University of Liege – Faculty of Applied Sciences ... · Coupling with other systems Atmospheric...

1

Engineers facinggroundwater challenges !

University of Liege – Faculty of Applied Sciences

ArGEnCo – Architecture, Geology, Environment and Constructions

GEO³ - Geotechnologies, Hydrogeology and Geophysical Survey

Prof. Dr. Ir. A. Dassargues

Bucharest, June 15, 2012

A thirsty planet…“Water is a necessary commodity to mankind and

groundwater is the largest available source of fresh water” (Walton, 1963)

New developments induce new problems and …new efforts to solve them

Society is more and more demanding in quantified answers, reliable estimations, and risk assessment, … for decisions(linked to … confidence interval of answers)

Groundwater field measurements

Groundwater modelling

Groundwater modelling: equations ? not a problem !

equations are mostly found and written …

how far are they properly describing the reality ?(Fossoul et al., CMWR, 2010)

conceptual choices are never perfect stress factors less known than often

considered parameters are still badly known scale issue calibration/validation, inverse modelling need for robustness, sensitivity analysis need uncertainty assessment CPU time /parallel processing

Groundwater modelling: conceptualization/parameterization issues

2

Groundwater modelling: solving techniques and computer efficiency Solving techniques have improved and are

improving in efficiency in accuracy for non-linear problems for highly heterogeneous parameters/properties for fractured rocks Coupling with many other processes

Physical, biochemical processes Coupling with other systems

Atmospheric / Climatic / Ocean models Social models/systems Economical models/systems

… Integrating !

Groundwater modelling: a few thinking

Reductionism vs Integrationism (W.Wood, 2012) Reductionism = to understand fundamental biological

and physical processes by eliminating interacting effects careful experimental design and choices of boundary and initial conditions

Integrationism = the emerging challenges of large-scale and coupled problems in soil, water, and energy

requiring extrapolation of well-known, small scale behaviours into larger scale models

numerical models has made possible the shift ! huge step forward to integrate the models of various

water disciplines on a common platform accommodate social and economic models, this aspect is

definitely a work in progress

Groundwater modelling: a few thinking … on the other hand What is the actual role of modelling ?

… to assess likelihood(Doherty, 2011)

A) complex models can become cumbersome they take too long to run poor for parameter estimation and uncertainty analysis but good receptacles for expert knowledge

B) simple models are fast and stable expert knowledge may be vague or nonexistent too far from the reality

optimal compromise between simplicity and complexityis needed

Groundwater modelling: coupling in the conceptual model

Large importance of the adopted conceptual model

Groundwater flow

GeomechanicsSolute transport

Thermic effects / geothermy

Physical and Biochemical reactions

+ others

3

Different conceptual coupling:example for solute transport

Mobile GW

Adsorption/ desorption

Advection, dispersion, diffusion and degradation

Grains /Matrix+ Immobile

water

Mobile GW Advection, dispersion, diffusion and degradation

Degradation

Adsorption/ desorption + immobile water effect

Grains /solid matrix

Adsorption/ desorption

Degradation

Immobile GW

immobile water effect

Degradation

Groundwater modelling: a few examples linked to practical questions

Nitrate contamination … was it only a diffuse contamination ? Not so sure !

(P. Biver, PhD, 1993)


River – GW interactions: arrival of a pollutant in the river

advection, dispersion, diffusion degradationadsorption

RM GWM

(Carabin & Dassargues, CMWR, 2000)


River – GW interactions: arrival of a pollutant in pumping wells

advection, dispersion, diffusion degradationadsorption

(Peeters et al., Env.Geol., 2003)

4

Groundwater modelling: seawater intrusion

2000 m

100 m 90 m

1000 m

Land

Recharge

Ocean

)0*( C

dmqn /105.5 40

n

C

0

n

C

1C

1C

zh

0nq

0nq

0C

(Carabin & Dassargues, WRR, 1999)







5

Groundwater modelling: heat transport… cooling of a future office building

(Fossoul et al., CMWR, 2010)


200 m³/h (nearly the critical value) Maximum global rate for a continuous pumping




Groundwater modelling: 2D vertical reactive transport for a contaminated site

NH4+

(Ammonium)

