Coupled insights from the palaeoenvironmental, historical and archaeological archives to support social-ecological resilience and the sustainable development goals

Palaeo-data has been extensively used to explore a variety of environmental changes (figure 1; table 2; Mills and Jones 2021), providing regional and global scale information about abrupt change due to both external forcing and non-linear responses to climate change (Williams et al 2011). Investigated changes include natural and anthropogenic vegetational changes (Ruddiman 2003, Kaplan et al 2010, Stephens et al 2019, Ellis et al 2021), temperature (e.g. PAGES2k Consortium 2012), hydroclimate (e.g. Steiger et al 2018), ocean acidification (Hönisch et al 2012), first human impacts on fresh surface water resources (Dubois et al 2018), groundwater variability (Gouramanis et al 2010), disturbance including fire (Mooney et al 2011, Codding et al 2014, Bliege Bird and Bird 2021), changes in pH and eutrophication (Smol et al 2001a), salinity (Smol et al 2001b), agricultural initiation and diversification (Barthel et al 2013, Guttmann-Bond 2010), human colonisation and settlement (Rolett and Diamond 2004, Seara et al 2020), greenhouse gas emissions (Masson-Delmotte et al 2013; indicators 9.4.1 and 13.2.2; table 3) and elemental and particulate contamination (Rose 2015, Chen et al 2016, 2020). These types of environmental changes have affected ancient societies such as the Khmer in Cambodia, the Akkadians in Mesopotamia and lowland Maya of southern Mexico and northern Central America (Weiss et al 1993, Hodell et al 1995, Buckley et al 2010). Although not referenced in relation to the SDG indicators, palaeo-information has already proven useful in water resources management and scenario planning (Smith et al 2007, Phillips et al 2009, Gurrapu et al 2022), stakeholder inclusion (Kerr et al 2022) or in improving risk or uncertainty estimates around extreme events (Lam et al 2017).

Importantly, placing recent extreme events described as ‘unprecedented’ over documented historical timeframes, like for example, the 2004 Indian Ocean Tsunami (Janakew et al 2008) or the southwestern North American megadrought (Williams et al 2022), into a long-term context is crucial for improving analyses of recurrence and/or frequency, magnitude (e.g. Klinger et al 2011, Lam et al 2017, Wilhelm et al 2019, Allen et al 2020). It is also useful for better understanding modes of environmental or social recovery and adaptative resilience (Wingard et al 2017). In this context, palaeo-data also provides the baseline canvas against which to evaluate the degree to which increasing human modifications of the environment have exacerbated hazards and, specifically, their contribution to hazard cascades (e.g. the 2018 Palu Earthquake; Bradley et al 2019).

Operationally, the SDGs rely on a variety of indicators against which to measure progress. Defining appropriate baselines for these indicators can be difficult, with many indicators relying on short-term baselines firmly rooted in the most recent decades. This means they may be premised on fundamentally flawed assumptions that a short and recent period sufficiently represents ‘average’ conditions. For example, Target 6.6 (‘By 2020, protect and restore water-related ecosystems’) relies on a 2000–2004 baseline to evaluate Indicator 6.6.1, ‘Change in the extent of water-related ecosystems over time’, and a 2016–18 baseline to specifically assess the extent of inland wetlands (www.unstats.un.org/sdgs/metadata/files/Metadata-06-06-01a). Target 6.6 is far from being achieved at the global or national levels (Convention on Wetlands 2021, van Denter 2021).

We use these indicators to discuss a number of issues associated with a baseline grounded in short-term data.

To do this, we selected ten areas hosting Ramsar-listed wetlands (www.ramsar.org/) and extracted the average reconstructed hydroclimate data (self-calibrating Palmer Drought Severity Index; scPDSI) from tree-ring based drought atlases (Cook et al 2007, 2010, 2016, Palmer et al 2015, Stahle et al 2016) for a 3° × 3° area around each wetland. For each area we then generated a probability distribution based on 10 000 five year (for the 2000–2004 baseline) and three year (for the 2016–18 baseline) bootstrapped means drawn from the 605 years in common across all drought atlases (1400−2005 CE). For each area, average values for 2000–2004 and 2016–18 were compared with their respective probability distributions to see how unusual conditions for the 2000–2004 and 2016–18 periods were (figure 2).

