Advanced reactor technologies are being developed at an accelerated pace, offering many benefits over existing technologies, including reduced costs and enhanced safety. Nonetheless, advanced technologies will continue to be exposed to a range of natural hazards, including those associated with evolving hydrometeorological processes. Currently, risk-informed decision-making (RIDM) metrics used throughout the industry (e.g., annual core damage frequency) do not explicitly account for non-stationarity in hazard conditions, hindering the use of insights from probabilistic risk assessments (PRAs) or generation risk assessments (GRAs) in design, siting, and other decisions related to evolving processes. Moreover, incorporating non-stationarity into RIDM related to safety and operations is challenging because there is significant uncertainty about how changing hydrometeorological conditions will influence natural hazards, and associated uncertainty characterizations tend to feature qualitative rather than quantitative probabilities. This study describes ongoing work to compare the use of various approaches (including deterministic scenario-based approaches, adaptive engineering approaches, and time-dependent risk measures) for non-stationary RIDM, using a representative case study focusing on their efficacy in flood wall design for nuclear installations. We evaluate the performance of these approaches using an expected cost objective function and a simplified hazard-to-damage mapping function for specialized evaluation at a location of interest. The study is intended to inform the adaptation or development of non-stationary RIDM approaches for evolving conditions.