Coupled Human and Natural Systems: A Novel Framework for Complexity Management

Abstract

Coupled human and natural systems (CHANS) represent dialectic interaction between human and nature subsystems. This dynamic interaction involves a prominent level of complexity stemming from the uncertain interrelation between the systems and the incorporated subsystems. The complexity within CHANS includes reciprocal effects, nonlinearity, uncertainties, and heterogeneity. Although many researchers have highlighted the significance of understanding the nature of the coupling effect, most of the prevailing literature emphasises either human or natural systems separately, while considering the other as exogenous, despite evaluating the reciprocal and complex interrelations. The current review utilises the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA). It focuses on synthesising the prevailing literature on the CHANS framework in several disciplines, focusing on the approach, findings, limitations, and implications. The review comprises 56 relevant articles, found through Endnote and Covidence database searches. The findings identify the dominant complexity character as reciprocal effects and feedback loops, confirming the complex interactions between human and natural systems. Furthermore, the review provides evidence surrounding the significance of developing an analytical framework that can better explain the complex connections between humans and nature, as it provides a comprehensive understanding of CHANS and their potential impacts.

1. Introduction

As the human population continues to rise globally, so does the impact of people on natural systems. Despite the current restrictions and policy measures that limit humans’ exploitation of nature, many societies’ anthropogenic activities are still damaging the long-standing viability and balance between ecological, social, and biogeochemical systems. To better understand these challenges, further investigation is required surrounding the coupled human–environment interaction, whereby anthropogenic activities alter the environmental system, leading to outcomes that will ultimately affect human societies. In turn, these adverse impacts can potentially lead people to be more conservative and protective of nature, thus causing what is known as complex feedback loops [1]. The complexity of coupled human and natural systems (CHANS), also termed social–ecological Networks (SEN) and human–earth systems stems from the important feedback across social and environmental dimensions. Not only does CHANS science and concept care about the broad and direct interaction between humans and ecosystems; it also goes beyond this to tackle the internal processes and integrated patterns within the system more holistically and unitarily [2,3]. In addition, analysing the complexity of the human environment establishes future relationships between SEN, which guides various models and conceptual approaches [4]. Much research has been conducted on how people feel about the environment. Although the relationship between humans and natural systems has been identified, how to understand and assess this relationship to further predict the future is unclear [5].

CHANS are systems in which human and natural systems interact. Although humans have harmonised with the natural environment since the beginning of human civilisation, the intensity of these interactions has increased exponentially since the Industrial Revolution in the late 18th century [6]. CHANS represent a holistic and integrated perspective to combine human and natural systems and illuminates their complex feedback mechanism. As CHANS involve various natural and human processes interacting, this poses a challenge for researchers to understand and analyse the dynamics of both systems [7]. CHANS highlight complex interactions that generate major implications for the world’s sustainable economic development goals.

Integrated human–nature interactions can be designed through several coupling techniques and those intended to uncover hidden connections such as spillovers and feedback [8]. Understanding the relationship and the associated patterns and their dynamic complex interaction is crucial to address the vulnerability of human society and natural ecosystems, including complex phenomena like climate change and the extreme weather events that are associated with it [2]. The interaction between humans and nature has proven to be complex and involves numerous social, economic, and cultural variables across spatial and temporal contexts [9]. This complexity stems from the dynamic interaction patterns that have been exceptionally intensified over time and are damaging the natural environment and humans’ livelihoods. Therefore, a proper interpretation of the dialectic interaction that is the human–nature nexus could contribute to the establishment of precise and equitable policies and regulations that could govern this complex system. Recently, this concept of CHANS has become a core topic of research and reflects the growing interest in reframing the relationships for better and more effective management strategies.

Moreover, socioeconomic transformation and dynamic changes throughout society have a significant impact on the interaction between humans and nature [10]. This highlights the need to harness complexity through traditional research and put the focus on immersing complex theories into place. Conversely, research frameworks that explain complex theories have enhanced the effectiveness of policy formulations, as they have been enriched by relevant and more accurate modelling that better represents the complex interactions and linkages [10]. As a result, both scientists and policymakers have now acknowledged the significance of integrating a comprehensive understanding of CHANS. Hence, this review highlights and evaluates the recent CHANS models within the context of intricate aspects that can be applied further to an enhanced conceptual framework to develop an analytical framework. In this context, this article aims to provide a systematic review of the recent research that has intended to develop the CHANS framework in several disciplines whilst focusing on the modelling approach, findings, limitations, and implications to uncover complex characteristics. Therefore, the objectives of this review are listed below:

• to uncover complex characteristics using reviewed CHANS frameworks,
• to identify the dominant character using the thematic analysis and Pareto analysis, and
• to identify and evaluate CHANS models used for visualizing complexity characteristics.

The paper begins with a literature review including a descriptive analysis of CHANS research development, in order to identify and present the complex characteristics of CHANS. This is followed by a thematic analysis to identify critical themes, a Pareto analysis to analyse the frequency of complex characteristics, a CHANS subsystem analysis, a summarisation of the CHANS model used for representing complexity and connecting subsystems (flow chart visualization in Figure 1), a critical reflection on the current approaches, recommendations for future research, and a conclusion.

2. Research Method

2.1. Literature Review

A systematic literature review (SLR) was utilized as the research methodology to synthesise the existing knowledge about CHANS. The studies included in this review exclusively delineate the steps and functions of developing a conceptual framework that depicts the structural characteristics of coupled human and natural systems or otherwise a distinct system of human or nature.

The SLR is followed by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework, which consists of four phases: identification, screening, eligibility, and inclusion for review [11]. This work used the PRISMA 2020 checklist, which is provided in the Appendix A (

Table A2. PRISMA checklist (adapted from [70]).

