I. J.

Mathematical Sciences and Computing, 2022, 4, 1-14

Published Online on October 8, 2022 by MECS Press (http://www.mecs-press.org/)
DOI: 10.5815/ijmsc.2022.04.01

Mathematical Based Implicit and Explicit Finite

Difference Techniques for Solving the Ground
Water Flow Equations Using Spreadsheets
Farzin Salmasi
Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Tabriz 51666 IRAN

Mohammad Taghi Sattari

Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Tabriz 51666 IRAN
Email: mtsattar@gmail.com (M.T.S.)

Halit Apaydin
Department of Agricultural Engineering, Faculty of Agriculture, Ankara University, Ankara 06110 TURKEY

Received: 19 February 2022; Accepted: 13 May 2022; Published: 08 October 2022

Abstract: In countries with arid and semi-arid climate such as Iran with water constraints, the use of groundwater
resources is very important. There are various mathematical based methods and software packages for modelling
groundwater resources. This paper uses groundwater flow problems to illustrate possible approaches for providing the
environment of active teaching. Mathematical models supported by software applications facilitate the gaining of an
insight into the physical behaviors by investigating a host of scenarios and events but they are poor in training critical
thinking for encapsulating the hardcore mathematical equations describing the problems. Whilst software engineering
has transformed the intellectual capitals accumulated between the 20 th century and the middle of the 21th century into
working tools, it has the drawback of encapsulating core mathematics away from common experience of the students
and practitioners. This diminishes critical thinking in a world of increasing risks and ought to be taken a serious side
effect of software engineering. This paper suggests a solution by building up a library of solvers using spreadsheets,
with the effect that the encapsulated knowledge of building modelling solvers can permanently be brought to life in
education with the active learning culture. Implementation was carried out in the same way for steady state flow as well
as explicit 2D and 3D finite difference approximation for transient flow. This study raises concern about the
encapsulated body of knowledge contributed to the emergence and the establishment of modelling software applications
since 1980. This body of knowledge comprise a deeper understanding of equations of often partial differential equations
describing physical problems, as well as their numerical transformation into systems of equations and their subsequent
properly- and improperly posed systems of equation in terms of their assumptions and quality conditions. The outcome
is the emergence of a cookbook mentality among the new breed of mathematical modelers without any critical thinking.
The results revealed that spreadsheet can be used with the aid of the Solver function. This idea capitalized on the
capabilities of the net-generation and opens up the possibility for the emergence of bottom-up open source modelling
platforms.

Index Terms: Critical thinking; explicit; groundwater modelling; finite difference; implicit; spreadsheet

1. Introduction

Gaining an insight into reality and invoking critical thinking in the learners mind should be considered as two of
important pedagogical issues, particularly in a world with increasing risks. Mathematical models supported by software
applications facilitate the gaining of an insight into the physical behaviors by investigating a host of scenarios and
events but they are poor in training critical thinking for encapsulating the hardcore mathematical equations describing
the problems. Arguably, an insight is a comprehensive understanding within a given framework of mind. Among
different definitions of critical thinking, Scriven and Paul [1] define it as skilfully conceptualizing, applying, analyzing,
synthesizing, and/or evaluating information as a guide to belief and action.
This paper takes critical thinking as the ability to search and shift from one reasoned framework of mind to another.
This paper views both insight and critical thinking as complementary, where one does not replace the other. Although
modelling has become indispensable, the emerging modelling practices nurture the cookbook mentality among the

This work is open access and licensed under the Creative Commons CC BY License. Volume 8 (2022), Issue 4
 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

