Floods are 2D

1
Floods are 2D

• Doug Alsdorf
• Ohio State University School of Earth Sciences,
  Byrd Polar Research Center
• Paul Bates
• University of Bristol School of Geographical
  Sciences

Funding from CNES, JPL, NASAs Terrestrial
Hydrology and Physical Oceanography Programs,
and the Ohio State Universitys Climate, Water,
Carbon Program
2
Amazon
Siberia
New Orleans
Ohio
Ohio R.
3
Outline

• Measurements of natural floods
• Hydrodynamic modeling of floods
• A couple of recommendations

4
Water Balance on a Floodplain
P
E
h at t3
h at t2
h at t1
Terre-Firme
Floodplain
River Channel

Q

h
q
q Qin Qout ?S

L

x

t
Continuity Equation q lateral inflow (P-E), Q
floodplain discharge h water height, L
reach length t time, x space
5
Interferometric SIR-C Measurements of Water Level
Changes
Interferogram
Amplitudes
Alsdorf et al., Nature, 2000
6
That was 1D and recessional flow, what about 2D
and inundation?
How does water flow through this
floodplain? Which channels convey the most
water? Where does water reside the longest?
7
Conventional Idea of Floodplain Inundation
8
Measurements of Floodplain Inundation
Localized, complex patterns of dh/dt Sharp dh/dt
aligned with many channels Purus flood wave is
apparent
9
Flow Directions Estimated from Continuity
Localized, complex patterns of dh/dt with sharp
dh/dt aligned with many channels indicates flow
to floodplain arriving via channels and emptying
to one side. Purus flood wave supplies water.
10
Measurements of Floodplain Inundation
Broad, simpler patterns of dh/dt Sharp dh/dt
aligned with fewer channels Amazon flood wave is
apparent
11
Flow Directions Estimated from Continuity
Broad, simpler patterns of dh/dt with sharp dh/dt
aligned along fewer channels implies diffuse flow
across entire floodplain. Amazon flood wave
supplies water.
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27

• Congo
• Repeat-pass interferometric SAR for dh/dt
• The Congo River is the second largest in the
  world, But, there are more operating stream
  gauges in Delaware County, Ohio than in all of
  the Congo Basin.
• Thanks to
• Hahn Chul Jung

28

• Congo Wetlands No clear relationship between
  topography and dh/dt (more data are needed)

dh/dt
SRTM DEM
red topography blue dh/dt
29
Flood dynamics

• Large low amplitude waves
• 1-1000km in length
• lt1 hour to 6 months duration
• Low slope 1-100cm km-1
• Gradually varied flow
• Above bankfull stage waves spread in 2D over
  floodplains
• Complex shallow water inundation dynamics
• Major control on wetland biogeochemistry and
  carbon cycling

30
Hydrodynamic models

• Principal tool for assessing flood risk
• Provide dynamic predictions of water depth and
  velocity
• Horizontal scales of 1-1000 m
• Temporal scales of 1-60 s over events lasting up
  to 1 year
• Domain sizes of 1- 100,000 km2
• Can be 1, 2 or 3 dimensional (but consensus that
  floodplain inundation is at least a 2D process)

31
Model data needs

• Boundary conditions
• Discharge and stage at river gauging stations
• Topography
• Ideally LiDAR (lt10m spatial resolution, lt10cm
  rmse vertical accuracy), but can also use SRTM
  for large rivers
• Calibration/validation data
• Measurements of water height and flood extent
• Used to calibrate model friction parameters

32
Existing measurements of surface water dynamics

• Limited to.
• Point gauging stations
• Very small numbers of consistent inundation
  images
• Satellite altimetery (gauges)

33
Data limitations to modelling

• When we have distributed cal/val data we can do
  great modelling
• But ..
• Existing gauges only test bulk flow routing
• Allows modellers to get away with 1D codes
• We only have data to test 2D model performance at
  10-15 sites globally
• Even here often only 1 flood extent image per
  event
• Doesnt allow us to test 2D model dynamics
• Lack of sufficient cal/val data means that many
  flood models suffer from high uncertainty

34
Carlisle, UK 10m model vs. ground survey
RMSE on water depth 0.32 m
35
Upton on Severn, UK 18m model vs airborne SAR
Model fit 89
36
What data do we want?

• Must have .
• Flood images at O(1-2) more sites than we
  currently have
• Need flood extent at 100m or less
• Water elevations to centimeter level accuracy
• Will make 2D modelling routine globally
• For a smaller number of sites we need multiple
  SWOT images through events
• Test model dynamic predictions
• Will lead to the development of better modelling
  tools
• Would be nice
• Discharge measurements in ungauged rivers
  accurate to 25
• Better global floodplain terrain data
• SRTM is all we have, but this has 5m scale
  vertical noise at 3 arc-second resolution
• Ideally need a global floodplain DEM with
  decimeter scale vertical errors

37
Will SWOT do this?

• Yes
• Dont need to image all floods, just a sufficient
  number
• Dealing with whole river reaches (10-1000km) so
  exact orbit repeat may not be necessary to image
  a single flood multiple times
• Pixel size and water height/slope accuracy within
  specification
• Can recover discharge for ungauged rivers
• Better floodplain terrain data may be a fantastic
  side product of the mission
• But may incur additional processing costs

Credit Karen Wiedman
38
Hydrodynamic modelling key message

• Routine application of hydrodynamic models
  appropriate to simulating floodplain inundation
  is currently prevented by a lack of observed 2D
  flood extent and water height data that can be
  used to calibrate such schemes. SWOT will provide
  these data and allow a step change in our ability
  to model floods.

39
A couple of recommendations
If you need more precise measure-ments of
natural events on Earth's surface, get into
space. Nature Water flow across floodplains is
more complex than implied by 1D point based
measurements. Flow paths and water sources are
not fixed in space and time, rather vary with
flood water elevations. Hydrodynamic models show
promise for improving our understanding of
floodplain hydraulics. We need high-resolution
topography and 2D mappings of h, dh/dt, dh/dx,
and inundated area.
40
Thanks
Wheeling Island
41
Additional Slides
42
Surface Water Interferometric Altimeter Concept

• Ka-band SAR interferometric system with 2 swaths,
  50 km each
• WSOA and SRTM heritage
• Produces heights and co-registered all-weather
  imagery
• 200 MHz bandwidth (0.75 cm range resolution)
• Use near-nadir returns for SAR altimeter/angle of
  arrival mode (e.g. Cryosat SIRAL mode) to fill
  swath
• No data compression onboard data downlinked to
  NOAA Ka-band ground stations

These surface water elevation measurements are
entirely new, especially on a global basis, and
thus represent an incredible step forward in
hydrology.
Courtesy of Ernesto Rodriguez, NASA JPL