The Mannheim case built on the national climate Impact Chains (IC) included in the German Adaptation Strategy (Umweltbundesamt, 2016, 2019), to avoid developing local Impact Chains from scratch. The German National Impact Chains have been co-produced by the experts who developed the VS, by stakeholders from German Federal Agencies and Ministries, and by domain experts. These national ICs characterize the possible climate impacts on a national level, clustered into 15 fields of action. The Impact Chains visualize components of climate risk as a diagram. This graphical representation includes (1) direct physical impacts of climate related hazards on exposed system elements, (2) the sensitivities of the exposed system elements, (3) the nature of the damage, and (4) impacts of damage to system elements on other system elements. Indirect impacts may propagate and thus form ICs and are not restricted to one field of action or one sector.

For assessing the specific risk of two hazards in the Mannheim case, only a subset of the national Impact Chains was needed. That is, we extracted the relevant subset of components from the national Impact Chains regarding the selected hazards, the sectors represented by the participating stakeholders, the related national fields of action, and all related sensitivities and relational information. The original layout of this true subset of the national Impact Chain has been optimized for result documentation and reference, but it is not ideal as a basis for further elaboration. In the next step, the layout of this subset was transferred into a suitable online collaboration tool and collaboratively enriched with regional risk factors (exposures, sensitivities, capacities, stressors, impacts). The resulting Impact Chain was subsequently improved and validated in several post-processing cycles.

Based on the stakeholders' oral feedback and completed questionnaires, we believe that this method was a successful and efficient way to conduct IC-based CRVA, mitigating the often-criticized time demand of the approach. Furthermore, the multi-stakeholder regional CRVA fostered information exchange and awareness regarding climate risk and adaptation opportunities in this heterogeneous group.

The benefits of this approach included (a) a faster start and general time saving by starting with concrete examples, and (b) a resulting regional Impact Chain that is consistent with the national ICs (Umweltbundesamt, 2016) and does not “re-invent the wheel.” We believe that the latter point could be an advantage for planning subsequent adaptation measures and for acquiring their funding, because national funding is indicated per national field of (climate) action. Therefore, building local climate risk analysis on national Impact Chains and structuring elements of risk (exposure, sensitivity, capacity, impact) per national field of action facilitates the identification of funding opportunities. The regional IC, in turn, can serve as a starting point and context for local CRVA. The approach of nested scales allows to establish consistent links between local, regional, and even national adaptation measures. However, while national Impact Chains were available for Germany, globally this availability is still an exception. In the future, standard Impact Chains for various sectors and systems could be the basis for local adaptations.

The current Impact Chain notation style has limitations in capturing the dynamic nature of climate risk, which is influenced by complex interrelationships between system factors spanning socio-economic, political, and environmental realms (Menk et al., 2022). Factors contributing to risk are summarized into a final risk score that does not feed back into the system, failing to account for the feedback loops that shape risk over time.

To address this limitation, the Salzburg case study employed a Causal Loop Diagram (CLD), which depicts causal flows within a system and allows for a more nuanced understanding of the dynamic interrelationships between risk factors (Figure 4). Using CLDs to map a risk system acknowledges the balancing or reinforcing effects of these factors, providing a more comprehensive view of the risk landscape. The study used a combination of stakeholder engagement, spatial data analysis, and literature review to create a CLD for drought risk, highlighting the role of water availability as a major driver of risk, and identifying the wider range of factors that influence agricultural success or failure.

The CLD-based approach was well-received by stakeholders, who appreciated the larger system understanding it provided and the identification of entry points for medium- and long-term planning. The study suggested that the use of CLDs can support the development of more effective and sustainable adaptation measures. However, the study also recognized that more extensive CLDs should be developed and presented as comprehensible sub-systems before being combined to create a bigger picture of the risk system.