Ca2+

(Calcium)

Cl-

(Chlorine)

(B. Haerens PhD, 2004)

6

Groundwater modelling: 2D vertical reactive transport for a contaminated site

C6H6O(Phenol)

O2

(Dissolved Oxygen)

NO32-

(Nitrate)

(B. Haerens PhD, 2004)

Integrated modelling at regional scale a few thinking

“Modelling large-scale effects from small-scale influences is a delicate act” (Bear, 2011)

Data management Conceptualization Upscaling parameters Calibration objectives Sensitivity analysis Range of the results

Integrated modelling at regional scale: data management and conceptualization

HydroCube

Model :ConceptualisationEquations

(Ph.Orban, PhD, 2009)

Data management : Database and GIS

(P.Wojda, PhD, 2009)

The HFEMC method is a new and pragmatic approach which allows:

combining, in a single model and in a fully integrated way, different mathematical and numerical approaches

keeping the advantages of the spatial representation using finite element mesh

The HFEMC method was implemented in the SUFT3D (Saturated and Unsaturated Flow and Transport in 3D) finite element code

Integrated modelling at regional scale: Hybrid Finite Element Mixing Cell method

(P.Orban, PhD, 2009)

7

Example of cutting in subdomains: RWM021


(Brouyère et al., MWE, 2009)Wildemeersch et al., JoH, 2010)

SUFT3D: Saturated – Unsaturated Flow and Transport in 3D Control volume finite element method (CVFE) For large scale applications

Flexible discretization / meshing approach Mathematical models of various (increasing) complexities

for flow and transport (Hybrid Mixing Cell Finite Element approach)


(P.Orban, PhD, 2009)

• Limit of the model similar to the limit of the hydrological basin• Boundary conditions• Heterogeneity of the chalk


(Orban et al., JoCH, 2010)



8

Cost-effectiveness analysis Benefit analysis

Integrated modelling at regional scale: with socio-economic values

(Orban et al., JoCH, 2010)(With BRGM, C. Hérivaux) -

2

4

6

8

10

12

14

2000 2010 2020 2030 2040 2050

Dis

cou

nte

d d

amag

e (M

€/ye

ar)

DamageNo programme

Program B (-32%) : +192 M€Program C (-41%) : +207 M€

-

10

20

30

40

50

60

70

80

2000 2010 2020 2030 2040 2050

Nit

rate

co

nce

ntr

ati

on

(m

g/l)

Abstraction points

Reservoirs (galeries)

No programme

-

10

20

30

40

50

60

70

80

2000 2010 2020 2030 2040 2050

Nit

rate

co

nce

ntr

ati

on

(m

g/l)

Abstraction points

Galleries

Programme B

-

10

20

30

40

50

60

70

80

2000 2010 2020 2030 2040 2050

Nit

rate

co

nce

ntr

ati

on

(m

g/l)

Abstraction points

Galleries

Programme C

-

2

4

6

8

10

12

14

2000 2010 2020 2030 2040 2050

Dis

cou

nte

d c

ost

s/ b

ene

fits

(M

€/y

ear)

Benefits

Cost of the programme of measuresProgramme B

192 M€

141 M€

-

2

4

6

8

10

12

14

2000 2010 2020 2030 2040 2050

Dis

cou

nte

d c

ost

s/ b

enef

its

(M€/

year

)

Benefits

Cost of the programme of measures

207 M€

221 M€

Integrated modelling at regional scale: with socio-economic values


(With BRGM, C. Hérivaux)

Guidelines for large-scale groundwater flow model calibrationde with respect to end-users’ expectations

Objective of the study Performance criteria to privilege

General evolution of base flow rates NSEq

Maximum value of base flow rates PEq

General evolution of hydraulic heads RMSh

Hydraulic head variations HHVEh

Hydraulic head maps RMSh

31(S. Wildemeersch, PhD, 2012)

Integrated modelling at regional scale: … impact of climate changes on GW quantity

T(°C)

2010 2040 2070 2100 2010

HydrologicalModel

2085

Climate changescenarios

Stochastic Weather Generator

Groundwater levels

T(°C)

2010 20852010 2040 2070 2100

Classical downscalingmethod

(P. Goderniaux, PhD, 2010)