Zoom In Zoom Out Reset image size

This comparison highlights two key points. Firstly, if it can be assumed that ‘average conditions’ are optimal, these baseline periods are not optimal in many locations (figure 2; Higgs et al 2014, Falk et al 2019). Both periods were very dry for western Mexico, western Tajikistan and eastern Australia. Therefore, on the basis of these baselines, apparent progress (expansion) may occur simply due to the natural occurrence of wetter conditions regardless of any management interventions. Conversely, choosing an abnormally wet a baseline period can lead to conclusions that declines have occurred when in fact a return to drier conditions is simply part of natural variability rather than associated with any management intervention. For eastern Mexico, southern Vietnam, southern New Zealand, eastern China and southern Scandinavia, relative conditions during the two periods differed greatly. These five cases illustrate how high levels of interannual variability, and /or significant influence of multi-decadal climate oscillations—such as in Australia (Power et al 1999, Peel et al 2004)—make it more likely that a five- or three year period will fail to reflect average values. Only for southern Spain were approximately average conditions experienced in both baseline periods in the context of 605 years of data (figure 2). Various hydroclimate reconstructions further demonstrate that more severe and/or protracted droughts and more severe floods than those observed over the past century have previously occurred (Baker 1998, Cook et al 2007, 2016, Wilhelm et al 2013, Palmer et al 2015, Stahle et al 2016, St George et al 2020, O’Donnell et al 2021, Ionita et al 2021, Cook et al in review, and references therein), or, in the case of South America, that recent hydroclimate variability is unprecedented over the past 600 years (Morales et al 2020).

Secondly, a universal baseline for Indicator 6.6.1 ignores the spatial heterogeneity of the impacts of natural climate variability and change (figure 2; Willis and Bhagwat 2009, Peterson et al 2013, Blaquez et al 2015, Dearing et al 2015, Campbell 2016, Falk et al 2019). This may result in potentially unrealistic comparisons across regions and inappropriate policy prescriptions. Regionally specific baselines will better contexturalise risk, and hence vulnerability to events relevant for specific regions (e.g. floods in low-lying areas or variation in major climate systems like ENSO). These two issues demonstrate the importance of considering how boundary conditions change over both temporal and spatial scales when aiming to build resilience (Gillson et al 2021; figure 1). Data over long time frames is also required to assess social-ecological impacts of nested climate events (Harris et al 2018) and projected cascading crises (IPCC 2022).

Moving (Folke et al 2010) and/or baselines premised on periods when the environment has already been heavily altered can also be highly problematic (Falk et al 2019, Gillson et al 2021). For example, palaeo-data over 7000 years indicates that a 1985 baseline against which wetland salinity for one wetland in the Australian Murray-Darling Basin was measured was far too high. This inadvertently contributed to ecological collapse rather than improved resilience (Gillson et al 2021). In some cases, scale-dependent notions of resilience rather than a single reference point may be more appropriate because it cannot be assumed that recent conditions have been optimal for a particular system (Falk et al 2019). Building resilience requires flexibility, an openness to learning and an understanding of the slow variables underlying system dynamics (Folke et al 2010, Dearing et al 2015). Palaeodata can capture temporal lags, internal and external variability to which slow variables respond over long time frames (Wang et al 2012 amongst others) thus providing a clear rationale for the serious consideration of pre-instrumental-era records, especially in relation to SDGs 6, 9, 11, 13, 14 and 15.

