What emergent dynamics have resulted from strong coupling between human activities and physical processes at the coastline?

Engineered changes to a coastline impact natural processes, and in turn these altered natural processes influence future engineered changes (Werner and McNamara, Reference Werner and McNamara2007). Once linked, such mutual influences play out over years to decades. For example, a common form of engineered coastal change, beach nourishment, changes rates and patterns of coastal erosion, which eventually influences the timing and size of the next nourishment event. Other examples include coupling between engineered dunes or seawall construction with natural processes such as dune growth and overwash. Faster timescale processes, however – transient rip current evolution or tourists buying a round of miniature golf – are not dynamically coupled across the human/natural boundary at which the respective systems interact. Nor are long timescale processes, such as tectonically driven coastal change or political revolutions. Strong coupling and associated nonlinear interactions between human activities and coastal processes at intermediate timescales are what provide the dynamical ingredients for potential emergent behaviors. As coastal systems have only been strongly coupled in this way since the wake of the Second World War and requisite empirical records are lacking, emergent behaviors must be explored and investigated with numerical models.

McNamara and Werner (Reference McNamara and Werner2008a, Reference McNamara and Werner2008b) were the first to explicitly model strong human–coastline interactions and show emergent phenomena resulting from that coupling. Emergence resulted from a destabilized response of the human-altered barrier island (relative to its natural counterpart) to impacts of a rising sea level. The instability manifested as episodic cycles of resort development and fortification, with alternate areas of collapse and (re)construction varying in both space and time. Subsequent work showed another form of emergent behavior: chaotic shoreline evolution in a model coupling economically optimized but spatially myopic nourishment cycles with alongshore sediment transport (Lazarus et al., Reference Lazarus, McNamara, Smith, Gopalakrishnan and Murray2011). Although the complete story arc of these emergent behaviors that play out over many decades to centuries – cyclical boom and bust in coastal real-estate, chaotic shoreline change along managed coastlines – has yet to be observed outside a numerical model (McNamara and Werner, Reference McNamara and Werner2008a, Reference McNamara and Werner2008b), we are witnessing a progression through some of the early plot points.

Other strongly coupled human–landscape systems have also shown indications of related dynamics, such as the emergent behavior associated with large and low-frequency disaster events in channelized river systems (Criss and Shock, Reference Criss and Shock2001) and wildfires at the wildland–urban interface (Radeloff et al., Reference Radeloff, Helmers, Kramer, Mockrin, Alexandre, Bar-Massada, Butsic, Hawbaker, Martinuzzi, Syphard and Stewart2018). How unique are coastal examples of disaster dynamics beyond the particulars of coastal settings, or do their dynamics translate across other systems? Are there other forms of human–landscape emergence, and when will we see them manifest – if we have not already?

What is necessary to dynamically influence the coastal system on long timescales, when the future fate of the system is forced by sea level?

Despite significant human-engineered alterations to barrier-island coastlines in time and space, there is nothing that current or near-future technology can do to change the fact that sea level will be rising for many decades to come, and in some places rising very fast – indeed, so fast that various coastal locales around the world will be inundated and perhaps cease to exist, in what collectively will constitute a catastrophic environmental and social disaster. In an irony of policy, risk reduction by hazard defense is likely exacerbating this outcome (Armstrong et al., Reference Armstrong, Lazarus, Limber, Goldstein, Thorpe and Ballinger2016; Lazarus et al., Reference Lazarus, Limber, Goldstein, Dodd and Armstrong2018). There is a stark contradiction between current economically driven engineered practices along peopled coastlines and the long-term inevitability of a rising ocean. This contradiction is delaying necessary planning and decisions regarding how to proactively adapt to the future world of higher sea level (Keeler et al., Reference Keeler, McNamara and Irish2018).

