2.4.5.3 Key methodological advances

Thus far, methodologies for identifying cascades have focused on conceptual mapping (Klose et al., 2021) or different modelling approaches, which have their shortcomings (Juhola et al., 2022). The soft modelling CLD approach that we used here has allowed us to identify causal pathways for tipping cascades from the Earth system to social systems and show how tipping cascades can be identified in complex systems. There are also examples of further developing CLD modelling towards the inclusion of stakeholders (Sohns et al., 2021; Inam et al., 2015). Inclusion of stakeholder knowledge may yield insights on what the potential thresholds are for tipping cascades. Further method questions include what are the key feedback loops and nonlinear dynamics that can lead to cascades across scales in different systems and how can these dynamics be quantified and integrated into models and assessments. Here, complex systems approaches grounded in network science, agent-based modelling and evolutionary approaches could be especially useful, as they directly capture feedbacks and restructuring of a system based on its changing elements, and could explicitly treat the emergence of different tipping dynamics (Filatova et al., 2016).

In addition to advancing modelling, there is a need to connect the models to empirical evidence, despite there being reservations in terms of how well they could be suited for early warning (see Chapter 2.5) or other governance purposes. For this, monitoring programmes are needed to gather data over time, keeping in mind that it is not always clear which variables are meaningful. When gathering empirical data on cascades, it is important to note that quantifying cascading impacts is challenging due to measurement and monitoring. For example, dependencies on infrastructure systems can be far away from the affected area. As such, cascading impacts would not be apparent using traditional risk assessments (Lawrence et al., 2020). Furthermore, the extent to which this information could be gathered in real time and acted upon presents another set of challenges.