Results 1  10
of
10
and J. CameronEstimating interchannel observation error correlations for IASI radiance data in the Met Office system
, 2012
"... error correlations for IASI radiance data ..."
(Show Context)
Conditioning and Preconditioning of the Minimisation Problem in Variational Data Assimilation
, 2011
"... Many numerical weather prediction (NWP) centres around the world implement a variational data assimilation (Var) scheme to find the initial state of the atmosphere, called the analysis. The analysis is used as the initial conditions for a numerical forecast model. For an accurate weather forecast an ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Many numerical weather prediction (NWP) centres around the world implement a variational data assimilation (Var) scheme to find the initial state of the atmosphere, called the analysis. The analysis is used as the initial conditions for a numerical forecast model. For an accurate weather forecast an accurate analysis is essential. Var is formulated as a optimization problem and is solved by a series of minimisations of linear leastsquare cost functions. The speed of convergence of these minimisations and the sensitivity of the analysis to perturbations are dependent on the condition number of the Hessian of the leastsquares cost function. A small condition number of the Var Hessian is essential for an accurate forecast. Many NWP centres perform a control variable transform (CVT) in order to solve a preconditioned Var (PVar) scheme. In this thesis we consider the conditioning of Var and PVar in detail by deriving new theoretical bounds on the condition number of the Var and PVar Hessians. Using the bounds we show that the Var Hessian is illconditioned when the error covariance matrix of the prior estimate is illconditioned. We also show that preconditioning with the CVT produces a significant reduction in the condition number of Var. Additionally, we show using the theoretical bounds that the condition number of the PVar Hessian is reduced if we increase the spacing of observations, reduce the accuracy of the observations and reduce the number of observations. We demonstrate these results numerically for both a simple oneparameter periodic system and the Met Office PVar scheme. We also demonstrate that the CVT produces a significant increase in the convergence rate of the conjugate gradient method used to solve the Var scheme.
A Practical Method to Estimate Information Content in the Context of 4DVar Data Assimilation. I: Methodology.
 Journal of Geophysical Research,
, 2011
"... Abstract Data assimilation obtains improved estimates of the state of a physical system by combining imperfect model results with sparse and noisy observations of reality. Not all observations used in data assimilation are equally valuable. The ability to characterize the usefulness of different da ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Abstract Data assimilation obtains improved estimates of the state of a physical system by combining imperfect model results with sparse and noisy observations of reality. Not all observations used in data assimilation are equally valuable. The ability to characterize the usefulness of different data points is important for analyzing the effectiveness of the assimilation system, for data pruning, and for the design of future sensor systems. In the companion paper [Sandu et al.
Using Observations at Different Spatial Scales in Data Assimilation for Environmental Prediction
"... Observations used in combination with model predictions for data assimilation can contain information at smaller scales than the model can resolve. Errors of representativity are errors that arise when the observations can resolve scales that the model cannot. Little is known about representativity ..."
Abstract
 Add to MetaCart
(Show Context)
Observations used in combination with model predictions for data assimilation can contain information at smaller scales than the model can resolve. Errors of representativity are errors that arise when the observations can resolve scales that the model cannot. Little is known about representativity errors, and consequently they are currently not correctly included in assimilation schemes. The aim of this thesis is to understand the structure of representativity error, and investigate if the assimilation can be improved by correctly accounting for representativity error. Thefirstapproach is to usean existing method that assumes that the model state is a truncation of a high resolution truth. Using the KuramotoSivishinky equation as the model, it is shown that representativity error is correlated. It is also shown that the correlation structure depends not on the number of observations but the distance between them. The representativity error is also affected by the observation type and model resolution. Using the same method representativity error is calculated for temperature and specific humidity fields from the Met Office high resolution model. This shows that representativity error is more significant for specific humidity than temperature and that representativity error is
School of Mathematical and Physical Sciences
, 2012
"... Data assimilation with correlated observation errors: analysis accuracy with approximate error covariance matrices by ..."
Abstract
 Add to MetaCart
(Show Context)
Data assimilation with correlated observation errors: analysis accuracy with approximate error covariance matrices by
School of Mathematical and Physical Sciences
, 2013
"... Scheduling satellitebased SAR acquisition for sequential assimilation of water level observations into flood modelling by ..."
Abstract
 Add to MetaCart
(Show Context)
Scheduling satellitebased SAR acquisition for sequential assimilation of water level observations into flood modelling by
School of Mathematical and Physical Sciences
, 2012
"... assimilation scheme with a coastal area morphodynamic model of Morecambe Bay by G.D. Thornhill, D.C. Mason, S.L. Dance, A.S. Lawless, N.K. Nichols and H.R. ForbesIntegration of a 3D Variational data assimilation scheme with a coastal area morphodynamic model of ..."
Abstract
 Add to MetaCart
assimilation scheme with a coastal area morphodynamic model of Morecambe Bay by G.D. Thornhill, D.C. Mason, S.L. Dance, A.S. Lawless, N.K. Nichols and H.R. ForbesIntegration of a 3D Variational data assimilation scheme with a coastal area morphodynamic model of
EBB EEEEEDECBEEE !"#EE BB$ECEDEAB%EB&E
, 2012
"... Data assimilation with correlated observation errors: experiments with a 1D shallow water model ABCDEFB ..."
Abstract
 Add to MetaCart
Data assimilation with correlated observation errors: experiments with a 1D shallow water model ABCDEFB