Topic	Description
Title	Identify the report as a systematic literature review
Abstract	Include the title as a systematic review, introduction, methods, results, discussion and other information including funding
Introduction	Describes the rationale in the context of existing knowledge, objectives and the research questions
Methods	Provides eligibility criteria, data sources, search strategy, selection and data collection process, collected data items, bias assessment and data synthesis method.
Results	Describe the detailed process of search and selection process, ideally presented using a flow diagram, study characteristics, biases, and results of the studies including the syntheses process
Discussion	Provides interpretation of the results, limitations of included studies, limitations of the review process, and practical implications
Other information	Provides financial and non-financial support, interests of review authors, availability of data and other materials

). The review was based on the PRISMA approach (Figure 2). Due to the extensive coverage of articles, the study’s article identification was completed using two databases, including Web of Science and Scopus. The initial literature review was based on these keywords/terms: (Complex*) AND (coupled human and natural OR Coupled human–natural OR Human–earth OR integrated human–earth OR social–earth OR human–environment OR social–ecological OR SES) AND system. Methodological restrictions and spatial and temporal scales were disregarded for this search. An initial search was implemented with peer-reviewed articles AND in the year range (2012–2022) AND written in English by using search filters. Inclusion criteria considered addressing the influence of complex coupled human–environment and representing a conceptual or analytical framework. Exclusion criteria consisted of two major factors: (1) not addressing the complex interaction between humans and nature; and (2) elaborating on either humans or nature as a separate system, rather than emphasising the complex connection between them. From both database searches, a total of 5201 articles were identified. The number of articles identified after applying filters was 3858.

The author, title, source, and abstract were exported to Endnote. Next, the exported list was transferred into Covidence for “screening”. Applying the third step of the systematic review process, abstracts were screened for articles related to the complexity of coupled human–environment interaction and 105 articles were identified. The full text of these selected 105 articles was further reviewed by reading the full text. The articles which addressed the influence of coupled human–environment interaction and had either a conceptual or analytical framework were identified as eligible.

Ultimately, applying quality assessment resulted in 56 articles (

Table A1. List of selected articles for the SLR.

Authors and Year	Reference Number
An et al. (2014a)	[6]
Aspinall and Staiano (2017)	[59]
Baker et al. (2022)	[20]
Beckage et al. (2018)	[27]
BenDor et al. (2014)	[29]
Biggs et al. (2015)	[39]
Biggs et al. (2012)	[40]
Bueno (2012)	[24]
Boumans et al. (2015)	[58]
Calvin and Bond-Lamberty (2018)	[28]
Carter et al. (2014)	[47]
Chaffin and Scown (2018)	[45]
Chen and Liu (2014)	[67]
Chin et al. (2014)	[21]
Cockburn et al. (2018)	[43]
Crevier et al. (2019)	[22]
Ding and Wei (2022)	[34]
Espinoza-Cisneros (2018)	[69]
Epstein et al. (2020)	[36]
Filatova et al. (2016)	[55]
Galvani et al. (2016)	[1]
Giuliani et al. (2016)	[66]
Hossain et al. (2020)	[46]
(Howard and Livermore, 2021)	[38]
Huber et al. (2013)	[30]
Hull et al. (2015)	[16]
Iwamura et al. (2016)	[49]
Jeffers et al. (2015)	[23]
Kibria et al. (2022)	[64]
Kline et al. (2017)	[54]
Kramer et al. (2017)	[50]
Li et al. (2018)	[19]
Liu (2017)	[8]
Liu et al. (2015a)	[53]
Liu et al. (2015b)	[52]
Liu et al. (2020)	[10]
Liu et al. (2021)	[61]
Lynch et al. (2014)	[41]
Meyfroidt. (2013a)	[31]
Morzillo et al. (2014)	[44]
Motesharrei et al. (2017)	[17]
Pacheco-Romero et al. (2020)	[65]
Polhill et al. (2016)	[56]
Refulio-Coronado et al. (2021)	[35]
Ruiz et al. (2020)	[60]
Sarkar et al. (2021)	[42]
Shi and Ling (2022)	[33]
Solé and Ariza (2019)	[25]
Spies et al. (2014)	[32]
Stokols et al. (2013)	[48]
Tesfatsion et al. (2017)	[68]
Thornton et al. (2017)	[26]
Tonini and Liu (2017)	[62]
Wandersee et al. (2012)	[37]
Wang et al. (2018)	[9]
Zvoleff and An (2014)	[63]

) to be studied in the review. Quality assessment of the identified 56 articles was established by answering the following questions adapted from [12].

Q1.

Is the research methodology clearly described in the study?

Q2.

Is the data collection method explicitly described in the study?

Q3.

Are the data analysis steps clearly stated in the study?

The articles identified in the previous step were assessed based on the above three questions. If a study completely fulfilled the criteria mentioned in the three questions, then that study was assigned the highest score of 2. If a study partially fulfilled the criteria, then that study was assigned a score of 1, and a score of 0 was assigned to the studies that did not meet the criteria. The maximum possible score for a study was a score of 6 (3 × 2). If the achieved score was greater than or equal to 5, then the study was ranked high quality. Medium-quality studies had a score of 3 or 4 while low-quality studies had a score of less than or equal to 2. Article data, including references, journal name, publication year, CHANS models, and key findings were extracted from the included articles and organized in a Microsoft Excel sheet. A thematic analysis was then undertaken and subsequently grouped the critical themes of the reviewed articles.

2.2. Trustworthiness of Systematic Literature Review (SLR)

According to [13,14], the trustworthiness of the SLR was maintained using credibility, dependability, transferability, and an audit trail. In accordance with the process, credibility was maintained while being transparent about the review process and summarising the step-by-step process to ensure that the process is clear to the reader. Biases were controlled by engaging all the authors in the article review process using Covidence, and Covidence enabled shared access and gaining the approval of all authors towards selecting an article. All the authors ensured that the inclusion and exclusion criteria were applied to finalise an article to be considered. To maintain transferability, a detailed description of the process was provided. For dependability, a clear PRISMA approach, which is a standard literature review approach, was used. Any disagreement was resolved among the authors through discussion. To keep audit trails, the selection criteria were available on Covidence, and the article data were kept in a Microsoft Excel sheet of the collected data to enhance consistency and transparency.

3. Results and Discussion

3.1. Descriptive Analysis of Selected Articles

Annual Development of CHANS Research

CHANS research first set out to investigate whether there is a relationship between humans and the natural system. The biophilia theory argued that human development is more related to the natural system, since humans evolve in nature [15]. In the past forty years, research on understanding the nature of interaction as well as highlighting the complexity and the dynamic links between humans and nature has increased. In early 2000, further research built on this theme and studies on the human–ecosystem relationship showed a spike.

