Bias correction in data assimilation

Slides:



Advertisements
Similar presentations
Bias correction in data assimilation
Advertisements

ECMWF/EUMETSAT NWP-SAF Satellite data assimilation Training Course
Numerical Weather Prediction (Met DA) The Analysis of Satellite Data lecture 2 Tony McNally ECMWF.
Enza Di Tomaso* and Niels Bormann ECMWF *EUMETSAT fellow
Introduction to data assimilation in meteorology Pierre Brousseau, Ludovic Auger ATMO 08,Alghero, september 2008.
RADCOR for US Sondes Dr. Bradley Ballish NCEP/NCO/PMB 10 March 2011.
© The Aerospace Corporation 2014 Observation Impact on WRF Model Forecast Accuracy over Southwest Asia Michael D. McAtee Environmental Satellite Systems.
Data Assimilation Andrew Collard. Overview Introduction to Atmospheric Data Assimilation Control Variables Observations Background Error Covariance Summary.
Maturation of Data Assimilation Over the Last Two Decades John C. Derber Environmental Modeling Center NCEP/NWS/NOAA.
Slide 1 IPWG, Beijing, October 2008 Slide 1 Assimilation of rain and cloud-affected microwave radiances at ECMWF Alan Geer, Peter Bauer, Philippe.
ECMWF CO 2 Data Assimilation at ECMWF Richard Engelen European Centre for Medium-Range Weather Forecasts Reading, United Kingdom Many thanks to Phil Watts,
Xin Kong, Lizzie Noyes, Gary Corlett, John Remedios, Simon Good and David Llewellyn-Jones Earth Observation Science, Space Research Centre, University.
ECMWF – 1© European Centre for Medium-Range Weather Forecasts Developments in the use of AMSU-A, ATMS and HIRS data at ECMWF Heather Lawrence, first-year.
Data assimilation of polar orbiting satellites at ECMWF
Recent activities on utilization of microwave imager data in the JMA NWP system - Preparation for AMSR2 data assimilation - Masahiro Kazumori Japan Meteorological.
Diagnosing Climate Change from Satellite Sounding Measurements – From Filter Radiometers to Spectrometers William L. Smith Sr 1,2., Elisabeth Weisz 1,
1 Detection and Determination of Channel Frequency Shift in AMSU-A Observations Cheng-Zhi Zou and Wenhui Wang IGARSS 2011, Vancouver, Canada, July 24-28,
Slide 1 Sakari Uppala and Dick Dee European Centre for Medium-Range Weather Forecasts ECMWF reanalysis: present and future.
ECMWF NAEDEX 2012 – ECMWF Status Report – Stephen Engilsh ECMWF Status Report Stephen English ECMWF.
Lessons on Satellite Meteorology Part VII: Metop Introduction to Metop Instruments The sounders with focus on IASI The GRAS instrument The ASCAT scatterometer.
Satellite Bias Correction for CFSRR Haixia Liu, Russ Treadon, Robert Kistler, John Derber, Suru Saha and Hua-lu Pan Nov. 7, 2007 with input from Jack Woollen,
COSMO General Meeting, Offenbach, 7 – 11 Sept Dependance of bias on initial time of forecasts 1 WG1 Overview
Reanalysis: When observations meet models
Land Surface Analysis SAF: Contributions to NWP Isabel F. Trigo.
Assimilation of Satellite Radiances into LM with 1D-Var and Nudging Reinhold, Christoph, Francesca, Blazej, Piotr, Iulia, Michael, Vadim DWD, ARPA-SIM,
USE OF AIRS/AMSU DATA FOR WEATHER AND CLIMATE RESEARCH Joel Susskind University of Maryland May 12, 2005.
Hyperspectral Infrared Alone Cloudy Sounding Algorithm Development Objective and Summary To prepare for the synergistic use of data from the high-temporal.
Jinlong Li 1, Jun Li 1, Christopher C. Schmidt 1, Timothy J. Schmit 2, and W. Paul Menzel 2 1 Cooperative Institute for Meteorological Satellite Studies.
