Introduction

From moths to amphibians to rays, a global biodiversity crisis is underway on land and in the sea1,2. As a consequence, many of the services that nature contributes to human well-being have been in decline since the 1970s2. Human use of land and sea, direct exploitation, climate change, pollution and alien invasive species are the main anthropogenic drivers of biodiversity loss3. In response to this crisis, the Kunming-Montreal Global Biodiversity Framework (GBF) set an ambitious plan to protect biodiversity through 4 Goals and 23 Targets4. Within the framework, the role of ecological connectivity (hereafter referred to as connectivity) in delivering societal goals is clearly recognized. Goal A aspires that “the integrity, connectivity and resilience of all ecosystems are maintained, enhanced, or restored, substantially increasing the area of natural ecosystems by 2050”. Connectivity is also an explicitly stated element in Targets 2, 3 and 12, with some elements in other targets being indirectly related to connectivity (e.g. pathogen spillover in Target 5 and managing the introduction of alien species in Target 6).

Although the ocean contains more than 95% of the habitable volume of our planet, all indicators of connectivity included in the GBF have either been proposed or implemented on land or in rivers but not in the sea. A report of expert and CBD (Convention on Biological Diversity) Party consultations proposed relevant indicators for measuring ecological connectivity across the post-2020 global biodiversity framework; it suggested that indicators for “coastal/marine” and “migratory” connectivity were not currently available and needed to be developed (Table 2 in ref. 5). However, many studies provide clear, quantitative measures of connectivity in the ocean for species ranging from macroalgae and invertebrates to vertebrate megafauna, including birds (e.g. refs. 6,7,8). Further, the importance of connectivity is increasingly being considered in marine conservation (e.g. refs. 9,10,11,12).

We advocate for the inclusion of existing, widely used approaches that quantify coastal/marine and migratory connectivity to address the gap in the GBF and avoid a potential spill over to other regional or global frameworks that might adopt or seek to inform GBF indicators. We assess all proposed indicators of connectivity in the GBF for their applicability to marine systems, but also provide examples of methods already practiced by ocean scientists and managers. To provide context, we include an overview of what connectivity is and how it is measured, and some of the differences in considering connectivity on land and in the sea.

What is connectivity and how is it measured?

In its broadest sense, connectivity is described as linkages between spatially distinct populations, communities, ecosystems, or habitats; these linkages are achieved by movement of organisms or dispersal of their propagules, as well as flows of nutrients or energy13,14. As species’ distributions are becoming increasingly fragmented15, replenishment of their populations is imperative for facilitating persistence and promoting resilience16. Maintaining, enhancing or restoring connectivity can promote replenishment. Similarly, the flow of materials and energy among ecosystems will sustain the healthy functioning of our planet. However, increased connectivity can also be harmful, for example by facilitating the spread of invasive species or pollutants.

Two distinct types of connectivity are most frequently considered, each type reflecting different elements of possible connections, and each measured in different ways. In the first type, elements of connectivity are related to physical aspects of habitats or the configuration of a network. These elements are collectively termed structural or landscape/seascape connectivity. In the second type, elements focus on movement and its outcomes (e.g. self-replenishment, animal movement between reproductive and foraging areas, population growth, population or species persistence). These are collectively termed functional connectivity. The type of connectivity to be considered depends on the established conservation goals or targets. Regardless of the type, implementation of connectivity in decision making is constrained by the complexity of representing connectivity in quantitative conservation objectives that are linked to high-level conservation goals (e.g. persistence and recovery)14.

In the context of the GBF, the underlying role of connectivity in meeting Goal A is relatively tangible (i.e. it is feasible to derive quantitative objectives): it is expected to contribute to an increase in the area of different ecosystems, reduce species’ extinction rates or risk, enhance genetic diversity and contribute to the resilience of wild species. However, linking SMART (specific, measurable, achievable, realistic and time bound) objectives and targets to these high-level goals will be challenging. The challenge is exacerbated by the fact that Targets 2, 3, and 12 aspire to “enhanced” connectivity without providing a reference or context that can be used to derive quantitative objectives. Decisions on the relevant elements and appropriate targets for connectivity are not straightforward and can require knowledge that may not exist or even be attainable. An added complexity is introduced by certain fundamental differences in aspects of connectivity between the terrestrial and marine realms.