It was evident that the number of published papers increased over time. The selected 56 articles were published between 2012 and 2022 inclusive. Figure 3 shows the publication trend by year. The trend indicates a fluctuation in the number of articles from 2012 until 2022. There was a noticeable spike in the year 2014. From 2014 to 2018, there was a decline in the number of articles. The average number of articles was around five per year. The years with the lowest number of articles were 2012 and 2019, with two articles each.

The categorization of short-listed articles consisted of six major types: integrated models/conceptual models, reviews, coupling examples, commentaries, linking tools, and comparative analyses (Figure 4). The integrated models/conceptual models comprised the articles with CHANS modelling and conceptual framework. Review articles represented reviews on CHANS from the prevailing literature. Articles with coupling examples described the methodology for modelling frameworks. The commentaries provided commentary related to CHANS and modelling approaches. Linking tools consisted of processes for modelling frameworks, and the comparative analyses compared various CHANS models and their frameworks.

3.2. Critical Themes Emerged from Thematic Analysis

3.2.1. Complexity Characteristics

• Reciprocal Effects and Feedback Loops

One of the critical components of CHANS emphasises identifying feedback among human and natural systems, which forms when impacts from one system on another system feedback to impact the initial system [16]. Feedback effects of human and natural systems have shown how changes in the earth system, in turn, have reciprocal effects on the human system and vice versa [17]. Reciprocal effects have become the source of one of the greatest modern challenges, because it is unclear how to feed a growing population while sustaining what remains of biodiversity and ecosystem services [1], Systems with bidirectional feedback often result in nonlinear dynamics that can result in unexpected outcomes [17]. Additionally, the coupled interconnections and feedback between human and environmental systems exacerbate environmental degradation [18,19]. COVID-19 is also an outcome of feedback loops and cascading interactions among the connected systems [20].

It is evident that to understand the dynamics of either system, earth system models are required to be coupled with human system models through bidirectional couplings representing the positive and negative feedback that exists in the real systems [17,21]. The growth of the human system has changed its relationship from an empty world to a full world because the human system is much larger than it was, and both systems are essential to be coupled interactively to account for their feedback on each other. CHANS explicitly address complex interactions and feedback between human and natural systems. In comparison with traditional ecological research, which often neglects the human impacts, CHANS evaluates both human and natural components, including their interaction [2]. People and nature interact reciprocally and form feedback loops. Attempting to conceptualise linked processes enhances the that the two are free to interact, each providing feedback to the other [22]. The feedback mechanism depicts the changes in systems that occur through negative and positive feedback loops that either minimise or amplify the response in a variable to maintain the stability of the system [23]. These cannot be viewed directly but may be hypothesised from current knowledge and tested using time series. Despite differences across the study areas and approaches/models used, there is a shared common thread in identifying feedback that helps to characterize the complexity [16]. Understanding the contribution of human–nature feedback to system resilience and adaptability is also significant within the context of the complexity of the systems [24,25]. For example, residents in Wolong use wood for cooking and heating, and they collect wood from the bamboo forests which are the staple food supplier for endangered giant pandas. The bamboo wood collection has led to substantial deterioration in forests and panda habitats. Furthermore, ref. [26] examined how feedback between climate and land will alter energy, agriculture, land and carbon in the future, while [27] explored how interactions between the risk of extreme events and behaviour could alter future climate change. Changes in human behaviour can influence emission levels, leading to feedback effects. In addition, the feedback can contribute to changes in Global Mean Temperature (GMT) which result from human actions [28].

Development near private forest lands improves future development by opening up markets and impacting landowner attitudes positively [29]. For example, socioeconomic changes and climate changes may accelerate the separation between closed forests and open grassland in ecosystems, which will then lead to losses in biodiversity [30]. The role of feedback in land-use change, in which humans and/or organisations change behaviours while responding to changes in terms of the accessibility of land, in turn, determines future land-use patterns [31]. Another example is when fire suppression was initiated in some forests; this strategy influenced increased fuel loads and greater fire risk over the long term [32].

Social–ecological systems (SES) is another term used to describe human and nature interactions, and consists of a few key subsystems such as resource units, users, and governance systems. Analysing SES is necessarily prone to the studies of feedback effects of the systems [33]. SES is a complex adaptive system that illustrates how over a long period, internal interaction between social, economic, and ecological subsystems will be changed [34]. SES helps to explore interdisciplinary insights into complex human and natural interactions [35]. SES characterises the structure of complexity by analysing a wide range of social, ecological, and institutional factors [36].

In addition, it is important to examine people’s perceptions of the environment and livelihood to better understand CHANS-related complexities such as feedback loops. For this reason, integrative approaches are of high use in the pursuit of understanding the complexities of CHANS [37]. The understanding of complex interactions between human society and the environment is significant because it facilitates a well-balanced policy-making approach. Feedback, as a complex characteristic, is vital among the others and should be integrated into CHANS modelling to enhance accuracy [38].

• Surprises/ Uncertainty

Uncertainty is an integral characteristic of CHANS. First, CHANS themselves are continuously evolving and changing in response to external and/or internal changes. Second, uncertainty stems from interactions between the components of the system. A third source of uncertainty arises from social dynamics, which play a critical role in predicting desired social and ecological outcomes [39,40,41]. When complexity is not understood, people may be surprised at the outcomes of human–nature couplings [2]. It has been found that the feedback effect has the potential to alter both human and earth systems; however, there is significant uncertainty in the outcome due to both human and natural system dynamics [28]. The nature of reciprocity has led to unprecedented outcomes. Uncertainty is also an important challenge in producing future behaviour and scenarios with CHANS models [17]. Given the complexity of CHANS, the emergence of unprecedented uncertainties is inevitable during a pandemic, such as COVID-19. Uncertainties mainly arose due to limited current knowledge about the virus and how people across the globe would modify their behaviours in response to the pandemic [42]. Surprises due to uncertainties are the major reason for the emergent requirement of managing CHANS following an interdisciplinary approach [43]. The aforementioned feedback can cause surprises to take place. For example, backyard habitat restoration initiated by humans to enhance wildlife presence will not only cause positive feedback, with increases in bird presence, but can also increase unwelcome animals such as bears. This surprising impact can in turn promote negative feedback in the future [44]. Complex coupled SES poses a significant risk to global sustainability due to the increasing uncertainties facing the entire world [45]. Modelling human–environment systems presents challenges mainly due to the uncertainties in system components [37].

