A Lagrangian uncertainty quantification approach to validate ocean model datasets

This work presents a methodology to measure how well the material transport produced by different ocean models aligns with observational data, using their trajectories as a basis for comparison. To this end, recent results that relate an uncertainty metric to invariant dynamic structures are used. T...

Descripción completa

Detalles Bibliográficos
Autores: García-Sánchez, G., Agaoglou, M., Smith, E.M.C., Mancho, A.M.
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:España
Institución:Consejo Superior de Investigaciones Científicas (CSIC)
Repositorio:DIGITAL.CSIC. Repositorio Institucional del CSIC
OAI Identifier:oai:digital.csic.es:10261/420763
Acceso en línea:http://hdl.handle.net/10261/420763
https://www.scopus.com/inward/record.uri?eid=2-s2.0-105004208281&doi=10.1016%2Fj.physd.2025.134690&partnerID=40&md5=8944ae9e24d515cef2fb1f52aee819ce
Access Level:acceso abierto
Palabra clave:Invariant dynamical structures
Material transport
Ocean models
Uncertainty quantification
Ocean engineering
Dynamic structure
Dynamical structure
Invariant dynamical structure
Invariant dynamics
Lagrangian
Observational data
Ocean model
Uncertainty
Uncertainty quantifications
Lagrange multipliers
Descripción
Sumario:This work presents a methodology to measure how well the material transport produced by different ocean models aligns with observational data, using their trajectories as a basis for comparison. To this end, recent results that relate an uncertainty metric to invariant dynamic structures are used. These connections shed light on how to implement statistical averaging strategies to systematically assess the quality of the ocean data set and its performance in terms of Lagrangian transport. The method is applied using both reanalysis and analysis data in the North Atlantic, where observed drifter trajectory data serve as benchmarks for validation. To assess the reliability of the proposed methodology, it is tested alongside a comparable, purpose-built example conducted under controlled conditions within the same region. We present evidence that the proposed methodology provides valuable information on model performance on different spatial and temporal scales. © 2025 The Authors