1.1 Introduction

Ecosystem collapse in response to rapid changes in environmental conditions has been an integral part of the evolution and dynamics of the biosphere and marine ecosystems through geologic time. Paleo records reveal rapid ecosystem collapse, recovery, and shifts in response to gradual and abrupt climate changes over decades to centuries. Documented changes include rapid transitions between forest, shrublands, and grasslands in the northern and southern borders of the boreal biome since the end of present interglacial (Jasinski and Payette 2005; de Lafontaine and Payette 2011), and continental mass mortality and ecosystem collapse due to drought in Australia in the early twentieth century (Godfree et al. 2019). Deep-sea benthic community collapses have occurred in response to millennia-long climate events over the past 20,000 years (Yasuhara et al. 2008)

Current trends in climate change are now dominated by increasing emissions of greenhouse gases from human activities arising from the combustion of fossil fuels, land clearing, and food production (Friedlingstein et al. 2019; Saunois et al. 2020; Tian et al. 2020; Hong et al. 2021). The resulting human-driven climate change, particularly in recent decades, has increased the velocity of change in the climate mean state and the frequency and/or intensity of many climate extremes well above historical levels (Loarie et al. 2009; Reboita et al. 2015; Fischer and Knutti 2016; Frölicher and Laufkötter 2018; Smale et al. 2019; Aghakouchak et al. 2020). These changes are setting the conditions for more rapid and abrupt ecosystem dynamics, some of which are already being observed in many ecosystems across the world, and that are expected to intensify as the climate continues to warm.

Ecosystem collapse is associated with the crossing of critical thresholds not only as a result of gradual climate changes, but more often due to abrupt climate extremes or compounded effects of multiple disturbances occurring at greater than historical frequencies. Although single well characterized thresholds are associated with some abrupt changes, for instance, thermal thresholds for many marine coastal ecosystems (e.g. seagrass meadows, kelp forests, coral reefs), it is the combination of multiple climate, weather, and human disturbances (compound events) that more commonly lead to abrupt ecosystem responses. Compound events might include one or more extreme climate events occurring simultaneously or successively, or the combination of climate extremes with underlaying ecosystem conditions that amplify the impact of disturbances (e.g., vulnerability from pollution), or the combination of multiple progressive disturbances (Seneviratne et al. 2012).

1.2 Defining Ecosystems Collapse

We adopt a broad definition of ecosystem collapse building on previous efforts (Keith et al. 2013; Lindenmayer et al. 2016; Cumming and Peterson 2017; Bland et al. 2018) as when the transformation of ecosystem identity occurs through altered structure, function, or biodiversity, is irreversible or persists for much longer than past dynamics, and happens abruptly or at rates well above historical trends.

Note the difficulty in requiring changes to be “irreversible” or permanent over long timeframes, which requires both the time for that permanency to be established and a clearly defined end point sufficiently discrete to characterize the change of identity (Keith et al. 2013). A “collapsed” ecosystem could have the capacity to recover, and in some instances, active restoration can play a role (Fig. 1.1). However, for those collapse in which anthropogenic climate change is a dominant driver, the long-term prognosis is more likely of a permanent transition. The reason is because global warming is likely to intensify for decades, then stabilize at best in the strongest mitigation scenarios, but not revert to prior levels during this century (Collins et al. 2013).

Fig. 1.1
figure 1

Several classes of temporal trends of drivers or states. The focus of this book is on abrupt (persistent or temporary) and ongoing changes, with underlying long-term and often linear climate trends. Redrawn and modified from Ratajczak et al. (2018)

In characterizing ecosystem collapse, it is important and helpful to mention the framing and features of previously published definitions. (Keith et al. 2013) define ecosystem collapse as a transition beyond a bounded threshold in one or more variables that define the identity of the ecosystem; collapse as a transformation of identity, loss of defining features, and replacement by a novel ecosystem. (Lindenmayer et al. 2016) call for three key conditions to be met, with changes being: (1) irreversible or time- and energy-consuming to reverse, (2) widespread, and (3) undesirable in terms of impairing ecosystem services or major losses of biodiversity. (Cumming and Peterson 2017) take the broadest view of collapse from a biophysical, social-ecological, and complex systems perspective, and suggest that collapse can also be viewed as the opposite of ecosystem resilience with collapse occurring when resilience has been lost. They define collapse through four criteria adapted here for ecological systems: (1) The identity of the ecological system must be lost by the disappearance of key system components and interactions; (2) Loss of identify should happen quickly relative to regeneration times and turnover rates of identify-defining components of the system; (3) Substantial losses of ecological capital occur; and (4) Consequences must be lasting, persisting longer than a single generation or much longer than dynamics typical for the system.

Transitions to ecosystem collapse or components of it that are described in the literature include regime shifts, critical transitions, thresholds, tipping points, hysteresis, alternative states, rapid loss of ecological integrity and state change, among others (Scheffer et al. 2001; Folke et al. 2004; Rocha et al. 2015, 2018; Vasilakopoulos et al. 2017; Munson et al. 2018; Ratajczak et al. 2018; Duke et al. 2019). Although some of these transitions and components of collapse can be clearly ascribed to the ecosystem responses presented in this book, others cannot because the dynamics are too recent and therefore lack the necessary long-term observations; this is particularly true for relatively long-lived ecosystems such as forests.

1.3 Observed Dynamics as They Occur

This book (Canadell and Jackson 2021) presents a collection of case studies across the world with ecosystem collapse-like dynamics over the past few decades and particularly over the most recent decade (Fig. 1.2). We have chosen ecosystem dynamics for which human-driven climate change and variability appear to have been the dominant or at least an important cause of the observed abrupt change, often in combination with other natural and human perturbations. Although it is not always possible to attribute the specific dynamics or collapse events to human-induced climate change, the description of drivers, their trends, and the broader context in which those dynamics occur help to establish causal links to the influence of climate change.