practitioners and students. Arguably, this is not healthy and therefore measures are needed to invoke and train critical
thinking among the students (Kyrpychenko et al. [2]).
One way of invoking critical thinking in teaching mathematical modelling is presented by using examples from
groundwater modelling. Groundwater flow problems are governed by partial differential parabolic equations, the
solutions of which are currently by numerical methods. Software engineering has been effective in transforming the
solution capabilities into versatile tools but their solvers often tend to comprise of a single solution method, as their
productions require considerable investments. Models built in one modelling platform are not often transferrable to
other platforms. Thus, the learners cannot train his critical thinking by trying different solution approaches but this is
feasible in the spreadsheet environment.
This paper seeks spreadsheet programming as a convenient environment for the training of critical thinking in the
researchers and the student. Simplified groundwater problems are used to obtain solutions of the governing equations
without involving any particular programming like MATLAB or FORTRAN. It shows one way to effective students’
learning by simplifying the underlying mathematics through dropping out tedious components in favor of arriving at
sensible solutions. In this way, it becomes feasible (i) to focus on flow patterns of practical problems caused by
different boundary conditions by inserting a sink (well) or source (recharge) on flow domain, (ii) to assessing their
effects on water table or sensitivity against hydraulic parameters of transitivity (T) or storivity (S), or (iii) trying out
different solution schemes to invoke critical thinking.
Learning and teaching are evolving. With a focus on recent situations Jewell [3] argues that in an applied science,
there exists a natural tension between the study of fundamental scientific theory and instruction in the application of
analysis and design methodologies within undergraduate engineering curricula. Most engineering courses are structured
to emphasize the relevant physical, chemical, and biological processes that are then reinforced by studying specific
problem solving skills applied to systems of engineering interest. Over the years, a wide range of courses of thought has
emerged comprising theories, concepts or philosophically-oriented doctrines.
If there can be a consensus on the most effective way of learning and teaching, it is the one that there is no such a
single theory. Nonetheless, each of these methods has a positive contribution but they also tend to overlook the central
themes promoted by other theories for being unable to embrace pluralism in their methodology. This is because they all
focus on finding the one best technique but this is just a cultural illusion and a legacy of philosophical doctrines.
Arguably, the best technique is pluralistic integrating all of these techniques on their positive contributions and wary of
their inter-conflicts. Pluralism is feasible by evolutionary thinking but this is not the focus of this paper.
Whilst using modelling software applications can train the user’s ability on gaining an insight into physical
problems, their limitations in training alternative techniques is obvious; spreadsheets overturns this limitation by
enabling the users to test different solvers and create new solvers, albeit with a limited capability, as discussed below.
Spreadsheets were originally created to help accountants to solve their accounting problems. Due to their intrinsic
ease-of-use, they have been extensively applied in many different fields [4], [5]. They are also widely used as a training
tool in engineering. For example, Excel spreadsheet has been used in simulating engineering systems, such as logic,
networks, control systems and antenna array design [6]. Spreadsheet has also been used in the field of machine design,
such as design of worm gear geometry, determination of spur gear form factors, aircraft structure analysis [7]. Karahan
[8] studied one-dimensional advection–diffusion equation (ADE) with finite differences method using implicit
spreadsheet simulation (ADEISS). Two examples which, have the numerical and analytical solutions in the literature,
are solved in order to test the ADEISS performance. Spreadsheets are typically programmed in Visual Basic to
customize their use for solving complicated engineering problems.
Dehghan and Mohammadi [9] used mesh less techniques and high dimensional partial differential equations to
solve the tumor growth model. Also they used a semi-closed finite difference method based on the Crank-Nicolson
scheme, and the other based on open Runge-Kutta time integration. Panigrahi and Velusamy [10] developed a transient,
multi-phase enthalpy based computational model to analyze the streaming and freezing characteristics of the molten
fuel in a blocked fuel subassembly (SA) during total flow blockage (TFB) at the foot of the SA. The model adopts both
implicit and explicit type of finite difference techniques employing a variable grid system for boundary tracking.
Macías-Diaz and Morales [11] used a stochastic differential equation in material refraction modelling. The developed
model considered that the spread of cracks on solids includes a deterministic and stochastic component. Sattari et al. [12]
predicted groundwater level using support vector regression (SVR) and M5 tree model in Ardebil Plain in Iran. They
used monthly groundwater level data from 24 piezometers for a 17‐year period (1997 to 2013). SVR and M5 models
predicted groundwater level with R = 0.996 and R = 0.983, respectively.
Modelling software applications have been integrated to engineering training in various ways and the challenge of
its incorporation into the teaching curricula was recognized even in the early days. Cryer [13] argued in the past in favor
of a productive role for scientific calculators, equation solvers, mathematics packages, spreadsheet applications,
commercial analysis software, and programming assignments, but the selected tool depends on the course context and
technology. Weiss and Gulliver [14] discussed the use of spreadsheets to analyze various hydraulic design projects.
They illustrate the use of spreadsheets as a tool to analyze practical problems, not only to teach valuable engineering
analysis skills but also to enhance users’ computer skills for preparing them to the challenges to be faced professionally.
Huddleston [15] discussed the use of spreadsheet tools to introduce students to fundamental concepts of
computational fluid dynamics by using an illustration from open-channel hydraulics. Huddlesston et al. [16] uses Excel