Abstracting a complex Impact Chain into a simplified indicator requires careful attention to weighting and aggregation methods. To aggregate indicators into a composite indicator, the Vulnerability Sourcebook recommends weighted arithmetic aggregation. This method multiplies individual indicators by their weights, sums them, and then divides by the sum of their weights (Fritzsche et al., 2014). If individual indicators show extreme negative values, an alternative method, weighted geometric aggregation, may be used (as already suggested in the VS as a side-note). This method is popular in indicator construction and decision-making, as it is a prioritization tool in Analytic Hierarchy Process (Krejčí and Stoklasa, 2018) and other multi-criteria analyses. Weighted geometric aggregation limits substitutability between risk factory due to its bias toward low values. Notably, the Human Development Index has shifted from the arithmetic to the geometric method.

This methodological advancement to IC-based CRVA involved the use of reverse weighted geometric aggregation instead of the weighted arithmetic aggregation recommended by the VS. The reverse weighted geometric aggregation method assigns greater weight to particularly high-risk factors and avoids the risk of low-risk factors compensating for them in the final risk score, which can result in an underestimation of risk. In the Paris case study, risk scores resulting from arithmetic and reverse weighted geometric aggregations were compared. The results consistently showed that the scores produced by the reverse weighted geometric method were higher, particularly in cases where sub-indicators had relatively high dispersion among coefficients.

We argue that for most of the applications using the risk value obtained by IC-based CRVA it is more favorable to overestimate, than to underestimate, risk. Therefore, we suggest using reverse geometric aggregation, to shift the bias toward high values instead (Guillaumont, 2009). Given that the method accounts for the interdependencies of the system and produces higher scores, we believe that it is relevant to apply it. To ensure clarity and consistency, we recommend providing a more detailed explanation of the method in the next edition of the VS. Alternatively, the quadratic average method could be used to address this issue.

Stakeholder knowledge and needs do not necessarily result in one Impact Chain that everyone can agree on, due to possibly diverging interests and opinions (Schneiderbauer et al., 2020).

The Upper Rhine case utilized TRIZ (Theory of Inventive Problem Solving) Inventive Design Method, which was originally developed in the engineering domain, and combined it with IC-based CRVA to address problems and find solutions in contradictory situations. The method involves breaking down a problem into its various components and is easily integrated with the IC-based CRVA framework. However, unlike Impact Chain development, this method is primarily focused on identifying solutions, particularly adaptation measures. The idea is to identify partial solutions for various parts of the problem. For instance, for a problem such as low waters inhibiting barges to transport regular volumes of freight, a partial solution, such as barges carrying less freight, could be identified. Partial solutions that are straightforwardly quantifiable are usually easy to resolve. However, this method also accounts for situations where different stakeholders hold differing “worldviews” about a problem and potential solutions. The method captures these worldviews in a special software, highlights contradictions, and then a consensus is sought in the next step.

Enriching IC-based CRVA with the TRIZ method improved participatory Impact Chain development and to identify and select indicators and adaptation measures. The approach involved a combination of workshops and individual interviews to account for the issue of individual stakeholders dominating discussions and influencing the general opinion. Working with the TRIZ software enabled a cross-breeding of qualitative and quantitative methods. This increased the perceived impartiality of the methods and results among stakeholders, which helps counteract the issue that simple qualitative methods are often considered insufficient or biased by stakeholders (Cheek et al., 2004).

However, the approach faced some challenges. Contradictory situations, especially on inherently subjective issues such as the level of social acceptability of a partial solution, were not always easy to solve. This difficulty resulted in a marginalization of qualitative problem dimensions, which are difficult to weight but are nonetheless essential to the identification of viable solutions. Additionally, the process was time-consuming and required deep involvement of the stakeholders, as well as a certain level of expertise to work with the TRIZ software. Nonetheless, the approach's combination of sociological and engineering expertise made it possible to convince stakeholders of the reliability and robustness of the results.

