Arctic spatially distributed climate dynamics prognostics verification by Gaidai multivariate reliability method: a holistic approach

This study benchmarks and cross-validates the recently developed Gaidai multivariate environmental risk assessment concept by performing a spatiotemporal analysis of Arctic regional wind speed measurements. The latter is an essential contemporary research topic, given global climate change concerns and discussion on merchant navigation via the Northern Sea Route (NSR). Multi-dimensional structural and environmental dynamic systems, whether physically observed or numerically simulated over a representative time frame, are particularly appropriate for the Gaidai risk appraisal approach, as it imposes no limitations on system or data sample dimensionality. This work demonstrates that, with raw in-situ-measured environmental data (e.g., climate dynamics), it is feasible to conservatively estimate natural hazard risks for a selected geographic location within a reasonable forward-looking time frame, even with a relatively limited data sample. Intricate nonlinear cross-correlations among diverse environmental system dimensions (components or covariates), along with the high dimensionality of the system (data sample), hinder the application of existing reliability and natural hazard forecast approaches. Two distinct geographic locations along NSR were selected for this investigation, where NASA had recently measured daily wind speeds. Ocean wind speeds represent a complex environmental system, driven by climatic dynamic processes, that is challenging to model due to its nonlinear, multidimensional, and cross-correlated characteristics. Developing state-of-the-art, robust approaches for environmental-system natural-hazard assessment is a fundamental aspect of contemporary naval and offshore structural analysis, given the operational requirements under adverse weather conditions. The proposed multidimensional natural hazard assessment concept may be employed in the planning and development of resilient coastal stations or ports, particularly those situated on Arctic Ocean shores and thus vulnerable to extreme weather events. This study is purposely limited to a bivariate (2D) environmental system to cross-validate results against the existing 2D reliability scheme. Indeed, the Gaidai multivariate reliability method was extensively presented by the authors, but its cross-validation against existing reliability methods remains meaningful.