Testing a particle filter to reconstruct climate changes over the past centuries
(2010) MeteoClim symposium — Location: Brussels, Belgium (5.November.2010)
Files
No attached file found for this publication.
Details
- Authors
- Abstract
- Particle filtering is a relatively new and promising approach for data assimilation in climate models. It is a challenging task to develop a non-degenerative particle filter for a large-scale geophysical model using a relatively small number of particles. We implement a particle filter in a climate model focusing on 100 years time scale. Several tests are performed with particle filtering using pseudo-observations obtained from twin experiment with the coupled climate model LOVECLIM, as well as using real-data observations over the last century. These tests demonstrate that at a reasonable cost, it is possible to obtain a model output correlated well with the observations at the large scales.
- Affiliations
Citations
Dubinkina, S., Goosse, H., Sallaz-Damaz, Y., Crespin, E., & Crucifix, M. (2010). Testing a particle filter to reconstruct climate changes over the past centuries. MeteoClim symposium, Brussels, Belgium. https://hdl.handle.net/2078.5/206266