September 16, 2008

R. Edward BeighleyCivil, Construction and Environmental EngineeringSan Diego State University

SWOT Hydrology WorkshopSWOT Hydrology Workshop

The Ohio State UniversityThe Ohio State University

The need for water surface slopes and channel The need for water surface slopes and channel width-depth relationships; an example from the width-depth relationships; an example from the

Amazon hydrologic-hydraulic modelingAmazon hydrologic-hydraulic modeling

• Overview of Hydrologic-Hydraulic Model

• Model parameterization in Amazon Basin– Channel hydraulic geometry and slope

• Model results– Hydrographs– Total water storage changes

• Links to SWOT mission– Channel/floodplain geometry and slope

SWOT Hydrology WorkshopSWOT Hydrology WorkshopThe Ohio State UniversityThe Ohio State University

Hydrologic-Hydraulic ModelHydrologic-Hydraulic Model

Large scale Hydrologic-Hydraulic model based on physics

• 1-D vertical water/energy balance

• 1-D lateral surface and subsurface kinematic wave routing

• 1-D Muskingum-Cunge channel and floodplain routing with lateral split of flow to/from channel and floodplain

• Irregular computational grid based on Pfafstetter units

• Model tracks soil water, surface runoff, subsurface runoff, channel and floodplain storage

Beighley, R.E., K. Eggert, T. Dunne, Y. He, and V. Gummadi, (in review). Simulating hydrologic and hydraulic processes throughout the Amazon River Basin, Submitted to Hydrological Processes.

Hydrologic-Hydraulic ModelHydrologic-Hydraulic Model

1-D vertical water/energy balance model

Model tracks soil storage in root zone and Canopy storage (minimal);

passes excess surface (qs) and root zone (D) waters to routing model

Sc Su P Es Ec ET Qs D

Subsurface store/flux w/ routing model

Hydrologic-Hydraulic ModelHydrologic-Hydraulic Model

• 1-D lateral surface and subsurface kinematic wave routing – Surface runoff– Subsurface runoff– Re-surfacing of subsurface runoff

• Subsurface separated into two zones (upper and lower) to capture depth profile of horizontal conductivity

Hydrologic-Hydraulic ModelHydrologic-Hydraulic Model

• 1-D Muskingum-Cunge channel and floodplain routing– with exchange between channel and floodplain

• Each model unit approximated as "Open Book"– Contains 2 planes, 1 channel, 1 floodplain– Channel and floodplain receive lateral inflow flow from planes 1,2

plus any upstream inflows

Hydrologic-Hydraulic ModelHydrologic-Hydraulic Model

• Computational grid based on Pfafstetter basins; self-replicating system of 9 units

• Units can be small for high drainage density or large to min computing resources

• Amazon Basin-Level 4 (5,189 units) at hourly time-step for period 2001 into 2008

(Max No. of Units = 9Level)

Hydrologic Modeling Performed at Hydrologic Modeling Performed at Plane Scale - Split PFAF unitsPlane Scale - Split PFAF units

Model parameters spatially average on split PFAF units

Routing models approximate split PFAF units as rectangles

5,189 channels: Channel length = 26 km (0.1 to 890 km)Channel width = 83 m (6 to 3,630 m)Floodplain width = 25 m (0.1 m to 55 km)

10,378 planes: Plane area = 110 km2 (<1 to 28,000 km2)

Model requires a representative channel Model requires a representative channel cross-section for each model unitcross-section for each model unit

(e.g., 5,189 needed for Amazon-L4)(e.g., 5,189 needed for Amazon-L4)

Channel Width (1 km)

Floodplain Width (>30 km)

Need method for determining channel and floodplain widths and depths throughout basin

Can obtain widths from remote sensing, BUT what do channels look like below surface: triangular, rectangular or some other shape?

What is bankfull depth & width (i.e., separate channel from floodplain)?

