Concepts underpinning the planetary boundaries framework are being incorporated into multilateral discussions on sustainability, influencing international environmental policy development. Research underlying the boundaries has primarily focused on terrestrial systems, despite the fundamental role of marine biomes for Earth system function and societal wellbeing, seriously hindering the efficacy of the boundary approach. We explore boundaries from a marine perspective. For each boundary, we show how improved integration of marine systems influences our understanding of the risk of crossing these limits. Better integration of marine systems is essential if planetary boundaries are to inform Earth system governance.
The planet is subject to increasing anthropogenic impacts and is exhibiting global environmental change at an accelerating rate, eroding the natural capital that sustains human wellbeing and prosperity1. The challenge of understanding these large-scale changes and their consequences for human wellbeing led to the development of a set of planetary boundaries by refs 2,3 to guide Earth system governance. These boundaries identify key biophysical limits (Box 1; Table 1); it is proposed that by staying within these limits, humanity may reduce the risk of crossing thresholds that could lead to devastating and potentially irreversible environmental change, ensuring the maintenance of critical ecosystem services2,3.
The planetary boundaries framework has generated significant research interest, particularly within the Earth systems governance literature, for example ref. 4. Moreover, the ideas underpinning the framework have been incorporated into multilateral discussions and agreements regarding sustainability, such as the Sustainable Development Goals5. This level of engagement suggests that the planetary boundaries narrative has the potential to shape future environmental policy6 and technological innovation7.
Planetary boundaries integrate knowledge across the biophysical sciences and have been intimately linked to analyses of the Great Acceleration (see glossary Supplementary Note 1)1. To date, much of the research literature has focused on terrestrial social–ecological systems, with less emphasis placed on marine systems (Supplementary Note 2; Supplementary Fig. 1). In light of (1) the global spatial dominance of marine ecosystems; (2) the fundamental ecological differences between marine and terrestrial biomes8; (3) the increasing human pressures on the world’s oceans9; and (4) the critical role marine systems play in supporting human wellbeing, particularly in developing nations10, this imbalance seriously hinders the efficacy of the planetary boundaries framework in supporting Earth system governance. Here, we provide guidance for redressing this imbalance. We explore research to support characterization of planetary boundaries for a blue planet. We discuss the ways in which the various boundaries interact, and options for assessing these interactions to provide a more integrated and holistic understanding of global environmental change. Finally, we articulate a research agenda to support implementation of the framework to enhance environmental governance.
Characterizing boundaries for a blue planet
To develop the boundaries for a blue planet, we outline how the existing boundaries3 could be amended through integration of concepts, processes and data that are applicable in marine systems (Fig. 1; Table 2). We highlight the potential implications of including marine systems in relation to our risk of crossing specific boundaries, and where important research gaps exist. Here, we explore four boundaries in more depth, highlighting how broadening integration with marine research has significant implications for boundary characterization: (1) land-system change to show how the scope of a boundary might be expanded to encompass marine systems; (2) biogeochemical flows to explore how additional marine perspectives might support more robust tracking of regional issues in a global boundary; (3) biosphere integrity to highlight a key knowledge gap; and (4) human-appropriated net primary production (HANPP), a new terrestrial boundary proposed by ref. 11 to provide an example of terrestrial–marine integration for a boundary that follows a different strategy to the original boundary framework of ref. 12. Our focus on these boundaries is illustrative but also pragmatic, as we believe modifications could be achieved over relatively short timescales because the necessary datasets or underlying knowledge are already in place.
The land-system change boundary addresses links between habitat and climate3. Vegetation cover mediates climate through carbon storage, and by affecting the transfer of moisture and energy at the Earth’s surface13. Habitat change that shifts vegetation type alters carbon sequestration rates, albedo and evapotranspiration (see glossary Supplementary Note 1), and is likely to drive significant climatic changes, with deforestation — particularly of boreal and tropical forests — estimated to contribute most to these shifts (Table 2)14. However, this boundary and the underlying analyses do not account for the influence of marine biomes on climate, including that of ice, seagrass and mangroves (see ref. 15 for the description of vegetation classification that underpins the land-system change boundary), nor do they account for how ocean–atmosphere coupling may counteract the effect of forest loss on climate14,16.
