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.

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


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


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)


Percentage of data used for each observation type in a global model main run

%


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


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.


RMS Error reduces when all ATOVS data are used.

Large peak seen in the T+24 forecasts is less pronounced.


T+48 pressure forecast error difference plots


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


EARS Station Coverage

24 hours

of data


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


Example of the Extra DataGlobal data

assimilated in main run

EARS and Global data assimilated in main run

NOAA16 & NOAA15

NOAA16


Difference in Analysis Increments from the Main Forecast Runs


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


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)


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.

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.

Documents