Climate change is expected to have significant impacts on social groups, with those with known vulnerabilities likely to be disproportionately affected and others becoming newly vulnerable. However, there is limited understanding of which groups will be affected and how, as well as the economic, social, and physical factors that drive their vulnerabilities. Social aspects of vulnerability are often difficult to measure and quantify (Hudson et al., 2019), and while the IC-based CRVA approach identifies knowledge, technology, institutions, and the economy as key drivers of vulnerability, it provides only a limited understanding of vulnerable groups, with generic representation of women and disadvantaged groups. To address this limitation, a method to assess social vulnerability was searched (see Birkmann, 2013 for an overview) that would engage stakeholders in a co-production process and align with the IC-based CRVA approach.

The Halmstad case explored flood hazards triggering cascades in the infrastructure system and their impacts on vulnerable groups in Halmstad Municipality, Sweden (Barquet et al., 2022). An exploratory scenario approach was adopted which is a form of storytelling that forecast how evets unfolds over time (Kok et al., 2007). A transdisciplinary process was designed in which key stakeholders from Halmstad Municipality were engaged using surveys, interviews, focus group discussions, and workshops (for further details see André et al., 2022; Englund et al., 2022). Exploratory scenarios helped to overcome issues related to data confidentiality as stakeholders could discuss a hypothetical scenario instead of sharing sensitive information on infrastructures and their services. It also allowed adding information on social aspects of risk, normally not considered when developing an IC-based CRVA. In an online workshop, scenarios with key stakeholders were co-developed to forecast cascading effects in flooding events and their impacts on people. The exploratory scenario and the Impact Chain were used as boundary objects to help stakeholders understand the social dimension of risk. Using digital tools Miro and MentiMeter, stakeholders identified possible infrastructure disruption and mapped social groups at risk in case of a disruption. Following the workshop, we conducted online interviews with the same stakeholders to refine scenarios and identify additional vulnerable groups.

Exploratory scenarios were found to be a valuable addition to the IC-based CRVA framework. While scenario planning is already widely used to support climate risk decision-making (Brown et al., 2016; Star et al., 2016; Flynn et al., 2018), in our case study, it helped overcome challenges associated with limited data availability and measuring intangible risks. It supported stakeholders in considering low-frequency but high-impact risks where historical data is scarce. The exploratory scenario also led to various co-benefits, such as increased awareness, simplified complexity, and possibly also improved decision-making support. In line with previous research, the boundary object improved the quality of discussion and data (Star and Griesemer, 1989). Our case study indicates that the IC-based CRVA framework is flexible and adaptable to suit diverse stakeholder needs and objectives.

Although the available guidelines to IC-based CRVA emphasize the importance of including relevant local stakeholders in the assessment process, specific guidance for addressing specific groups of local stakeholders is rare. For mobilizing industrial stakeholders, it is essential to understand how they manage operational risk. Typically, this is an integral part of their business continuity management or enterprise risk management. We concluded that results of any CRVA conducted for a participating corporate stakeholders need to be integrable with their risk management approach. One way to achieve this is to pinpoint climate risk factors to business units, activities, and processes, since these are the places in which operational risk management procedures are anchored in. An established model for describing these business elements is the Value Chain (Porter, 1985). Porter divided a company's activities into primary and supporting activities. Value Chain diagrams describe these activities in graphical form (Figure 5). The further division of activities into units, areas, and interlocking processes that take place in these areas and across areas offers opportunities for analysis and optimization, including risk assessment. Since Business Continuity Management (BCM) is typically process- or unit-oriented, the mapping of climate risk factors (sensitivities, capacities) and intermediate impacts (knock-on effects from direct physical impacts) to Value Chain C elements is suited to inform BCM managers on climate specific risk.

For implementing the Value Chain CRVA for a single business as the last phase of the Mannheim case study, we built upon a previously developed Impact Chain. We started by validating the relevant climate risk factors for the corporate stakeholder, the Mannheim Large Powerplant, ruling out factors that were only relevant for other participating stakeholders. We continued with eliciting the company's Value Chain, then pinpointing the remaining climate risk factors to the Value Chain elements and identifying suitable measurable indicators for each factor. Furthermore, we conducted a quantitative risk assessment focused on the economic effect of low waters of the Rhine River on costs of transporting freight, i.e., coal, via inland waterway.