2 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

illustrations to enable users’ analysis while still requiring enough manual development to reinforce the underlying
engineering principles. His argument is that analysis commonly results in nonlinear differential or algebraic equations
or systems of equations but the users’ computational capability is a limiting factor. According to him, this impedes
learning but he suggests that it can be overcome through instructors’ diligence by balancing the need to emphasize the
engineering system physics versus numerical complexity using worksheet capabilities.
Pandit [17] provides several examples with application of spreadsheet in the areas of fluid mechanics, hydraulics,
hydrology and storm water management in a workbook. Niazkar and Afzali [18] present two engineering examples in
their paper using Excel spreadsheet. These examples are (i) earthquake data bases for computing peak ground
acceleration (PGA) with highly nonlinear equations and (ii) the calculation of gradually varied flow condition in
channels, in this Excel spreadsheet example, a first ordinary differential equation governs the value of water depth along
the channel. In order to solve this equation, a finite difference scheme, e.g. Euler’s method can be applied. Divayana et
al. [19] evaluated a computer-based evaluation application called the CIPP-SAW.
This study examines the use of commonly available spreadsheet package to analyze groundwater flow problems to
solve nonlinear systems of equations. The application of this technology is an efficient way to enable researchers to be
engaged with solving relatively complex engineering problems while minimizing the computational burden. Built-in
linear and nonlinear system functions are commonly available in commercial spreadsheet programs, providing an
affordable alternative to more complicated or expensive software. These facilities enable researchers to analyze realistic
applications while still requiring manual development of the governing equations to reinforce their underlying
engineering principles.
The remain of the article is structured as follows: hydraulics of groundwater, parabolic groundwater equations,
boundary conditions, spreadsheet implementation of groundwater problem, iterative calculations, explicit 2D finite
difference approximation, explicit 3D finite difference approximation, implementation of the 2D explicit solver (steady
state flow), implicit 2D finite difference approximation, discussion and conclusions.

2. Hydraulics of Groundwater

The customary research practice prior to 1950 was to seek analytical solutions for practical problems (i.e. to solve
the equations for any point in space and time) by imposing a set of approximations but the advent of computational
capabilities in the late 1950s and 1960s turned the tide and set the pace for the research efforts on establishing the proof-
of-concept for nearly all of the equations available then by sweeping away these limiting assumptions. Furthermore, the
advent of software engineering since the 1980s in the consultancy industry has transformed these proof-of-concepts into
successful working tools.
Consider the example of transient groundwater flows, which are described by partial differential parabolic
equations describing flows of a wide range of engineering problems. Their solutions depend on the specification of
boundary conditions and initial conditions, as well as the specification of a number of parameters including storitivity
and permeability. The goal of the solution is the determination of the free surface but the solution strategies have gone
through paradigm shifts. In the past, it was customary to simplify the equations and solve them by analytical techniques.
These techniques had restrictive assumptions and therefore their solutions were not of practical significance. The advent
of computers since the 1950s and their applications to groundwater problems in the 1960s-1980s removed these
restrictions but the emergence of software engineering since the 1980s has given rise to commercial and the public
domain software applications.
The software applications often induce the cookbook mentality among the engineers and scientists hampering
direct experience of solving the equations, essential for learning. This paper reflects the past research mindsets and
promotes their reactivation in pedagogy as the most convenient way to transform passive into active learning by using
spreadsheets. This section presents one possible solution to support a better education of the students on groundwater
problems by using spreadsheets.
2.1 Parabolic Groundwater equations
Consider the general one dimensional (1D) transient flow equation in a confined aquifer, in which the unknown
variable is time dependent and referred to as unsteady, non-equilibrium, or transient problems. Their governing
equation combines the continuity equation with the dynamic equation (the Darcy law) for an aquifer and expressed as
[20]:

