EUMETSAT04 04/2004 © Crown copyright Use of EARS in Global and Regional NWP Models at the Met Office Brett Candy, Steve English, Roger Saunders and Amy Doherty Satellite Applications, Met Office
EUMETSAT04 04/2004 © Crown copyright
Use of EARS in Global and Regional NWP Models at the Met
Office
Brett Candy, Steve English, Roger Saunders and Amy Doherty
Satellite Applications, Met Office
EUMETSAT04 05/2004 © Crown copyright
Introduction
Operational forecast model data cutoff
Impact of lost ATOVS data
Results of experiments
Use of EARS data at the Met Office
Plans for the future
EUMETSAT04 05/2004 © Crown copyright
Met Office NWP Models Global
– Global ATOVS from NESDIS
– 110 min cutoff (main run)
UK Mesoscale– Local ATOVS reception
from West Freugh, Scotland
– 120 min cutoff
European Mesoscale– Global ATOVS from
NESDIS
– 180 min cutoff (currently)
EUMETSAT04 05/2004 © Crown copyright
Percentage of data used for each observation type in a global model main run
%
EUMETSAT04 05/2004 © Crown copyright
Arrival Times of Data
• North Atlantic Region
• Six-hour window 25/02/2004 09:00-15:00
window
0
10
20
30
40
50
60
70
80
90
100
9 10 11 12 13 14 15 16 17 18
Global EARS
Main Run Update Run
First Overpass
EUMETSAT04 05/2004 © Crown copyright
No Cutoff Experiment
All available ATOVS data were used in the Global model and results compared to a control run in which no data arriving after the cutoff were assimilated.
Results showed the positive impact of the missing data.
EUMETSAT04 05/2004 © Crown copyright
RMS Error reduces when all ATOVS data are used.
Large peak seen in the T+24 forecasts is less pronounced.
EUMETSAT04 05/2004 © Crown copyright
Case Study26th May 2003 North Pacific
Cyclone
No cutoff experiment predicts a deeper cyclone, closer match to verifying analysis (979 hPa)
982 hPa
978 hPa
EUMETSAT04 05/2004 © Crown copyright
Using EARS in the Global Model Trial ran for 3 weeks from the 24th February 2002
NOAA15 & 16 EARS data from seven stations
Number of Additional Obs in Northern Hemisphere
0
1000
2000
3000
4000
5000
6000
7000
8000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Trial Day
EARS+Global
Global
EUMETSAT04 05/2004 © Crown copyright
Example of the Extra DataGlobal data
assimilated in main run
EARS and Global data assimilated in main run
NOAA16 & NOAA15
NOAA16
EUMETSAT04 05/2004 © Crown copyright
Forecast Impact Overall Neutral Benefit in the NH Extra-Tropics
Positive Benefit when significant extra data assimilated
500 hPa height field
0
10
20
30
40
50
60
70
80
90
24 48 72 96 120 144
Forecast Range (hours)
RM
S E
rro
r (m
)
EARS+Global
Global
EUMETSAT04 05/2004 © Crown copyright
Use in Operational NWP Ears included in latest package of changes to operational system
Initially using 3 stations:
– Tromsø (Norway), Maspalomas (Canaries), Edmonton (Canada).
Package also includes:
– First use of AIRS radiances
– Other ATOVS processing changes
» Introduce AMSU radiances from AQUA
» Move to RTTOV-7 forward model
» More data over land
Pre-op trial indicates rise of 3 % on a global index (equivalent to yearly
improvement from all NWP changes)
EUMETSAT04 05/2004 © Crown copyright
Conclusions The use of late satellite data can occasionally have
large forecast impacts.
Using the EARS data to fill in voids from the global
data results in positive forecast benefit.
Near future: use of EARS data in Euro model.
Additional application: AMSU precipitation imagery for
duty forecasters.