Conclusion Longer reaches smaller errors. However, wider rivers, which are less affected by random errors, had a more stable response to reach length.

Sinuosity method less variability in reach length than hydraulic controls. Few short reaches produced by the HC method could dominate pass-based error statistics, especially for narrow rivers under low discharges.

Biases caused by terrain layover dominated height error statistics for the Sacramento River. Correction methods eliminating pixels with high likelihood of layover could result in improvement

Synthetic SWOT measurements2

Geophysical data processing

Interferogram formation

Height reconstruction

Instrument and media errors

Builds perfect interferograms (no noise)

Adds noise to received power

Adds noise to interferograms (right and left)

Classifies the surface (detects bodies of water)

Reads satellite orbit

Constructs baseline origin coordinates, velocity vector, etc

Builds swath grid

Classificationerrors

Position and height errors

Jet Propulsion Lab’s SWOT Hydrology simulator

Data requirementsTerrain DEM (high resolution ~3m for realistic layover simulation) DEM of the water surface (e.g. derived from a hydraulic model) Water mask

Desirable characteristics: Simulations covering a variety of discharges with a long domain spamming several reaches (50 km and longer)

AcknowlegementsFunded by NASA SWOT Algorithm Definition Team contract to the Ohio State University.

We would like to acknowledge Brent A. Williams, Dani Esteban-Fernandez, and Eva Peral for their support operating the SWOT instrument simulator, Toby Minear for the original HEC-RAS model of the Sacramento River, and Guy Schumann and Alessio Domeneghetti for the original hydraulic model of the Po River.

Results4

Simulated overpasses

Gro

un

d t

rack

Ne

ar

ran

ge

Fa

r ra

ng

e

Legend

SwathWater depth (m)

11

0.1

Elevation (m)

60

-2

0 3015Km

Scale

Sacramento River study area Po River study area

60

-25

28

0

Swath

Elevation (m) Water depth (m)

Far

0 20 40Km n

ear

near

Gro

und

track

Simulated overpasses

Sacramento River

Average width: 102 m to 150 m3 3Discharges:118 m /s to 510 m /s

Simulation length: 6 months9 overpasses

Po River:

Average width: 340 m to 655 m 3 3Discharges: 700 m /s to 4770 m /s

Simulation length: 1 year14 overpasses showing both swaths

Heig

ht

RM

SE

, cm

8

10

12

14

16

18Sacramento River

10km reaches20km reachesHydraulic controlsSinuosity

Decreasing terrain layover

Heig

ht

RM

SE

, cm

12.5

13.5

14.5

15.5Sacramento River

Slo

pe R

MS

E, cm

/km

0

1

2

3

4Predicted errorriver widths:

50 m 150m

Height, width, slope errors

1.5

2.5

3.5

4.5

Po River

Arbitrary lengthHCSinuosity

0

1

2

3

4Predicted errorriver widths:

200 m 400m

0.05

0.07

0.09

0.11

0.13

0.15

Average reach length, km0 5 10 15 20 25

Wid

th r

RM

SE

0.03

0.04

0.05

0.06

0.07

0.08

0.09Sacramento River

Average reach length, km0 5 10 15 20 25

0.05

0.07

0.09

0.11

0.13

3Average discharge, m /s500 1500 2500 3500

Major out-of-bank flow

Wid

th r

RM

SE

0.04

0.06

0.08

0.10

3Average discharge, m /s100 200 300 400 500

Slo

pe R

MS

E, cm

/km

1.0

1.5

2.0

2.5

3.0

3.5Narrower river, more impact on shorter reaches (HC)

0.5

1.5

0.7

0.9

1.1

1.3

2

3

4

5Po River

Errors per overpass Errors per reach lengthStatement of the problem

10 km Gaussian averaging window:High resolution

returns are noisy

Reach averaging Weights

10 km averaging window

Dam

Water surface

Retrieved points

Common smoothing technique

1

How much averaging should we apply?

How do we avoid averaging different things?

Questions:

-How much is needed?

-At which point lowering resolution no longer pays off?

-How do we detect hydraulic features?

Pros: Noise reductionCons: Loss of spatial information

Improper averaging could lead to biases

e-mail:frasson.1@osu.edu

Impacts of river segmentation strategies on reach-averaged product uncertainties for the upcoming Surface Water and Ocean Topography (SWOT) mission

Renato P.M. Frasson, Rui Wei, Toby Minear, Stephen Tuozzolo, Alessio Domeneghetti, Michael Durand H21F-1485

Surface Water and Ocean Topography mission

Objective: detect and measure water surface elevation

Expected Launch: 2021Repeat period: 20.86 daysLatitudes from 78ºS to 78ºNResolution: 10 m to 60 m x 6 mHeight accuracy: 10 cm*Slope accuracy: 1.7 cm/km*

2*when averaged over >1km120 km

20 km

6m

60 to 10 m

50 km

Gro

und

track

Swath characteristics

Methodology3

Initial Centerline

Centerline

Water Depth [m]16.75

0Terrain DEM [m]

60

-1.96

# RiverObs Node

RiverObs Refined

Pixel CloudClassification

Land near Water

Water near Land

Interior Water

0 1000Meters

500

#

#

#

#

#

#

#

#

#

#

#

#

#

#

#

##

##

#

###

#

#

#

#

#

#

#

#

Simulated pixel cloud River nodes

Vectorization of the pixel cloud

Simulation length

Sacramento River- 152 km- 747 nodes

Po River

- 133 km total 113 km without gaps- 349 nodes

Reach definition strategies

Hydraulic controls

Estimate concavity of the water surface

Find discontinuities in the water surface

Break up reaches at inflection points

Merge short reachesBreak up reaches at discontinuities and

swath edges

Sinuosity

Calculate sinuosity

Add reach boundaries at the local maxima/minima of Merge short reaches

Break up reaches at discontinuities and

swath edges dSdx

Flow distance, km

Wate

r ele

vatio

n, m

5

10

15

20

25

30

35

40Sacramento

0 20 40 60 80 100 120 140 160

Positive concavity

Negative concavity

Discontinuity

Swath edge

Flow distance, km

Po River

-2

0

2

6

10

12

14

16

0 20 40 60 80 100 120 140

8

4

8

Swath edge

Swath edge

Reaches with similar HC and sinuosity (falling within 2 km) boundaries

HC and Sinuosity boundaries within 1km

Equally spaced nodes (~200m) containing:

- Node ID- Water surface elevation- elevation uncertainty- River width- width uncertainty- Connectivity information- Spatial information- Data quality flags

Sacramento River

Hydraulic Controls18 Reaches, average length 8.3 km

Sinuosity17 Reaches, average length 8.9 km

Po River:

Hydraulic Controls12 reaches, average length: 9.8 km

Sinuosity14 reaches, average length: 8.4 km

Sacramento

1

2

3

4

5

0 20 40 60 80 100 120 140 160

Sinuosity

Reach boundaries

Swath edge 2

3

3

2

1

Flow distance, km

Po River

0 20 40 60 80 100 120 140

Sin

uosi

ty

Flow distance, km