9


T(°C)

2010 2040 2070 2100 2010

HydrologicalModel

2085



Groundwater levels

T(°C)

2010 20852010 2040 2070 2100




T(°C)

2010 2040 2070 2100 2010

HydrologicalModel

2085



Groundwater levels

T(°C)

2010 20852010 2040 2070 2100




T(°C)

2010 2040 2070 2100 2010

HydrologicalModel

2085



Groundwater levels

T(°C)

2010 20852010 2040 2070 2100



Dynamical downscaling

Regional Climate Models (RCM)

General Circulation Models (GCM)

Statistical downscaling

Climate Scenariosat the scale of the catchment

Hydrological models Finite Element Code HYDROGEOSPHERE

surface – subsurface integrated model



(Therrien & Sudicky, JoCH, 1996)

10

INTERCONNECTED PROCESSES


More realistic representation

Surface and subsurface data for calibration

Better representation of groundwater recharge

Crucial for climate change impact studies(P. Goderniaux, PhD, 2010)

The Geer basin is modelled using HydroGeoSphere


(Goderniaux et al., JoH, 2009)

Water exchangesbetween nodesat each time step

SUBSURFACE DOMAIN

3D variably‐saturated flow

(Richards' equation)

SURFACE DOMAIN

2D surface water flow

(Diffusion‐wave approximation of the Saint Venant equations)

Precipitations Actual ET = f(PET, soil moisture, root & evap. Depth, LAI, canopy)

At each time step and node


(Goderniaux et al., JoH, 2009)

Climate change scenarios are appliedon the Geer basin model

Groundwater levels

RCM : Arpege_h

All 6 RCMs

MEAN groundwater levels

(Goderniaux et al., WRR, 2011)

11

Results are achieved for all observation wellsand all RCMs


Confidence interval of the computed change in groundwater levels (h)

For the Geer basin hydrological model

‐ 95 % linear confidence intervals around predicted values

‐ Calculated over 4 years of HIRHAM_H

‐ Use of UCODE_2005 (USGS)

Predicted groundwater levelsand 95 % confidence intervals

Change in groundwater levelsand 95 % confidence intervals


Confidence interval of the computed surface flow rates

Absolute surface flow ratesand 95 % confidence intervals

Change in surface flow ratesand 95 % confidence intervals

→ Uncertainty more important for winters!


mean value variogram γ(d) = f( 2, λ)

d

γ

Geostatistics and groundwater modellingvariogram, kriging and cond. simulation

(C. Rentier, PhD, 2003)

12

)x(z )x(SNCjz+

x x

x

z (x)

xx1 x2 x3

_)x(SNC

jz )x(SC

jz =

Geostatistics and groundwater modellingconditional simulation


Kriging Error

P1

P2Pr1

Pz2Pz3

Pr3

Pr2

Pz1

Pz5

Pz6Pz4

h

x

x

xx K

ρ

Co-conditional simulation:application for protection zone


t = 24 hours t = 50 days

Determinist 3D model

Determinist 2D model

(0,5) (0,5)

Co-conditional simulation:application for protection zone


… applied here on the geological heterogeneity of Brussels Sands

do not rely on variogram capturing structure by a

‘training image’ borrows the multiple points

patterns from the training image

still relies on the forgotten assumption of ergodicity and stationarity

(Jeff Caers, U. Stanford)

≈ 10 m

≈ 3 m

Multiple-points geostatistics and groundwater modelling

(Huysmans et al., JoH, 2008)

13

Field observations Pictures and geological

sketches Subjective distinction

between clay-rich bottom sets/distinct mud drapes and sandy zones

Determination of lamination angle

≈ 16 m

Multiple-points geostatistics and groundwater modelling

(Huysmans et al., JoH, 2008)

Multiple-point facies realizations and intrafacies permeability simulation

3 example facies realizations out of 150 realizations

3 example hydraulic conductivity realizations out of 150 realizations

(Huysmans & Dassargues, HJ, 2009)

GW vulnerability: where engineers have saved geologists and geographersThe start no clear definition relative concept often causes are mixed with results

(i.e. contaminated zones are considered as vulnerable)

multi-criteria analysis but with which weighting, units, objective function ?

