360 5. DISCUSSION The outcomes of this work highlight that advances in early-warning capability depend as much on organisational design as on analytical innovation. Transdisciplinary research consistently shows that complex, high-stakes problems are best addressed when teams establish a shared language, build a collaborative culture, and integrate diverse expertise through structured interaction-conditions that have been shown to be critical for success in real-world transdisciplinary environments. The two technical studies underpinning this work demonstrate how Slow Movement Analysis (SMA), combined with multi-interval processing and structured feature analysis, substantially improves the detection of low-velocity deformation—revealing precursory trends well before conventional real-time radar systems can respond. In parallel, AI-supported movement-pattern grouping and geotechnical validation show that these slow movements can be reliably differentiated into stable, pre-failure, and failure patterns, enabling clearer operational interpretation. In this work, AI-driven data-analysis tools were used to interpret thousands of movement patterns that would be impossible to assess manually. By analysing long-term displacement, velocity, and acceleration behaviours across different geological domains, these tools identified areas with similar deformation characteristics and highlighted where slow, emerging trends aligned with known precursors to instability— providing earlier, more confident insight for geotechnical decision-making Figure 3. AI‑Derived Movement Behaviour Clusters (Calderon, Barnes, 2025) This allowed engineers to clearly differentiate between normal background movement and early indicators of concern, providing a more confident basis for early-warning thresholds and enabling geotechnical teams to focus attention where it matters most, embedding these outcomes into the existent monitoring system as a complementary tool rather than modify the current systems.
RkJQdWJsaXNoZXIy MTM0Mzk2