Improving this dire forecast will require a fundamental change in the economic system that drives short-timescale profit extraction from coastal systems (Smith et al., Reference Smith, Slott, McNamara and Murray2009; Gopalakrishnan et al., Reference Gopalakrishnan, Smith, Slott and Murray2011; Gopalakrishnan et al., Reference Gopalakrishnan, Landry, Smith and Whitehead2016). Extractive motive is the constant, long-timescale, goal-oriented process that has come to dynamically dominate the “global” human–coastal coupled system. A change to this long-timescale driver – one that would, for example, prioritize benefits over many generations rather than just slivers of a single one – would fundamentally weaken the currently self-reinforcing positive feedback between risk reduction and short-term market profit (Keeler et al., Reference Keeler, McNamara and Irish2018; Lazarus et al., Reference Lazarus, Limber, Goldstein, Dodd and Armstrong2018; Lazarus, Reference Lazarus2022b) that is distracting us from the planetary environmental drivers that really matter.

Another way to fundamentally change the long-timescale evolution of the prevailing human–coastal coupled system driven by short-term profit considerations would be to promote relational interactions among systemic variables. Relational interactions stand in direct contrast to extractive interactions. For example, having humans promote justice (Kimmerer, Reference Kimmerer2013) for a sand dune or beachscape in the same manner one would a person (Stokstad, Reference Stokstad2022): that is, legally acknowledging intrinsic value (Nordstrom, Reference Nordstrom1990) and affording landscapes some rights (e.g., Kolbert, Reference Kolbert2022) would be one way to create – or restore – a relational dynamic between human and natural entities. This is hardly farcical: consider surfers or coastal bird watchers who have done their part to fight for the sustained existence of surf spots and nesting areas in tidal flats – or, for that matter, Indigenous cultures whose practices of environmental sustainability succeeded for centuries to millennia.

Why do events that should warn us about the future and offer a chance to reset lead to decisions that increase systemic fragility?

Relatively regular events such as coastal storms, hurricanes, sea level anomalies, and high-tide or sunny-day flooding cause destruction of the built environment or interrupt its typical functioning. If we know that events cause disruption and that the frequency of disruption might increase, then why do these events not function as canaries in a coalmine? Paradoxically, destruction along human-altered coastlines often leads to a doubling-down on the built environment – increased rebuilding – that in turn leads to more money, more homes, and more lives impacted by subsequent storms. An example comes from the intensity of the built environment, quantified by building footprint size, along the US Atlantic and Gulf coasts (Lazarus et al., Reference Lazarus, Limber, Goldstein, Dodd and Armstrong2018). Destructive hurricanes ultimately result in buildings that are larger than they were before the hurricane, a phenomena that we – and others (e.g., Godschalk et al., Reference Godschalk, Brower and Beatley1989) – have heard referred to anecdotally as “storm destruction leads to urban renewal.” This phenomenon – to Build Back Bigger – is likely related to Burby’s (Reference Burby2006) “safe development paradox” and White’s (Reference White1945) “levee effect,” where measures meant to mitigate risk from natural hazards tend to backfire and promote further development, and in coastal settings can be observed with respect to beach nourishment (Armstrong et al., Reference Armstrong, Lazarus, Limber, Goldstein, Thorpe and Ballinger2016). Individual buildings or subsets of larger communities may be rebuilt on elevated pilings or with wind-resistant roofs (Highfield et al., Reference Highfield, Peacock and Van Zandt2014) – engineering adaptations intended to reduce short-term fragility – but the longer-term, cumulative, emergent dynamic is one of exacerbated exposure and greater systemic fragility to future hazard.

Whether all of these paradoxes and effects can be neatly collapsed into a single unifying frame remains to be seen. We can think of several hypotheses as to why this increased fragility occurs: in all cases, there is still money to be made (e.g., McNamara and Keeler, Reference McNamara and Keeler2013); there is a threshold in terms of event frequency that has not been crossed or a misconception of true risk (Turner and Landry, Reference Turner and Landry2022); the suppression of actuarially fair insurance uptake because of disaster assistance expectations (Landry et al., Reference Landry, Turner and Petrolia2021); risk tolerance of residents can vary, or a resident’s benefit of living in a place outweighs the risk; migration is complex or not an option (because of reasons that are financial, emotional and/or social); there are emotional and/or cultural reasons to remain (i.e., place attachment; Costas et al., Reference Costas, Ferreira and Martinez2015); the current cultural memory, or perhaps market memory, of past events (Hallstrom and Smith, Reference Hallstrom and Smith2005) is not long. More work could be done to examine systemic fragility along the coastline, explore whether other systems beyond coastal examples express similar dynamics, and investigate the root causes of these dynamics. These issues can also be explored from a climate justice lens (e.g., Hino and Nance, Reference Hino and Nance2021). Answering these questions would likely inform coastline prediction, help us better understand human–coastal coupled systems, and could yield usable information for policy interventions.