Used 82 Streamflow Stations in Amazon Basin1,000 to 4.7M km2

1

0

0

0

1

1

0.8

0.2

0.4

0.8 0.2 10

100

1,000

10,000

100 1,000 10,000 100,000 1,000,000 10,000,000Drainage Area (km2)

Ban

kfu

ll C

han

nel

Wid

th (

m)

Hydraulic GeometryHydraulic Geometry(from in-situ and remotely sensed data)(from in-situ and remotely sensed data)

• Based on Leopold and Maddock (1953) Hydraulic Geometry relations w = aQb d = cQf v = kQm

• Determined bankfull discharge and hydraulic geometry: (a,c,k) and (b,f,m)

• Developed relations for channel/floodplain characteristics & drainage area

QQbb = 0.096A = 0.096A0.950.95

WWbb = 2.36A = 2.36A0.470.47

DDbb = 0.25A = 0.25A0.340.34

WWff = 0.017A = 0.017A0.960.960.5

Q = D(1/f)

Q = D2

Floodplain width vs. upstream drainage areaFloodplain width vs. upstream drainage area(does not capture local features)(does not capture local features)

Current approach, does not fully capture local variations NOT governed by drainage area alone

Reduced floodplain widths impact model results by increasing flow depths and velocity to handle discharge

Hess et al., 2003Our simulated channel and floodplains

4

8

12

16

0 1 2 3 4 5

0

10

20

30

0 1 2 3 4 5

20

40

60

80

0 1 2 3 4 5

0

20

40

60

80

0 1 2 3 4 5

ModelGauge

0

10

20

30

40

50

0 1 2 3 4 5

Model Results - based on 34 gauges 2001-2005Model Results - based on 34 gauges 2001-2005– TRMM 3B42, roughness: channel = 0.04; floodplain =0.07– Mean annual mass error = 6.2%; peak error = 5.7%

Qx10 m3/s 5K km2

5

10

15

20

25

30

0 1 2 3 4 5

Qx100 m3/s 41K km2

Qx1,000 m3/s 460K km2

Qx1,000 m3/s 1.1M km2 Qx10,000 m3/s 2.1M km2 Qx10,000 m3/s 4.7M km2

22

23

24

25

1585 1595 1605 1615

4

5

6

7

1740 1750 1760 177050

55

60

65

70

750 760 770 780

8

12

16

375 385 395 405

0

1

2

3

4

1125 1135 1145 1155

Model-Daily

Gauge-Daily

Model-Monthly

Gauge-Monthly

15

20

25

30

35

1535 1545 1555 1565

30 day windows of discharge30 day windows of discharge• 1,000 to 100,000 km2 (discharges variability - 1 to 5 days)• >100,000 km2 (discharge variability - 5 to 10 days)• 0.001 to 0.0001 m3/s/km2/day

Qx10 m3/s 5K km2

Qx100 m3/s 41K km2

Qx1,000 m3/s 460K km2

Qx1,000 m3/s 1.1M km2

Qx10,000 m3/s 2.1M km2 Qx10,000 m3/s 4.7M km2

Monthly storage changes based on mean storage from Apr 2002 to Dec 2003

Basin-wide monthly S ranges from +/- 5 cm

Storage changes account for- root zone moisture- sub-surface/surface runoff- channels and floodplains

Results show importance of both landscape and channel/floodplain stores

Monthly Monthly S (cm)S (cm)Jan-Dec 2005Jan-Dec 2005

Water Storage Components• Rooting zone soils - 20%

• Subsurface routing - 40%

• Channel/floodplains - 40%

Monthly S +/- 5 cm

Yr-Yr monthly variability 2.5cm(due to timing of seasonal precip)

Distribution of ChangesDistribution of Changesin Total Water Storagein Total Water Storage

-10

-5

0

5

10

J F M A M J J A S O N D

Cha

nge

in M

ont

hly

Wat

er S

tora

ge

(cm

)RZ

SS

CH/FP

Total

Repeat measurements from SWOT will provide hydraulic geometry!– Will provide critical data at ungauged locations (or improve in-situ data)

Need SWOT slopes: slopes from SRTM must be adjusted based on assumed range in bankfull channel velocity and in-situ data– Slopes adjusted based on

Manning's velocity (0.3-2.3 m/s)

– Median reach length = 25 km0.1 to 900 km

– Median Slopes

SRTM - 113 cm/km (1-9,000)

Adjusted - 26 cm/km (1-750)

Current Challenge/LimitationSWOT to the Rescue?

0.1

1

10

0.1 1 10 100

Velocity from SRTM-Slope (m/s)

Mod

el V

eloc

ity (

m/s

)