1. SMOS-BEC – Barcelona (Spain) Revealing Geophysically-Consistent Spatial Structures in SMOS Surface Salinity Derived Maps Marcos Portabella, Estrella Olmedo, Justino Martínez, Antonio Turiel SMOS Barcelona Expert Centre Pg. Marítim de la Barceloneta 37-49, Barcelona SPAIN E-mail: smos-bec@icm.csic.es URL: www.smos-bec.icm.csic.es

2. SMOS-BEC Outline of the talk SMOS L2 standard product Accounting for non-linearities New product design Geophysical assessment with singularity analysis Conclusions

3. SMOS-BEC SMOS standard L2 product (I)

4. SMOS-BEC SMOS standard L2 product (II) Xi 2 /Nm>1.25 ALL THESE POINTS ARE FILTERED Nm=33 Nm=97 Nm=137 Xi 2 /Nm>1.25 ALL THESE POINTS ARE FILTERED Xi 2 /Nm>1.25 ALL THESE POINTS ARE FILTERED Comparison of distribution functions (real->grey theoretical->green): PORTION OF DISGARDED POINTS IS NOT ALLWAYS THE SAME

5. SMOS-BEC SMOS standard L2 product (III) 5 / 10 Nm=33Nm=97 Nm=137 Comparison of density functions (real->red theoretical->blue): >95% <5% BAD QUALITY GOOD QUALITYGOOD QUALITY GOOD QUALITYGOOD QUALITY GOOD QUALITYGOOD QUALITY >95% <5% BAD QUALITY

6. SMOS-BEC Accounting for nonlinearities (I) 1.- Floor error: smaller on 1st Stokes parameter

7. SMOS-BEC Accounting for nonlinearities (II) 2.- Land sea contamination and 18-day subcycle: 18-day is the (approx.) repeat subcycle of the satellite; land-sea contamination has this inner cycle also (J. Tenerelli, private comm.) From 9-day averages From 18-day averages Annual meanIntra-annual variabilityInter-annual variability

8. SMOS-BEC Accounting for nonlinearities (III) 3.- Climatology may induce some biases

9. SMOS-BEC Accounting for nonlinearities (IV) 4.- Some tails, many ripples  Nodal sampling Average reduction of std. dev. of 0.7 K

10. SMOS-BEC Accounting for nonlinearities (V) 5.- Histograms of single retrievals of SSS have greater dispersion than expected 18-day single-angle SSS histograms

11. SMOS-BEC Design of a new product (I) Taking into account the discussed non-linear effects, we have designed a new L3 product: Daily OTT Derived from 1 st Stokes parameter 18-day, 0.5º resolution Nodal sampling applied SSS is computed averaging all single-angle SSS retrievals 3 psu around the mode of their distribution. We have compared it with the standard DPGS product

12. SMOS-BEC Standard 18-day binned map (CP34-BEC)

13. SMOS-BEC Mean around the mode

14. SMOS-BEC Standard 18-day binned map (CP34-BEC)

15. SMOS-BEC Mean around the mode

16. SMOS-BEC ZoneDescriptionLatitudeLongitude GlobalTropics and mid-latitudes 60S-60NAll TropicTropics 30S-30NAll NPacA region of the North Pacific 45N-60N170E-140W Z122A region of the South Eastern Pacific 30S-0N150W-120W Z124A region of the South Western Tropical Pacific 24S-10S165E-165W Z126Equatorial Oceans 10S-10NAll Z131Southern Ocean 60S-40SAll Z132Intertropical Pacific 5N-15N110W-180W OTTZone used for the OTT computation 45S-5S140W-95W Comparison with Argo Real-Time Mode ARGO comparison GlobalTropicNpacZ122Z124Z126Z131Z132OTT CP34- BEC (18-d) Mean -0.19-0.20-0.62-0.09-0.10-0.220.02-0.21-0.05 Std 0.380.330.690.200.280.330.530.260.23 Nbuoys 3720222166151372798566125295 Mean around mode Mean 0.03-0.05-0.170.070.04-0.080.43-0.060.25 Std 0.560.491.110.250.380.430.800.330.28 Nbuoys 4243260288155386842530125297

17. SMOS-BEC Geophysical assessment via SA (I) What is singularity analysis? All ocean scalars are submitted to the action of flow advection (the same for all them) plus other dynamic effects (specific). The ocean is a turbulent flow, both in 3D (vertical mixing, small scales of order of meters) and in 2D (horizontal dispersion, important at sub-mesoscale and greater scales). In a turbulent flow, advection creates sharp changes and irregularities in all scalars, which may be of small amplitude but are well characterized by a dimensionless scalar field: the singularity exponents field (denoted by h). Singularity analysis is a sophisticated mathematical technique to extract singularity exponents from maps of a given scalar.

18. SMOS-BEC What is SA useful for? Singularity exponents of different ocean scalars must correspond, so they can be used to assess the quality of different products Geophysical assessment via SA (II) SST SSS  h = 0,15

19. SMOS-BEC

20. Standard 18-day binned map (CP34-BEC)

21. SMOS-BEC Mean around the mode

22. SMOS-BEC

24. Standard 18-day binned map (CP34-BEC)

25. SMOS-BEC Mean around the mode

26. SMOS-BEC Conclusions Many on-linear effects impact SSS retrievals in SMOS. We have made an effort to reduce those effects and to create new maps as less biased as possible. When compared to Argo, the new maps are globally less precise (greater. std. dev.) but more accurate (smaller bias) Singularity analysis reveals consistent geophysical structures in the new products.

27. SMOS-BEC GUI Singularity Analysis 27 / 10 http://cp34-bec.cmima.csic.es/CP34GUIWeb/ Singularity Analysis Web Service is now operational (registered users)

28. Last meeting during 2 nd SMOS Science Conference (May 2015) SMOS-Mission Oceanographic Data Exploitation SMOS-MODE www.smos-mode.eu info@smos-mode.eu SMOS-MODE supports the network of SMOS ocean-related R&D