How can we best test our ideas and models (numerical, conceptual) beyond the weak “test” of confirming that models match reality?

Models of human–coastal systems often require evaluation to determine if their results are able to offer useful explanations of observed phenomena. Evaluation typically takes the form of confirmation: authors display real-world and model results side by side and discuss the match (qualitatively and quantitatively) using past system states. Note that many coastal models are often developed to understand the future dynamics that could occur under certain sets of possible conditions. A useful exercise might be to develop a platform where future predictions can be tested – either for an entire domain or for key sets of variables. Coastal models could be deployed online so that future scientists could monitor the results in real-time. Just as NOAA provides both tide gauge data and tide predictions and therefore allows anyone to observe, in real-time, the match or mismatch. Similar work has also occurred in the climate modeling community focused on assessing past model predictions (e.g., Rahmstorf et al., Reference Rahmstorf, Perrette and Vermeer2012; Hausfather et al., Reference Hausfather, Drake, Abbott and Schmidt2020). We anticipate that observing how coastal models perform in prediction, and also analyzing in what conditions they fail, would be instructive. Model failure often points to missing processes, missing linkages, or other insights. Displaying real-time predictions is of course fraught: it would need to be clear to users and observers that these are not operational tools for forecasting. But such a service would likely be very useful to future researchers and would be worth any bruising to modelers’ egos. Adjusting our conception of model testing to include the idea of online, continuously running models, where anyone can observe model strengths and weaknesses, could be a worthwhile cultural sea-change for coastal science.

In addition to observing predictions from grid-based models, effort could be invested in determining and tracking a reduced set of emergent variables. The now classic example of this idea from geomorphology is the bedform models of Werner and Kocurek (Reference Werner and Kocurek1997, Reference Werner and Kocurek1999), where bedform dynamics is understood in the context of pattern defects and crestline orientation. It remains unclear how and if coastal models can be distilled to a reduced set of emergent variables. A set of emergent variables could be predicted and tracked through time, plotting them on relevant phase spaces, and then try to observe if trajectories on the phase space match modeled behavior. Furthermore, the observed trajectories of emergent variables could be used to understand the dynamics of the system (i.e., Cristelli et al., Reference Cristelli, Tacchella and Pietronero2015).

What does instability in human–coastal coupled systems look like, and how do we know when the system is unstable?

Sea level will be so high at some future date that many human–coastal systems will be forced to change significantly relative to their current state, and may drive many coastal communities to collapse. Will we know how close to collapse we are? The critical slowing down (CSD) interpretation of impending drastic system change, and its related analytic tool set for detecting early warning signals in empirical data, have been applied to a wide variety of dynamical systems with a mix of success (Wang et al., Reference Wang, Dearing, Langdon, Zhang, Yang, Dakos and Scheffer2012; van de Leemput et al., Reference van de Leemput, Wichers, Cramer, Borsboom, Tuerlinckx, Kuppens, van Nes, Viechtbauer, Giltay, Aggen, Derom, Jacobs, Kendler, van der Maas, Neale, Peeters, Thiery, Zachar and Scheffer2014) and failure (Boettiger and Hastings, Reference Boettiger and Hastings2012; Wagner and Eisenman, Reference Wagner and Eisenman2015). Unfortunately, CSD tools can be overextended beyond their mechanistic utility. Unless the system of interest has a long-term steady state that is a fixed point – more specifically, a system in which all observed variability is imposed externally – then using CSD is akin to diagnosing acute anxiety with a thermometer. If some observed variability arises from intrinsic, internal dynamics, as is characteristic of coupled systems, then CSD tools may not illuminate any early warnings of critical instability. In our context, strongly coupled human–coastal systems are unlikely to be amenable to CSD probes.

