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

Many governments and organisations are currently aligning many aspects of their policies and practices to the sustainable development goals (SDGs). Achieving the SDGs should increase social-ecological resilience to shocks like climate change and its impacts. Here, we consider the relationship amongst the three elements—the SDGs, social-ecological resilience and climate change—as a positive feedback loop. We argue that long-term memory encoded in historical, archaeological and related ‘palaeo-data’ is central to understanding each of these elements of the feedback loop, especially when long-term fluctuations are inherent in social-ecological systems and their responses to abrupt change. Yet, there is scant reference to the valuable contribution that can be made by these data from the past in the SDGs or their targets and indicators. The historical and archaeological records emphasise the importance of some key themes running through the SDGs including how diversity, inclusion, learning and innovation can reduce vulnerability to abrupt change, and the role of connectivity. Using paleo-data, we demonstrate how changes in the extent of water-related ecosystems as measured by indicator 6.6.1 may simply be related to natural hydroclimate variability, rather than reflecting actual progress towards Target 6.6. This highlights issues associated with using SDG indicator baselines predicated on short-term and very recent data only. Within the context of the contributions from long-term data to inform the positive feedback loop, we ask whether our current inability to substantively combat anthropogenic climate change threatens achieving both the SDGS and enhanced resilience to climate change itself. We argue that long-term records are central to understanding how and what will improve resilience and enhance our ability to both mitigate and adapt to climate change. However, for uptake of these data to occur, improved understanding of their quality and potential by policymakers and managers is required.


