Data Assimilation with Local Translation Error Analysis

• Activity Number: 425 - Contributed Poster Presentations: Uncertainty Quantification in Complex Systems Interest Group
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Uncertainty Quantification in Complex Systems Interest Group
• Author(s): Kazuyuki Nakamura*
• Companies: Meiji University
• Keywords: Data assimilation; Particle filter; Sequential Bayesian filter

Abstract

Data assimilation is the technique which aims at combining numerical simulation model with observation data. Several types of sequential Bayesian filters are used in the data assimilation field, but the way to choose which algorithm should be used is unclear. To provide guidance for the choice of the algorithm, we tested the use of local translation error (LTE) approach. LTE is an index for quantifying local determinism of chaotic systems that are related to noise sensitivity and nonlinearity of systems in a short period. In our presentation, we will show the application result of the LTE analysis in data assimilation with Lorenz 96 model. We will also discuss the applicability of the LTE analysis for data assimilation.