Connectivity in the sea compared to on land

While movement on land occurs mostly on two dimensions (except for birds; and see ref. 17), movement in the ocean occurs mostly in three directions (i.e. including depth). Being considered a more “open” system, the rates of import and export from particular locations or ecosystems are greater in the ocean than on land18, and often facilitated by ocean currents. Resistance to movement tends to be weaker in the ocean than on land. However, resistance can be associated with specific oceanographic (e.g eddies, upwelling, strong unidirectional currents, temperature fronts) or physical features (e.g. islands, seamounts, mid ocean ridges). Some of these features have sharp boundaries (either physical impediments or strong oceanographic gradients) akin to many on land (e.g. deep sea canyons), while others are more difficult to circumscribe (e.g. the Sargasso Sea). These features may impede passive dispersal, restrict long-distance migration pathways, or define patches of rich feeding grounds, and locations for mating or spawning, and thus increase resistance to movement in the sea.

For many species in the ocean, the ability for and distance of movement (which ultimately determines connectivity) differs among life-history stages. Many invertebrates and fish produce many small, weakly swimming younger stages that are dispersed passively by currents. This strategy results in longer potential dispersal distances (100s-1000s of kilometers) and higher influence of external sources on population replenishment in the ocean than on land18. For migratory species, the distribution, connectivity, and physiological capacity to exploit life in three dimensions can also vary across life-history stages19. For example, young seabirds are often more exploratory than adults because they do not need to return to a nest site. In contrast, young elephant seals may be unable to dive as deep as adults while their lungs develop.

These differences between realms have implications for spatial management and conservation, and thus for the indicators used to measure connectivity in each realm. Overall, the great dispersal potential of many marine species often requires consideration of connectivity and its integration in conservation on larger spatial scales in the ocean than on land. The spatial arrangement of management or conservation areas based on structural connectivity can be designed similarly on land and in the ocean. However, because movement in the ocean is not constrained by the physical continuity of suitable habitat, permeability is greater and resistance lower, making structural connectivity a poor proxy for functional connectivity. Consequently, higher emphasis is placed on metrics of functional connectivity, such as (i) the measurement of transport of propagules and understanding the processes that facilitate or inhibit transport; and (ii) the size of a management or conservation area that would ensure self-replenishment18. For migratory connectivity, emphasis is placed on the links among sites that are important to the life histories of animals (for example, connectivity between reproduction areas and post-breeding areas), and measuring the degree to which breeding populations of a species, and individuals within a population, stay separated or mix throughout their migratory cycles (as defined by ref. 20).

Measuring connectivity in the sea

Seven indicators listed in the GBF have been used or proposed to estimate connectivity directly (Table 1: Bioclimatic Ecosystem Resilience Index – BERI; Dendritic Connectivity Index – DCI; Parc connectedness – Parc; Protected Area Isolation Index – PAI; Protected Areas Network metric – ProNet; Protected Connected Index - ProtConn and ConnIntact). BERI, ConnIntact, Parc and ProNet use environmental maps of biodiversity or habitats to evaluate whether configurations (distribution and distance) of habitats or management units allow for possible connections (structural connectivity). DCI, PAI and ProtConn combine these maps with estimates of dispersal or migratory movement to calculate probability of movement among suitable habitats or management units (functional connectivity). Two other indicators (Biodiversity Habitat Index - BHI; Relative Magnitude of Fragmentation - RMF) do not estimate connectivity directly because they are based on degree of fragmentation, which, although it has an inverse relationship with connectivity, on its own does not estimate functional connectivity.

Table 1 Approaches for measuring connectivity in the sea, including relevant indicators proposed for monitoring progress towards the Goals and Targets of the Kunming-Montreal Global Biodiversity Framework (GBF)
Full size table

The seven proposed connectivity-specific indicators (Table 1) have been developed for and tested in the terrestrial realm, and of these DCI has been applied only to riverine systems. To our knowledge, the applicability of these indcators to the marine realm has not yet been evaluated in the peer reviewed literature. Some of these indicators are developed specifically using global terrestrial datasets of the distribution of biodiversity composition from land cover (BERI, Parc, BHI, RMF), based on which resistance surfaces to mammal movement have been calculated (PAI).