like an Eurovision song contest !!!! - hydraulic conductivity: 10 points- topography: 6 points- depth to water: 8 points…

Overlay and index methods

- estimated protective effect of the overlying layers in each pixel of the map;

- ‘physical attributes’ overlaid in a GIS with weighting coefficients given empirically (or conventionally);

- classification of the ‘vulnerability index’ values;

- nice colored maps

• DRASTIC, DRASTIC modified, SINTACS, EPIK, GOD, German method, PI, …

• Main problems

- mixing semi-quantitative approximation of physical processes with pure empirical (local conventional) agreements;

- not to be validated nor compared

Vulnerability mapping: the empirical start

(R. Gogu, PhD, 2000)

(Gogu & Dassargues, Env. Geol., 2000)

14

(R. Gogu, PhD, 2000)


(Gogu & Dassargues, HJ, 2000) (R. Gogu, PhD, 2000)

M e u s eM e u s eM e u s e

BRUSSELSBRUSSELSBRUSSELS

LIEGELIEGELIEGE

N E T H E R L A N D SN E T H E R L A N D SN E T H E R L A N D S

G E R M A N YG E R M A N YG E R M A N Y

L U X E M B U R GL U X E M B U R GL U X E M B U R G

F R A N C EF R A N C EF R A N C E

B E L G I U MB E L G I U MB E L G I U M

NAMURNAMURNAMUR

N E B L O NN E B L O NN E B L O N0 50 100 km0 50 100 km0 50 100 km

M e u s eM e u s eM e u s e

BRUSSELSBRUSSELSBRUSSELS

LIEGELIEGELIEGE

N E T H E R L A N D SN E T H E R L A N D SN E T H E R L A N D S

G E R M A N YG E R M A N YG E R M A N Y

L U X E M B U R GL U X E M B U R GL U X E M B U R G

F R A N C EF R A N C EF R A N C E

B E L G I U MB E L G I U MB E L G I U M

NAMURNAMURNAMUR

N E B L O NN E B L O NN E B L O N0 50 100 km0 50 100 km0 50 100 km0 50 100 km0 50 100 km0 50 100 km

Swallowhole water catchment area

Perennial streamsTemporary streams

Water supply galleries

Legend

N

§%

0 1000 2000 3000 Meters

N

A - A' Geological cross-section

#

##

#

#

#

#

#

#

#

#

# ##

#

#

#

#

#

#

§

§

§

§

§

§

§

§

%

%

%%

%

%

%

%

I

I

A'

A

250

260

265

200

255

230

220

240

190

275

270

280

180

205

270

230

250

240

260

250

260

26

0

220

250

275

265

240

250

255

255

23

0

240

250

210

265

25

0

200

255

240

250

240

270

F2

F1

P1

P0

P3

P2

P10

P16P17

P18

210

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu

Geological strata Alluvial depositsTertiary formationsNamurian - H1bNamurian - H1aVisean (Vise unit)-V2a, V2b, V2cVisean (Dinant unit)-V1Upper Tournaisian - T2Lower Tournaisian - T1 Strunian - Fa2dUpper Famennian - Fa2a, Fa2b, Fa2cLower Famennian - F1a, F1b, F1c

Piezometric head contour lines for the carbonate aquifer (1998)# Wells and Piezometers

Syncline axeAnticline axe

Figure 4.3 Geology and hydrogeology of the study area

Syncline axeAnticline axeWells and PiezometersPiezometric head contour lines for the carbonate aquifer (1998)

Alluvial depositsTertiary formationsNamurian - H1bNamurian - H1aVisean (Vise unit) - V2a, V2b, V2cVisean (Dinant unit) - V1Upper Tournaisian - T2Lower Touraisian - T1Strunian - Fa2dUpper Famennian - Fa2a, Fa2b, Fa2cLower Famennian - F1a, F1b, F1c

Geological strata:

LEGEND:

Water supply galleriesPerennial streamsTemporary streamsSwallowhole water catchment area

Geology and hydrogeology of the study area




Legend

N

§%

0 1000 2000 3000 Meters

N


#

##

#

#

#

#

#

#

#

#

# ##

#

#

#

#

#

#

§

§

§

§

§

§

§

§

%

%

%%

%

%

%

%

I

I

A'

