APPLICATIONS OF SEA-LEVEL PRESSURE

Transcription

1 APPLICATIONS OF SEA-LEVEL PRESSURE RETRIEVAL FROM SCATTEROMETER WINDS Jérôme Patoux, Ralph C. Foster and Robert A. Brown OVWST meeting Seattle, November 20, 2008

2 Pressure retrieval: how does it work? model function U 10 PBL model V GR gradient wind equation p fi t a pressure fi eld by least-squares minimization swath of pressure gradients

3 How good are the resulting pressure fields? Compare the pressure difference between two buoys with the corresponding pressure difference in the UWQS pressure swath W 160 W 150 W 140 W 130 W 60 N N N 30 N N 10 N

4 UWQS vs. buoy ECMWF vs. buoy UWQS BPG (hpa) Buoy BPG (hpa) R 2 = a= 0.4±0.3 b=1.083± Buoy BPG (hpa) R 2 =0.995 a= 0.1± b=0.982± ECMWF BPG (hpa) 40 (a) Buoy BPG (hpa) 40 (b) Buoy BPG (hpa)

5 W 140 W 120 W 80 W 70 W 60 W 50 W 40 W 50 N 60 N N N N 20 N N 10 N Repeat the comparison with all possible pairs of NDBC buoys over the QS period.

6 UWQS vs. buoy ECMWF vs. buoy UWQS BPG (hpa) Buoy BPG (hpa) R 2 = a=0.5±0.0 b=1.014± Buoy BPG (hpa) R 2 =0.989 a=0.3± b=0.980± ECMWF BPG (hpa) 40 (a) Buoy BPG (hpa) 40 (b) Buoy BPG (hpa)

7 W 160 W 150 W 140 W 130 W 60 N N N 30 N N 10 N How do the UWQS pressure swaths compare to the ECMWF global sea-level pressure fields? Calculate the rms difference between ECMWF sea-level pressure fi eld and QS swaths that fall within one hour of synoptic time. 2 N.H. rms (hpa) 1 S.H. Tropics Wind vector cell

8 How do the UWQS sea-level pressure spectra compare to ECMWF? Zonal wavelength (km) Zonal wavelength (km) SOUTH PACIFIC TROPICAL PACIFIC /3 5/ Power (hpa 2 ) UWQS ECMWF 3 UWQS 10 1 Power (hpa 2 ) ECMWF Zonal wavenumber (cyc/km) Zonal wavenumber (cyc/km)

9 For more details, see: Patoux, J., R. C. Foster and R. A. Brown (2008): An evaluation of scatterometer-derived oceanic surface pressure fi elds, J. Applied. Meteor. Clim., 47,

10 Application #1: Can the UWQS slp fields help improve weather forecasting? Implementation at the Ocean Prediction Center (Joe Sienkiewicz, Joan Von Ahn) OPC QS GFS UWQS Apr 2005

11 The UWQS sea-level pressure fi elds contain mesoscale information that is absent from NWP analyses. Example: secondary lows (frontal waves? mesoscale lows? mesocyclones?) W W 60 N S N S

12 Application #2: Can the UWQS slp fields help improve midlatitude cyclone statistics (intensity, structure, tracks, etc.)? 0 30 W 30 E 60 W 60 E 90 E 90 W 120 W 120 E 150 W 150 E 180

13 Maybe... but building cyclone tracks from scatterometer-derived pressure swaths alone is impractical, because cyclone centers can fall in the gaps N (a) W 140 W (b) 40 W 20 W 0 20 S N 40 S N S

14 QS-modified ECMWF sea-level pressure - 27 Jan :00 UTC/02:56 UTC 30 S 60 S Original 30 S 60 S Inject UWQS mesoscale pressure information into ECMWF analysis by: decomposing the two pressure fi elds by wavelet analysis swapping the high wavenumber wavelet coeffi cients reconstructing the ECMWF pressure fi eld enhanced by UWQS mesoscale detail Modified 30 S 60 S UWPBL 0 30 E 60 E 90 E 120 E 150 E W 120 W 90 W 60 W 30 W 0