So what are some of the symptoms we might expect to observe as human–coastal coupled systems head toward drastic change? And how do we see them in observed data? These systems contain a tangle of nonlinear interactions between human and natural processes, yet of the many complex ways these systems interact their steady state is one that is a small subset of their theoretically possible configurations. To invoke the formal terminology of dynamical systems: human–coastal coupled systems exist in attractors. Some characteristic features of this attractor state are dense populations, significant investment in erosion mitigation, immobile infrastructure, and high property values. For any system to find itself in a steady state attractor there must be dissipative processes acting. Dissipation is an umbrella term for dynamics that reduce differences in system states, which is how a system can find itself in a subset of its possible states (Nicolis and Nicolis, Reference Nicolis and Nicolis1995). If an external perturbation kicks the system away from the attractor, the dissipative processes drive it back. As a stable systemic configuration becomes less stable, a symptom of that change is that dissipation will reduce. There are ways to measure the loss of dissipation (Williams and McNamara, Reference Williams and McNamara2021), but they have yet to be applied to empirical observations from coastal systems – human-altered or natural. Measurement of observed dynamical instability in coastal systems is an intriguing challenge. As dissipation in human–coastal coupled systems is reduced, a qualitative symptom that a system is spending more time outside its attractor might include, for example, cycles of destruction and repair, even during otherwise modest storm events – as occurs along low-lying road networks on reaches of the North Carolina Outer Banks.

What are the dynamical differences between current human practices along coastlines and how humans interacted with coastlines in the distant past?

The key word here is dynamical. Fluvial and tidal meanders were long perceived as fundamentally different physical phenomena, but viewed through the right scaling lens their dynamics reflect strong geometric and kinematic similarities (Finotello et al., Reference Finotello, Lanzoni, Ghinassi, Marani, Rinaldo and D’Alpaos2018). Ancient and pre-modern coastlines of course differed from present-day coastlines in material and societal ways. We are not advocating direct comparisons of practices – a relative accounting of populations and infrastructural footprints and feats of engineering. Rather, what insights into systemic stability and resilience might emerge from Indigenous histories of coastal settings, from coastal and marine archeology, from palaeontological analysis of environmental change over several millennia? (And who will benefit from these insights, and how? What measures will ensure that this knowledge regarding past human coastal alterations, particularly where it derives from Indigenous sources, is not a process of further resource extraction?)

If assumptions of the scientific mainstream get dismantled slowly, slowly, then all at once (Kuhn, Reference Kuhn1962), then coastal science has its own spaces to watch. One is Indigenous fisheries. For example, oyster shell middens are physical relics of socially complex, ecologically intensive fisheries that persisted for millennia (Reeder-Myers et al., Reference Reeder-Myers, Braje, Hofman, Elliott Smith, Garland, Grone, Hadden, Hatch, Hunt, Kelley, LeFebvre, Lockman, McKechnie, McNiven, Newsom, Pluckhahn, Sanchez, Schwadron, Smith, Smith, Spiess, Tayac, Thompson, Vollman, Weitzel and Rick2022). Embedded in their strata, the geographies of their spatial distribution, and in wider contextual evidence related to middens are dynamical signatures indicative of a stable, strong attractor for this social–ecological coupled system: so what were the system states, behaviors, and dynamics that sustained such stability? As conventional management approaches to fisheries management have struggled to deliver long-term sustainability in fisheries stocks (Wilson et al., Reference Wilson, Acheson, Metcalfe and Kleban1994; Acheson, Reference Acheson2006; Wilson, Reference Wilson2006) – or, for some species, failed to prevent ecological disaster (Berkes et al., Reference Berkes, Hughes, Steneck, Wilson, Bellwood, Crona, Folke, Gunderson, Leslie, Norberg, Nyström, Olsson, Osterblom, Scheffer and Worm2006) – there is growing interest in understanding, adopting, and adapting the structures of alternative, apparently long-lived systems. If and how these alternative systems that appear to foster ecological resilience become embedded in or replace conventional fisheries practices remains to be seen – but the apparent shift in discourse toward social–ecological dynamical stability over long timescales is itself an interesting development.

