University of Miskolc

Faculty of Earth Sciences and

Engineering
Institute of Environmental
Management

HYDROGEOLOGI
C MODELLING
PROCESSES
Geohydrology. Water Resources
Management

Yetzabbel Flores Carpio

 CONTENT
• General Groundwater Modelling Process Workflow
• Conceptual model
• Geometry
• Boundary conditions
• Sinks and sources
• Numerical simulation
• Code selection
• MODFLOW –USG
• Package
• Processes

• Calibration and Sensitivity analysis

• Prediction and uncertainty analysis

2 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES

 GENERAL GW-MODELLIG PROCESS WORKFLOW

Steady state
Darcy’s Law (Momentum balance)+
Equation of continuity (Mass balance)
Real Transient state
porous Darcy’s Law (Momentum balance)+
media Field data Equation of diffusion
measurements

Hydrogeologic
REAL
WOR
CONCEPTUAL QUANTITATIVE model
MODEL HYDROGEOLOGY
D RESULTS
Equivalen
t Calibration,
porous validation and Flow net
Hydrogeological
media framework
sensitivity Water budget
Border conditions
analyses Contamination
Sink and sources
plume
Travel time
Water bodies
iteraction
3 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 CONCEPTUAL MODEL
a. Geometry
Geological
• Geological framework data
collection
• Upscaling
• Heterogeneities and anisotropy
• Structural, sedimentologic and stratigraphic
interpretation Stratigraphic
HYDROESTRATIGRAP and
HIC UNIT Sedimentolog
• Hydrogeological parameterization DEFINITION y
Interpretation
• Hydraulic conductivity
• Horizontal
• Vertical
• Anisotropy factor 3D Geological
Model
• Storage coefficient Construction

• Specific yield (unconfined aquifers)

• Effective porosity
Figure 1. Workflow for the construction of the geometry of the
system from the field geological data collection until the
parameterization of the hydrogeologic units. (Author, 2020)
4 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 Figure 2. Geologic cross- section based on stratigraphic and sedimentologic interpretation of the borehole logs. Three
main geological facies are interpreted, showing that the screened section of the wells have correlation with a fluvial
sedimentary environment where thick sand layers intercalated with tiny clay layers are expected. (Author, 2019)

5 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES

 Vertical Speci
Specific Porosi
Horizontal K Vertical K Anisotro fic
Storage ty
py Yield

Code
Descriptio
Material
n
(m/d) (m/d) (kh/kv) (1/m) (-) (-)

Facies
Loess loess 11 4.32E-03 3.60E-03 10 1.00E-04

0.1 0.12
Clay 1 silty clay 5 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Red sand 17 4.32E+00 2.88E+00 10 1.00E-05
0.2 0.25
Clay 2 silty clay 6 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Magenta silty sand 12 4.14E-01 3.45E-01 10 1.00E-04

Lacustrine
0.16 0.18
Clay 3 carbonate silty clay 7 2.59E-04 1.73E-04 10 5.00E-03
0.05 0.08
Pink silty sand 16 1.73E+00 1.38E+00 10 5.00E-04
0.2 0.25
Clay 4 silty clay 8 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Light Pink medium 10 2.62E+00 1.75E+00 10 3.00E-05
sand 0.19 0.2
Clay 5 silty clay 9 6.05E-04 6.05E-04 10 5.00E-03
0.05 0.08
Channel Complex coarse sand 1 1.40E+01 7.78E+00 10 1.00E-05
1 0.19 0.2
Overbank 1 silty clay 13 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Channel Complex medium 2 2.62E+00 1.45E+00 10 3.00E-05
2 sand 0.22 0.25

Fluvial
Overbank 2 clayey silt 14 8.64E-04 7.85E-04 10 2.00E-03
0.09 0.1
Channel Complex coarse sand 3 3.82E+00 1.91E+00 10 1.00E-05
3 0.25 0.28
Overbank 3 silty clay 15 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Channel Complex medium 4 4.14E+00 2.07E+00 10 3.00E-05
4 sand 0.2 0.25
Underlying Silt 1 silty clay 21 6.05E-04 5.50E-04 10 5.00E-03
0.05 0.08
Underlying Sand silty sand 18 4.14E-01 2.76E-01 10 5.00E-04
1 0.16 0.18