As a reference for progress towards the relevant SDGs, establishing appropriate means of measuring progress against indicators has enormous importance. This task requires a sound grasp of spatial and temporal variability across scales and the complexity of direct, indirect and lagged effects upon which global, regional and local processes act and respond to anthropogenic change (indicator 13.3.1). This highlights a need for much greater palaeo-literacy by planners and decision makers, and such palaeo-literacy is an important part of an inclusive education about climate change (SDG indicator 13.3.1; table 3). Improved palaeo-literacy would support development and implementation of global, national, regional and local policies that encompass pre-industrialisation environmental conditions, natural versus non-natural variability and trajectories, resilience and buffering capacities, and rates of recovery post-disturbance (e.g. Rockstrom et al 2009; table 3). Palaeo-data would also be useful in the global south where observational data is scant or of very short duration.

Palaeoclimate data informs us that abrupt changes or reaching major tipping points will have extensive climate impacts. For example, changes in the Atlantic Meridional Overturning Circulation, affect the west African and east Asian monsoons, the Amazon basin, and contribute to heat build-up in the Southern Ocean with cascading impacts on the Antarctic ice-sheet, major fisheries and food production (Dahl et al 2005, Hu et al 2015). By itself, however, palaeoclimate data does not elaborate on the resilience of past societies to abrupt change. Extensive historical and archaeological data from across the Holocene (the last ∼11 700 years) yields significant insight, however (Brovkin et al 2021). ‘Abrupt’ climate change can occur across a variety of temporal (e.g. tens to hundreds of years) and spatial (i.e. local, regional, and global) scales. Additionally, as responses of social-ecological systems to abrupt change can occur over much longer time frames than decadal (e.g. Spate 2019), it is highly relevant to consider a variety of time scales.

The overarching lesson that can be drawn from historical and archaeological records is that social-ecological responses to abrupt change are always context dependent, with vulnerability and exposure to even moderate climate shocks mediated by social and political institutions. They often result in marked social change even if some delay occurs (e.g. Staubwasser and Weiss 2006, O’Brien et al 2007, Hegmon et al 2008, Campbell 2016, Nelson et al 2016, Wang et al 2016, Flohr et al 2016, Allcock 2017, Challinor et al 2017, Danti 2018, Di Cosmo et al 2018, Haldon et al 2018, Bal 2019, Frenkel 2019, Kleijne et al 2020, Yang et al 2019, Peregrine 2020, Burke et al 2021, Degroot et al 2021). Moderate shocks such as the Little Ice Age and Late Antique Little Ice Age were associated with widespread famine and disease, repeated harvest failure in many regions, geopolitical shifts, regional migration, major changes in land use and changing religious inclinations (see Gunn 2000, Høilund Nielsen 2005, Nunn et al 2007, Pfister 2009, Löwenborg 2012, Bondeson and Bondesson 2014, Tvauri 2014, Degroot 2015, Price and Graslund 2015, Büntgen et al 2016, Campbell 2016, Sadowski 2020). Yet, in many other cases, societies proved resilient to abrupt (whether over decadal or centennial scales) climate change (Yang et al 2019 and references therein, 2021, Degroot et al 2021). Through analysis of the cluster of volcanic eruptions occurring between 1637 and 1646, during the final stages of the Thirty Years’ War (1618–1648), Stoffel et al (2022) offer a textbook example of difficulties in attributing political instability, harvest failure and famines solely to volcanic climatic impacts. This example shows that it is time to move past reductive framings in which climate (and environment more broadly) either is or is not deemed an important contributor to major historical events. Below we briefly outline some specific points that repeatedly arise in the historical and archaeological literature that are relevant to the SDGs (figure 3).

Zoom In Zoom Out Reset image size

Our purpose here has been to demonstrate to policy makers and managers that together, palaeo data, archaeological and historical records point to a number of key factors that promote resilience and are relevant to the SDG framework and its implementation. We draw on the cited examples to outline three fundamental lessons from long-term memory.