( )
- (1)

where, S is storivity defined as the volume of water released from storage per unite area of aquifer per unit decline in
head; ( ) is sink/source used to simulate both distributed and point sources (positive values) and sinks (negative
values) with units of length per unit time; is transmissivity and K is hydraulic conductivity. Transmissivity is defined
as T=KD, where D is the porous medium depth. For more details, see Bear [21], among others.

Volume 8 (2022), Issue 4 3

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

2.2 Boundary Conditions

The solution of Eq. (1) requires the specification of boundary conditions to constrain the problem and to make
solutions unique. The different types of boundary conditions are (a) head is known for surfaces bounding the flow
region (Dirichlet conditions), (b) flow is known across surfaces bounding the region (Neumann conditions), (c) some
combination of (a) and (b) is known as the Robin Boundary Conditions (mixed conditions).
In water engineering problems, the Dirichlet boundary condition occurs when one side of the domain is maintained
at a fixed water level. For example the upstream slope of an earthen dam has the Dirichlet condition, because all nods in
the upstream slope is normally specified in terms of head: ( ), where z = elevation at each node from
datum, p is water pressure above each node and is specific gravity of water [22].
One of the most common boundary conditions in water engineering problems is the no-flow boundaries (Neumann
conditions). Examples are (i) dam foundations normally resting on an impervious bed rock, (ii) cut-off walls below the
dams implemented for reducing seepage flow.
2.3. Spreadsheet Implementation of Groundwater Problem
Many problems of the physical sciences are solved by a set of governing equations but their solutions for practical
problems are often obtained by numerical techniques. These techniques transform the governing equations into a system
of equations represented in matrix formats. The automatic solutions of the matrices of often nonlinear equations are
obtained by further transforming them into an algorithmic fashion, in which the solutions are obtained through iterations.
The bird’s eye view of the complete solutions is that the boundary conditions set the iterative solutions running through
the algorithmic equations connected to one another as a chain with the following possibilities: 1D (e.g. steady solutions);
2D (e.g. 1D in time plus 1D in space); 3D (e.g. 1D in time plus 2D in space) and 4D (e.g. 1D in time plus 3D in space).
Mathematically minded researchers can understand the working of such mathematical complexity and the aim here is to
lay down the solutions for those of less keen on mathematics but rife for a critical understanding of the whole process,
as discussed below.
2.4. Iterative Calculations
Iterative solutions of systems of equations are also feasible in spreadsheets, as they offer both grids and inbuilt
iterative facilities. However, this medium of solving equations has its peculiar features, setting its strength and
weaknesses, as discussed in this section. The matrix format of equations is not used in iterative spreadsheet calculations
but the finite difference equations are applied directly to its each and every grid nodes, where a grid in the physical
system is seen as a cell in the spreadsheet and within the domain of the solution, the adjacent cells are interconnected.
This is a tedious process, as the appropriate equation has to be written for each cell. However, they are easily copied
through by using the copying and pasting facilities of spreadsheets. These facilities automatically increment row and
column numbers.
It is recommended to deactivate the iterative calculation feature of spreadsheets while defining the equations to
avoid the temporary warning messages related to circular references, and to reactivate this iterative feature defining all
the equations. The iterative spreadsheet calculations are based on the numerical concept of successive relaxation (SR),
which applies to the solution of both linear and nonlinear system of equations.
2.5. Explicit 2D finite difference approximation
2.5.1. An Explicit 2D Solvers (Transient flow)
For an aquifer of infinite width of space steps of , the second order finite difference scheme may be used to
discretize the continuous equations for groundwater flows expressed by Eq. (1) in space and a first order finite
difference scheme discretized in time, as follows (Wang and Anderson 1982):
The solution for 3D; h=f(x,y,t):