A

250

260

265

200

255

230

220

240

190

275

270

280

180

205

270

230

250

240

260

250

260

26

0

220

250

275

265

240

250

255

255

23

0

240

250

210

265

25

0

200

255

240

250

240

270

F2

F1

P1

P0

P3

P2

P10

P16P17

P18

210

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu







Geological strata:

LEGEND:





Legend

N

§%

0 1000 2000 3000 Meters

N


#

##

#

#

#

#

#

#

#

#

# ##

#

#

#

#

#

#

§

§

§

§

§

§

§

§

%

%

%%

%

%

%

%

I

I

A'

A

250

260

265

200

255

230

220

240

190

275

270

280

180

205

270

230

250

240

260

250

260

26

0

220

250

275

265

240

250

255

255

23

0

240

250

210

265

25

0

200

255

240

250

240

270

F2

F1

P1

P0

P3

P2

P10

P16P17

P18

210

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu








Legend

N

§%

0 1000 2000 3000 Meters

N


#

##

#

#

#

#

#

#

#

#

# ##

#

#

#

#

#

#

§

§

§

§

§

§

§

§

%

%

%%

%

%

%

%

I

I

A'

A

250

260

265

200

255

230

220

240

190

275

270

280

180

205

270

230

250

240

260

250

260

26

0

220

250

275

265

240

250

255

255

23

0

240

250

210

265

25

0

200

255

240

250

240

270

F2

F1

P1

P0

P3

P2

P10

P16P17

P18

210

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu








Legend

N

§%

0 1000 2000 3000 Meters

N


#

##

#

#

#

#

#

#

#

#

# ##

#

#

#

#

#

#

§

§

§

§

§

§

§

§

%

%

%%

%

%

%

%

I

I

A'

A

250

260

265

200

255

230

220

240

190

275

270

280

180

205

270

230

250

240

260

250

260

26

0

220

250

275

265

240

250

255

255

23

0

240

250

210

265

25

0

200

255

240

250

240

270

F2

F1

P1

P0

P3

P2

P10

P16P17

P18

210

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu







Geological strata:

LEGEND:




Geological strata:

LEGEND:



Geological strata:

LEGEND:



Geology and hydrogeology of the study area

NN SS

NéblonNéblonOcquierOcquierstreamstream

SwallowholeSwallowholeHigh High levellevel

Low levelLow level

AA A’A’

1 km1 km

Strong fractured Strong fractured zonezone

± 35 m± 35 m

SwallowholeSwallowholeSpringSpring

LEGENDLEGEND::

ALLUVIAL DEPOSITS UPPER TOALLUVIAL DEPOSITS UPPER TOURNAISIAN MAJOR FOLD AXISURNAISIAN MAJOR FOLD AXIS

TERTIARY SANDYTERTIARY SANDY--CLAY DEPOSITS LOWER TOURNAISIAN CLAY DEPOSITS LOWER TOURNAISIAN GROUNDWATER LEVELGROUNDWATER LEVEL

NAMURIAN NAMURIAN STRUNIANSTRUNIAN

VISEAN VISEAN UPPER FAMENIANUPPER FAMENIAN

LOWER FAMENIANLOW ER FAMENIAN

NN SS



Low levelLow level

AA A’A’

1 km1 km


± 35 m± 35 m


LEGENDLEGEND::






NN SS



Low levelLow level

AA A’A’

1 km1 km


± 35 m± 35 m


NN SSNN SS



Low levelLow level

AA A’A’AA A’A’

1 km1 km


± 35 m± 35 m


LEGENDLEGEND::






LEGENDLEGEND::






LEGENDLEGEND::







(Gogu et al., Env. Geol., 2003) (R. Gogu, PhD, 2000)

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu

High vulnerability

Moderate vulnerability

Low vulnerability

Very high vulnerability




Degree of vulnerability

Legend

N

1000 0 1000 2000 3000 Meters

Final vulnerability map using EPIK method (new classification - Doerfliger et al. 1999)

High vulnerabilityModerate vulnerabilityLow vulnerability

LEGEND:



Final vulnerability map using EPIK method1000 0 1000 2000 3000 Meters

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu

High vulnerability


Low vulnerability






Legend

N

1000 0 1000 2000 3000 Meters

Final vulnerability map using EPIK method (new classification - Doerfliger et al. 1999)


LEGEND:




LEGEND:




Final vulnerability map using EPIK method1000 0 1000 2000 3000 Meters1000 0 1000 2000 3000 Meters1000 0 1000 2000 3000 Meters

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Very low vulnerability

Low vulnerability



Legend

Figure 4.39 Final vulnerability map using DRASTIC methodand the commonly used final classes of vulnerability

LEGEND:



Very high vulnerabilityHigh vulnerabilityModerate vulnerabilityLow vulnerabilityVery low vulnerability

Final vulnerability map using DRASTIC method - commonly used classes of vulnerability

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability


Low vulnerability



Legend


LEGEND:




#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability


Low vulnerability



Legend


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability


Low vulnerability



Legend


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability


Low vulnerability



Legend


LEGEND:




LEGEND:






Final vulnerability map using DRASTIC method - commonly used classes of vulnerability

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Low vulnerability



Legend

Figure 4.40 Final vulnerability map using DRASTIC method and the classes of vulnerability defined by Kumar and Engel (1997)

Very high vulnerabilityHigh vulnerabilityModerate vulnerabilityLow vulnerability

LEGEND:



Final vulnerability map using DRASTIC method - (classes of vulnerability defined by Kumar and Engel 1997)

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Low vulnerability



Legend



LEGEND:



#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Low vulnerability



Legend


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Low vulnerability



Legend


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu




N

1000 0 1000 2000 3000 Meters


High vulnerability

Low vulnerability



Legend



LEGEND:





LEGEND:




Final vulnerability map using DRASTIC method - (classes of vulnerability defined by Kumar and Engel 1997)

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability


Low vulnerability



Legend

N

1000 0 1000 2000 3000 Meters

Figure 4.43 Final vulnerability map using GOD method


LEGEND:



Final vulnerability map using GOD method

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability


Low vulnerability



Legend

N

1000 0 1000 2000 3000 Meters



LEGEND:



#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability


Low vulnerability



Legend

N

1000 0 1000 2000 3000 Meters


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability


Low vulnerability



Legend

N

1000 0 1000 2000 3000 Meters


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability


Low vulnerability



Legend

N

1000 0 1000 2000 3000 Meters



LEGEND:




LEGEND:




Final vulnerability map using GOD method

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu

Extreme vulnerability


High vulnerability


Low vulnerability



Legend




N

1000 0 1000 2000 3000 Meters

Figure 4.42 Final vulnerability map using ISIS method

LEGEND:


Extreme vulnerabilityVery high vulnerabilityHigh vulnerability

Moderate vulnerabilityLow vulnerabilityVery low vulnerability


Final vulnerability map using ISIS method

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu



High vulnerability


Low vulnerability



Legend




N

1000 0 1000 2000 3000 Meters


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu



High vulnerability


Low vulnerability



Legend




N

1000 0 1000 2000 3000 Meters


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu



High vulnerability


Low vulnerability



Legend




N

1000 0 1000 2000 3000 Meters


LEGEND:





LEGEND:








Final vulnerability map using ISIS method

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability



Low vulnerability


Legend

N

1000 0 1000 2000 3000 MetersFigure 4.41 Final vulnerability map using German method


LEGEND:



Final vulnerability map using The German method

#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability



Low vulnerability


Legend

N



LEGEND:



#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability



Low vulnerability


Legend

N


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability



Low vulnerability


Legend

N


#

# ##

#

#

#

#

#

#

Ama

Oneu

Bende

Vervoz

Borlon

Bonsin

Ouffet

Ocquier

Chardeneux

Bois-et-Borsu





High vulnerability



Low vulnerability


Legend

N



LEGEND:




LEGEND:




Final vulnerability map using The German methodIt appears not serious that ‘scientists’ would provide so different results … with the same data on a single case-study !!!


