© Crown copyright Met Office UM 4D-Var Regional Reanalysis Progress Richard Renshaw, Stephen Oxley, Adam Maycock, Peter Jermey, Dale Barker, Tom Green, DingMin Li
Jan 29, 2016
© Crown copyright Met Office
UM 4D-Var Regional Reanalysis ProgressRichard Renshaw, Stephen Oxley, Adam Maycock, Peter Jermey,
Dale Barker, Tom Green, DingMin Li
© Crown copyright Met Office
Contents
Technical highlights
First full reanalysis 2008/9
Validation
Developments for 2013/14
© Crown copyright Met Office
WP2.1 Building capacity for advanced regional data assimilation
orography
12km grid
480 x 384
reanalysis period 2008/2009
© Crown copyright Met Office
Technical Highlights
• Capability to generate ODBs
© Crown copyright Met Office
ODB – obs monitoring database
• ODB stores observations + qc + O-B + O-A + ...
• Established ECMWF database + utilities
• Array of tools available “for free”
• Metview macros (quick look)
• Obstat (detailed stats / graphics)
© Crown copyright Met Office
Technical Highlights
• Capability to generate ODBs
• Able to archive reanalysis fields in ECMWF mars
© Crown copyright Met Office
First run 2008/9
© Crown copyright Met Office
4 Parallel Streams
2008 2009
A
BC
D
with 1 month overlap for spin-up
© Crown copyright Met Office
How long to spin up ?
rms screen temperature
© Crown copyright Met Office
How long to spin up ?
rms surface pressure
© Crown copyright Met Office
Var Resolution
UM 12km
Var 24km
© Crown copyright Met Office
Var Resolution
4DVar run time:• 36km 1 node hour• 24km 3 node hours• 12km 20 node hours
UM T+241.5 node hours
© Crown copyright Met Office
Var Resolution: 24 vs 36 km
© Crown copyright Met Office
Observations• Surface (SYNOP, buoy, etc) incl visibility• Upper air (sonde, pilot, wind profiler)• Aircraft• AMV (‘satwinds’)• GPS-RO• Scatterometer winds• ATOVS• AIRS• IASI• GPSRO• MSG clear sky radiances
© Crown copyright Met Office
Bias correction ofsatellite radiances
Initial reanalysis:
Radiances processed, not assimilated
Final reanalysis:
Radiances assimilated
monthly bias statistics
© Crown copyright Met Office
Validation
Peter Jermey
© Crown copyright Met Office
Verification - Results
+2.7 wtd skill diff
+6.6 wtd skill diff
+3.2 wtd skill diff
•Verifying at T+6 – a good analysis should produce a good short range forecast
Jan 08 Jul/Aug 08
Sept/Oct 09
© Crown copyright Met Office
Statistics and
EXTREME Statistics!
© Crown copyright Met Office
First run 2008-2009
© Crown copyright Met Office
Statistics
Standard Statistics• Mean• Std Dev• Range
Extreme Statistics• Max of Daily Max• Max of Daily Min• Min of Daily Max• Min of Daily Min• Icing days, Frost days, Summer days, Tropical NightsSummer Days
Count of days for which daily max T>25 degrees
• Maximum count of consecutive dry/wet days
• Count of wet days, count of days with precip above 10mm, 20mm.
Extreme Statistics are defined as ‘core indices’ of climate change by The joint CCI/CLIVAR/JCOMM Expert Team on Climate Change Detection and Indices.
Wet Days A day on which precip is greater than 1mm
• Average precip on wet days, max precip on 5 consecutive days, maximum precip on a single day, total precip, precip above 95th and 99th percentiles.
Percentiles For base period (ERA) 1961-1990
• Percentage of days where max temp > 90th percentile etc…
•We calculate these statistics for ERA and EURO4M and compare with observations statistics from European Climate Assessment & Dataset ECA&D http://eca.knmi.nl
© Crown copyright Met Office
Validation of climate indicesPeter Jermey
© Crown copyright Met Office
Validation of climate indicesPeter Jermey
© Crown copyright Met Office
July 2008
© Crown copyright Met Office
July 2008 Floods
ERA
EURO4M
ERA
EURO4M
RMS
Mean
73
69 -36
-37
Max of 5 Daily Precip/mm
mm
•EURO4M closer to Obs•Both models not as wet as Obs
•EURO4M more detail
© Crown copyright Met Office
Developments for 2013
© Crown copyright Met Office
Cloud assimilation
NAE assimilates 3D cloud fields from nowcasting system
(combines satellite imagery + surface reports)
EURO4M will have to rely on using surface reports directly
© Crown copyright Met Office
Wattisham, 00Z 2012/03/13AAXX 13004 03590 11238 83504 10064 20060 30240 40352 53002 60001 71022 886// 92350 333 55/// 20411 84703 86706 88708
Cloud from SYNOP reportsPeter Francis
© Crown copyright Met Office
Wattisham, 00Z 2012/03/13AAXX 13004 03590 11238 83504 10064 20060 30240 40352 53002 60001 71022 886// 92350 333 55/// 20411 84703 86706 88708
Cloud from SYNOP report
84703 4 oktas Stratus, height 90m
© Crown copyright Met Office
Wattisham, 00Z 2012/03/13AAXX 13004 03590 11238 83504 10064 20060 30240 40352 53002 60001 71022 886// 92350 333 55/// 20411 84703 86706 88708
Cloud from SYNOP report
84703 4 oktas Stratus, height 90m
86706 6 oktas Stratus, height 180m
88708 8 oktas Stratus, height 240m
© Crown copyright Met Office
Precipitation assimilation
• Operational UK models assimilate radar rainrate (latent heat nudging)
• For EURO4M, aim to assimilate raingauge accumulations
© Crown copyright Met Office
Precipitation assimilation
Plan
Use E-Obs gridded daily precipitations
Keith Ngan, Andrew Lorenc, Richard
© Crown copyright Met Office
Precipitation assimilation
Plan
Use E-Obs gridded daily precipitations
System to disaggregate 24hr accumulations to 6hrs
Var outer loop with spin-up problems minimised
(analysis increments trigger rain in model).
