Towards an OTT strategy: analysis of OTTs made from November 2010 reprocessed L1c

www.argans.co.uk

SMOS L2 Ocean Salinity

Towards an OTT strategy:

analysis of OTTs made fromNovember 2010 reprocessed L1c

20 October 2011

ARGANS

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Background:•November 2010 L1c reprocessed by Deimos L1PP v500, single slope AM

•OTT made using v500 L2OS library

•Orbit selection:

• ascending or descending

• > 104 grid points in South Pacific, latitude 5 to 45 south

•Grid points filtered:

• ocean (> 200km from land)

• no ice, missing ECMWF or heavy rain

• wind speed > 3m/s & < 12 m/s

•Measurements filtered:

• for sun glint: L1 flags sun point/tails/fov/area; L2 flag high sun glint

• for moon glint: L1 moon point flag; L2 moon specular direction flag

• for galactic glint: Fm_gal_noise_error/GN1 or Fm_high_gal_noise/GN2

• for L1 or L2 RFI, & high TEC gradient

• ignore measurements with L1 Fm_border_fov or Fm_error flags

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Selecting orbits for OTTs

Grid points in ascending/descending regions

OTT selected when n > 104

44 ascending, 44 descending

Ascending OTT region:121±16W @ 5S, 111±16W @ 45S

Descending OTT region:112±16W @ 5S, 123±16W @ 45S

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

OTT measurement filtering

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Ascending orbit OTTs, November 2010: X pol

Drift +0.15 K / 30 days2 week cycle/wave?

NIR8/11/2010

NIR23/11/2010

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Ascending orbit OTTs, November 2010: Y pol

Drift +0.15 K / 30 days

NIR8/11/2010

NIR23/11/2010

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Descending orbit OTTs, November 2010: X pol

AFFOV drifts away from EAFFOV after 18 November

NIR8/11/2010 NIR

23/11/2010

Outlier: ascending orbit!

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

Descending orbit OTTs, November 2010: Y pol

NIR8/11/2010

NIR23/11/2010

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

ascending 20101102T123646 descending 20101102T014617 merged

Typical OTT: 2 November

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

9 ascending OTT 10 descending OTT 19, both ascending + descending

Average OTTs, 1-7 November 2010

Vertical lines: sun tails?

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

ascending 20101102T123646 descending 20101102T014617 merged

Vertical line is because OTT algorithm filters L1c sun tails & sun fov flags, set around xi = -0.25 to 0.3 as the sun moves during the month – maybe we shouldn’t filter?

Typical OTT: 2 November: measurement count after filtering

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

ascending 20101102T123646 descending 20101102T014617 merged

Typical OTT: 2 November: std(err)/ra

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

9 ascending OTT 10 descending OTT 19, both ascending + descendingSmoother/closer to 1.0: lower sigmaSun & tails impact descending OTT

Averaging OTTs, 1-7 November 2010: std(err)/ra

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

6% measurements rejected due to high galactic noise(1.6% GN1, 5.4% GN2)

Maybe galactic noise error/bias should be propagated into OTTs?

OTT pathology: 20101101T004509

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

20% measurements rejected because L1c sun glint area flag set(3% for sun tails)

L1c sun tails & area flags ignored in L2OS – ignore in OTT?

OTT pathology: 20101130T142711

Snapshot from L1c

Sun tail

www.argans.co.uk/smos

SMOS L2 Ocean Salinity

84% measurements rejected due to L2 detected RFI (actually poor model-data fit):ignore L2 rfi flag when making OTT?

12 bad L1c orbits, 20101110 & 20101130: problem with L1 calibration?

OTT pathology: 20101110T140533

All L1c snapshots look similar

SSS