15 QS-modified ECMWF sea-level pressure - 27 Jan :00 UTC/02:56 UTC 30 S 60 S Original 30 S 60 S Modified 30 S 60 S UWPBL 0 30 E 60 E 90 E 120 E 150 E W 120 W 90 W 60 W 30 W 0

16 Patoux, J., X. Yuan and C. Li (2009): Satellite-Based Midlatitude Cyclone Statistics Over the Southern Ocean. Part I: Scatterometer-Derived Pressure Fields and Storm Tracking, J. Geophys. Res., in press W 40 W 40 S 60 W 40 W S 996 (a) (b) (c) In the enhanced UWPBL sea-level pressure fi elds: 5-10% more cyclone centers are identifi ed the cyclones are slightly deeper than in ECMWF 8% tracks are initiated at least 6 hours earlier, while 7% tracks are extended by at least 6 hours the UWPBL sea-level pressure fi elds contain ~1% more spectral energy than ECMWF

17 Fluxes associated with a January 2003 storm 02 Jan Jan Jan Jan Jan S 60 S pressure (mb) S 60 S 300 E 330 E 330 E E E 0 30 E 60 E stress (10 1 N m 2 ) Patoux (2007)

18 Yuan, X., J. Patoux and C. Li (2009): Satellite-Based Midlatitude Cyclone Statistics Over the Southern Ocean. Part II: Tracks and Surface Fluxes, J. Geophys. Res., in press. SPATIALLY INTEGRATED FLUXES SPATIALLY AVERAGED FLUXES Whole ocean 0.4 Inside storms Stress (10 12 N) Inside storms Stress (N m 2 ) Whole ocean 0 (a) (b) When the air-sea fl uxes are calculated from the enhanced UWPBL sea-level pressure fi elds: the spatially integrated stress magnitude is 7.8% higher inside cyclones and 2.1% higher over the entire Southern Ocean the spatially averaged stress magnitude is 4.6% higher inside cyclones and 3.1% higher over the entire Southern Ocean

19 Application #3: Can we assimilate the UWQS slp fields in a NWP model? Bob Atlas, Joe Ardizzone...? (4:30... stay tuned...)

20 Application #4: If we re-derive the surface winds from the UWQS sea-level pressure fields, how do they compare with the original QS winds?

21 Compare QS/DIRTH/UWPBL surface winds with buoy winds RAIN FREE RAIN FLAGGED W 160 W 150 W 140 W 130 W 60 N N N 30 N N 10 N Mean QS speed (m s 1 ) QuikSCAT DIRTH UWPBL Mean QS speed (m s 1 ) Direction difference standard deviation (a) (b) (d) (e) Direction difference standard deviation Buoy speed (m s 1 ) Buoy speed (m s 1 )

22 Compare QS/DIRTH/UWPBL surface winds with buoy winds W 160 W 150 W 140 W 130 W 60 N N N 30 N N 10 N Vector correlation (rain-free) QS DIRTH UWPBL UWPBL (dir. only) All buoys North Pacifi c North Atlantic Vector correlation (rain-flagged) QS DIRTH UWPBL UWPBL (dir. only) North Atlantic = 2, perfect vector correlation

23 Calculate the rms difference between ECMWF and QS/DIRTH/UWPBL surface wind speed and direction NORTHERN HEMISPHERE SOUTHERN HEMISPHERE Speed std dev (m s 1 ) QS DIRTH UWPBL 20 Dir std dev (m s 1 ) Wind vector cell Wind vector cell

24 Compare QS/DIRTH/UWPBL and ECMWF spectra EDGES SWEET SPOTS NADIR Frequency (cyc./25 km) Frequency (cyc./25 km) Frequency (cyc./25 km) Power (m 2 s 2 ) QS Power (m 2 s 2 ) 10 0 DIRTH UWPBL EC Frequency (cyc./25 km) Frequency (cyc./25 km) Frequency (cyc./25 km)

25 Application #5: We can use the sea-level pressure retrieval as an independent measure of performance (ASCAT, Oceansat-2, etc.) 31Mar07 22:42 UTC 120 E 150 E W 31Mar07 17:27 UTC 120 E 150 E W 60 N 60 N 30 N 30 N S 30 S 60 S 60 S pressure (hpa)

26 THANK YOU!