Another space is in the deliberate human alteration of coastal environments, for which archeological analyses keep winding back the clock. The oldest known seawall, dated to 7,500–7,000 before present, sits on the Carmel Coast of Israel, and reflects “the extensive effort invested by the Neolithic villagers in its conception, organization and construction.” However, the authors remark, “this distinct social action and display of resilience proved a temporary solution and ultimately the village was inundated and abandoned” (Galili et al., Reference Galili, Benjamin, Eshed, Rosen, McCarthy and Horwitz2019, p. 1). The abandoned city of Nan Madol, a UNESCO World Heritage Site in the Federated States of Micronesia, includes a high-walled complex of nearly 100 artificial islands and canal system built atop a coral reef flat (McCoy et al., Reference McCoy, Alderson, Hemi, Cheng and Edwards2016; Comer et al., Reference Comer, Comer, Dumitru, Ayres, Levin, Seikel, White and Harrower2019). Nan Madol was a dynastic seat for several hundred years, into the 17th century; the technological means by which the complex was constructed remains unresolved (Pala, Reference Pala2009). Elsewhere, new insights are emerging regarding Māori settlement of Aeotearoa (New Zealand), suggesting rapid responses among the Māori to shifts in environmental conditions (Bunbury et al., Reference Bunbury, Petchey and Bickler2022). Ancient and historical cultural sites are a helpful reminder that human–coastal coupled systems have emerged (and been abandoned) before, with dynamics that may parallel or diverge from modern systems in ways we cannot know if we do not ask.

How will we address the chronic, latent, cumulative problem that even minor destruction along developed coastlines causes significant environmental pollution?

The epigraph Tweet from the official account of the Cape Hatteras National Seashore refers to an event in May 2022, widely shared on social media and picked up by international news outlets, in which two unoccupied beachfront houses in Rodanthe, on the Outer Banks of North Carolina, USA, collapsed and broke apart during a day of heavy but not atypical surf conditions. Another house in Rodanthe had collapsed in February. In both cases, hazardous debris was soon bobbing around hundreds of meters offshore, and washing up on beaches over 20 km away. (Crist, Reference Crist2022; Fausset, Reference Fausset2022; Gleeson, Reference Gleeson2022; NPS, 2022a, 2022b; Price, Reference Price2022). In statements released by the National Park Service, the public was both warned of the hazard posed by the debris field and “invited to help clean up” (NPS, 2022a, 2022b).

These particular houses are only the most recent in a long list of such collapses, and they are hardly unique to the private-property peccadillos of the US barrier coast. When the fragility of market-driven human–coastal coupled systems (see Question “Why do events that should warn us about the future and offer a chance to reset lead to decisions that increase systemic fragility?”) results in their eventual failure, that failure will manifest in part as the abandonment of built infrastructure. An inevitable consequence of abandonment, therefore, is pollution. To clear an abandoned built environment – not a building, but a town, a city – and not replace it with new infrastructure is laughably cost-prohibitive (certainly over politically delicate timescales). That means whatever we see now in the coastal zone will still be there, left to get torn apart by decades of storms: beach houses, with garages full of solvents and paint and weedkiller and septic tanks somewhere under the sand; motel units and hotel blocks and strip malls and box stores; roadbeds and utility wires and storm drainage and everything else constructed that people live in and among (Weisman, Reference Weisman2007). This manifestation of coastal pollution – one derived directly from patterns of market-driven real-estate development on low-lying coastal floodplains – is distinguished from, but not unrelated to, more conventional and ubiquitous forms of coastal pollution, including agricultural runoff, sewage discharge, and the exposure of waste-storage landfill sites deliberately sited in areas prone to coastal erosion (Rabalais et al., Reference Rabalais, Diaz, Levin, Turner, Gilbert and Zhang2010; Nicholls et al., Reference Nicholls, Beaven, Stringfellow, Monfort, Le Cozannet, Wahl, Gebert, Wadey, Arns, Spencer, Reinhart, Heimovaara, Santos, Enríquez and Cope2021; Tuholske et al., Reference Tuholske, Halpern, Blasco, Villasenor, Frazier and Caylor2021).

