Skip to main content

Reflections on cross-impact balances, a systematic method constructing global socio-technical scenarios for climate change research

Climatic Change volume 162, pages 1705–1722 (2020)Cite this article

Abstract

Experiences with an algorithmic technique—cross-impact balances (CIB)—for exploring scenarios rather than relying solely upon expert intuitions are discussed. With CIB, two types of uncertainty for climate change research have been explored: (1) socio-technical uncertainties not represented explicitly in integrated assessment models (sometimes called “context scenarios”) and (2) sampling the space of possible futures to model. By applying CIB retrospectively and prospectively to two global socio-economic scenario exercises for climate change research (the Special Report on Emissions Scenarios and the Shared Socioeconomic Pathways), CIB proved instructive in two ways. First, CIB revealed system behaviors that were not obvious when social variables, such as quality of governance, were not captured explicitly by integrated assessment models. Second, CIB can algorithmically rank different plausible futures according to their self-consistency. These two capabilities have raised awareness about the limitations of accepting what may be “obvious” to model, as practices that focus solely on quantitative variables or rely upon intuitions for scenario analysis may result in detailed analyses of only a subset of important policy-relevant futures. From these experiences, systematic methods like CIB are recommended in conjunction with more detailed modeling to develop integrated socio-technical scenarios in energy-economy research.

This is a preview of subscription content, access via your institution.

Access options

Buy single article

Instant access to the full article PDF.

USD 39.95

Price includes VAT (USA)
Tax calculation will be finalised during checkout.

Notes

  1. Ritchey (2018) provides a typology of scientific modeling methods (see Ritchey’s Table 2). In this typology, CIB can be situated as a dynamic model employing discrete variables whose relationships are directed (i.e., CIB models contain dependent and independent variables). Variable interrelationships are quantified on an ordinal scale, and interrelationships can represent mathematical, logical, or non-causal relationships (e.g., correlations). CIB collects data in a pairwise fashion, so it resembles an acyclic model. However, cyclic connectivity between variables (i.e., circular feedback) is uncovered through the impact-balance calculation that determines whether scenarios are internally consistent. In Ritchey’s hierarchy of scientific models, CIB has greater information content than morphological analysis and occupies modeling typology spaces proximal to Bayesian Networks. A key difference between CIB and Bayesian Networks, however, is that CIB does not employ probabilistic connections between variables.

  2. In principle, CIB is a Turing machine, which means that it can represent any finite-state system (Weimer-Jehle 2009).

  3. In CIB, internal consistency is the criterion for selecting a small number of scenarios from a large set of possibilities. Tietje (2005) makes a similar argument for consistency analysis as part of Formative Scenario Analysis (Scholz and Tietje 2002). Börjeson et al. (2006) insinuate this purpose as well in their discussion of consistency techniques for scenario analysis (with the methods of cross-impact analysis and morphological analysis provided as examples). However, as summarized by Alcamo (2001), the characteristics of good scenarios include not only the internal consistency but also the ability to broaden the understanding of experts and decision makers. Focusing on the latter characteristic, proponents of diversity analysis (Amanatidou et al. 2016; Carlsen et al. 2016) argue that a formal approach to maximizing differences between alternative scenarios, rather than internal consistency, can be used to select a small number.

  4. Bell (1997) provides a helpful ontological explanation for why holistic judgments about plausible futures overwhelm human judgment. After acknowledging that simple individual events have frequentist and subjective probabilities, he adds that complex events are “the intersection of two (or more) causal chains” (p. 153). Such intersections are akin to “being at the right place at the right time” and are the simultaneous event combinations that CIB systematically searches for. In CIB, the complete matrix of cross-impacts acts as a database embedding causal chains for each state of each scenario driver in the system under study.

  5. How scientific inquiry manages to result in knowledge with predictive power is explored by philosophers of science. A key claim is that objective methods may foster objectivity in study results (Reiss and Sprenger 2017), which, in turn, may be useful for policy recommendations. Objectivity is an ideal never fully achieved; nevertheless, most philosophers have argued that objectivity in science is worth preserving and maximizing. As thoroughly discussed by Lloyd and Schweizer (2014), CIB better exemplifies objectivity across multiple dimensions compared with the well-established scenario approach of Intuitive Logics.

  6. Absar and Preston (2015) demonstrate narrative downscaling for SSP extensions to the Southeastern USA using the Factor-Actor-Sector (FAS) framework (Kok et al. 2006). Similar to CIB, FAS specifies each scenario element (a factor, actor, or sector) relevant at each spatial scale and may make use of influence diagrams to track gross interactions between elements. However, beyond this, methodological similarities between FAS and CIB end. Absar and Preston, as well as Kok et al., noted challenges with FAS for the analysts (or stakeholder study participants) to exercise their judgment to maintain vertical and horizontal (i.e., across-scale and within-scale) consistency for multi-scale scenarios. Similar to statements made by Alcamo acknowledging the limitations of SAS, both Absar and Preston as well as Kok et al. noted that it is unlikely that FAS would yield reproducible multi-scale scenarios.

  7. Dooley et al. (2018) also criticized the lack of explicit policy assumptions for ensuring food security. This criticism may be misplaced however. In later studies (van Vuuren et al. 2017b), modeling teams used the SSPs to replicate the Representative Concentration Pathways (van Vuuren et al. 2011). Through agreed-upon SSP quantifications and qualitative scenarios, modeling teams harmonized assumptions about the size of the global population as well as global trends for meat consumption, food waste, and restoration of marginal lands. All of these factors have significant implications for how much land is required for food security (Boysen et al. 2017). Setting policy constraints for food security on modeling exercises a priori as advocated by Dooley et al. would not only be artificial but also likely to skew IAM calculations regarding the land cover required to feed different possible worlds.

  8. Importantly, practices of inclusion also involve diversity, in reflecting both a diversity of worldviews (e.g., van Asselt and Rotmans 2002) and a diversity of participants meaningfully engaged (Beck and Mahony 2018; Yamineva 2017). Current efforts toward transparency invite more diverse participation in scenario-based research through intellectual contributions, but it remains to be seen how much diverse participation will increase.

  9. Another option is to pair different models embodying complementary “expertise”; see Trutnevyte et al. (2014).

References

Download references

Acknowledgements

The author acknowledges Witold-Roger Poganietz and Ricarda Scheele for helpful discussions and three anonymous reviewers for feedback that improved the manuscript.

Funding

This essay was invited by Wolfgang Weimer-Jehle in connection with the Helmholtz Alliance ENERGY-TRANS. Dr. Schweizer thanks the Helmholtz Association for supporting a visiting professorship with the Alliance.

Author information

Authors and Affiliations

  1. Department of Knowledge Integration, University of Waterloo, Waterloo, Ontario, Canada

    Vanessa J. Schweizer

Authors
  1. Vanessa J. Schweizer
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Vanessa J. Schweizer.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This article is part of a Special Issue on 'Integrated Scenario Building in Energy Transition Research' edited by Witold-Roger Poganietz and Wolfgang Weimer-Jehle.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Schweizer, V.J. Reflections on cross-impact balances, a systematic method constructing global socio-technical scenarios for climate change research. Climatic Change 162, 1705–1722 (2020). https://doi.org/10.1007/s10584-019-02615-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10584-019-02615-2

Keywords

  • Scenarios
  • Socio-economic
  • Socio-technical
  • Cross-impact
  • Energy
  • Climate change