Hou/JTST NASA GEOS-3/TRMM Re-Analysis: Capturing Observed Rainfall Variability in Global Analysis Arthur Hou NASA Goddard Space Flight Center 2.
RADIOSONDE TEMPERATURE BIAS ESTIMATION USING A VARIATIONAL APPROACH Marco Milan Vienna 19/04/2012.
1 Bias correction in data assimilation Dick Dee ECMWF Meteorological Training Course Data Assimilation and Use of Satellite Data 11 May 2011.
ECMWF reanalysis using GPS RO data Sean Healy Shinya Kobayashi, Saki Uppala, Mark Ringer and Mike Rennie.
Infrared Sounding Data in the GMAO Data Assimilation System JCSDA Infrared Sounding Working Group (ISWG) 30 January 2009.
Global vs mesoscale ATOVS assimilation at the Met Office Global Large obs error (4 K) NESDIS 1B radiances NOAA-15 & 16 HIRS and AMSU thinned to 154 km.
The assimilation of satellite radiances in LM F. Di Giuseppe, B. Krzeminski,R. Hess, C. Shraff (1) ARPA-SIM Italy (2) IMGW,Poland (3)DWD, Germany.
© Crown copyright Met Office Assimilating infra-red sounder data over land John Eyre for Ed Pavelin Met Office, UK Acknowledgements: Brett Candy DAOS-WG,
June 20, 2005Workshop on Chemical data assimilation and data needs Data Assimilation Methods Experience from operational meteorological assimilation John.
Slide 1 Investigations on alternative interpretations of AMVs Kirsti Salonen and Niels Bormann 12 th International Winds Workshop, 19 th June 2014.
ECMWF/EUMETSAT NWP-SAF Satellite data assimilation Training Course Mar 2016.
The development of the NSST within the NCEP GFS/CFS
Indirect impact of ozone assimilation using Gridpoint Statistical Interpolation (GSI) data assimilation system for regional applications Kathryn Newman1,2,
ECMWF/EUMETSAT NWP-SAF Satellite data assimilation Training Course
What is atmospheric radiative transfer?
All-sky assimilation of microwave sounder radiances
ECMWF/EUMETSAT NWP-SAF Satellite data assimilation Training Course
Copernicus Atmosphere Monitoring Service
Joint GRWG and GDWG Meeting February 2010, Toulouse, France
Tony Reale ATOVS Sounding Products (ITSVC-12)
Radio occultation (RO) and its use in NWP
ECMWF/EUMETSAT NWP-SAF Satellite data assimilation Training Course
European Centre for Medium-Range Weather Forecasts
Data Assimilation Training
Bias correction in data assimilation
Impact of Traditional and Non-traditional Observation Sources using the Grid-point Statistical Interpolation Data Assimilation System for Regional Applications.
Vicarious calibration by liquid cloud target
Weak constraint 4D-Var at ECMWF
Cristina Lupu, Niels Bormann, Reima Eresmaa
Stéphane Laroche Judy St-James Iriola Mati Réal Sarrazin
Impact of hyperspectral IR radiances on wind analyses
The ECMWF weak constraint 4D-Var formulation
FSOI adapted for used with 4D-EnVar
Use of ATOVS at DAO Joanna Joiner, Donald Frank, Arlindo da Silva,
Initialization of Numerical Forecast Models with Satellite data
Infrared Satellite Data Assimilation at NCAR
Item Taking into account radiosonde position in verification
New DA techniques and applications for stratospheric data sets
Impact of aircraft data in the MSC forecast systems
Why use NWP for GSICS? It is crucial for climate and very desirable for NWP that we understand the characteristics of satellite radiance biases Simultaneous.
Assimilation of MW data in The C3S ERA5 Reanalysis
Atmospheric reanalysis at ECMWF
Presentation transcript:

Bias correction in data assimilation Niels Bormann, based on material from Hans Hersbach and Dick Dee Meteorological Training Course Data Assimilation 30 March 2017

Overview of this lecture In this lecture the variational bias correction scheme (VarBC) as used at ECMWF is explained. VarBC replaced the tedious job of estimating observation bias off-line for each satellite instrument or in-situ network by an automatic self-adaptive system. This is achieved by making the bias estimation an integral part of the ECMWF variational data assimilation system, where now both the initial model state and observation bias estimates are updated simultaneously. By the end of the session you should be able to realize that: many observations are biased, and the characteristics of bias vary widely depending on the type of instrument, distinguishing model bias from observation bias is often difficult, the success of an adaptive system implicitly relies on a redundancy in the underlying observing system.

Everyone knows that models are biased. Not everyone knows that most observations are biased as well. So… where is the bias term in this equation? model background constraint observational constraint

Outline Introduction Biases in models, observations, and observation operators Implications for data assimilation Variational analysis and correction of observation bias The need for an adaptive system Variational bias correction (VarBC) Extension to other types of observations Limitations due to the effects of model bias

ECMWF Meteo-France DWD Model bias: Systematic Day-3 Z500 errors in three different forecast models ECMWF Meteo-France DWD The above figure shows systematic Z500 errors at D+3 for three different models (ECMWF, Meteo-France and German Weather Service). The first point to notice is that at D+3 the different models show similar systematic errors. While this is perhaps not too surprisingly for the ECMWF and Meteo-France model (some of the model development is shared), the similarity between the ECMWF and the DWD model is astonishing. Summarizing, the above results show that different models have similar systematic error characteristics. Different models often have similar error characteristics Period DJF 2001-2003

Model bias: Seasonal variation in upper-stratospheric model errors T255L60 model currently used for the ERA-Interim reanalysis 40hPa (22km) 0.1hPa (65km) Summer: Radiation, ozone? 40hPa (22km) 0.1hPa (65km) Winter: Gravity-wave drag?

Observation bias: Radiosonde temperature observations Daytime warm bias due to radiative heating of the temperature sensor (depends on solar elevation and equipment type) Mean temperature anomalies for different solar elevations observed – ERA-40 background at Saigon (200 hPa, 0 UTC) Bias changes due to change of equipment

Observation and observation operator bias: Satellite radiances Monitoring the background departures (averaged in time and/or space): Constant bias (HIRS channel 5) Diurnal bias variation in a geostationary satellite Obs-FG bias [K] nadir high zenith angle Bias depending on scan position (AMSU-A ch 7) NH Trop SH Air-mass dependent bias (AMSU-A channel 14) Obs-FG bias [K]

Observation and observation operator bias: Satellite radiances – sources of bias Monitoring the background departures (averaged in time and/or space): HIRS channel 5 (peaking around 600hPa) on NOAA-14 satellite has +2.0K radiance bias against FG. Obs-FG bias [K] Same channel on NOAA-16 satellite has no radiance bias against FG. Obs-FG bias [K] NOAA-14 channel 5 has an instrument bias.

Observation and observation operator bias: Satellite radiances – sources of bias Different bias for HIRS due to change in spectroscopy used in the radiative transfer model: Obs-FG bias [K] Other common causes for biases in radiative transfer: Bias in assumed concentrations of atmospheric gases (e.g., CO2, aerosols) Neglected effects (e.g., clouds) Incorrect spectral response function …. Channel number Old spectroscopy New spectroscopy METEOSAT-9, 13.4µm channel: Drift in bias due to ice-build up on sensor: Sensor decontamination Obs –FG Bias

Implications for data assimilation: Bias problems in a nutshell Observations and observation operators have biases, which may change over time Daytime warm bias in radiosonde measurements of stratospheric temperature; radiosonde equipment changes Biases in satellite radiance measurements and radiative transfer models Biases in cloud-drift wind data due to problems in height assignment Models have biases, and changes in observational coverage over time may change the extent to which observations correct these biases Stratospheric temperature bias modulated by radiance assimilation This is especially important for reanalysis (trend analysis) Data assimilation methods are primarily designed to correct small random errors in the model background Systematic inconsistencies among different parts of the observing system lead to all kinds of problems

Implications for data assimilation: The effect of model bias on trend estimates Most assimilation systems assume unbiased models and unbiased data Biases in models and/or data can induce spurious trends in the assimilation