Today, forests represent about 7% of the Earth’s surface17, and potentially up to 13% historically14, an area matched by continental shelves (6.3%) and a percentage far outweighed by the coverage of marine biomes (70.9%). Critically, just as forest biomes influence regional carbon and energy fluxes14, mangroves and other marine systems are characterized by biogeophyscial processes that influence climate18 (Fig. 2). More importantly, this influence is of sufficient magnitude to warrant consideration in analyses of the impacts of habitat change on climate. For example, several coastal marine habitats have the highest carbon sequestration rates of any habitat on the plant (for example, salt marshes: 218 ± 24 g C m−2 yr−1 (mean ± standard error) versus tropical forests: 4.0 ± 0.5 g C m−2 yr−1)19,20. Furthermore, it is estimated that deforestation is driving emissions of 1.2 Pg CO2 yr−1 (ref. 21), whereas degradation of coastal wetlands (mangroves, seagrasses and marshes) alone is estimated to be driving emissions of 0.12–1.0 Pg CO2 yr−1 (ref. 22), despite these wetlands covering <1% of the Earth’s surface. Similarly, the difference in albedo between boreal forest and grasslands (0.08 versus 0.2) is smaller than the difference between sea ice and open ocean (0.1–0.81 versus 0.07)23,24. In light of the large-scale habitat changes occurring in the coastal environment, modifying the land-system change boundary to incorporate marine systems is likely to significantly influence our understanding of the current risk of experiencing large-scale climatic effects from habitat modification.
Expanding the scope of the land-system change boundary to include marine biomes would require an alteration to the existing control variable (forest cover remaining), or addition of sub-boundaries demarcating the loss of marine habitats. Relative ice cover may be a useful sub-boundary across land and sea. Focusing more specifically on habitats unique to the oceans, due to the wide variety of marine biomes, a control variable such as three-dimensional (3D) structural complexity of the habitat25,26 or area of seabed undisturbed by anthropogenic activities such as seabed mining, coastal hardening or fishing27,28 may be appropriate. Biome-specific boundaries have been set for land-system change; similar biome-specific boundaries could be set for marine systems, for example ‘acceptable’ loss of 3D structure on coral reefs may differ to that considered ‘acceptable’ for kelp forests. Such approaches are already getting attention as part of the European Union Habitats Directive and Marine Strategy Framework Directive29, which has put significant effort into determining appropriate components for its aggregate ‘sea-floor integrity’ index (see glossary Supplementary Note 1)30, and the Integrated Ecosystem Assessment process in the USA31, which includes more than 30 potential habitat indicators. Integrating sub-boundaries on land and in the ocean to produce a coherent ‘Earth surface change’ boundary is likely to prove challenging but represents an essential development in the planetary boundaries framework.
While some of the boundaries (for example, climate) are linked to global scale tipping points, others represent processes whereby regional scale change accumulates to such a magnitude that there are global consequences (Table 2). The biogeochemical flows boundary, which is expressed as two sub-boundaries (nitrogen (N) and phosphorous (P)), represents such aggregative regional scale effects2. The P boundary explicitly engages with cross-scale issues by incorporating a regional boundary that recognizes heterogeneity in both nutrient inputs and the absorptive capacity of freshwater systems32,33. However, the integration of regional information is inconsistently applied across systems and in relation to N. Currently, the N and P boundaries purely focus on marine system change on the global scale (see Box 1, Table 1 and Table 2)3. Background marine biogeochemical regimes are highly heterogeneous (horizontally and with depth), driving differences in biogeochemical cycling, primary productivity and trophic pathways34,35. These differences cause spatial variability in the vulnerability of marine systems to anthropogenic nutrient flows36, and suggest the need for a more nuanced treatment of this boundary to account for regional marine effects that are consistent with the existing regional treatment of the P boundary in relation to freshwater systems. Importantly, a vast literature exists exploring the biogeochemistry of coastal and oceanic waters that could underpin such an extension to the boundary. For example, data are available on the export of nutrients from watersheds and submarine ground water9,33,37. Furthermore, both ecosystem modelling and empirical research are supporting regionally derived water quality forecasts36 and increased understanding of the variability in nutrient biogeochemistry within the world’s oceans. The implications for altered primary productivity, food web structure, ecosystem function and resilience, and societal wellbeing have also been explored38,39,40. Given that eutrophication is one of the most frequently observed causes of ecosystem regime shifts globally (see glossary Supplementary Note 1; www.regimeshifts.org), it seems that, of all the boundaries, scientists are best placed to provide quantified values for regional marine biogeochemical boundaries.
