The Workshop on Flow-dependent aspects of data assimilation was from 11 to 13 June 2007.
Description
The workshop considered advances in data assimilation methods that address the flow-dependence of the analysis problem. Current assimilation methods for operational numerical weather prediction rely on time-averaged covariance statistics that may be close to optimal on average, but which are quite incorrect in extreme situations such as intense baroclinic development, strong organised convection and tropical cyclones. Such cases are of particular interest, both to weather forecasters and to the general public, because of their unusual nature, high intensity, and impact on society. Flow dependence is important in less extreme situations too. Properly accounting for the day-to-day variation of error statistics, as well as their anisotropies and inhomogeneities, has the potential to significantly improve analysis quality.
Flow dependence would also allow more effective use of the available observations, and facilitate quality control that could retain extreme observations. Topics for the workshop included: ensemble-assimilation methods, state-dependent modelling of background-error statistics, flow-dependent quality control, data selection, and targeting of observations.
Programme
Introduction and working group reports
Presentations
| Flow dependence in global data assimilation systems | |
|---|---|
|
Ideas for adding flow-dependence to the Met Office VAR system Andrew Lorenc (Met Office) |
|
|
Covariance modelling in a grid-point analysis Jim Purser (NCEP) |
|
|
Use of analysis ensembles in estimating flow-dependent background error variances Mike Fisher (ECMWF) |
|
| Ensemble approaches | |
|---|---|
|
Ensemble Jb modelling Mark Buehner (Environment Canada) |
|
|
Impacts of flow-dependent background-error covariances in the NCEP Global Forecast System Jeff Whitaker (NOAA/ESRL) |
|
|
Use of analysis ensembles in estimating flow-dependent background error variances Lars Isaksen (ECMWF) |
|
|
Implementation of the LETKF on the NCEP Global Forecast System Istvan Szunyogh (University of Maryland) |
|
| Change of variable approaches | |
|---|---|
|
Flow-dependent transforms Ian Roulstone (University of Surrey) |
- |
|
A regime-dependent balanced control variable based on potential vorticity Ross Bannister (University of Reading) |
|
|
Background errors in HIRLAM variational data assimilation Magnus Lindskog (HIRLAM) |
|
|
Diabatic omega equation Luc Fillion (Environment Canada) |
|
|
Humidity control variable and total water Elias Holm (ECMWF) |
|
| Implementation issues | |
|---|---|
|
Spatial filtering of analysis ensemble statistics Olivier Pannekoucke (Meteo-France) |
|
|
Representation of inhomogeneous, non-separable covariances by sparse wavelet transformed matrices Andreas Rhodin (DWD) |
|
|
Ensemble forecasting and flow-dependent estimates of initial uncertainty Martin Leutbecher |
|
Proceedings
|
A regime-dependent balanced control variable based on potential vorticity R N Bannister |
|
|
A variational assimilation ensemble and the spatial filtering of its error covariances: increase of sample size by local spatial averaging L Berre |
|
|
Ensemble-based background-error covariances in variational data assimilation M Buehner |
|
|
Case dependent implicit normal mode balance operators S Lee |
|
|
The sensitivity of analysis errors to the specification of background error covariances M Fisher |
|
|
Humidity control variable and supersaturation E V Holm |
|
|
Use of analysis ensembles in estimating flow-dependent background error variance L Isaksen |
|
|
Ensemble forecasting and flow-dependent estimates of initial uncertainty M Leutbecher |
|
|
Background errors in HIRLAM variational data assimilation M Lindskog |
|
|
Ideas for adding flow-dependence to the Met Office VAR system A Lorenc |
|
|
Covariance modelling in a grid-point analysis R Purser |
|
|
Representation of inhomogeneous, non-separable covariances by sparse wavelet- transformed matrices A Rhodin |
|
|
Flow-dependent transforms I Roulstone |
|
|
The Local Ensemble Transform Kalman Filter and its implementation on the NCEP global model at the University of Maryland I Szunyogh |
|
|
Impacts of flow-dependent background-error covariances in the NCEP Global Forecast System J Whittaker |
|