• Nonlinearity

Nonlinearity stems from reciprocal relationships. Nonlinear systems often show dynamics that would be missed if bidirectional interactions between subsystems were not modelled [17]. Furthermore, besides the dynamic relationships such as feedback in CHANS, identification and understanding of other dynamic relationships such as nonlinearities are required. For example, it is important to investigate how long vegetation cover can reduce run-off for before it also reduces resilience, which can cause sudden and massive floods in response to climate change [46].

• Resilience

Coupled human–environment systems have the significant characteristic of adaptability, which includes the capacity to avoid thresholds and adapt to changes in systems [19,47]. CHANS are not just complex, but also adaptive [1]. The vulnerability associated with feedback effects should also be fundamentally included in CHANS studies to identify the adaptive behaviour of both the social and human systems [28]. The ability of dynamic models such as CHANS to capture various interactions of complex systems, their potential to adapt as the real system changes, their ability to model coupled processes, and their flexibility to incorporate other approaches render them as versatile systems [17,48]. Furthermore, the resilience of human systems depends on the adaptability of socio-economic and environmental conditions. Low levels of resilience can set the systems up to fail in response to sudden, surprising, changes in ecosystems that are difficult or impossible to reverse [2].

The capacity of human and natural systems to absorb disturbance and rearrange their function, structure, identity, and feedback refers to resilience. Maintaining resilience is an important prerequisite for ensuring the longer-term sustainability of a human–natural system [48]. For instance, the dependence of the human population on natural resources is predominant in maintaining the stability and resilience of both systems [49]. Resilience is the main concept required to understand how SES respond to changes by maintaining the status quo amid a change or disturbance [35].

• Heterogeneity

Human–nature interactions vary across space, time, and organizational units [2]. Similarly, CHANS are not static, nor perpetual, but change over time. Human and natural systems are complex and heterogeneous, and this heterogeneity impacts on understanding processes, causes, effects, and various mitigating factors of human and natural systems across spatial and temporal scales. This becomes a distinguishing feature of CHANS research [50]. Modelling CHANS is more indirect and complex in developed countries than in developing countries. This is especially because of the growing human population and their heterogeneous inputs towards the natural system [2].

3.2.2. Modelling Challenges Due to Complexity

A major challenge in CHANS research is to develop analytical methods that effectively depict the interactions between human and natural systems considering temporal and spatial organizational scale variations. The challenges consist of combining dissimilar data (ex: biophysical/social, quantitative/qualitative [50]. Linking the natural and human components can also be a challenge when incorporating spatial heterogeneity [46,51]. One of the other challenges in integrated human–earth system modelling is selecting the appropriate level of complexity. Models can range from simple to highly complex, and finding the right balance is crucial for an accurate representation of the system dynamics. Describing human behaviour and its relationship with the Earth system is a complex task. Human behaviour is influenced by various factors, and capturing this complexity in models is a challenge. Data availability and consistency pose challenges for integrated modelling. Integrated models require data from multiple sources, and ensuring the quality and consistency of these data can be difficult [28,52]. Growing populations continuously pose a significant risk to sustaining the ecosystem balance, and this has become a challenge in CHANS modelling as well [1,53].

Reciprocal interactions as a complex characteristic present a major challenge as to how to represent human behaviour as influenced by natural factors, and how to combine this into CHANS models. Furthermore, distilling complex human system processes into a simple model is a difficult task [54]. Uncertainty is another complex characteristic posing a significant challenge to modelling future scenarios in models. The complexity of CHANS showcases the importance of a holistic approach to understanding and managing a system to endeavour to predict surprises stemming from uncertainties. Critical sustainability challenges require strong collaboration between earth and social scientists to develop effective CHANS models that will ultimately result in policies and measures [17]. Most CHANS show nonlinear behaviour and feedback between social and environmental systems, which can be considered a significant challenge for modelling [55,56].

3.3. Pareto Analysis of Complexity Characteristics

Pareto analysis is a statistical tool for ranking data by the frequency of occurrences ordered in descending order. Pareto charts also classify data using the “80/20 rule” if 80% of defects result from only 20% of probable causes. The “vital few” factors represent a substantial amount (80%), while the “useful many” factors represent the remaining 20%.

Pareto analysis was applied in the present study to categorize data from the literature based on the frequency of occurrences from most frequent to least frequent [57].

Table 1. Summarizing the complexity aspects/examples of reviewed CHANS models.