Most of the connectivity indicators proposed in the GBF monitoring framework conceivably could be applied to marine ecosystems with a number of caveats. Global classifications of marine biodiversity have led to delineating 10 s of ecoregions for surface pelagic and mesopelagic zones21,22, yet most of the species’ distributions within these three-dimensional ecoregions remain unknown. Efforts to expand this knowledge in some ecoregions are ongoing23, but a complete description remains elusive in the foreseeable future. Detailed global maps of ecosystem cover of the entire seabed do not currently exist, except for corals (https://allencoralatlas.org/). A complete map of seabed features is targeted for completion by 2030 (https://seabed2030.org/our-mission/), and some of these features may prove useful proxies for associated ecosystem types. More likely is mapping of targeted ecosystems (e.g. seagrass beds, macroalgal beds, deep-water corals) or physical features (e.g. seamounts, upwelling regions) at specific locations or regional scales (e.g. https://pacific-data.sprep.org/dataset/global-distribution-seamounts-and-knolls,24,25). Overall, the indicators listed in the GBF could be used for individual species at local and regional scales, but not for entire ecosystems. On global scales, the indicators could be used for coral reef ecosystems, as well as for migratory species with global coverage such as seabirds (https://www.birdlife.org/) or turtles8. Several of the indicators are based on estimations of resistance surfaces which may not be relevant for all contexts in the marine realm. In the absence of specific measures, in-water distance could be used as a proxy for resistance to connectivity for larval dispersal, although it does not incorporate the asymmetry and variation in ocean flows.

Multiple approaches are used to quantify structural and functional connectivity in coastal and marine ecosystems and for migratory animals (Table 1). Approaches to protecting structural connectivity often focus on important physical features (e.g. seamounts, frontal zones) or areas (e.g. adjacent coastal habitats) that provide connectivity26. Very often approaches in the marine realm target elements of functional connectivity27. Functional connectivity has been estimated for networks of populations, habitats, ecosystems or marine protected areas. Specifically, by calculating the strength and magnitude of connections, the contribution of connectivity to replenishment and viability of populations, and the importance of particular nodes in the network can be assessed9,28,29. Network-related analytical approaches for these calculations range in complexity from simple rules of thumb to numerical modelling depending on availability of biological (e.g. larval traits, migration patterns, genetic structure) and environmental (e.g. ocean currents, maps of physical features) data30,31. Noting that ocean dynamics and, therefore connectivity, vary on the order of minutes to centuries and meters to tens of thousands of kilometers32, indicators of functional connectivity for multiple spatio-temporal scales need to be incorporated into the GBF. Importantly, the anticipated effects of climate change on both biological traits (e.g. larval duration) and environmental conditions (e.g. stratification, ocean circulation) also need to be incorporated into the estimations of functional connectivity33.

Recommendations

To meet the goal of halting or reversing biodiversity loss, we must integrate connectivity more effectively in the Goals and Targets of the Global Biodiversity Framework for all realms. The GBF has a fundamental role in promoting both ecosystem integrity and resilience, and the attention of nations, intergovernmental organizations, ENGOs, practitioners and academia will be focused on informing and measuring the indicators agreed upon for the various targets. The absence of indicators of marine and coastal connectivity and migratory connectivity must be urgently addressed in the GBF. Potential inadequacies resulting from the terrestrial origin of the currently proposed indicators can be tackled by engaging expertise in marine connectivity and taking into consideration the differences between the terrestrial and marine realm. Specifically, the proposed indicators need to be tested on marine systems to identify the taxonomic and spatial scope of their potential application. For example, PAI has been developed specifically for terrestrial mammal movement/migration and its application to marine mammals is untested. PARC (based on ecological similarity among units) and BERI (based on species turnover) have been developed using exclusively global terrestrial biodiversity datasets, which are unavailable for the marine environment. In fact, other than for seabirds or turtles, estimates of connectivity on global scales or for entire communities have rarely been attempted to date in the ocean. Most importantly, we must strive towards the inclusion of indicators of functional connectivity, as they are more relevant for the 90% of the planet beyond the coastlines and afford consideration of the processes involved in movement and migration of individuals, genes and species. Indicators of functional connectivity are necessary for the design of effective networks of protected areas that are resilient to future stressors from climate, habitat loss and resource extraction. Further, functional connectivity can facilitate the establishment of targets that are specific, measurable, achievable, relevant and time bound. Although some of the indicators proposed in the GBF could conceivably be calculated for networks in the marine environment (ProtConn, ConnIntact, ProNet), they currently estimate structural connectivity. One way forward is to adapt indicators to use more metrics that can be calculated from graph theory, such as metrics of centrality of nodes or MPAs in a network (e.g. betweenness or Page Rank) or inflow, outflow, self-retention or self-recruitment (Table 1). Importantly, the lack of marine and coastal connectivity indicators is emblematic of a consistent and pervasive neglect of attention to the global ocean. Given its size, consequent ecosystem diversity, and immense role in providing services to humans and the health of our planet, the global ocean should not be an ocean apart from the Global Biodiversity Framework.