Accounting for regional marine biogeochemical flows goes beyond simply producing a more comprehensive boundary for human-derived N and P. Within the marine system, the importance of other elements when considering the biogeochemical flows boundary has been recognized3, but not explored further. The marine biogeochemistry literature could inform the addition of other sub-boundaries such as iron (Fe) and silicon (Si). For example, Fe is regionally limiting within marine waters and Fe budgets are significantly impacted by anthropogenic disturbances39,41. Furthermore, regional enrichment patterns of N, P, Si and Fe within marine systems have broader-scale climatic and biodiversity implications42,43. Integration of these different components into the existing framework would help support an improved understanding of how the biogeochemical, climate and biosphere integrity boundaries interact to delimit a ‘safe operating space’ for humanity.
So far, the focus of the biogeochemical boundary has been on bottom-up anthropogenic drivers such as the addition of fertilizers. However, top-down effects such as the influence of fisheries exploitation on biogeochemical cycles are likely to have impacts that are of sufficient magnitude to affect the behaviour of this boundary44. While the indirect aspects of these effects are currently poorly understood, they are, along with the effects of fishing on ecosystem structure and function, a topic of burgeoning interest. Just as with hunting on land, fishing has the potential to influence geochemical cycling by disrupting ecosystem functioning via the removal of key species, the redistribution of relative biomass across trophic levels and the dilution of some nutrient mediating processes45,46. The magnitude of these effects and how they can be mitigated warrant further research and consideration within the existing boundary framework.
Biosphere integrity: functional diversity
The biosphere integrity boundary was identified by ref. 3 as one of the two core planetary boundaries (along with climate change) as it is central to the state of the Earth system — crossing either of these boundaries may shift the Earth into a new state. This boundary focuses on the persistence and functioning of the biosphere. Persistence is underpinned by genetic diversity, whereas function is determined by the diversity of functional traits (Table 2)3. Characterizing the functional diversity sub-boundary has proved to be particularly challenging because of the lack of a suitable control variable47. The biodiversity intactness index (BII), which estimates the proportion of biodiversity found in intact ecosystems that remains within a corresponding human-impacted ecosystem, was presented as a stopgap measure. Recently, BII has been estimated for terrestrial systems, using modelled intact area biodiversity as a baseline3,48. Yet, there have been no studies estimating BII for marine systems. There is potential to broaden the coverage of BII studies to include the world’s oceans using marine wilderness baselines (for example, ref. 49). However, this may prove challenging for less well studied marine ecosystems, and species-level metrics such as BII are only indirectly linked to function. As a result, we suggest directing efforts towards developing a more appropriate control variable using a trait-based metric that is more closely tied to the functions provided by communities47,50. Size- and trait-based modelling of marine communities suggests one potentially robust avenue for exploring trends in the functional composition of communities from the Holocene into the Anthropocene51. Such an approach has three key advantages: (1) it allows estimation of undisturbed baseline states that are not reliant on wilderness areas that may be subject to anthropogenic disturbance, and as such, are not representative of ‘Holocene-like’ conditions48,52; (2) the relative lack of focus on describing species within marine systems may prove problematic when attempting to estimate biodiversity change using empirical observations alone53; and (3) recent integration of a range of traits into size-based modelling explicitly allows for estimation of changes in function, historically and in response to future anthropogenic impacts54.
Modelling changes in the functioning of ecosystems is an important step, but there is also the need to choose specific indicators for the control variable to monitor changes empirically. This is a complex undertaking, but simulation testing, such as in fisheries indicator research, could be used to understand which are the most informative indicators, to gain insights into current status, trends and thresholds in function55,56. For example, ecosystem models have been used in simulations to test the efficacy of a range of indicators in the context of the effects of fishing and ecosystem state. The models are used to represent the ecosystem and its perturbation (for example, fishing pressure and climate change scenarios) and to generate ‘data’. Indicators estimated from these data are then compared against the trajectories in the ecosystem model to see how well they capture the true levels of change55.
Regardless of the control variable used, accounting for marine systems in the functional diversity sub-boundary is likely to have significant consequences for our understanding of the current state of play. A previous study48 estimated that nearly 60% of terrestrial systems have crossed the proposed functional diversity boundary to some degree, based on BII. Accounting for marine biomes, which dominate the Earth’s surface, may give a very different picture.
