SWOT Surface Water and Ocean Topographysome notes from the latest Science Team meeting in June 2018

Nadia Ayoub

with inputs fromRosemary Morrow

Outline

1. SWOT mission: general characteristics and main scientific goals2. SSH signal at high resolution: physical content, present nadir altimetric data3. SWOT in the coastal ocean4. SWOT CAL/VAL

from swot.jpl.nasa.gov

SWOT – 2D view of the surface water topography

- KaRIn altimeter• SAR - alongtrack• Interferometry – crosstrack

- effective resolution over the oceans: ~15 km

- conventional nadir altimeter

- launch in 2021

SWOT – scientific objectives (oceanography)

primary goal

mesoscale and submesoscaledynamics in the open ocean

- observability of small-scales fromaltimetry

other expected progresses at the end of the mission improve the accuracy on estimates of

tides at global scales reduce uncertainties on the geoid

- impact of small-scales dynamicson horizontal and vertical transport

- ocean cascade from large to small scales

- observability of sub-surface variability from SSH

secondary objective

coastal dynamics

SWOT – Orbits and products

fast sampling orbit – 1-day repetitivitymonths 1 to 6CAL/VAL phase

kml files available onwww.aviso.altimetry.fr

SWOT – Orbits and products

Science orbit – 21-day repetitivity

figure à inclure

No. Obs

in 21 days

courtesy of R. Morrowkml files available on www.aviso.altimetry.fr

SWOT – Products

Two families of products:

1. hydrological (high-resolution) products• available over the continents and in a narrow coastal band (defined in a ‘HR mask’)• 4 sub-categories pixel-cloud, river, lakes, raster • pixel-cloud resolution: a few tens of meters (across and along track)• raster resolution ~100-250 m• accuracy : a few tens of centimeters

2. oceanographic (low-resolution) products3 sub-categories: - basic oceanographic product (2 km resolution/posting)

- expert low resolution product (2 km resolution/posting)- expert high resolution product (500m resolution/ 250m posting)

basic oceanographic product over the whole ocean + continents

data accuracy depends on• resolution (the higher resolution, the lower accuracy)• onboard processing• accuracy on the geophysical corrections (post-processing)• the location within the swath

CFOSAT : Chinese-French Oceanic SATellitemonitoring of the ocean surface wind and waves

Sentinel -3BJason CS/Sentinel 6 – 2020 and 2026SWOT

+ constellation missions ?

2008 Jason 2

2013 SARAL

2018

20202021

Past, current and future altimetric missions

from www.aviso.altimetry.fr

2016 Sentinel 3A

Present altimetry at small-scales: high-resolution nadir altimetry, SAR data

Mean SSH anomaly spectrum (May-June 2012) in the Tropical Pacific

hump

instrumental white noise

3 km

Corruption of J2 data below ~100km:

- instrumental white noise seen as a 3 cm plateau for wavelengths < 3km

- spatially coherent error= ‘spectral hump’ due to inhomogeneity in the backscatter properties of the oceansurface

“sophisticated editing, filtering, or postprocessing schemes are needed 1) to mitigate the spectral hump and 2) to keep a good coverage of altimeter measurements”

Main contributor to measurement errors at small-scales(from Dibarboure et al., 2014)

Present altimetry at small-scales: high-resolution nadir altimetry, SAR data

Mean SSH anomaly spectrum (May-June 2012) in the Tropical Pacific

Benefit of SAR measurements (from Dibarboure et al., 2014)

no spectral hump due to a smallerfootprint

error contamination less than 50% of the signal at scales as small as 30km

30 km70-80 km

Present altimetry at small-scales: high-resolution nadir altimetry, SAR data

Ka-band vs Ku-band altimetry at high resolution(from Bonnefond et al., 2018)

Jason2 data at 20 Hz (~300m)Sentinel-3A at 20 HzAltika/Saral at 40 Hz (~125m)

SLA power spectrum withspecific editing and corrections

KuSARKa

Better observability of small-scales( ~30-100 km) in Altika dataNo hump because of a smaller and statistically more homogeneous footprint

30 km

Physical content of the SSH signal at small-scales

« in the 15-200 km range of scales pertinent to SWOT, the geostrophically balanced motion may lose its dominance and be overtaken by the unbalanced wave motions, including near-inertial flows, internal tides, and inertia–gravity waves » (Qiu, 2018)

- large geographical and seasonalvariability

- results from MITGCM simulations

- Vergara and Morrow findconsistent results from nadir altimetry

from Qiu et al., JPO, 2018

major issue addressed by the SWOT science team:

how to separate geostrophic flow from internal gravity waves signals at scales ranging

between 15 km and 200 km ?