Keith Ngan, Andrew Lorenc, Richard
© Crown copyright Met Office
n
ib
airi
scan xfccbias1
)(
Variational Bias Correction
Airmass-dependent bias correction of satellite radiances (based on Harris and Kelly, 2001)
Currently coeffs c are calculated off-line monthly
VarBC will give smooth and automatic updating
Code is in place – hope to trial in 2013
(DingMin Li, Andrew Lorenc , Dale Barker)
© Crown copyright Met Office
Collaboration – Cross-Validation
Compare our reanalysis against:• SMHI• ERA• Obs climatologies
Peter
© Crown copyright Met Office
Summary
• Initial reanalysis is run (2008-9)
• new validation tools
• Production reanalysis will be better:
• satellite radiances
• surface cloud
• ODBs and mars archive
• A final reanalysis aims to include later developments
• precipitation assimilation
• variational bias correction
© Crown copyright Met Office
Questions ?
© Crown copyright Met Office
Extra slides...
© Crown copyright Met Office
model orography
ERA-Interim
Model T255 (80km)
Var T159 (125km)
Met Office
Model 12km
Var 24km
© Crown copyright Met Office
ERA-Interim vs EURO4M
• 12km, 70 levels• 24km 4D-Var• 6-hour analysis window• assimilate:
• conventional obs incl vis• satellite radiances• GPS (ground & RO)• Cloud• Precipitation
• Initial state and boundary conditions from ERA-Interim analyses
• T255 (80km), 60 levels• T159 (125km) 4D-Var• 12-hour analysis window• assimilate:
• conventional obs• satellite radiances
© Crown copyright Met Office
Observation processing
Corrections to radiosonde temperature, surface pressure,
- use same as UKMO Global
Rejection lists
- use old UKMO Global and NAE lists
© Crown copyright Met Office
Observations from ECMWF
• Surface (SYNOP, buoy, etc) incl visibility• Upper air (sonde, pilot, wind profiler)• Aircraft• AMV (‘satwinds’)• ATOVS• AIRS• IASI• GPSRO• MSG clear sky radiances
© Crown copyright Met Office
Observations from MetDB
• Ground-based GPS• Scatterometer winds
© Crown copyright Met Office
Verification
• Description of Work: Temperatur
e,Wind,
Water Vapour,
Surf. P., Surf. Radiation Budget,
Earth Radiation Budget,Cloud
Properties,SST, Precip,Snow Cover, and
Visibility.
Rel. Hum.
• Skill Scores (as in NWP Index):
Skill =(Per(Anal))2 – Fc2
(Per(Anal))2
• ETS as (in UK Index):
Tot/Base
Total Cloud
Cloud Base
3, 5, 7 octs.
100m, 500m, 1500m Precip 0.5mm, 1mm, 4mm
Vis (1.5m) 200m, 1000m, 4000m
© Crown copyright Met Office
February 2009
© Crown copyright Met Office
February 2009 Snow
Min of Daily Min Temp/degrees
Max of Daily Precip/mm
ERA
EURO4M
© Crown copyright Met Office
February 2009 Snow
Min of Daily Min Temp/degrees
ERA
EURO4M
ERA
EURO4M
RMS
Mean
2.8
2.3 0.9
0.3
•EURO4M closer to Obs•EURO4M larger bias•EURO4M and ERA warmer than obs
© Crown copyright Met Office
February 2009 Snow
ERA
EURO4M
ERA
EURO4M
RMS
Mean
7.5
6.2 -0.45
-5.6
Max of Daily Precip/mm
•EURO4M closer to Obs•EURO4M smaller bias•EURO4M and ERA dryer than obs
© Crown copyright Met Office
Reports of 24hr accum precip
© Crown copyright Met Office
12Z Reports of 6hr precip
© Crown copyright Met Office
00Z Reports of 6hr precip
© Crown copyright Met Office
15Z Reports of 6hr precip
© Crown copyright Met Office
Covariances
NAE
Old covariances from Global model
Horizontal length scales are ‘guesses’
EURO4M
Covariances calculated (NMC method) in CVT
Horizontal length scales also from CVT
Marek, Gordon, Jean-Francois
© Crown copyright Met Office
CVT covariances: psi horizontal length scales
350km
100km
vertical mode (1 to 70)
© Crown copyright Met Office
CVT covariances: psi horizontal length scales
350km
100km
vertical mode (1 to 70)
NAE 180km
© Crown copyright Met Office
Cov: CVT vs NAE