You are here: Home / Research / Groups / Echanges Côte-Large / Scientific projects / FES2012 / Data assimilation

Data assimilation

by ECOLA last modified Feb 11, 2013 04:11 PM

Data preparation:

  • data decimation (altimetry-decimate)
    • decimation from error wavelength
    • editing from GLORYS-v2
    • along-track outliers detection
  • detidor (tide gauges)

 

Data decimation

We aim to automatically process the proper data density for data assimilation from an appropriate criterion, such as model error length scales. The error wave length can be estimated from the difference between  the prior solutions and FES2004 or GOT4.8.

 

M2.FES2012GOT48.jpg

M2 tide - FES2012 prior versus GOT48

 

Data editing

We aim to eliminate data that can be contaminated by non-tidal SSH signal. Error estimates coming from the harmonic analysis is a first guess of the level contamination. We wish to double the editing criterion by examining non-tidal SSH signal energy at aliased tidal frequencies from GLORYS products. The next two plots is an illustration for the M2 tide.

 

filtered60RMS.jpeg

~60 days filtering

TP60atlas.jpeg

~60 days harmonic analysis

Ocean circulation ssh signal at M2 TP/J1/J2 aliased frequency

 

M2GLORYS.jpg

GLORYS-v2 ~60 days harmonic analysis

300 km smoothing

M2TPJ1J2.jpg

TP/J1/J2 M2 error estimates

 

 

 

Data error estimate, contamination part

Similarly to data editing, final error estimates will be based on both harmonic analysis diagnostics and GLORYS estimates

 

Data error estimate, data depletetion/processing part

In some areas (such as coastal areas), loss of data or data processing limitations can seriously degrade the quality of the harmonic constants. The smoothness of along-track data will be examined to detect outliers.

 

 

 

Document Actions

logo cnes logo IRD Logo université de Toulouse Logo université Paul Sabatier Logo CNRS
Logo bibliothèque OBS Logo Observatoire Midi Pyrénées