Introduction
Projected increases in the frequency and/or intensity of climate-related extremes and the imminent threat of abrupt changes and tipping points (Cai et al 2016, Steffen et al 2018, Lenton et al 2019, Brovkin et al 2021, IPCC 2021 increase the exigency of understanding the nature of social-ecological resilience to past change. Tipping points represent an irreversible shift from one climate regime to another, and, along with climate extremes and generally abrupt climate change (but not necessarily tipping points), their occurrence will have highly significant implications for adaptive resilience of social-ecological systems (for definitions, see table 1). Adaptive resilience refers to the ability of a system to return to a similar but not identical state to the previous one; an ability to absorb shocks while maintaining function (Folke et al 2004, Walker et al 2004, Peregrine 2021. The 2030 Agenda for Sustainable Development program of action can be viewed as a response to issues impeding progress towards improved resilience. Essentially, it aims to facilitate transformations required to enhance sustainability and implicitly, adaptive resilience (Andrijevic et al 2020), through critical transformations (Sachs et al 2019).
As part of the 2030 Agenda, the sustainable development goals (SDGs) comprise 17 non-legally binding goals (United Nations 2015a) consisting of 169 targets that are assessed against pre-specified indicators. These goals are a mixture of 'planetary' (SDGs 6,(13)(14)(15) and 'social' (SDGs 1-12, 16-17) goals. By design, the goals overlap so as to provide seamless coverage of the key issues facing humanity and the environment. For example, Target 1.5, 11.b and 13.1 cover the remit of climatic and other natural hazards under different guises, Goal 1-Poverty alleviation, Goal 11-Safe cities and Goal 13-Combating climate change. Closely related to the Intergovernmental Panel on Climate Change reports (IPCC 2021(IPCC , 2022, SDG13 specifically pinpoints the need for urgent action to combat climate change and its longterm effects and those of climate-related hazards. It also recognises the need for widespread implementation of the Sendai Framework for Disaster Risk Reduction (United Nations 2015b). Many international conventions, treaties and agreements are aligned with the SDGs (e.g. the Ramsar Convention, www.ramsar.org/).
Ostensibly, achieving the SDGs should improve social-ecological resilience to both abrupt climate changes and the persistent and growing impacts of anthropogenically-induced climate change. However, the impacts of the COVID-19 shock on progress towards the SDGs demonstrates the complexity of interrelationships, conflict even, amongst the goals. While the pandemic has had negative impacts on progress towards social SDGs, planetary health temporarily improved (United Nations 2020) before a rapid return to deteriorating planetary health as economies re-opened (Sachs et al 2021). This raises fundamental questions about the robustness of the SDG framework for improving resilience to anthropogenic climate change (Skene 2021). The fact that taking urgent action to combat climate change (SDG13) presents major challenges to 35 of the 37 OECD countries (Sachs et al 2021) adds to this concern. The interaction amongst SDGs, social-ecological resilience and climate change and its impacts, can be represented as a positive feedback loop (figure 1) in which the direction of flow is mediated by social and political structures and organisation.
Historical, archaeological and palaeoenvironmental data are pivotal to scholarship on the history of climate and society (Guillet et al 2017, Degroot et al 2021. As the only natural laboratory we have, they provide critical insight into responses of the physical environment, social and political organisation, religious practices, diet and agricultural practices to complex and abrupt change (figure 1). We argue that these long-term records can make a central contribution to understanding, and developing measures of, resilience and progress towards resilience (Berkes et al 2000, Folke et al 2002, Gómez-Baggethun et al 2013, Weiberg and Finnè 2018, Petzold et al 2020. Insights from these records should help shape policy approaches to implementing the SDGs, not least because local, regional and national framings of climate change impacts are commonly constructed in light of historical precedents (e.g. the fall of the Roman Empire). A more specific level of utility is the contribution long-term memory can have to developing, or understanding what constitutes, appropriate indicators and baselines for the SDGs (figure 1). This is especially relevant because while the SDGs may be considered multi-decadal in their outlook, the dynamics of physical and social systems are underpinned by 'slow variables' such as, for example, soil health, the education or health system, or water quality. Further, understanding the likely reactions of these slow variables to interventions, or 'fast variables' (Walker et al 2012), also requires information that extends beyond recent decades. It is therefore not possible to build resilience to change, or to adequately identify where thresholds for tipping points exist if these slow variables are not well understood (Folke et al 2010).
Reference to, or integration of, palaeoenvironmental, archaeological or historical records in the formulation of the SDGs or their indicators, however, is currently lacking. Collectively, these records provide warnings of the social-ecological costs of, and stories of long-term social-ecological resilience to, past abrupt change. This long-term data provides policy-relevant information to all three vertices of the feedback loop (figure 1) and their lack of consideration highlights the need to demonstrate how and why they deserve serious consideration by Table 1. Working definitions of terms used in this manuscript. Note that there are a number of different versions of resilience (Walker et al 2004, Folke 2006, Folke et al 2010, Cote and Nightingale 2012, Wilson et al 2013, Fedele et al 2019.

Term Description
Tipping point The passing of a threshold at which small changes can lead to nonlinear change processes driven by internal system dynamics and that lead to a different system state. These changes can, but do not always occur much faster than changes in the relevant forcing (Williams et al 2011, Brovkin et al 2021. Realisation of impacts may take time (Dearing et al 2015, Kopp et al 2016.

Adaptive resilience
The ability of a system to return to a similar but not identical state to the previous one; an ability to absorb shocks while maintaining function (Walker et al 2004 Changes due to factors internal to the system will typically be locally/regionally heterogeneous (Williams et al 2011). Abrupt change may occur over longer (e.g. multi-decadal-centennial) or shorter (annual-decadal) time scales. It may also occur as a result of nested processes or press and pulse pressures (Harris et al 2018) that may be largely due to internal or a mixture of external and internal factors. Slow variables Slow changing variables (relative to fast variables) within a system (Walker et al 2012). Generally controlled by external drivers, but also by intrinsic drivers.

Fast variables
These types of variables control the dynamics of a system (Walker et al 2012). Vulnerability Predisposition to be adversely affected by a change, includes sensitivity/ susceptibility to harm and lack of capacity to adapt (IPCC 2022). Exposure Livelihoods, species, ecosystem, environmental function, service and resources, infrastructure or economic/social/cultural assets that could be adversely affected by change (IPCC 2022).