Much of the medieval town of Dunwich, England – “Britain’s Atlantis” on the eastern of England – sits in the nearshore: a dramatic example, albeit from the 13th century, of coastal abandonment following a series of major storm impacts and repeated disruptions to trade infrastructure (Sear et al., Reference Sear, Bacon, Murdock, Doneghan, Baggaley, Serra and LeBas2011, Reference Sear, Murdock, LeBas, Baggaley and Gubbins2013; Enfield, Reference Enfield2022). In the past 900 years, more than 300 coastal settlements in the North Sea basin have been abandoned as a result of coastal flooding and erosion (Sear et al., Reference Sear, Murdock, LeBas, Baggaley and Gubbins2013). What can we discern and learn about modern human–coastal coupled systems from reconstructing dynamics of abandonment, and the environmental artifacts and evidentiary legacies that remain? And what might reconstructing dynamics of settlement and abandonment teach us about possible future environmental impacts of human–coastal coupled systems?

What externalities exist beyond directly linked interactions between human and natural processes at a given location?

Numerical modeling experiments have suggested complex dynamics arising between neighboring beach towns that nourish out of sync (Williams et al., Reference Williams, McNamara, Smith, Murray and Gopalakrishnan2013; Gopalakrishnan et al., Reference Gopalakrishnan, McNamara, Smith and Murray2017). The experiments essentially demonstrated that a town could get caught out relative to its neighbors, nourishing more frequently, and therefore at greater expense, while its neighbors benefitted from lateral diffusion of nourishment sand for which they did not have to pay: a dynamic of “suckers” (the frequent nourishers) and “free-riders” (the lucky neighbors) (Williams et al., Reference Williams, McNamara, Smith, Murray and Gopalakrishnan2013). An earlier deliberately simplified numerical model of spatially extended nourishment dynamics showed that unless every town alongshore nourished simultaneously, then the system devolved into chaotic patterns of nourishment, such that no town could optimize net benefits from nourishment over time (Lazarus et al., Reference Lazarus, McNamara, Smith, Gopalakrishnan and Murray2011). Another numerical modeling exercise explored the possibility that some towns will be forced by their relative spatial geography to nourish more frequently than others, widening disparities in the sustainability and precarity of towns that can afford to nourish and those that cannot (McNamara et al., Reference McNamara, Murray and Smith2011).

On a planetary scale, these are all relatively local externalities – and they are all economic. Other local externalities are ecological, such as the largely unknown consequences of long-term, repeated beach nourishment on beach and nearshore marine ecology (Peterson and Bishop, Reference Peterson and Bishop2005). But still other externalities are both more diffuse and ensnaring. The economic sector arguably driving archetypal human–coastal coupled system dynamics is tourism, which has two troubling consequences. One is the emergence of a “gilded trap,” in which a single economic sector becomes so lucrative that it displaces all others (Steneck et al., Reference Steneck, Hughes, Cinner, Adger, Arnold, Berkes, Boudreau, Brown, Folke, Gunderson, Olsson, Scheffer, Stephenson, Walker, Wilson and Worm2011; Lazarus, Reference Lazarus2017), resulting in a highly precarious local dependence on a market increasingly exposed to disruptive shock – whether geophysical, such as a natural hazard event, or economic, such as the effectively instantaneous cessation of tourism triggered by the COVID-19 pandemic (Lazarus, Reference Lazarus2022b). Another consequence is the homogenization of the “beach town” – characteristics of the specific location may vary, but the provision of local amenities is largely the same around the world: hotels, condos, restaurants, beach chairs and umbrellas for hire. If all beach towns are essentially alike – and if tourist consumers expect them to be essentially alike – then all beach towns are similarly vulnerable to the same dynamical traps: positive feedbacks that drive negative social and/or socio-economic consequences that themselves reinforce the trapping feedback, making the trap difficult to disrupt (Lazarus, Reference Lazarus2022b). These patterns raise the question of how human–coastal coupled systems are both driven by, and manifestations of, the infrastructure of global value chains (Tsing, Reference Tsing2004; Gereffi, Reference Gereffi2018) – and what that relationship to globalization means for the evolution of human–coastal coupled system dynamics.