Since the publication of the initial framework, HANPP has been proposed as a new strategy that could potentially replace a number of the original boundaries because it is relatively straightforward to measure and integrates many of the other interacting boundaries; primary production is influenced by habitat type, climatic conditions, availability of carbon dioxide, nutrients and freshwater, and in turn supports biosphere integrity11. The studies proposing HANPP as a boundary have focused purely on terrestrial systems11,57, despite marine and terrestrial primary production being approximately equal in magnitude58, research suggesting that similar proportions of productivity flow to fisheries on continental shelves as is appropriated by humans on land59,60, and the oceans being comparable to land as a carbon sink61. However, it should be noted that the productivity estimates focus on slightly different characterizations of primary productivity, for example marine estimates focus on surface waters and do not account for spatial variability in the vertical patterns in primary productivity. Furthermore, terrestrial estimates represent biomass accumulation, whereas those in the ocean represent new production. If HANPP is to be used as a replacement, integrative planetary boundary, addition of marine productivity is key and is achievable with existing knowledge (Supplementary Note 3), although further work will be needed to harmonize terrestrial and marine estimates to ensure they represent equivalent metrics. Nonetheless, preliminary mapping of terrestrial and marine HANPP highlights the highly heterogeneous distribution of society’s appropriation of primary productivity; the regions of least concern, such as the open ocean, and the regions where limits are being approached both on land and in the sea, for example Southeast Asia (Fig. 3; Supplementary Note 3). Where fisheries catches are approaching productivity limits, the problem is likely to be exacerbated by climate change62. Moreover, some of the geo-engineering solutions proposed as technological options for addressing climate change involve the direct modification of marine ecosystems and production, for example via ocean fertilization63, highlighting that planetary boundaries in the oceans may face large and increasingly pressing challenges.
Understanding boundary interactions
The planetary boundaries encompass some that represent structural changes that impact important processes (for example, land-system change that impacts climate), whereas others represent the processes themselves (for example, climate)64. This dual nature of the boundaries was initially a pragmatic approach, but it also highlights the close coupling of Earth systems. The consequence is that the different boundaries cannot be viewed in isolation. Boundary interactions drive considerable uncertainty regarding the direction, scale and rate of change likely to be observed in Earth systems, and the potential reversibility of any undesirable changes65,66. Thus, the modifications we have suggested to the framework need to be explored against the backdrop of these interactions. Highlighting this need may seem redundant, but, while the issue of boundary interaction has been discussed (38% of papers reviewed in Supplementary Note 2), it has rarely been dealt with explicitly (but see ref. 66). As a result, we suggest that modifying the boundaries to more comprehensively account for marine ecosystems should not simply mean setting new, static boundary values. Rather, the planetary boundaries framework as a whole might be better served by boundary ‘windows’ that cover a range of possible values. Such an approach allows for uncertainty around the behaviour of individual boundaries and their interactions with each other, as these influence where boundaries should be best placed to avoid undesirable changes to the Earth system66. Furthermore, it builds on the delineation of a zone of uncertainty around each boundary2. A previous study66 explores how boundary windows might work in relation to interactions between the climate change and land-system change boundaries. This slightly more reflexive version of the planetary boundaries framework does not preclude effective management; rather, it is more in-line with a precautionary approach due to the high uncertainty surrounding the behaviour of Earth system processes67. Moreover, it is directly in-line with adaptive management processes best suited to acting under uncertainty66, where actions and targets are updated as new information becomes available68.
Exploring boundary interactions and uncertainty will be largely reliant on systems modelling that accurately captures biophysical–human feedbacks69. There is a growing suite of ecosystem models arising from both aquatic and terrestrial disciplines, which provide robust ways of investigating the importance of different biophysical processes, and their non-linear interactions and feedbacks across scales70,71. Model ensembles (see glossary Supplementary Note 1) allow for comparisons among model outputs72, providing a powerful approach to explore risk in relation to boundary dynamics, how human responses can modify those dynamics, and our likelihood of crossing interacting boundaries. Management strategy evaluation, a tool initially developed for fisheries management73, is now being used much more widely to explore virtual worlds under a range of different scenarios74. The modelling challenges going forward are many (for example, handling uncertainty, improved capacity to represent cross-scale processes, provision of pragmatic approaches that can be applied with fewer resources), but among the greatest is the inclusion of human influences, and how they have added to the interconnected nature and interdependency of the boundaries.