We investigated just one indicator due to resource limitations. However, we produced a tangible, actionable decision support for one element of climate risk, which the stakeholders appreciated. Users of this approach should be aware that businesses may keep certain information and data private. In our case, we worked with disclosed surcharges instead of undisclosed total costs. Bringing these results into the Value Chain of businesses reduces entry barriers for the implementation of adaptation measures, as one could incentivize organizations to invest in climate change adaptation by bringing the topic into regular business continuity practice.

Climate change risks have become increasingly important for (large) financial investments in the real estate sector. Yet, meaningfully integrating climate risk information into investment portfolios remains difficult. To address this, the Netherlands case explored potential climate risks affecting two real estate assets in Utrecht and Rotterdam using IC-based CRVA. The case study aimed to assess and find effective ways to secure assets against flood risks by combining knowledge co-production and scientific information to support financial decisions.

To engage industry stakeholders and gather data for a climate risk assessment, the case utilized semi-structured interviews and a workshop. Real estate and finance companies, as well as consultancy companies, were involved as stakeholders and end-users of the final risk assessment. The knowledge co-production process was central to the assessment, and the aim was to advance the IC-based CRVA's user interface. The case study relied on open-access data collection and interactive maps showing current and future flood risks in the Netherlands for hazard and exposure data. However, data at the asset scale were not readily available in open-access databases, so close working partnerships with stakeholders from real estate companies were needed to obtain this data.

Feedback analysis of the process showed that knowledge co-production was well appreciated and the dialogue and interaction helped refine the companies' needs. Some information requested by the companies were impossible to meet with available climate knowledge, but through continuous engagement enabled by co-production, these expectations were addressed after explanations were given. However, knowledge co-production presents key challenges such as the subjectivity of some data collected that could impact the quality of the results. The process was laborious with a series of back and forth with stakeholders including workshops, interviews, meetings and phone calls. As a result, applying the IC-based CRVA framework became complex and time consuming.

Uncertainties in data and the selection of weighting and aggregation processes make it difficult to confidently turn results into adaptation action. Incorporating qualitative indicators and knowledge gaps can also be challenging and may jeopardize the validity of the assessment. Even with extensive stakeholder and expert involvement, our understanding of the factors which influence climate risk remains imperfect (Booysen, 2002; Gall, 2015; Gawith et al., 2016). Thus, there is a need to account for uncertainty in the development of Impact Chains, that remains unaddressed by the current methodological framework which relies heavily on data that may not be available. Some initiatives include uncertainties management strategies in CRVA, such as the probabilistic impact risk model software CLIMADA (Aznar-Siguan and Bresch, 2019) or CAPRA (Cardona et al., 2012). However, although they are valuable tools that provide a way to consider uncertainty measures in different risk components, they are restricted to specific software that requires some level of expertise, thus, limiting its usability.

The Balearic Islands case developed a general formalism for integrating uncertainties into IC-based CRVA. The formalism uses probability density functions (PDFs) instead of scalar quantities for weights and indicator values. PDFs are propagated through the entire Impact Chain using a Monte-Carlo approach, resulting in a final risk PDF (Figure 6; Melo-Aguilar et al., 2022). The PDFs can be defined freely, but a Gaussian function is a common choice, where the amplitude of the PDF is determined by the assigned uncertainty. In the Balearic Islands case, indicator-related uncertainties were estimated from data (e.g., spread of climate projections for future temperatures) or the standard deviation of time series, while weight-associated uncertainties were established through an Analytical Hierarchy Process survey with stakeholders (Melo-Aguilar et al., 2022). The formalism can be implemented by developing the suitable computer codes (e.g., in Python or MATLAB) or using the UNTIC tool (untic.pythonanywhere.com) which can be used without much technical expertise.