How do technological changes impact human–coastal coupled systems?

Coastal infrastructure is an ancient technological phenomenon (Gillis, Reference Gillis2015), but like many symptoms of the Anthropocene, the scale and rate of its present proliferation are unprecedented. The extent of shoreline hardening globally is unknown, but Gittman et al. (Reference Gittman, Scyphers, Smith, Neylan and Grabowski2016) estimate that in the USA, seawalls, breakwaters, and other hard structures have replaced more than half of all natural shorelines. In a forward-looking global analysis, Floerl et al. (Reference Floerl, Atalah, Bugnot, Chandler, Dafforn, Floerl, Zaiko and Major2021) predict a 50–76% expansion of coastal infrastructure within the next 25 years, particularly in the vicinity of coastal urban centers. Bugnot et al. (Reference Bugnot, Mayer-Pinto, Airoldi, Heery, Johnston, Critchley, Strain, Morris, LHL, Bishop, Sheehan, Coleman and Dafforn2021) likewise project a 23% increase in the physical footprint of coastal and marine built structures between 2018 and 2028. These assessments reinforce what Nordstrom (Reference Nordstrom1994, Reference Nordstrom2004), in synthesizing observations of human-altered coastal geomorphology from around the world, saw as the “inexorable transformation of the coast to a human artifact” (Nordstrom, Reference Nordstrom1994, p. 510).

The escalating economic costs (to say nothing of environmental costs) associated with current methods of coastal defenses (Temmerman et al., Reference Temmerman, Meire, Bouma, Herman, Ysebaert and De Vriend2013) – which are there to protect coastal built environments from systemic disruption – are reminiscent of the “cycles of innovation” problem in sustainability science, as described by West (Reference West2017). “To sustain open-ended growth in light of resource limitation” and avoid systemic collapse, West explains, “requires continuous cycles of paradigm-shifting innovations” (West, Reference West2017, p. 416). However, because open-ended growth in human and technological systems is nonlinear – indeed, superexponential – “the time between successive innovations has to get shorter and shorter. Thus paradigm-shifting discoveries, adaptations, and innovations must occur at an increasingly accelerated pace” (West, Reference West2017, p. 418).

In the approximately seven millennia since the advent of the seawall (Galili et al., Reference Galili, Benjamin, Eshed, Rosen, McCarthy and Horwitz2019), the fundamental innovation in engineered coastal protection must be beach nourishment (NRC, 1995) and its variations, such as sediment bypassing by pumping (Castelle et al., Reference Castelle, Turner, Bertin and Tomlinson2009) and meganourishment (Stive et al., Reference Stive, De Schipper, Luijendijk, Aarninkhof, van Gelder-Maas, Van Thiel de Vries and Ranasinghe2013). But beach nourishment – the deliberate replacement of sand from a nonlocal source to mitigate chronic shoreline erosion – is energy-intensive, and sea-level rise will drive up the requisite volumes of nourishment deliveries even where sand is abundant (de Schipper et al., Reference de Schipper, Ludka, Raubenheimer, Luijendijk and Schlacher2021). The further irony of current modes of coastal protection, hard and soft, is that the emissions produced in their creation are contributing to the environmental forcing they are intended to counteract. While the next technological innovation in human–coastal coupled systems is unknown and unknowable in detail, the trajectory of technological innovation may be predictable (Haff, Reference Haff2014). At present, that trajectory appears to be describing an ever-increasing rate of consumption (of physical space, of materials for hazard protection) that demands provisioning – but the technological limits of the system cannot keep pace. In the coming decades, will we witness a finite time singularity in human–coastal coupled systems – that is, when nonlinearly increasing demand for a resource becomes infinite within a finite period of time (Johansen and Sornette, Reference Johansen and Sornette2001; West, Reference West2017)? And in the absence of a technological innovation, will we witness a bifurcation in human–natural coastal systems: the aggressive preservation of some, but the abandonment of many?