Complexity Characteristics of CHANS	Frequency	Relative Frequency	Cumulative Frequency
Reciprocal effects and feedback loops (1) The impact of hazards increases adaptability (balancing loop). (2) Increasing the level of income impacts land use changes (reinforcing loop). (3) Hazard damages decrease the level of income (balancing loop). (4) Economy level impacts on increasing population and vice versa (reinforcing/ balancing loop). (5) The economic level increases health and population (reinforcing loop). (6) COVID-19 restrictions impact on reducing CO2 emissions. (7) COVID-19 restrictions impact on reducing fuel consumption. (8) COVID-19 restrictions impact on reducing environmental noise. (9) COVID-19 restrictions increase the quality of air and water. (10) COVID-19 restrictions’ impact on increasing CC (climate change) will reduce biodiversity including natural habitats. (11) Conceptual CHANS can directly analyse the reciprocal interactions of humans and nature. (12) Humans and nature interact with each other and create complex feedback loops. (13) ABM examples show a reciprocal effect on the intense use of fuelwood, which diminishes natural habitats. (14) CHANS research identifies the reciprocal interactions between human and natural systems. (15) CHANS represent reciprocal relationships and cross-scale interactions. (16) There are types of both indirect and direct feedback, which are intertwined with each other. (17) CHANS involves complex interconnections and feedback among human and natural subsystems. (18) CHANS explicitly recognizes the interactions and feedback among human and natural systems. (19) CHANS reveals dynamics with reciprocal feedback loops. (20) Telecoupling toolboxes can describe and quantify dialectic socio-economic and environmental linkages. (21) Earth-system modelling consists of both one-way and two-way feedback systems. (22) One of the key aspects of CHANS frameworks is to analyse feedback among human and natural systems. (23) Human–nature feedbacks constitute a key component of coupled systems. (24) Considering reciprocal causality or feedback is a key to understanding CHANS. (25) Feedback combines the human and natural systems. (26) Complexity in CHANS can be understood through feedback. (27) CHANS highlights reciprocal interconnections among biophysical and socio-economic variables. (28) Bidirectional coupling represents the feedback effect. (29) Intracoupling frameworks depict reciprocal effects, including feedback, in CHANS. (30) CHANS are identified as having reciprocal and feedback effects. (31) Social-ecological interactions are identified with feedback effects. (32) SES should break down into interconnected nodes to identify the complex effects such as feedback effects.	32	0.4000	0.4000
Nonlinearity (1) Nonlinearities of each human and environmental subsystem are addressed within the coupled human–environment system resilience. (2) Human activities have a nonlinear impact on natural habitats. (3) The relationships within and among coupled systems are nonlinear and there exist threshold points. (4) ABM examples describe that beyond the threshold fuelwood collection, natural habitats are unresponsive. (5) CHANS emerge with nonlinear characteristics. (6) Considering multiple spatial scales when integrating systems identifies their nonlinear relationships. (7) CHANS describes the human–environment system as a nonlinear complex system. (8) Incorporating feedback into CHANS modelling increases the nonlinearity. (9) Bidirectional feedback often produces nonlinear dynamics. (10) Metacoupling frameworks can address complex features such as nonlinearity. (11) Social–ecological interactions are nonlinear. (12) CHANS are identified as having characteristics that impact on the system structure, including nonlinearity. (13) Social–ecological systems management is defined as a “wicked problem” due to complex characteristics such as nonlinearity. (14) Evaluating SES interactions requires complex thinking due to inherent nonlinear characteristics.	14	0.1750	0.5750
Resilience (1) The conceptual framework describes the processes that should be applied when developing climate policies to reduce risks and increase the level of adaptability. (2) The model can identify the priority areas for policy development when planning to avoid crises like COVID-19. (3) The model can represent flows and processes that cause system changes in which spatial, temporal, and organizational scales can be analysed. (4) The telecoupling process can link distant coupled systems to develop complex and integrated systems that can be guided for the implementation of policies. (5) The ability to retain the balance of human–nature coupling. (6) ABM examples show that land use change can alter the fertility rate of people, flora, and fauna. (7) CHANS facilitate concepts for assessing vulnerability and enhancing adaptive capacity to increase resilience. (8) CHANS show dynamics linked to resilience. (9) Adaptive capacity describes the level of resilience. (10) CHANS are identified as having characteristics that impact on the system structure, including resilience. (11) Resilience is a characteristic of socio–ecological systems. (12) Complex SES should embed resilience thinking.	12	0.1500	0.7250
Surprises (Uncertainty) (1) Short-term socio-economic shocks (financial crisis). (2) Biophysical shocks (floods, precipitation). (3) Uncertainties about COVID-19 due to the limited knowledge about the virus and absence of a vaccine. (4) System, process, and scale dynamics. (5) Surprising outcomes from the human–nature coupling. (6) Reciprocal interactions often result in surprises. (7) CHANS show dynamics with surprises. (8) Feedback can cause surprises which highlight the complexity of CHANS. (9) Metacoupling frameworks can address uncertainty. (10) CHANS are identified as having characteristics that impact on the system structure, including surprises. (11) Social–ecological systems management is defined as a “wicked problem” due to complex characteristics such as uncertainty.	11	0.1375	0.8625
Heterogeneity (1) Conceptual models evaluate the geographical heterogeneity and variability in environmental conditions. (2) Drivers of the urban community, which represents both human and environment subsystems, have various responses and resilience disturbances; these should be addressed within urban resilience transformation management contexts. (3) The telecoupling process followed by the initial conceptual framework addresses the heterogeneity of the human system, including socio-economic and demographic characteristics and vice versa. (4) There can be substantial variables inside both human and nature subsystems. (5) The ABM example represents the heterogeneity of all variables inside each sub model. (6) CHANS, including their interactions, are complex and heterogeneous. (7) Recent development of Integrated Assessment Models (ISMs) has tended to expand system boundaries to increase the heterogeneity of components. (8) ABM is able to investigate CHANS’ heterogeneity. (9) CHANS characteristics tend to be heterogeneous. (10) CHANS are identified as having characteristics that impact on the system structure, including heterogeneity. (11) Heterogeneity throughout the world impacts on the complexity of SES.	11	0.1375	1.0000
Total	80	1.0000	1.0000

below summarizes the reviewed articles including complexity aspects under the main complexity characteristics to scrutinize characteristics evident in the short-listed CHANS models. Frequency refers to the number of occurrences, that is, the summation of complexity aspects and/or examples under each characteristic. Relative frequency refers to the respective frequency of a characteristic divided by the total number of frequencies, which equals 80. Cumulative frequency refers to the summation of the relative frequency assigned to each characteristic. Whether it is a case-specific CHANS or a general overview of CHANS, each incident explores the complex characteristics and is categorized accordingly. All complexity characteristics of CHANS are listed under five main categories: reciprocal effects and feedback loops, resilience, surprises (uncertainty), heterogeneity, and nonlinearity. This work scrutinized occurrences of the main complex characteristics. Of the five characteristics, three accounted for 72.50% of the cumulative percentage, which can be described as the “vital few”. Two characteristics accounted for 27.50%, and can be described as the “useful many”.

The Pareto chart in Figure 5 depicts four vital characteristics, namely, reciprocal effects, feedback loops, nonlinearity, and resilience, used to classify CHANS complexity. Reciprocal effects and feedback loops denote the highest score, which indicates the complex interaction between human and natural systems. Further, nonlinearity was scored as the second most significant character. Resilience accounted for the third vital characteristic, because coupled systems have different levels of resilience, and this denotes the ability of systems to retain the same structure even after various disturbances occur. In summary, all three characteristics represent the complexity of CHANS, which mimics the requirement of having an integrated and interdisciplinary approach to analysing the coupled interaction. For this reason, future researchers are required to understand and acknowledge the vital characteristics within the context of complex CHANS. This will make it possible to manage a substantial balance of CHANS.

3.4. CHANS Subsystems Where the Complexity Originates from

Complexity stems from both subsystems. The biophysical subsystem comprises biophysical values such as hydrology, climate, and geographic components, which cause natural disasters, and in turn shape social factors such as economy, demographics, and social/cultural values. In general, subsystems of both the human and natural systems consist of several internal elements, some of them being key while the others are general elements [10,58]. Elements inside the subsystem are connecting, and subsystems connect to each other. Most importantly, there are more links between human and natural systems, and these are where the complexity originates from.

