Katrien van eerdenbrugh b reach a methodology for the assessment of data consistency

BReach: a methodology for the assessment of data consistency (in rating curve data)

Katrien Van Eerdenbrugh Ghent University Laboratory of Hydrology and Water Management

Data consistency in rating curve data

Objective method?

BReach (Bidirectional Reach)

1. Model selection 2. Sampling of the parameter space 3. Assessment of a quality measure

4. Assignment of tolerance degrees 5. Assessment of bidirectional reach 6. Identification of consistent data periods

BReach

Validation

𝑄 ≅ 𝑐(ℎ − ℎ0)𝑛

1. Model selection (rating curve)

steady state conditions uniform flow constant roughness simplified cross section

If

2. Sampling of the parameter space

𝑸 ≅ 𝒄(𝒉 − 𝒉𝟎)𝒏

Pappenberger et al., 2006

3. Assessment of a quality measure

3. Assessment of a quality measure

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

Pappenberger et al., 2006 Sorted h-Q data points

Para

met

er s

ets

Para

met

er s

ets

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

4. Assignment of tolerance degrees

Sorted h-Q data points

Para

met

er s

ets

10% data points

allowed with

quality=0

• Model structural uncertainty • Data outliers

5. Assessment of bidirectional reach

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

Left reach

98

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

Left reach

98

-

40 41 42 43 … 97 98 99 100 101

1 0 0 0 0 … 0 0.67 0.67 1 0

2 0 0 0 1 … 0 0 0 0 0

3 0 0 0.5 1 … 1 1 1 1 1

… … … … … … … … … … …

1200000 0 0 0 1 … 0 0 0 0 0.17

1200001 1 1 0.83 1 … 1 1 1 1 1

1200002 0 0.83 1 1 … 1 1 1 1 1

Left reach

98

-

42

…

-

27

30

Sorted h-Q data points

Para

met

er s

ets

10% data points

allowed with

quality=0

• Model structural uncertainty • Data outliers

10% data points

allowed with

quality=0

4. Assignment of tolerance degrees

6. Identification of consistent data periods

-1 m

-1 m +2 m

Validation: synthetic data

• Simulation with hydrodynamic model: 2 geometries • Selection of (random) transition date • Selection of Q/h results before and after transition • Add noise (observational uncertainty)

Capability to detect transition point?

Validation: synthetic data

Validation: observational uncertainty

Pappenberger et al., 2006

Validation: observ. Uncertainty (4x σ)

Validation: model deficiency

BReach: conclusions

Robust methodology • Validated • Little dependency of subjective choices • Flexibility

Possible applications: • Temporal variability • Dependency of a variable • NO assessment of parameter values

(near) future

BReach(t) – BReach(h): • Variety of Q stations (Flanders, UK, Sweden)

Other models: • Hydrological • Hydraulic

Questions

Are there any questions/remarks?

Locations for BReach analysis in Belgium?