Will knowing more about the dynamics of human–coastal coupled systems at intermediate timescales change the seemingly inevitable future?

How much carbon dioxide will be in the atmosphere over the course of the coming century is difficult to predict because that quantity depends on how human activities, energy technologies, and energy markets will evolve in that timeframe. Sea-level rise, however, has been set in motion – and there is no emissions scenario in which sea level will not force some low-lying human–coastal systems into a different kind of existence (Nicholls and Cazenave, Reference Nicholls and Cazenave2010; Wong et al., Reference Wong, Losada, Gattuso, Hinkel, Khattabi, McInnes, Saito, Sallenger, Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Girma, Kissel, Levy, MacCracken, Mastrandrea and White2014; Pörtner et al., Reference Pörtner, Roberts, Masson-Delmotte, Zhai, Tignor, Poloczanska and Weyer2019). Research into dynamics that may play out as communities and societies converge on this critical instability often alludes to potential policy implications. The Netherlands arguably leads the world in integrated, solutions-oriented considerations of human–coastal coupled systems under climate change (Kabat et al., Reference Kabat, Van Vierssen, Veraart, Vellinga and Aerts2005; Kwadijk et al., Reference Kwadijk, Haasnoot, Mulder, Hoogvliet, Jeuken, van der Krogt, van Oostrom, Schelfhout, van Velzen, van Waveren and de Wit2010). Their national policies for “climate-proofing” do not decouple them from the kinds of dynamics we have discussed here – but by engineering to 10,000-year timescales, they have gained themselves more time than most to problem-solve.

Some sense of what might be required for broader policy action to take long-term system collapse into account might be found in fisheries – another strongly coupled human–natural system (Ostrom, Reference Ostrom2009). For large-scale fishing activity subject to market pressures – so, neglecting small-scale locally governed fisheries with minimal record-keeping – not until fish stocks started to decline did policies get enacted to address the possibility of collapse (Smith and Wilen, Reference Smith and Wilen2002). This did not prevent some fisheries from collapsing – the North Atlantic cod fishery, most famously – but catch-limit policies were nevertheless a revelation compared to the predominant attitude of the early 20th century that the ocean contained a limitless supply of fish (Smith and Wilen, Reference Smith and Wilen2002). Transposing this onto human–coastal coupled systems suggests that a direct, problematic signal of instability may be needed to trigger enforceable, actionable policy changes. That signal may need to be as unmistakable as water ponding in coastal streets frequently enough to disrupt profit dynamics.

Markets may begin to signal looming trouble before policy has such a reckoning (McNamara and Keeler, Reference McNamara and Keeler2013). For example, it could be that as amenity value is lost with the encroaching sea or as insurance rates increase, coastal property values will fall. Once this happens and the tax base decreases, a Pandora’s box of infrastructure adaptation problems – all of them expensive – will be without a lid. Scattered communities – and countries – will see this eventuality sooner than others, long before policies are in place to address the circumstances. Echoing fisheries, this will probably be too late to save whichever locale is unwittingly the cod equivalent. However, the possibility remains that policy actions will yet be able to prevent the collapse of many, many coastal communities worldwide (Mach and Siders, Reference Mach and Siders2021) – and perhaps push human–coastal coupled systems toward a new attractor at intermediate timescales that is described by dynamics that are more relational than extractive. We imagine that a new dynamical attractor will likewise manifest at intermediate timescales of years to decades, but the timescale of the transition itself from one dynamical attractor to another – driven by a combination of environmental forcing and market behaviors – is unknown.