Research to address the subsystems to better understand the integrated links between humans and the environment was undertaken by [59]. This research modelled the detailed association between the components of the human and natural systems by breaking each system into sub-blocks. The developed model included two systems: human (socio-economic) and environment (biophysical). The former includes demographics, technological, economic, and governance sub-systems, while the latter consists of a coupled earth system. The model highlighted the fact that the relationship among the subsystems is better represented by flows as fluxes and changes. Therefore, more potential interactions can be visualized and analysed, taking into consideration wider and deeper connections that allow for a better understanding of the CHANS concept [59].

Further research intended to explore CHANS by illustrating the subsystems was also developed by [60]. Their research investigated the human subsystem with a distinct focus on climate change as an outcome of CHANS. The research identified a long matrix of around 33 sub-perception drivers incorporated into 7 main driver classes: education and awareness of scientific work; media exposure; the influence of corporations; ethnography; wealth; personal experience; and perception and demographics [60]. The methodology of the study offered an interdisciplinary view, which is helpful for policymakers, as the model comparably presented the seven main drivers that break into the sub-drivers.

According to

Table 2. Most common identified subsystem elements for presenting the complexity of CHANS.

Author, Year	Conceptual Model/Model Name	Identified CHANS Subsystem Elements
		Human System Elements	Natural System Elements
[46]	Coupled human and landscape model	Social institutions, economy, demographics	Climate, hydrology, geology
[42]	Coupled human and environment systems amid the COVID-19 crisis	Impacts of COVID-19: Health risk, economic recession, physiological adaptability, food scarcity/shortages	Impacts of the current crisis: CO2 emission reduction, noise reduction, improved quality of air and water
[59]	Model for land system changes	Demographics, governance, technology, economy, decision making	Land, ocean, atmosphere
[19]	Resilience model	Economy, production, energy	Land use, treatment, pollution
[61]	Schematic model	Social, economic, and cultural dimensions	Hydrological and climatic systems
[60]	Fuzzy cognitive mapping	Public perception under below drivers. CC education and awareness, anthropologic factors, demographics, financial capability, personal experience, impact of medias, influence of corporations	Climate change
[50]	Word cloud and box and Whicker analysis	Society and culture, understanding of science, behaviour and economics	Climate change, agriculture
[62]	Agent-based modelling (case study)	Tourism	Wolong National Nature Reserve (China) & wildlife
[9]	Coupled human and natural model	Social, political, and economic settings	Natural ecosystem
[2]	Conceptual representation of the CHNC	Population, economy, society, information, and other	Water, land, atmosphere, biodiversity, others
[63]	Agent-based modelling	Demographic factors of the population	Agricultural land use
[64]	Network diagram	The network considered six well-being dimensions and how they are connected, i.e., food sufficiency, livelihood security, physical health, stress level (mental), freedom of choice, and social cohesion
[54]	Agent-based model	Socio-economic factors for certain groups, including tribes and family forest owners	Biophysical factors of participant groups
[17]	Schematic model	Population demographics, transportation, industries	Global atmosphere including temperature, wind, rain, and CO2 emissions
[65]	Conceptual framework of SES	Human population dynamics, well-being and development, governance	Organic carbon dynamics, water dynamics, surface energy balance
[66]	Schematic representation of Adda River basin integrated model	Water demand and water supply decisions based on the cropping pattern	Adda River model
[67]	A framework in landscape studies based on CHANS concept	Socioeconomic system dynamics	Landscape ecology

, the social subsystems consist mostly of the changes in socio-economic scenarios due to population growth, economic crisis, pandemics, and so on. The natural subsystems consist of significant climate change, CO₂ emissions, changes in land use and land cover, water dynamics, and so on. Based on the purpose of CHANS models, various models pass various information from the natural to the human system and vice versa. More frequently, connecting subsystems among both systems are based either positive or negative feedback loops, which are evidenced by the commonality of reciprocal effect and feedback loops in CHANS models.

The subsystems are utilised to clearly represent the reciprocal effects of causal relationships. The thematic analysis and Pareto analysis also evidenced the frequent occurrences of representing the reciprocal effects and causal relationships via feedback loops.

3.5. Research Approaches and Conceptual Models in CHANS to Visualise Complex Characteristics

The current review revealed that the conceptual model has proven to be an effective tool to better visualize the complex interconnectivity within a system. It does this by identifying the relative weights of the system’s components and therefore highlights the prioritized areas for the implementation of viable management strategies [42]. Moreover, understanding the human–nature interaction seems to be key to both achieving human well-being and global sustainability goals, since the two systems are in a dialectic interaction [10]. This review has highlighted that refining the conceptual framework was accomplished through various approaches, and the current review explains how this was performed (Table 2).

Ref. [46] analysed the influence of interacting human and natural components with a particular focus on disasters triggered by natural events. A conceptual model of the coupled human–landscape system (CHLS) was developed for a specific spatial dimension. CHLS is based on the dynamic relationship between humans and nature. Key indicators of the human and natural systems as well as the links between them were depicted through the CHLS [46]. The development of a model, followed by a meta-analysis of the relevant literature to address the impacts of government COVID-19 policy interventions for the components of coupled human and environment systems (CHES) and complex interactions, explores the benefits of lockdowns and social distancing rules for the natural environment [42].

Moreover, the coupling approach is a general method employed for explaining complex mutual dependence and interactions between humans and nature [10]. Two-way coupling enables analysis of human–earth system feedback to better represent a system [28]. A conceptual framework of telecoupling, a sort of coupling mechanism, develops an integrated approach that explicitly evaluates the interactions of the coupled human and natural system [62]. Furthermore, the telecoupling toolbox, a set of developed toolsets, identifies and deeply explores five interconnected components of the telecoupling framework: agents, flows, systems, causes, and effects [68]. Each toolset is comprised of several customized Python and R scripts that qualitatively and quantitatively assess the interconnections of coupled human and natural systems. Coupled Human and Natural Cube (CHNC) is a novel framework that represents the coupled interactions between human and natural environments; it is a complex system consisting of social, economic, cultural, and natural aspects, and seeks to overcome the time-space compression challenge [10]. Absorbing the concepts of telecoupling, CHNC was devised. CHNC is a comprehensive and integrated conceptual framework. CHNC’s portrayal of the coupled human–nature system can be likened to a Rubik’s cube with four dimensions: space, time, organization, and appearance. Zooming in on the cube, there are several commanding elements. The colour of every surface of the small cube shows subsystems consisting of biophysical values. The evolutionary rule of CHNC proposes that when each side of the Rubik’s cube is the same colour, it shows the subsystems are coordinated.