Marine boundaries and governance
The global environmental governance challenges presented by the planetary boundaries concepts are well documented4, but have been seldom explored specifically for marine systems (Supplementary Fig. 1). Critically, including marine biomes within the framework exacerbates these challenges, as governance necessarily must account for both sovereign and common-pool resources. Furthermore, monitoring and enforcement within the high seas presents a unique set of governance challenges not experienced in terrestrial biomes75. Thus, accounting for broad-scale marine habitat change in an ‘Earth surface change’ boundary and enforcing policies to counteract these changes across remote ocean areas, is likely to be both difficult and contested. As such, we posit that increased integration of marine biomes into the planetary boundary concept may present larger and more immediate challenges to Earth systems governance than currently realized.
Responding successfully to the full suite of unique marine-specific challenges outlined here will require complementary efforts across multiple jurisdictions. This will be best achieved through collaborative, global, multi-actor governance networks that promote collective action, foster learning and nurture trust among diverse stakeholders5. Such participatory governance approaches will also allow decision-makers to capitalize on scale-specific knowledge (for example, local or cultural knowledge), so as to match governance responses to the scale of a problem76. In combination, these attributes enhance societies’ ability to respond adaptively to disturbances across multiple spatial scales77. The successful implementation of such networked approaches to governance will, however, require new and innovative institutional arrangements that actively foster participation and collaboration among stakeholders, and manage power imbalances among different actors78. Research is needed to explore the optimal institutional design for supporting the assessment and integration of planetary boundaries into global environmental governance initiatives.
The changing nature of the planetary boundaries, coupled with high degrees of uncertainty and boundary interactions, also necessitates innovative and forward-looking approaches to governance that allow decision-makers to be proactive in anticipation of global environmental change. In this regard, the emerging concept of ‘anticipatory governance’ may prove important79. Initially developed in the field of technology, anticipatory governance is defined as ‘a broad-based capacity extended through society that can act on a variety of inputs to actively steer society towards desired outcomes’80. This is achieved via stakeholder-inclusive foresight activities, such as scenario development, which allow participants to envision plausible futures, contingencies and consequences, and develop the necessary knowledge base for addressing the social, ethical and policy challenges associated with global environmental challenges81,82. Furthermore, by virtue of its proactive approach to environmental governance, anticipatory governance also allows for the early identification of the core capacities, technologies and enabling conditions that must be developed to underpin the transformative changes required83.
Planetary boundaries research has primarily focused on terrestrial systems, in part because there is a perception that so much remains unknown about the oceans. However, given the fundamental dependence of the global biosphere on oceanic components and processes, this imbalance threatens the integrity of the planetary boundaries approach and the usefulness of conclusions drawn from it. Thus, if we are to advance the framework from a useful heuristic to a set of guidelines for Earth system governance, we need to better account for marine systems. For some boundaries, such as HANPP, increased integration may simply mean addition of marine data. In other cases, expanding the scope of the boundary may be necessary; for example, moving from a land-system to an Earth surface change boundary. Marine research may also help to support characterization of boundaries where there is current uncertainty on how to proceed, for example, biosphere integrity. Moreover, a better understanding of how the boundaries interact is fundamental to operationalizing planetary boundaries, and marine research has important contributions to make in this regard. Most of the data and techniques necessary to initiate this increased integration of marine systems are available, but there is considerable scope for further work (Table 2). For example, here we focus on modification of the existing planetary boundaries. These boundaries are not the only options. Exploration of additional boundaries that describe biophysical processes inherent to marine systems and that are central to Earth system function, such as changes in vertical mixing and ocean circulation patterns, present a fertile arena for new research. Rather than putting the oceans to the side, the important knowledge gaps associated with those biomes signal the need for extensive additional research if planetary boundaries are to inform effective Earth system governance that supports societal well-being on a blue planet.
Terrestrial data used in global HANPP map were downloaded from https://www.aau.at/blog/global-hanpp-2000/. Marine data used in global HANPP map are freely available from http://dx.doi.org/10.4226/77/58293083b0515.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This project is supported by funding from the University of Tasmania and the Commonwealth Scientific and Industrial Research Organisation via the Centre for Marine Socioecology. R.A.W. acknowledges support from the Australian Research Council (Discovery project DP140101377) and E.J.M.-G. acknowledges a Pew Marine Fellowship. Thank you to R. Little for discussions relating to this paper. Availability of data used to produce Fig. 3 is described in the Supplementary Information.