Fig. 1.2
figure 2

A selection of the ecosystems showing collapse-like dynamics found around the world and described in the chapters of the book (Canadell and Jackson 2021). Photo credits provided in each chapter

Each chapter describes and characterizes individual events or regional ecosystem dynamics, the drivers of change, and the unprecedented nature of the change, when possible. Given what is known about the evolution of climate change and other disturbances in each region/ecosystem, authors explore the likely evolution of those ecosystems and the challenge to manage those ecosystems in the face of abrupt dynamics. A key component to understand possible future trajectories is to explore ways to build resilience in those ecosystems or facilitate desirable transitions between states.

For most case studies in this book, however, the collapse dynamics are too recent or still emerging to have observations available long enough to determine whether the changes are irreversible and, therefore, the likelihood of transitioning into an alternate or permanent state, degraded or novel.

The book covers three broad latitudinal regions encompassing most biome types including polar and boreal ecosystems, temperate and semi-arid ecosystems, and tropical/temperate coastal ecosystems.

1.3.1 Polar and Boreal Ecosystems

Chapter 2 (Bergstrom et al. 2021). The section on boreal and polar systems begins with a description of the collapse of an alpine ecosystem on sub-Antarctic Macquarie Island, about 650 km southwest from New Zealand. Initially driven by water stress linked to multi-decadal changes in climate, and followed by a novel pathogen, ecosystem decline has continued for more than a decade (Chap. 2).

Chapter 3 (Olefeldt et al. 2021). Rapid physical and biological changes are occurring in the vast permafrost region in response to the fastest rates of air and soil warming in the world. The thawing of permafrost observed in the region is leading to rapid landscape changes, with surface collapses of several meters affecting the physical substrate, hydrology, and aerobic conditions, to which plant community and ecosystems are rapidly responding.

Chapter 4 (Burrell et al. 2021) addresses a key process causing forest collapse in southern Siberia. The rapid warming of the region and lengthening of the fire season are leading to documented post-fire recruitment failure and replacement of large areas of forest ecosystems with grasslands.

Chapter 5 (Payette 2021) provides important paleo context for ecosystem collapse as a feature of the evolution of boreal ecosystems in North America over millennia, before humans began to alter the long-term climate trends. The interaction of climate, fire, and insects have been important in boreal regions in the past, continue to be today and likely to intensify in the future.

1.3.2 Temperate and Semi-arid Ecosystems

Chapter 6 (Bowman et al. 2021) provides a clear example of rapid transition of a relict alpine ecosystem in Tasmania threated by both climate change and increased fire activity. The chapter focuses on a specific event in which fire destroyed a palaeoendemic ecosystem dominated by the slow growing conifer Athrotaxis cupressoides with individuals hundreds of years old. This fire-sensitive ecosystem, including peat soils, has little or no capacity to survive fire or rapidly re-establish after it.

Chapter 7 (Lloret and Batllori 2021) explores the role of changing heatwaves and drought on forest dieback, particularly for semi-arid and temperate forests, and state shifts to shrubland and grassland ecosystems. The chapter frames ecosystem collapse within the broader context of successional, and novel ecosystems forming due to climate change, interacting with other natural and human impacts.

Chapter 8 (Ruthrof et al. 2021) further explores forest dieback and coastal marine ecosystem collapse in two regions with some climatic similarities, the southwestern United States and western Australia. The selected ecosystems include diverse forest ecosystems with dieback driven by complex interactions among drought, heat, and other disturbances, including fire and insect damage. For the marine ecosystems, coastal and algae blooms are largely driven by abrupt temperature changes brought about by increasingly stronger large-scale climate modes such as the El Niño/La Niña Southern Oscillation.

1.3.3 Tropical and Temperate Coastal Ecosystems

Chapter 9 (Duke et al. 2021) provides an overview of mangrove dynamics and their drivers, and examines the extraordinary mangrove dieback event in 2015 and early 2016 in which a 1500 km shoreline of mangroves in the Gulf of Carpentaria in northeast of Australia synchronously died. It was a complex phenomenon with multiple drivers including sea level drop, and changes in climate and weather.

Chapter 10 (Pratchett et al. 2021) covers mass bleaching events of the Great Barrier Reef in Australia that occurred in 2016, 2017, and 2020. The recurrent mass bleaching events took place with compounding impacts from previous bleaching events and other disturbances showing many characteristics of an ecosystem collapse profile.

Chapter 11 (Garrabou et al. 2021) presents a new analysis of recurrent and large-scale mass mortality events in Coralligenous communities of the Mediterranean Sea driven by marine heatwaves. The analysis covers the recovery of communities over years and explores plausible futures based on climate projections for the Mediterranean.

Chapter 12 (Wernberg 2021) presents the impacts of an exceptional marine heatwave in 2011 in Western Australia over 2000 km of seashore that resulted in the local extinction of 100 km of kelp forest. Turf algae took the place of the kelp forest aided by the elevated background ocean temperatures due to climate change that made recovery difficult. To date, the ecosystem has not recovered.

Chapter 13 (Serrano et al. 2021) presents an unprecedented event in which about 1300 km2 of seagrass ecosystems collapsed as a result of a marine heatwave in 2010/11 in Shark Bay, western Australia. The loss of the seagrass canopy resulted in the erosion of sediments and loss of large carbon stocks over the following 6 years.