evaluating the surface water ocean topography … · evaluating the surface water ocean topography...

19
Evaluating the Surface Water Ocean Topography Mission Hydrologic Observations Kostas Andreadis 1,2 , Michael Durand 2 , Sylvain Biancamaria 3 , Elizabeth Clark 1 , Ernesto Rodriguez 4 , Delwyn Moller 5 , Doug Alsdorf 2 , Dennis Lettenmaier 1 , and Nelly Mognard 3,6 1 University of Washington, 2 Ohio State University 3 Laboratoire d’Etudes en G´ eophysique et Oc´ eanographie Spatiales 4 NASA Jet Propulsion Laboratory, 5 Remote Sensing Solutions 6 Centre National d’Etude Spatiales IEEE IGARSS, Cape Town - 15 July 2009 Andreadis et al. - IGARSS 2009 SWOT Hydrology

Upload: dothuan

Post on 04-Aug-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

Evaluating the Surface Water Ocean TopographyMission Hydrologic Observations

Kostas Andreadis1,2, Michael Durand2, Sylvain Biancamaria3,Elizabeth Clark1, Ernesto Rodriguez4, Delwyn Moller5, Doug

Alsdorf2, Dennis Lettenmaier1, and Nelly Mognard3,6

1University of Washington, 2Ohio State University3Laboratoire d’Etudes en Geophysique et Oceanographie Spatiales

4NASA Jet Propulsion Laboratory, 5Remote Sensing Solutions6Centre National d’Etude Spatiales

IEEE IGARSS, Cape Town - 15 July 2009

Andreadis et al. - IGARSS 2009 SWOT Hydrology

The SWOT satellite

◮ Surface Water OceanTopography (SWOT)satellite mission

◮ Ka-band SARinterferometric system

◮ Two swaths, 10 and 60 kmon each side of the nadirtrack

◮ WSOA and SRTM heritage

◮ Produces heights andco-registered all weatherimagery

Andreadis et al. - IGARSS 2009 SWOT Hydrology

What type of measurements?

◮ SWOT will measure◮ Water surface elevation (h)◮ Temporal and spatial variability in heights (dh/dt and dh/dx)◮ Water inundation (water/no water)◮ Lake and reservoir storage (changes)

◮ Examples of similar measurements from other sensors

◮ SWOT will improve their heritage

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Amazon repeat-pass interferometric SAR

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Ohio River widths (LandSat-derived)

Courtesy: Jon Partsch

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Virtual mission motivation

◮ What will SWOT “see”?◮ Height and width accuracy of rivers, lakes, wetlands and

reservoirs◮ How many times will rivers be observed per orbit cycle?◮ How far upstream the river network will SWOT observe?

◮ How will SWOT estimate discharge?◮ What are the expected errors and what is the contribution of

different sources?◮ “Direct” retrieval and Data assimilation◮ Estimating other parameters (e.g. bathymetry, roughness)

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Ability of SWOT to observe storage change

Height and areal errorsvary with lake area

Results from PeaceAthabasca lakes

XX

XXXy

��

��

��)

��

��

��

���1

PP

PPPq

Courtesy: Hyongki Lee

Andreadis et al. - IGARSS 2009 SWOT Hydrology

A river characteristics global database

◮ In-situmeasurements ofriver characteristicsare sparse andinadequate

◮ Need to create arealistic global rivernetwork to evaluateSWOT observations

◮ Power lawregressions to relatedrainage area todischarge, width,depth

200˚ 240˚ 280˚ 320˚ 0˚ 40˚ 80˚ 120˚ 160˚

−60˚

−40˚

−20˚

20˚

40˚

60˚

80˚

0 100 200 300 400 500>

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Global map of number of observations per orbit cycle

200˚ 240˚ 280˚ 320˚ 0˚ 40˚ 80˚ 120˚ 160˚

−60˚

−40˚

−20˚

20˚

40˚

60˚

80˚

0 2 4 6 8 10>

Andreadis et al. - IGARSS 2009 SWOT Hydrology

What rivers will SWOT see?

◮ Rivers with withs > 50 m will beobserved

◮ We can evaluate the characteristicsof rivers observed

220˚ 230˚ 240˚ 250˚ 260˚ 270˚ 280˚ 290˚ 300˚

30˚

40˚

50˚

60˚

200 400 600 800 1000

Width (m)

280˚ 290˚ 300˚ 310˚ 320˚

−50˚

−40˚

−30˚

−20˚

−10˚

10˚

1000 2000 3000 4000 5000

Width (m)

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Estimating river discharge

◮ “Direct” retrieval

Q =1

nwS

12 z

53

◮ Fast method based on Manning’sequation

◮ Similar to SRTM-based dischargeestimation over Amazon

◮ Data assimilation

◮ Coupling of hydrodynamics and hydrologic models

◮ Channel and floodplain discharge: states

◮ Water surface elevations: observations

◮ Computationally more expensive

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Discharge “direct” retrieval

◮ Synthetic test over tributaries of the Ohio River basin

◮ Depth error propagates into discharge estimates

◮ However, discharge variations are accurate despite those errors

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Assessing errors in discharge estimation from SWOT

Used in-situ dischargemeasurements to examineerrors due to (i) temporalsampling and (ii) heighterrors

σ2m

Q2=

σ2t

Q2+

σ2s

Q2

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Assimilation schematic

◮ Synthetic experiment where true WSE and discharge aresimulated

◮ Expected height errors added to “true” WSE to generateSWOT observations

◮ Open-loop and Filter simulations use corrupted model inputs

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Data assimilation - Water depth

Water depth (in meters) mapsfor different simulations on 13March 1995 (06:00)

TRUTH

TRUTH - OPEN LOOP TRUTH - FILTER

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Data assimilation - Channel discharge

◮ Channel discharge profiles for the different simulations duringtwo update times

◮ Calculated using uncertain width, roughness and depth

13 March 1995 - 06:00 22 April 1995 - 06:00

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Estimating other river characteristics

◮ Exploit SWOT information content for indirect estimation ofriver characteristics

Estimation of channel roughness coefficient(Ohio River study domain)

Estimation of river bathymetry (AmazonRiver study domain)

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Bathymetric slope estimation

◮ SWOT can measure inundated area and total storage onfloodplains

◮ Knowing these through time, allows selection of correctchannel bathymetric slope

Andreadis et al. - IGARSS 2009 SWOT Hydrology

Ongoing work

◮ Expand data assimilation studies on different river basins (e.g.Mississippi, Ob)

◮ SWOT instrument simulation over entire continent

◮ Methods for retrieval of bathymetry (e.g. slope/width todepth stochastic models)

◮ Use of detailed in-situ measurements to evaluate the dischargeestimation error budget

◮ Understand the role of long-wavelength errors (e.g. slope,geoid) affect Amazon hydrologic characterization

◮ Quantifying the impact of errors in the context of dataassimilation (precipitation, lateral inflows)

◮ Optimization of hydrodynamics model for faster simulations

◮ Product delivery considerations (computational scalability etc)

Andreadis et al. - IGARSS 2009 SWOT Hydrology