all case studies above, here the costs and benefits of increased connectedness must be weighed: considering that hundreds of dams are being removed as part of the Great Lakes Restoration Initiative, we can anticipate salmon-mediated contamination of newly accessible upstream tributaries to increase in this metaecosystem. More generally this case study illustrates that the movement of materials across ecosystems (from lake to tributaries) may have important consequences for human welfare.

These five case studies illustrate how metaecology can help address important environmental issues more effectively across scales and at multiple levels of biological organization, but the wider use of metaecological approaches is still relatively rare in applied ecology. We conducted a formal bibliometric analysis of the applied ecology peer-reviewed literature to (i) quantify the pervasiveness of metaecological concepts and (ii) assess whether explicit metaecological frameworks are being used in studies of environmental change, based on their use of terminology, study design, and literature cited. Our analysis shows that very few of the surveyed papers used or referred explicitly to a metaecological approach (Box 1). Additionally, we analysed published guidelines for mitigating or managing the risk of the key drivers of environmental change, which generally revealed a widespread focus on local scale processes, and a lack of focus on spatial context (spatial hierarchy and heterogeneity) and connectedness (Box 2). Despite scientific studies (Box 1) and management actions (Box 2) seldom explicitly considered metaecology, they frequently incorporated two or all three of its key features. Connectedness (i.e. flow) was the least considered feature in the studies reviewed here, even though it is one of the key aspects to be addressed in management actions. Among drivers of environmental change, studies on land use change were more likely to incorporate connectedness while studies addressing pollution and overharvesting rarely acknowledged it.