Surface air temperature anomaly (oC) with respect to 1987-2001 Implications for data assimilation: ERA-40 surface temperatures compared to land-station values Surface air temperature anomaly (oC) with respect to 1987-2001 Based on monthly CRUTEM2v data (Jones and Moberg, 2003) Based on ERA-40 reanalysis Northern hemisphere

Outline Introduction Biases in models, observations, and observation operators Implications for data assimilation Variational analysis and correction of observation bias The need for an adaptive system The variational bias correction scheme: VarBC Extension to other types of observations Limitations due to the effects of model bias

Variational analysis and bias correction: A brief review of variational data assimilation Minimise background constraint (Jb) observational constraint (Jo) The input xb represents past information propagated by the forecast model (the model background) The input [y – h(xb)] represents the new information entering the system (the background departures) The function h(x) represents a model for simulating observations (the observation operator) Minimising the cost function J(x) produces an adjustment to the model background based on all used observations (the analysis)

Variational analysis and bias correction: Error sources in the input data Minimise background constraint (Jb) observational constraint (Jo) Errors in the input [y – h(xb)] arise from: errors in the actual observations errors in the model background errors in the observation operator There is no general method for separating these different error sources we only have information about differences there is no true reference in the real world! The analysis does not respond well to conflicting input information A lot of work is done to remove biases prior to assimilation: ideally by removing the cause in practise by careful comparison against other data

Satellite radiance bias correction at ECMWF, prior to 2006 Scan bias and air-mass dependent bias for each satellite/sensor/channel were estimated off-line from background departures, and stored in files (Harris and Kelly 2001) Error model for brightness temperature data: where Periodically estimate scan bias and predictor coefficients: typically 2 weeks of background departures 2-step regression procedure careful masking and data selection Average the background departures: Predictors, for instance: 1000-300 hPa thickness 200-50 hPa thickness surface skin temperature total precipitable water

The need for an adaptive bias correction system The observing system is increasingly complex and constantly changing It is dominated by satellite radiance data: biases are flow-dependent, and may change with time they are different for different sensors they are different for different channels How can we manage the bias corrections for all these different components? This requires a consistent approach and a flexible, automated system

The Variational Bias Correction scheme: The general idea The bias in a given instrument/channel (bias group) is described by (a few) bias parameters: typically, these are functions of air-mass and scan-position (the predictors) These parameters can be estimated in a variational analysis along with the model state (Derber and Wu, 1998 at NCEP, USA) The standard variational analysis minimizes Modify the observation operator to account for bias: Include the bias parameters in the control vector: Minimize instead What is needed to implement this: The modified operator and its TL + adjoint A cycling scheme for updating the bias parameter estimates An effective preconditioner for the joint minimization problem

Variational bias correction: The modified analysis problem The original problem: Jb: background constraint Jo: observation constraint Jb: background constraint for x J: background constraint for  Jo: bias-corrected observation constraint The modified problem: Parameter estimates from previous analysis A model for the observation bias

The need for an adequate bias model Prerequisite for any bias correction is a good model for the bias (b(x,β)): Ideally, guided by the physical origins of the bias. In practice, bias models are derived empirically from observation monitoring. Constant bias (HIRS channel 5) Diurnal bias variation in a geostationary satellite Obs-FG bias [K] Air-mass dependent bias (AMSU-A ch 10) nadir high zenith angle Bias depending on scan position (AMSU-A ch 7) 1.7 1.0 Obs-FG bias [K] 0.0 -1.0

Example 1: Spinning up new instruments – IASI on MetOp A IASI is a high-resolution interferometer with 8461 channels Initially unstable – data gaps, preprocessing changes

Example 2: NOAA-9 MSU channel 3 bias corrections (cosmic storm) 200 hPa temperature departures from radiosonde observations Variational bias correction smoothly handles the abrupt change in bias: initially QC rejects most data from this channel the variational analysis adjusts the bias estimates bias-corrected data are gradually allowed back in no shock to the system!