The first is the much-commented upon friction between SDG8 and part of SDG9 (industrialisation) with the planetary SDGs 6, 13–15 that has flow-on consequences for environmental justice (Hickel 2018, Menton et al 2020, Skene 2021). Evidence from the past shows that expansion of human activity has adversely impacted the environment through desiccation and deforestation, and that these impacts can be amplified by abrupt onset of adverse climate conditions (see Campbell 2016, Cook et al in review, Allcock 2017, Challinor et al 2017, Fei et al 2019, Mischke et al 2019, Stephens et al 2019). Apparently flourishing societies can persist beyond critical environmental tipping points despite their increasing vulnerability to collapse (Allen et al 2019, Weiberg and Finnè 2018, Scheffer et al 2021). A piecemeal focus on achieving individual SDGs ultimately ignores potential conflict inherent within the SDGs themselves and their fragility vis-a-vis climate extremes and natural hazards (Reichstein et al 2021).

Secondly, the SDGs are consistent with a view that social-ecological systems will readily adapt to abrupt climate change and its impacts given technological and economic constraints (e.g. Reilly and Schimmelpfennig 2000). However, the failure by the OECD countries to overcome major challenges to combating climate change, suggests our current direction around the feedback loop is anti-clockwise (figure 1), retarding progress towards several SDGs (cf IPCC 2022). In the past, abrupt climate changes have typically been associated with increased inequality (Scanlon 1988, Sheets 2020), and current climate change is reversing progress made towards greater equity, food and water security and improved health (Romanello et al 2021, IPCC 2022). Incremental changes in climate are also increasingly challenging agricultural potential, equality and health outcomes in many regions (Ramankutty et al 2002, Lesk et al 2016, Challinor et al 2017, Romanello et al 2021, IPCC 2022). Additionally, concerns exist that emissions overshoots will occur due to COVID-19 recovery plans while the epidemic continues to disproportionately affect the most disadvantaged (Romanello et al 2021). Without an applied understanding of long-term impacts of shocks, and long-term trajectories of change, adaptation, collapse and resilience, and why some societies have succeeded or failed in responding to these shocks, the capacity of the SDG framework to improve resilience over medium—long time frames may be compromised (see Quiggan et al 2021).

Thirdly, using universal shallow baselines that do not recognise inherent diversity in social-ecological systems against which to measure progress in relation to specific targets is likely to result in inappropriate measures of progress in many cases, and potentially environmental degradation (SDGs 6, 14–16; Gillson et al 2021). This will especially be the case when processes of change are underlain by long-term variability.

Projections indicate that within 50 years temperatures will move outside the narrow de facto human tolerance envelope of the past 6000 years (Xu et al 2020), emphasising the urgency of combating climate change. Climate change threatens the resilience that increased diversity and inclusion, improved equity and education, improved infrastructure, justice and a healthy physical environment can provide. Even moderate climatic downturns in the past have led to major societal decline. We must therefore ask whether the current configuration of societal and organisational structures and priorities, and changes embodied in the SDGs, sufficiently support actions to provide the resilience and willingness required to successfully address climate change (clockwise direction, figure 1). Or, will that structural configuration, priorities and the scale of climate change, overwhelm the resilience measures embodied in the SDGs (anti-clockwise direction, figure 1)? Our assessment here is a timely reminder of the power of the past to illuminate future directions as the SDGs are being increasingly translated into policies, actions and education agendas (Kelman 2017, Rees 2017, Stewart and Gill 2020). Although such long-term data cannot provide all answers, it does shine a critical light on what has and has not previously promoted social-ecological resilience and informs measures of progress.

In conclusion, we highlight four key messages:

• (a)  
  The relationship amongst climate change, the SDGs and resilience can be broadly considered a positive feedback loop (figure 1). To achieve progress towards the resilience, we need to travel in a clockwise direction.
• (b)  
  Variability and change over long time frames are inherent in natural, and human, systems. It is therefore essential to incorporate the information from the wealth of palaeo-records available into frameworks purporting to measure progress towards resilience.
• (c)  
  Analysis of historical and archaeological records over long time spans and in relation to specific events is critical to informing policies that aim to increase our resilience to the accumulating impacts of change.
• (d)  
  We need to very carefully assess what records of the past tell us about the potential conflict between planetary and some social goals. Where long-term records indicate persistent clashes in objectives, we need to be sufficiently bold to robustly address these challenges in order to avoid promoting an anti-clockwise journey around the feedback loop.