Hydrology Virtual Mission for WATERHM

1
Hydrology Virtual Mission for WATERHM

• Doug Alsdorf, Ohio State U.
• Mike Durand, Jim Hamski, Brian Kiel, Gina
  LeFavour, Jon Partsch
• Dennis Lettenmaier, U. Washington
• Kostas Andreadis, Liz Clark
• Delwyn Moller, JPL

2
What is the VM?

• Goal to define the spatial and temporal
  trade-offs when measuring storage changes and
  discharge.
• Goal accomplished by answering 3 key questions
• What are the spatial and temporal samplings of ?S
  and Q required to accurately constrain weather
  and climate models?
• Storage change plays a hydrologic role in many
  basins, but is knowing ?S sufficient, or is Q
  also required to accurately constrain the water
  balance?
• Can reach-to-reach discharge variations be
  accurately measured from space?

3
How will the VM work?

• We will conduct the following tasks to address
  the 3 questions
• Various orbital tracks will be overlaid on a
  global VIC model-based mapping of ?S and Q to
  determine the percentages of ?S and Q that can
  potentially be sampled. Rather than only knowing
  the percentages of water bodies missed, this will
  demonstrate the percentages of ?S and Q measured
  for each basins water balance as dictated by
  various orbits and sampling technologies.
• VM-I creates model h surfaces with errors
  expected from a spaceborne technology. These
  simulated h values will be used to construct
  storage changes in three selected basins, and
  resulting ?S values will be compared to VIC model
  supplied Q values (and Q values from Task 3).
• Our initial SRTM-based Mannings-method of
  estimating Amazon discharge will be expanded to
  other rivers and channel cross-sectional
  geometries will be investigated for constraining
  Q. An alternative method of estimating discharge
  will be constructed in a data assimilation of
  instrument simulated h surfaces (i.e., from Task
  2 above). Spatial and temporal sampling
  resolutions and errors will be investigated in
  both Q methods.
• The instrument simulator of VM-I will be enhanced
  (and applied in the Tasks 2 and 3) with layover
  identification, assessment of height accuracies
  as bounded by the use of the SRTM DEM for
  instrument calibration, and quantification of h
  averaging schemes.

4
The VM Concept

• The VM uses both model and real data, and uses
  complex and simple techniques.

Simple
Moderate

Q
z ?(h-bathymetry)
Complex
g(S0-Sf)
S0 bathymetric slope Sf friction or energy
slope, i.e., dh/dx
Synthetic World
Real World
VIC
LISFLOOD
Instrument Simulator
SRTM
Altimetry
InSAR
Imagery

• Use data assimilation to introduce errors (P,
  Z) and varying sampling resolutions to determine
  their impacts on estimating Q and measuring ?S.
• Use Mannings n approach to estimate Q and
  measure ?S.

• Use DA to estimate Q
• Mannings equation, Continuity to estimate Q

These are designed to assess Q from existing
satellite measurements.
Key all parameters can be measured from space,
but not depth (n is empirical).
5
We can measure width

• Uses classification based on 30m Landsat data

69 meters
Pavelsky Smith, algorithm in review
6
What does SRTM tell us?

• SRTM h accuracy is too coarse for reaches lt 100s
  kms, but Q is not too bad!

Q m3/s Observed SRTM Error Tupe 63100 62900 -0.3
Itapeua 74200 79800 7.6 Manacapuru 90500 84900 -6
.2
7
What does repeat-pass InSAR tell us?

• Floodplains have complex patterns of water flow,
  use continuity to get flux

8
First Results from Data Assimilation

• See next presentation on Ohio River data
  assimilation
• Preliminary feasibility test shows successful
  estimation of discharge by assimilating satellite
  water surface elevations
• Nominal 8 day overpass frequency gives best
  results effect of updating largely lost by 16
  days
• Results are exploratory and cannot be assumed to
  be general -- additional experiments with more
  realistic hydrodynamic model errors (Mannings
  coefficient, channel width etc), hydrologic model
  errors, and more topographically complex basins
  (e.g. Amazon River) are needed.
• Assumption that truth and filter models (both
  hydrologic and hydrodynamic) are identical needs
  to be investigated