Feedback loops are one of the main distinctive features that describe how humans both affect and are affected by nature, which contributes to CHANS complexity. Identification of several feedback loops through the Coupled Human and Landscape (CHL) model depicts the interdependencies within and between biophysical and human subsystems [46]. Some loops are reinforcing loops, while others are balancing loops. Reinforcing (positive) loops intensify the decreasing or increasing impacts of a system, while the balancing (negative) loops reduce or prevent the changing results of a system.

Table 3. Delineating the reviewed CHANS models including the mechanism used to depict complexity.

Author, Year	Modelling Approach	Mechanism
[46]	Coupled human and landscape model	A causal loop diagram depicts the complex dynamics including interactions and feedback between human and natural systems.
[42]	Coupled human and environmental systems amid COVID-19 crisis	A conceptual framework shows the complex interactions of coupled human and natural systems amid a pandemic.
[59]	Model for land system changes	A conceptual model of land presents the drivers and relevant processes among the sub-systems considering land as a coupled system.
[19]	Resilience model	The resilience transformation in complex human and environment systems was quantified through catastrophe theory and adaptive cycle methods.
[61]	Schematic model	The framework in the form of a schematic diagram represents an integrated system including complex interactions and feedback between human and natural systems.
[60]	Fuzzy cognitive mapping	Interaction among drivers including the direction, influencing capacity and the nature of the influence (either positive or negative) presents in the cognitive mapping in the form of a network diagram.
[50]	Word cloud and box and whisker analysis	The most significant CHANS complexity perspectives were identified.
[62]	Agent-based modelling (case study)	The applicability of the telecoupling toolbox is explained using a real-world example (Wolong National Nature Reserve).
[9]	Coupled human and natural model	The model shows the structure of coupled human and natural systems, dynamic interactions, and feedback.
[10]	Conceptual representation of the CHNC	CHNC is an extended version of CHANS that includes four dimensions, time, space, organisation, and appearance, to explain the coupling effects.
[63]	Agent-based modelling	The model explores the role of feedback between land use and demographic change.
[64]	Network diagram	The diagram shows the interactions between the community’s well-being and the ecosystem.
[54]	Agent-based model	The model investigates the characteristics of forest management and land users and their influence on wildfire behaviour.
[17]	Schematic model	The proposed framework, together with the feedback and drivers, focuses on policy interventions that can be implemented.
[65]	Conceptual Framework of SES	The framework represents three components, social system, ecological system, and interactions, with 13 dimensions of SES functioning.
[66]	Schematic representation of Adda River basin integrated model	The model shows the complex network of feedback between human and natural components by considering water supply and water demand decisions.
[69]	Conceptual framework	A social-ecological system framework was analysed for complexity using subsystems and variables analysis.

delineates various conceptual models.

Ref. [59] presented a conceptual model of the coupled human–environment system reflecting the rationale of land systems science. The model’s purpose was to develop a fully integrated and interdisciplinary understanding explicitly covering the interactions between human and natural systems. Catastrophe theory also describes the evolution of forms that are applicable in analysing dynamics. The theory can analyse the transformation of human and environmental systems between different equilibria [19]. Most importantly, the outcome of catastrophe theory is that transformations involve time. Those two critical characteristics enable the quantification of interactions between human–environmental systems. In one study, addressing the complex dynamics of factors influencing risks of natural disasters was based on a dynamic systems approach, through the use of a causal loop diagram. This approach proves the behaviour of a complex system. This approach enables, firstly, considering the entire system focusing on separate parts, and secondly, synthesizing complex interactions [46].

Apart from this, causal loop diagrams clarify complexity and help to scrutinize the cause-and-effect interaction, as well as the closed loops that can form either negative or positive feedback loops. Positive feedback loops, in other words reinforcing loops, tend to increase dynamic behaviour. Conversely, negative feedback loops or balancing loops act to reduce impact [46]. The representation of a causal loop diagram consists of biophysical and human subsystems [46]. As per Table 3, the CHANS model development was based on complexity characteristics, and the common complexity characteristic shown was the feedback loops using causal diagrams.

4. Critical Reflection

As the world’s human population continues to rise across the globe, so too does the impact of humans on the natural system. Even with added restrictions and policy measures that attempt to ensure that we do as little damage as possible to the environment, societies’ industrial activities pose a sustained threat to the natural world. This unplanned outcome has led to the disturbance of CHANS and results in severe and extreme weather and pollution-related events that need to be fully investigated and resolved. There is fast-growing research that aims to explore CHANS components so that the Intergovernmental Panel on Climate Change (IPCC) and especially the International Network of Research on Coupled Human and Natural Systems (CHANS-Net) can better investigate the complexity of the CHANS. Societies’ anthropogenic activities entail dangerous risks that extend beyond climate change and its symptoms (much larger and more persistent hurricanes, floods, bushfires, etc.), where knowledge regarding the complex interrelations of CHANS components is lacking.

The traditional approaches to understanding, managing, and predicting the future structure of CHANS are challenged by the dynamic interrelations and rapid changes that are moving ahead of the management curve. Hence, there seems to be a need to formulate such complex interrelations through integrated and interdisciplinary approaches. CHANS modelling does not intend to find worldwide solutions, but to offer an instructional and empirical tool and framework for analysing the linkages and to foresee potential patterns that can minimize the involved variables.

Unique properties of CHANS do not belong to either human or natural systems separately but stem from the nexus between both. The added layer of complexity stems from the unique characteristics, functions, and dynamics of each system. The inherent complexity and complex characteristics should be dealt with in an integrated and interdisciplinary way that recognises the dynamic interactions. Structural characteristics of CHANS, including nested hierarchy and most importantly the complementary relationships among human and natural systems, should be considered when developing interdisciplinary frameworks. Preliminary conceptual frameworks have proven to be related to developing analytical frameworks using quantitative methods. Even the elements of human and natural systems derived from existing conceptual frameworks can be utilized for further refinement and analysis.