Deltaic
Underlying Silt 2 clayey silt 22 7.78E-04 7.07E-04 10 2.00E-03
0.09 0.1
Underlying Sand silty sand 19 4.14E-01 3.45E-01 10 5.00E-04
2
Figure 3. A combination of 2D cross sections (as Figure 2) of different orientation permit Table Underlying
1. Hydrogeologic parameters used to hydraulically characterized
0.16 0.18
Silt 3 clay 23 5.18E-05 5.18E-05 10 9.00E-03
create a 3D interpretation of the geological geometry of the area, it is call a subjective the lithofacies, and translate them into hydrofacies. Not always one
0.03 0.04
geological approximation. In green color are represented the lacustrine facie, the yellow color lithofacies will correspond to one hydrofacie, because a criteria for a
corresponded to a fluvial facie, under it were deposited a deltaic facie. (Author, 2019) hydrofacie is based on their hydraulic properties. (Author, 2019)

6 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES

 b. Boundary conditions
They refer to hydraulic conditions along the perimeter of the problem domain. Exist
three possibilities to be applied:

• Specified head (Dirichlet conditions): A specified value of the hydraulic head

is setting up at the boundary location.

• Specfied flux (Neumann conditions): In this case the derivative of the head at

the boundary is specified.

• Head dependent flux boundaries (Cauchy or mixed condition): The flow

across the boundary is calculated from Darcy’s law using a gradient calculated as

the difference between a specified head outside the boundary and the head

computed by the model at the node located on or near the boundary

c. Sink and sources

They refer to hydraulic conditions into of the problem domain that cause effects on the
hydraulic head distribution.

• Wells: Into this type can be located production or recharge wells, and it systematic

periods of use with flow rate data.

• Surface water bodies: It make reference with location and hydraulic characteristics

of water bodies like rivers, streams, ponds, lakes or dams.

Figure 4. The 3D geometry of one aquifer system with 22 hydrofacies (Table
• Precipitation and evapotranspiration: The volume of aquifer recharge are 1) and four exploited aquifer layers is settled up to study how the hydraulic
head varies if 4 production wells are working with 500 m3/d under specified
conditioned for climatic and seasonal factors. During some periods will exist a intense head boundary conditions to the productive layers. 12 borehole log where
used for the interpretation of the geological geometry. (Author, 2019)
precipitation or it can have a higher evaporation rate, these characteristic must be
7 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 NUMERICAL MODEL
a. Code selection
After have a solid conceptual model of the problem the next step is

choose the correct software or code to solve it.

For complex geometric aquifers, a code that support the

UnStructure Grid is choose with main objective of a better

representation of aquifer’s geometry.

MODFLOW-USG is a version that use Control-Volume Finite-Difference

(CVFD) Approximation. In a CVFD model, spatial discretization can

include a combination of rectangular and square cells as well as

hexagons, triangles, and irregular shapes. Moreover, spatial

discretization can be different in each vertical layer. Figure 6. Plan view of finite-difference cells (n and m) in an
unstructured grid as used in the CVFD. (a) Connection lengths
The CVFD method is a type of finite volume. Formally, the flow are measured from the center of the cell. The flow area is
shown in gray. (b) Geometry correction of the cell conecxion to
equation is integrated over a small control volume V; and after, the bisect the shared face at the correct angle. (Panday et al.,
2013)
volume integral can be converted into a surface integral.

8 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES

 b. MODFLOW-USG

It is similar to MODFLOW–2005. Both

use stress periods and time steps for
temporal discretization.

An important difference between

MODFLOW and MODFLOW–USG is
cells shape ability where cells are
identified by node number when
used with an unstructured grid
dataset.

• Packages and process: Table 2

show a list of the packages and
process suported by MODFLOW-
USG. The most notable exception is
Table 2. List of packages and processes supported in Version 1 of
the Unsaturated Zone Flow (UZF). MODFLOW-USG documented by Panday et al. (2013)
9 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 CALIBRATION AND SENSITIVITY
ANALISIS
• Calibration is the process of adjusting parameters or
fluxes, such as hydraulic conductivity and recharge,
within reasonable limits to match observations. Normali
developed in two stages: trial-and-error history matching
and parameter stimation.
• Verification is a process of testing the calibrated model
by demonstrating that it can successfully predict a set of
observations not used previously for model calibration
• Sensitivity analysis quantifies the effect that a change
of a given parameter produce into the model results