Our proposed methodological advancement is robust and flexible, allowing to be integrated with different aggregation methods (for example the reverse geometric aggregation as applied in the Paris case) and a wide range of situations. It offers a new perspective on Impact Chains for integrating factors even when knowledge of their severity and role is limited. This allows to include all relevant factors and indicators without requiring data availability. This methodological advancement is valuable as current IC-based CRVA heavily relies on data, and when no data is available, indicators tend to be discarded (Kienberger et al., 2016), leading to biased results. Future work should focus on developing strategies to validate risk analysis, producing robust information that is as independent from subjective choices as possible. Calibrating the IC operationalization by computing the risk for past situations and comparing it to real-world impacts would increase the validity of the results.

The transboundary Germany case uses a global multi-regional input-output model (MRIO), called GINFORS-E (GWS, 2022a) to assess the impacts of transboundary climate risks on agricultural commodity markets. The model combines detailed agricultural information with integrated MRIO-based modeling of global economic developments, mapping the multinational interplay of demand, trade, and production for 20 crop commodities and eight livestock products.

The case study demonstrates that using a global and regional economic modeling approach can provide a dynamic assessment of the indirect economic impacts of climate change, including trans-regional impacts, and is thus valuable for evaluating TCRs in agriculture commodities. However, there is still room for further development of MRIO models to fully utilize the available information and involve stakeholders at all stages of the participatory Impact Chain development process. In the UNCHAIN project, this potential was not fully realized, highlighting the need for continued research in this area. The main advantage of this approach is its ability to project self-contained and consistent multi-national socio-economic developments over time, allowing for the assessment of the impacts of climate risks on trade and production in different regions.

In contrast, the national Germany case assesses the future monetary costs of domestic climate impacts in Germany under different socio-economic scenarios (a trend scenario, a sustainable scenario, and a dynamic scenario), by applying the national macro-econometric model PANA RHEI (GWS, 2022b). The model simulates the impacts of climate change on transportation, health care, and electricity sectors and the effects of possible adaptation measures. The results are presented using socio-economic indicators such as the gross domestic product, employment, or production, making them comparable so that they function as decision supporting information.

The main methodological challenges arose from quantifying effects of adaptation measures. Additionally, since the direct macroeconomic effects of climate impacts depend on many factors, it remains challenging to quantify the resulting monetary costs, especially for medium to long-term projections. As a result, more research is needed to strengthen the evidence base for individual effects of future adaptation measures on increasing resilience and avoiding direct economic. So far, no quantified national reference values have been established for (socio-) economic scenario analyses at the sector level. The more detailed representations of economic developments in macroeconomic models compared to most IAMs can be used to derive corresponding projections. Hence, it would be promising to obtain further detailed macroeconomic assessment results through complementary modeling studies of climate impacts, adaptation costs, and benefits for focal sectors in a joint project.

From a methodological perspective, both applications highlight similar challenges. Furthermore, as the estimated macroeconomic costs of climate impacts and adaptation measures vary significantly across individual socioeconomic scenarios, the case studies illustrate the need for dynamic risk assessments. Compared to earlier approaches that projected future adaptation actions and implied adaptation costs by applying traditional IAMs (see, for example, Agrawala et al., 2011), this approach differs methodologically in that it does not rely on any optimization assumptions. This implies that the modeling approach is not rooted in a highly abstract theoretical framework (like the “Ramsey optimal growth framework” applied by both models compared by Agrawala et al., 2011) as it does not intend to deduce any normative conclusions about “optimal” mitigation and adaptation measures. Instead, a positive analytical approach is followed by this approach as well as the by the MRIO Model GINFORS-E featured for transboundary climate risk assessments: Both models project the future evolution of inter-sectoral economic impact relationships from historical empirical observations under alternative socio-economic scenarios. Given these pathways projections, researchers as well as involved stakeholders can then examine in detail how different direct climate impacts and adaptation measures in individual sectors and/or world regions affect other sectors, and/or world regions economically.