Risk

+ probability of hazard + input function + mass of contaminant

Specific vulnerability

+ (bio)chemical behavior of the contaminant in the medium

Intrinsic vulnerability: inherent geologicaland hydrogeological characteristics that control the impulse response of the systemto a Dirac-type input of conservative contaminant

… for groundwater ?

“a hierarchical process starting with intrinsic vulnerability, then progressing to specific vulnerability, and finally to risk assessment when combining with hazard (i.e. potential pollution at the surface)”

(Brouyère et al., 2001)

GW vulnerability: more accurate definitions

15

Vulnerability mapping: more accurate definitions

adapted from Goldscheider, 2002

« Origin-Pathway-Target Model »origin

R E S O U R C E

S O U R C E

1st pathway:unsaturated zone

1st target: groundwater surface

2nd target:spring/well2nd pathway: aquifer

(Brouyère et al., 2001)

… the potential risk of pollution for a considered target :

1) if a pollution is likely to occur somewhere in the catchment, how long does it take to reach the target ?

2) which (relative) maximum concentration is to be reached ?

3) for how long could the target be polluted (above a given threshold) ?

Still needed: … a conventional agreement on the relative importance (weighting) of each of these criteria !!!

Vulnerability mapping: more “physically based”

z

cv

z

cD

zt

ce

m

it

0max* CCC ii

id

Vulnerability mapping: physically based but also feasible

(Popescu et al., ModelCARE, 2008)

In a Pressure-State-Impact causal chain,

… a generalized concept of groundwater vulnerability should reflect the easiness with which the groundwater system (the ‘state’) transmits pressures into impacts.

evaluating how a change in a given upstream factor (e.g. changes in groundwater recharge, surface contamination, etc.)has knock-on effects on downstream factors (e.g. base-flow to rivers, groundwater quality in a well, etc.)

= sensitivity of the ith downstream factor to a change in the jth upstream factor

dQ

dh

ctorUpstreamFad

FactorDownstreamdS

j

iij

)(

)(

Vulnerability mapping: generalized engineering concept

(Dassargues et al., IAHS n°327, 2009)

16

Vulnerability of Groundwaterchanges due to engineering involvement

Scale issues

(Lemieux et al., submitted)

solved by a differential approach (computed with a perturbation method or by using the derivative as a primary variable)

solved by a variational approach (adjoint operator method)

Vulnerability of Groundwaterchanges due to engineering involvement

Scale issues

(Lemieux et al., submitted)

Large progress …

groundwater models grew in complexity to tackle problems previously beyond our reach(Bredehoeft, 2010)

together with improvement of groundwater user interface (GUI) to input data and extract meaningful results

manual calibration no longer practical so that linear and nonlinear optimizing techniques

… but the model is not an end in itself, rather a powerful tool that organizes thinking and engineering judgment

… but always many challenges in perspective

Scale issues Multiple porous and permeable media in fractured/karstic

media Numerical progress FDifferences, FElements, FVolumes,

Meshless methods, … Efficiency for CPU time and active memory allocation and

management Real/actual confidence intervals linked to hierarchical

calibration objectives and conceptual errors Coupling with other processes linked to integrated approaches Automatic and better localized measurements Regional-Scale Monitoring Network for Risk-Informed

Groundwater Management

STILL A LOT OF APPLIED RESEARCH AND DEVELOPPING CHALLENGES FOR ENGINEERS

17

Thank you to UTCBand particularly to Prof Radu Drobot

and to all former students of the Tempus programs from 1992 to 1999 and to colleagues of the NATO SQUASH project from 2000-2004 ! Thank you !

www.argenco.ulg.ac.be/geo3

Acknowledgments to: P. Biver, R. Gogu, S. Brouyère, C. Rentier, V.T. Tam, B. Haerens, M. Huysmans, J. Batlle Aguilar, IC Popescu, D. Rocha, Ph. Orban, P. Wojda, R.

Rojas, L. Peeters, N. Gardin, P. Goderniaux, S. Wildemeersch, F. Fossoul, J. Gesels, P. Jamin, J. Couturier,

JM Lemieux, R. Therrien, and many others

University of Liege – Faculty of Applied Sciences ... · Coupling with other systems Atmospheric...

Documents

Transcript of University of Liege – Faculty of Applied Sciences ... · Coupling with other systems Atmospheric...