Figure 1.
Simplified positive feedback loop between the sustainable development goals framework, social-ecological resilience and climate change. Contributions of historical, archaeological and paleo-data are central to understanding past environmental change, including responses to climate change and social-ecological resilience. Palaeo-data can also provide input into developing suitable indicators for some targets. If proceeding in a clockwise direction (red line), achieving the SDGs should enhance social-ecological resilience which then supports action on climate change and limitation of warming which then enhances the ability to achieve the SDGs. However, an anticlockwise (purple line) direction indicates that progress towards the SDGs falters which then negatively impacts on social-ecological resilience and impairs the ability to limit anthropogenic climate change.
Escalating changes and reaching tipping points further undermines the ability to achieve the SDGs.
policy makers and managers. The need for longterm information is particularly acute if the resulting prognoses look beyond the most commonly . 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 shortterm 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 shortterm data.
To do this, we selected ten areas hosting Ramsarlisted wetlands (www.ramsar.org/) and extracted the average reconstructed hydroclimate data (selfcalibrating Palmer Drought Severity Index; scPDSI) from tree-ring based drought atlases ( 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   . 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 lowlying 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, scaledependent 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 palaeoliteracy 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.

The relevance of archaeological and historical information for the SDGs
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 icesheet, 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 socialecological 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 (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).

Learning, experimentation and innovation
Retaining, valuing, expanding and enriching cultural knowledge while encouraging innovation are fundamentally part of the SDG framework (SDGs 4 and 9, Target 13.3 and implicitly, SDGs 2-3, 6, 11-17; figure 3). Together, a wide range of palaeoclimate and archaeological records highlight the importance of learning and innovation. Changes in land and water management practices, crops grown, and technological change across many regions (e.g. the North Atlantic, Middle-East, Mediterranean, South America, Asia, Europe) in response to abrupt climatic downturns or sequences of downturns, changes in seasonality at decadal to centennial-scales throughout the Holocene contributed to resilience of many societies (Szczesny 2016, Marsh et

Diversity and inclusion
As a theme, broadening diversity and inclusion permeates the SDGs, both explicitly (SDGs 4-11, 14-15) and implicitly (SDGs 1-3, 12, 16-17 Sachs et al (2019) outline six critical and multifacetted transformations required to achieve the SDGs. These transformations require interrelated and complex long-term changes and well-coordinated implementation (Sachs et al 2019). In other words, a high degree of connectivity is required for the implementation of the SDGs. Extensive evidence demonstrates the importance of connectivity for resilience through cultivation of extensive trade, migration, knowledge and cultural networks that provided support in times of need ( of connective networks has been associated with a loss of resilience, decreased innovation and diversity and increased conflict (Nunn et al 2007, Hegmon et al 2008, Waldinger 2015, Sadowski 2020, Jariel 2021. Increasingly fragmented landscapes can lead to biodiversity loss from which other impacts cascade (Chase et al 2020). In some cases, however, increased flexibility has resulted in self-serving local elites (Campbell 2016).

Connectivity, flexibility and rigidity traps
Failure to manage complexity and interrelatedness through more favourable times, however, can contribute to rigidity traps ( Similarity in trajectories of societal decline or collapse across multiple societies and time periods highlights the potential dangers of our highly interconnected and interdependent modern systems. COVID-19 and the rapid spread of other pests and diseases pose challenges to this elevated interdependence, increasing our vulnerability to abrupt change (Li 2020). Failure of a single link in highly interconnected trade and production networks can create extensive disruptions, increasing vulnerability to shocks (Challinor et al 2017). Managing levels of connectivity and flexibility is particularly relevant for SDGs 2, 6, 8, 9, 12-15 (figure 3) to avoid promoting short term buffers that simply increase long-term vulnerability and reduce intergenerational equity (Lim et al 2018). High levels of complexity in administrative and implementation structures for the SDGs may be similarly problematic.

Discussion and conclusions
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 flowon 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 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 mediumlong 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 andGill 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.

Data availability statement
No new data were created or analysed in this study.