your name On the specification of the Background Error Covariance Matrix for Wave Data Assimilation Systems • Jesús Portilla
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
your name
On the specification of the Background Error Covariance Matrix for Wave Data Assimilation Systems
• Jesús Portilla
your name
Introduction
• Motivation for Data Assimilation• Model and observations usually don’t match
• Users tend to trust more in observations
• Some situations are simply too difficult to model
• Some areas have dense monitoring networks that is a pity not to use to improve model results
your name
Introduction
• Background errors determine the extent and the magnitude in
which observations get introduced into the model wave field
your nameObjective DA Statistical DA
a b o bx x K x x
2
2 2b
o b
K
• Statistical DA concept
your name
1 2
,,
i
o mi j j
i j
J Q x x
Optimization problem
• The DA scheme
• Variational (3DVAR, 4DVAR)
• Optimal interpolation
• Kalman Filtering
• Adjoint modelling
• Neural Networks
• . . .
error covariance matrixerror covariance matrix
0J
2 0J
3DVAR
your name
• Background error covariance matrix (BECM)
Covariance (Target)
Variances (can be estimated, e.g., triple co-location)Correlation coefficient
(can be estimated, e.g., via the R2)
i j
iji i
w w
w w
Greenslade, D.J.M. and I.R. Young, 2005: The impact of Altimeter Sampling Patterns on Estimates of Background Errors in a Global Wave Model, J. Atmos. Oc. Tech., 22, No. 12 pp 1895 – 1917.
your name
• The North Sea
Voorrips A.C., V.K. Makin, and S. Hasselmann., 1997: Assimilation of wave spectra from pitch-and-roll buoys in a North Sea wave model, J. Geophys. Res., 102 (C3), 5829-5849
Parametric error correlation length
(using wave height)
expa
d
L
3 / 2
200( )
a
L km
• Background errors (parametric)
K13
your name
• Some remarks about Background errors● Our current knowledge about the structure (“shape and dimensions”) of Background Errors is very poor.
• For consistent DA, wave conditions must be homogeneous, isotropic, and ergodic over the assimilation domain.
• The computation of the Background Errors should consider the wave spectrum as the reference variable and not integral parameters like Hs.
• Background Errors depend on wave climate, which in turn might be characterized by the existence of different regimes. For a proper specification of the BECM each wave system has to be considered independently. • The wave climate and therefore the BECM is point specific and season dependent.
your name
• Wave climate
MODEL
BUOY
wind sea
wind sea
swell
swell
• Buoy Hs = 4.2 m
• Model Hs = 2.7 m
• Matching observations and model spectra
your name
• is the partition spectrum
• The truth is emulated from WWIII model output
• Computation of the BECM
i j
iji i
w w
w w
2
221
analyzed true
ij
true true
S SR
S S
,S S f
your name
• The spectral correction model
@ @, * * ,analysis remote o true obs o oS f S f
energy correctionfrequency correction
direction correction
• Each wave system is corrected individually
• No assumptions are made about the wind-sea or swell condition
your name
• Calculating for two main wave systems (e.g., North Sea)
your name
• Calculating for two main wave systems (e.g., North Sea)
your name
Northerly system K13 Southwesterly system K13 Parametric (general)
• Calculating for two main wave systems (e.g., North Sea)
your name
• Summary
• A method for the computation of the BECM has been developed
• This method considers explicitly:
a) The local spectral wave climate
b) The spectral correction model to be applied
• Assumptions about the wind-sea or swell condition are not used
• Conclusions
• The developed method allows calculating the BECM objectively on statistical bases
• The computed BECM’s implicitly define the spatial domain where the conditions of isotropy and homogeneity are fulfilled
• The condition of ergodicity can be included for instance by computing BECM’s for each season
your name
References
Voorrips A.C., V.K. Makin, and S. Hasselmann., 1997: Assimilation of wave spectra from pitch-and-roll buoys in a North Sea wave model, J. Geophys. Res., 102 (C3), 5829-5849
Greenslade, D.J.M. and I.R. Young, 2005: The impact of Altimeter Sampling Patterns on Estimates of Background Errors in a Global Wave Model, J. Atmos. Oc. Tech., 22, No. 12 pp 1895 – 1917.
Thanks for your attention!
your name
Related Documents