The degree of difficulty of model calibration depends on the

amount and the quality of measured data, the complexity of
processes being simulated, and the complexity of the
conceptual model. Before proceeding to implement the
calibration have a robust and representative conceptual
model is important, as well as select the ranked targeds for Figure 7. General workflow for parameter stimation,
the process. the second phase of hytory matching for a model
designed for forecasting. (Anderson et al., 2015)
10 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 PREDICCION AND UNCERTAINITY
ANALYSIS
• Frecuently, the model’s purpose is addressed by making a forecast
of the response of the system to future conditions, it is related with
prediction which naturally reflect an uncertanity for future
estimations.
• Uncertainty analysis may be undertaken for these reasons:
• Illustrate (and quantify) how conceptual model limitations affect predictive
model results
• Quantify the impact of variation in parameter estimates and assumptions

For any forecast, two broad sources of uncertainty can be considered:

(1)uncertainty associated with the model itself: from

assumptions used in the conceptual model, measurement error in
observations used to calibrate the model, simplifications required by
calibration, and simplification error resulting from defects in the
parameterization of the model selected.

(2)uncertainty associated with accurate specification of future

conditions: arises when forecasts need to estimate future stresses Figure 8. Workflow to preform a basic uncertainty
analysis. (Anderson et al., 2015)
and properties.
11 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 SUMARY
The modelling process can be summarize in the next steps: (1) conceptual model construction, (2) code selection and
numerical modelling, (3) calibration and validation, and (4) prediction and uncertainty analysis.

Start the process with a robust conceptualization of the model will help in the next steps of the process to have a clear scene
about the phenomenon that will save time and resources in the following steps.

Nowadays, a plenty of possibilities of codes had been displayed on the hydrogeologic modelling scene. It lies on the modeler
criteria the selection of which one will preform better the conditions of a specific problematic, and he/she will pick it up to solve
the phenomenon onto the study.

When the numerical modelling stage is passed out, the calibration process must be carried out to assess the base model
performance. At this step, how appropiate the model is for our problem has been evaluated. Solutions for parameter
sensitivity, lack of data or inconsistences of the conceptual model must be apply to rich the level of fitting without lose of
reliability of the model.

After the appropiate base model is reached, the forcasting of the process is run. The forcasting as a predictig have assiciated
uncertanity which must be measure with the objective of report the reliability of the model and the probability of the presented
evolucion of the phenomenom in time.

The modelling process is an algorithm with a complex assemble of deterministic and subjective knowledge where the
uncertainty is accumulative throughout the stages, this fact supports the necessity of incorporating a variety of techniques
into the process with the objective of reducing the grade of uncertainty on the results.
12 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES
 BIBLIOGRAPHY
• Anderson, M. P., Woessner, W. W. and Hunt, R. J. (2015) Applied Groundwater Modeling. Simulation
of Flow and Advective Transport. Second edi, Applied Groundwater Modeling. Second edi. doi:
10.1016/c2009-0-21563-7.
• Coelho, C. D., Faria, A. C. S. and Marques, E. A. G. (2017) ‘Comparative analysis of different
boundary conditions and their influence on numerical hydrogeological modeling of Palmital
watershed, southeast Brazil’, Journal of Hydrology: Regional Studies. Elsevier, 12, pp. 210–219. doi:
10.1016/[Link].2017.05.006.
• Engelhardt, I. et al. (2014) ‘Complexity vs. Simplicity: Groundwater model ranking using
information criteria’, Groundwater, 52(4), pp. 573–583. doi: 10.1111/gwat.12080.
• Panday, Sorab, Langevin, C.D., Niswonger, R.G., Ibaraki, Motomu, and Hughes, J.D., 2017,
MODFLOW-USG version 1.4.00: An unstructured grid version of MODFLOW for simulating
groundwater flow and tightly coupled processes using a control volume finite-difference
formulation: U.S. Geological Survey Software Release, 27 October
2017, [Link]
• Recktenwald, G. W. (2019) The Control-Volume Finite-Difference Approximation to the Diffusion
Equation. Portland. Available at:
[Link]
• Wels, C., Mackie, D. and Scibek, J. (2012) Guidelines for Groundwater Modelling to Assess Impacts
of Proposed Natural Resource Development Activities. Available at: [Link]
• Welter, D. E. et al. (2015) ‘Approaches in Highly Parameterized Inversion : PEST ++ Version 3, A
Parameter ESTimation and Uncertainty Analysis Software Suite Optimized for Large Environmental
13 02/03/2025 HYDROGEOLOGIC MODELLING PROCESSES