Example 3: Fit to conventional data Introduction of VarBC in ECMWF operations

Outline Introduction Biases in models, observations, and observation operators Implications for data assimilation Variational analysis and correction of observation bias The need for an adaptive system Variational bias correction (VarBC) Extension to other types of observations Limitations due to the effects of model bias

Extension to other types of observations Current bias ‘classes’ in the ECMWF operational system: Radiances: clear sky/all sky, infrared/microwave, polar/geostationary Total column ozone: currently only OMI Aircraft data: one group per aircraft Ground-based radar precipitation: one group embracing US stations Other automated bias corrections, but outside 4D-Var: Surface pressure Radiosonde temperature and humidity Soil moisture (in SEKF surface analysis) Specific: ERA-Interim: VarBC for radiances only ERA-20C: the 20th century reanalysis using surface observations only MACC: atmospheric composition

VarBC for satellite radiances ~1,500 channels (~40 sensors on ~25 different satellites) Anchored to each other, GPS-RO, and all conventional observations Bias model: β0 + ∑βi pi(model state) + ∑βj pj(instrument state) (~11,400 parameters in total)

VarBC for ozone OMI, (SCIAMACHY, GOMOS, SEVIRI, GOME2, GOME in past) Anchored to SBUV/2 Bias model: β0 + β1 x solar elevation

VarBC for aircraft temperature For each aircraft separately (~5000 distinct aircraft) Anchored to all temperature-sensitive observations Bias model: β0 + β1 x ascent rate + β2 x descent rate Average temperature departures for the northern hemisphere during a 2-week period Aircraft Radiosonde Control Impact

Outline Introduction Biases in models, observations, and observation operators Implications for data assimilation Variational analysis and correction of observation bias The need for an adaptive system Variational bias correction (VarBC) Extension to other types of observations Limitations due to the effects of model bias

Limitations of VarBC: Interaction with model bias VarBC introduces extra degrees of freedom in the variational analysis, to help improve the fit to the (bias-corrected) observations: It works well (even if the model is biased) when the analysis is strongly constrained by observations: model abundant observations It does not work as well when there are large model biases and few observations to constrain them: model observations VarBC is not designed to correct model biases: Need for a weak-constraint 4D-Var (Laloyaux)

Limitations of VarBC: Interaction with model bias Example: Stratospheric temperature biases Using VarBC for all satellite radiances sensitive to stratospheric temperature leads to unrealistic drifts in the bias corrections due to model bias (red line) Additional anchoring is imposed through assimilating AMSU-A channel 14 without a bias correction (blue line) (channel sensitive to temperature around 1-5 hPa) (channel sensitive to temperature around 2-10 hPa)

Summary Biases are everywhere: Most observations cannot be usefully assimilated without bias adjustments Manual estimation of biases in satellite data is practically impossible Bias estimates can be updated automatically during data assimilation Variational bias correction works best in situations where: there is sufficient redundancy in the data; or there are no large model biases Challenges: How to develop good bias models for observations How to separate observation bias from model bias

Additional information Harris and Kelly, 2001: A satellite radiance-bias correction scheme for data assimilation. Q. J. R. Meteorol. Soc., 127, 1453-1468 Derber and Wu, 1998: The use of TOVS cloud-cleared radiances in the NCEP SSI analysis system. Mon. Wea. Rev., 126, 2287-2299 Dee, 2004: Variational bias correction of radiance data in the ECMWF system. Pp. 97-112 in Proceedings of the ECMWF workshop on assimilation of high spectral resolution sounders in NWP, 28 June-1 July 2004, Reading, UK Dee, 2005: Bias and data assimilation. Q. J. R. Meteorol. Soc., 131, 3323-3343 Dee and Uppala, 2009: Variational bias correction of satellite radiance data in the ERA-Interim reanalysis. Q. J. R. Meteorol. Soc., 135, 1830-1841 Han and Bormann, 2016: Constrained adaptive bias correction for satellite radiance assimilation in the ECMWF 4D-Var system. ECMWF Technical Memorandum 783. Feel free to contact me with questions: Niels.Bormann@ecmwf.int