Impacts of climate change are not confined or experienced only at a national and local level. According to the Intergovernmental Panel on Climate Changes' (IPCC) Sixth Assessment Report (AR6) Working Group III it is a problem that affects all parts of economy and society and requires actions across sectors and scales (13-16f; IPCC, 2022b). As in most European countries, Norwegian municipalities have largely been given responsibility for climate adaptation and focus so far has been mostly on adaptation to local climatic changes. However, a recent trend is shifting transboundary risks into focus, asking how Norway may be impacted by climate impacts elsewhere on the planet through trade, finance, people (tourism, health, and migration) and through biophysical processes. Therefore, it is important to have information about the patterns and magnitudes of climate risks in and between different regions and sectors, and methodological frameworks for the assessment of climate-related interactions between economic sectors and world regions have been promoted recently.⁶ Currently, most assessments of transboundary climate risks are conducted at the global or regional (cross-country), and to sometimes at the national level. However, there is a lack of information about how these risks affect the sub-national level of governance (Harris et al., 2022).

In UNCHAIN, two case studies addressed transboundary risks for Norway through stakeholder integration in participatory workshops.

The IC-based CRVA approach was employed to identify transboundary climate risks that are pertinent to food production in Norway, through consultation with key stakeholders in Klepp Municipality. By breaking down the climate risks into nodes and links, this method seemed to be well-suited for analyzing TCRs.

The IC-based CRVA framework was useful for communicating the concept of transboundary climate risks to local stakeholders, showing the differences, similarities, and possible interactions between traditional “local” risks and new “transboundary” risks. However, producing numerical indicators and indexes was too complicated and time-consuming for the case's scope, making it challenging to produce actionable knowledge for addressing transboundary risks at the local governance level. Focus from practitioners on TCRs is correlated to specific governance measures (i.e., requirements), a focus often lacking as policy mostly has focused on carbon footprint and “local” climate risks and the strengthening of resilience and adaptive capacity to such.

Practitioners tended to focus on local climate risks and building resilience and adaptive capacity to those risks, rather than on transboundary risks. However, it is important to link work and analyses on local and transboundary risks to avoid potential conflicts and unintended consequences. Local climate risks may be reduced at the expense of increasing transboundary risks, which was presented to local stakeholders. They responded that other means of adapting to local climate risks need to be investigated, which is also discussed in a recent report from the Environmental Protection Agency of Norway on climate change and food security (Bardalen et al., 2022).

A workshop was arranged to identify risks associated with salmon farming, and to develop Impact Chains that would span the inputs, production and markets involved. The participants were farmers who belonged to an industry cluster and who were responsible for addressing environmental issues in their respective companies. None of the companies had a specific climate change risk assessment plan developed yet. The identified risks ranged from environmental degradation, such as the potential collapse of soy production in Brazil, to geopolitical concerns that may or may not be triggered by climate change, and from increases in freight prices to a collapse in local salmon markets.

During the workshop the participants discussed and acknowledged the complexity and multiplicity of issues related to risk identification, and they recognized the need for clear risk ownership as the basis for decision-making and for tailoring measures to strengthen adaptability. The participants also recognized that timing is crucial when investing in improved adaptive capacity, as premature investments could create new risks. Therefore, there is a clear need for assessing the identified risk(s) and establishing risk ownership and regulatory guidance. These discussions highlighted the potential for improving risk assessment practices within individual producer organizations.

The application of an IC-based CRVA was found to be useful in discussing these complex issues. This approach facilitated a high level of co-production and allowed for a focus on the interconnections between different risks, across scales, geographical boundaries, and sectors. The overall experience suggests that a mixed-methods approach with knowledge co-production, literature analysis, media analysis, and assessment of industry strategies on environmental issues, is the most effective. The workshop revealed a certain sensitivity and awareness of the interconnectedness of climate change challenges across borders, which is often overlooked in more nationally-focused climate policy discussions. However, when it comes to identifying actionable steps, the industry stakeholders tended to focus on areas that are already prioritized, such as specific policy measures, obligations, and challenges. This underscores the importance of governance risk ownership in shaping decision-making, as confirmed by these co-production and workshop settings.