Climate change (CC) is currently a topic of major scientific interest, and it is dependent on various factors, called drivers. The quantification of the relative strength of those drivers across social, economic, geographical, political, and educational identities is challenging. CC introduces great uncertainty to our understanding of CHANS [50]. Human-induced anthropogenic CC and how its consequences affect the structure of coupled human and natural systems should be considered in CHANS studies. Although the impacts of CC are universal, the responses of natural and human systems do vary due to differences in the level of vulnerability, climate sensitivity, and adaptive capacity. Complex and iterative feedback of CHANS should be analysed based on human-induced anthropogenic CC and how the consequences of CC structure coupled human and natural systems, as well as the complex linkages between them. Furthermore, zooming in on the human subsystem, within the context of CC, can lead to significant discourse. Analysing social and economic factors that shape the responses to the natural system would be more beneficial. Integrated socio-economic aspects need to be emphasized in policy development and implementation.

The findings of this review are presented in the form of a framework (Figure 6) that can even be utilized for future research to understand and better manage the complexity of CHANS. The conceptual model development should be followed by the proper evaluation of interactions analysis of the factors inside the subsystems to propose complexity management mechanisms. In this way, complexity management of CHANS is becoming more robust and valid, and can be effectively utilised for policymaking. This conceptual framework is used in this paper to understand where the complexity of CHANS originates from. The conceptual framework consists of the first three layers, starting from the problem identification, and progressing the factor identification inside subsystems. The conceptual framework is suggested to further develop to integrate the analytical framework. The integration is proposed to be conducted using the interaction analysis of the identified factors inside sub-systems. The interaction analysis can then be further developed to visualise non-linear relationships between the factors. Ultimately, the interaction analysis would be beneficial in successful policymaking.

5. Recommendations for Future Research

Understanding and evaluating the complexity of CHANS and adopting adequate measures to manage severity is essential. Developing a conceptual framework considering the complex dynamics (interactions and feedback) provides a way to identify the interactions within and between the subsystems of CHANS. A conceptual and interdisciplinary model can guide the development of a numerical model that can quantify the impacts. Ultimately but importantly, zooming in on the subsystems of CHANS will lead to significant discourse as well as effective implementation of the policy.

The limitations of developing integrated and interdisciplinary approaches, including addressing inherent heterogeneity among the systems that explore human and natural systems, were delineated by most of the studies. In general, time-space compression and interactions constitute great challenges to traditional approaches in developing analytical frameworks of CHANS. Although the significance of CHANS and related research is recognized, there has not yet been an assessment of the most important CHANS research questions. The influence of CHANS in governance highlights a mismatch that should be addressed by well-devised policies. Understanding human–nature interactions, pathways of coping complexity, feedback and adaptive management, and social and natural integrated approaches are critically important when developing policies.

Although progress has been achieved to adapt to or even overcome some natural hazards, the current research lacks comprehensive interdisciplinary and integrated approaches that enable both natural and social aspects to be articulated simultaneously. The lack of integrated approaches that combine both socio-economic and biophysical aspects is one of the great challenges when developing ad-hoc disaster risk management schemes with natural catastrophes. Most importantly, the poor adaptability of human society to a pandemic arises primarily due to the poor understanding of the complex interconnections of CHES. Although CHNC is a novel framework, it is still impractical to obtain a complete unification of human and natural systems. Both systems fluctuate rapidly due to external and internal disturbances, which makes it even more complex. A great challenge in CHANS research is to develop analytical models that depict human and natural systems on a worldwide scale. Challenges include the articulation of contrasting data types, such as biophysical and social data, into one model. Conceptual models can serve to develop individual studies and especially process modelling for analytical frameworks. Integrating a statistical approach into a conceptual model can reinforce our understanding of the dynamic relationship between humans and nature, which is due to the availability of quantitative data. This in turn leads to more logical findings.

The identification of feedback loops from a causal loop diagram can be related to policy implications as the impact flow is clearly defined. For instance, understanding how the drivers of CC perception are interconnected should support the implementation of effective mitigation and adaptation policies. Integrated and comprehensive modelling approaches are essential because they address the ability to reduce risk and improve resilience. Socio-economic scenarios should be refined with a particular focus on social shocks such as economic crises. Having surveyed data, especially for community awareness and perception of natural disasters, to operationalize the conceptualized models is highlighted. Connecting human and biophysical components may require system dynamics modelling to be developed to quantify the relationship.

6. Limitations

Although this review provided a holistic overview of the complex characteristics of CHANS, the SLR was not without its limitations. The review excluded peer-reviewed articles written in languages other than English, articles published before 2012 or after 2022, and articles in the grey literature. Cohen Kappa statistics were not considered for a consensus to assess the articles for inclusion. In addition, this review did not provide methodological recommendations on the analytical framework to model CHANS, which could be addressed in future studies. Furthermore, the reviewed literature spanned across different geographic locations and adopted a wide range of research methods. Therefore, articulating the findings of the literature and presenting a thematic analysis may have some contextual limitations. To overcome the mentioned limitations, future research may compare the findings and conduct a more focused review.

7. Conclusions

CHANS are not static, nor perpetual, but change over time. The complexity stems from the dynamic nature of both the human and natural systems, hindering stability and balance. Hence, this review evaluates the complexity characteristics, subsystem components and CHANS model development in the shortlisted articles. Identified modelling approaches and conceptual models depict the complexity that exists in both human and natural systems. According to the thematic analysis and the Pareto analysis, among the five major complexity characteristics, reciprocal effects and feedback loops were more frequent compared to the other factors. The complexity stems from the subsystems of both systems, and can be visualised by connecting the subsystem components with relevant information passing, such as CO₂ emission. The capability of integrated models in better visualising the complexity has been commonly investigated via representing reciprocal effects and feedback loops via causal loop diagrams.

Although the CHANS concept was established including their conceptual frameworks and modelling approaches, it is difficult to integrate human and natural systems due to inherent uncertainties. The ability of an analytical framework to address dynamics and system integration is recommended for future research to consider. Although the planet is a single system, it consists of complex interactions between humans and nature. To be successful in CHANS modelling, synthesizing knowledge, methods, and data is essential. Finally, more research, more models, and more studies, including integrated feedback systems, are needed to robustly quantify the sign and magnitude of human–Earth system interactions.