Physical content of the SSH signal at small-scales – Internal waves

Physical content of the SSH signal at small-scales – Internal waves

nadir altimetry

HYCOM model

M2 internal tides signature in SSH(from Shriver et al. 2012)

Many efforts :

- to model internal gravity waves- to characterize coherent /incoherentinternal tides- ultimately to propose a method to filter out the coherent IT signal fromSSH

SWOT – in the coastal ocean

SWOT will bring a huge amount of observations in the coastal regions with: a much better resolution data closer to the coast a better accuracy

than current nadir altimetry

because SWOT targets both oceanic and hydrological processes large impact on estuaries-deltas studies are also expected very favourable scientific environment with both communities involved

links with modeling/DA studies• SWOT data will allow to test the meso/submesoscale dynamics at the global

scale as simulated by high-resolution (HR) models• in preparation to the mission, HR simulations help to better understand the

physical content of the future SWOT observations• assimilation of SWOT data expected to improve ocean circulation hindcasts and

forecasts provided the complex error budget is properly taken into account

• high-resolution data available only on continents and nearshore (but not everywhere !) definition of a HR mask

• the very large volume of data to download,process and store is the main constraint on the HR mask

• definition of the mask in progress (as for June 2018)

• possibility of a seasonal mask

• because of the complex geometry of the coast and due to the method of defining the mask, some coastal areas of interest may not be covered

SWOT – in the coastal ocean

Availability of high-resolution data in the nearshore

Laignel et al., SWOT ST meeting, June 2018

most studies: • spectrum averaged along long-segments• averages over large areas (several thousands of km)• averages in deep ocean

what about coastal ocean ? • shorter segment availables• smaller dynamics scales + non isotropy• higher error budgets

SWOT – in the coastal ocean

Physical content of the SSH signal (a spectral approach)

Example in the Bay of Biscay(M.L. Dabat, N. Ayoub,

R. Morrow, et al.)

Jason2 - 20 Hz Sentinel-3A - 20 HzAltika/Saral - 40 Hz

SWOT – in the coastal ocean

Physical content of the SSH signal (a spectral approach)

seasonal variability

CAL/VAL objectives (Bonnefond et al., 2018)• continuity between past, present and future missions • reliability between offshore, coastal and inland altimetric measurement• characterize varied geographically correlated errors in an absolute sense

SWOT – launch in Sep 2021 pre-launch campaigns (eg pilot projects in California) fast-sampling phase dedicated to cal/val Jan-Mar 2022 (but the exact time window

may shift) nominal phase post-launch campaigns

need of a multi-sites approach to test different dynamical regimes rely on a broader community and agency than the one involved in the ST

SWOT – CAL/VAL

several discussions/propositions under progress within the Science Team ‘adopt a cross-over’ initiative

Overview of possible 1-day sitesSWOT Launch Sept 2021

Fast sampling phase (1 day orbit) : Jan-Fev-Mar 2022

Science phase (global coverage, 21-day orbit) : 3 yr > Mar 2022

Proposed In SWOT CalVal plan Under discussion

courtesy of R. Morrow

next NASA/ROSES and CNES/TOSCA call in February or March 2019

Additional slides

Physical content of the SSH signal at small-scales

« in the 15-200 km range of scales pertinent to SWOT, the geostrophically balanced motion may lose its dominance and be overtaken by the unbalanced wave motions, including near-inertial flows, internal tides, and inertia–gravity waves » (Qiu, 2018)

- large geographical and seasonalvariability

- results from MITGCM simulations

- Vergara and Morrow findconsistent results from nadir altimetry

from Qiu et al., JPO, 2018

Lt

Lt

Seasonal wavenumber spectra for surface KEin the 500 km x 500 km box centered at 37°N, 164° E.

Workshop : Towards High-Resolution of Ocean Dynamics and Terrestrial

Surface Waters from Space

21 – 22 October 2010, Lisbon, Portugal

Carl Wunsch :

• Key point 1 : experience strongly suggests that models must be tested against observations on all time and space scales. One of the major issues facing oceanography is that high resolution models become fundamentally untestable, so we need more repeated observations at finer space and time scales.

• Key point 2 : new oceanographic instruments working on previously unmeasured time or spatial scales, have led to surprising, sometimes startling, discoveries that were not predicted.

7

courtesy of R. Morrow

Today’s Alongtrack altimetry technology

Conventional

Low

Resolution

Mode (LRM)

Signal & noise

averaged over

5-7 km radius

footprint

Synthetic

Aperture

Radar (SAR)

Mode –

delayed

doppler

300 m

alongtrack, 5-

7 km

crosstrack

radius

CR2-SAR

Sentinel-3 (2016)

J-CS (2020)

T/P – Jason

Envisat

SARAL 3

LRM waveform footprint = disc-shaped surface of the sea that is illuminated to create the altimeter waveform (its radius is approximately 10km for Jason-2 and 7 km for Cryosat-2 LRM);

leading edge footprint = surface of the sea that is illuminated when the backscattered energy increases on the first nonzero waveform gates (radius of the order of 1–3km on Jason-2).

Altimeter footprint geometry on a 3-km segment for (a) LRM and (b) SARM. Darker zones highlight the leading-edge footprint, providing the bulk of the SSH content through the leading edge of the waveform. The tiny black shape at the bottom of each subplot illustrates how a single SARM footprint (i.e., waveform) can be affected by isolated spurious reflections, whereas larger LRM footprints repeatedly sample the error over multiple waveforms (i.e., correlated error).

from Dibarboure et al., 2014

Physical content of the SSH signal at small-scales

« in the 15-100 km range of scales pertinent to SWOT, the geostrophically balanced motion may lose its dominance and be overtaken by the unbalanced wave motions, including near-inertial flows, internal tides, and inertia–gravity waves » (Qiu, 2018)

Seasonal wavenumber spectra for surface KEin the 500 km x 500 km box centered at 37°N, 164° E.

from Qiu et al., JPO, 2018

Lt

Lt