( ) (2.a)

The solution for 2D; h=f(x,t):

( ) (2.b)

where is head at point i,j and at time step: n+1. Explicit solutions are susceptible to instability unless they satisfy
Courant-Frederick-Levy (CFR) conditions given by:

⁄( ) (2.c)

4 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

2.5.2. Implementation of the 2D Explicit Solver (Transient flow)

Fig. 1 presents a test case comprising a simple example, and specified in terms of the following boundary
condition: ( ) and ( ) for h1=20 m, h2=15 m for an aquifer of infinite width and length of:
l=100 m. The initial condition is ( ) for .

Fig. 1. Boundary and initial conditions for flow at porous media: (1D: space, 1D: time)

The simple problem is solved by (2.b) using the explicit scheme for a 2D problem (1D in space), which must
satisfy the stability condition through: ⁄ ( ) . The solution of these equations engages the researcher with
the direct learning of the specification of the boundary and initial conditions and with implementing the solution of (2.b).
The worksheet for Fig. 1 uses (2.b) structured as follows: (i) the values of constant parameters are specified in a
table, comprising: t, x, T and S, as well as the time steps. (ii) Another table presents the solver, defining the spatial
layout of the test case for each time step with the grey cells specifying the Dirichlet boundary conditions. (iii) The 1D-
spatial domain is set up in the cells of one row defining the computational domain by defining the equations for each
cell at the computational domain. For this example, the process starts at cell C10. Its equation is displayed in the
“formula bar” in Table 1: fx =+C9+$B$5*(D9-2*C9+B9)/($B$2^2).
Evidently, cell C10 is interconnected to 3 cells of: B9, C9 and D9 at time minute. So the value of cell
C10 (at time=5 minute), depends on the value of 3 cells in the previous time, minute. Using the Drag-and-
Fill knob facilities, this equation is copied to the other cells and the computation area is filled up by the equations.
Because the value of head (h) at n+1 time depends only on the values of heads in previous time (n level), equation (2.b)
can be solved step by step in time. In other words, if value of head (h) at n+1 time level depends on both heads at n-1
and n+1 time levels, a system of equations at each time level can be formed for their solution.

Table 1. Different formula in cells after copying cell C10 to other cells
Cell name Formulas
C10 =+C9+$B$5*(D9-2*C9+B9)/($B$2^2)
C11 =+C10+$B$5*(D10-2*C10+B10)/($B$2^2)
C12 =+C11+$B$5*(D11-2*C11+B11)/($B$2^2)
. .
. .
. .
K28 =+K27+$B$5*(L27-2*K27+J27)/($B$2^2)
K29 =+K28+$B$5*(L28-2*K28+J28)/($B$2^2)

Notably, Cells $B$5 and $B$2 have fixed values as indicated by “$” sign in their left and right. After copying C10
formulas in all cells ranges C10 to K29, the value of cells $B$5 and $B$2 remain as per their set values at the leftmost
column. So after the completion of the computational area, the option for the iteration process is activated again and this
triggers the computation. The time taken depends on the defined tolerance, e.g. 0.0001. Using the drag-and-fill facilities
set up the equation, in the cells in the computational area and an example is displayed in Fig. 2.

Volume 8 (2022), Issue 4 5

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

Fig. 2. Excel spreadsheet for example of 1D explicit solver

Assuming that x=10 m T=0.02 m2/s = 1728 m2/day, S=0.002 and setting = 5 minutes, it can be shown that the
CFR stability condition of ⁄ ( ) is satisfied. Fig. 3 shows the results for the profile of the potentiometric
surface for several time steps.

20

19

18
h (m)

25 min
17 50 min
75 min
100 min
16
200 min
400 min
15
0 10 20 30 40 50 60 70 80 90 100

Distance (m)
Fig. 3. Potentiometric surface in different times

2.5.3. Further notes on implementing boundary conditions

One of common boundary conditions in groundwater problems is the no-flow boundary condition (Neumann
conditions), e.g. a dam resting on impervious bedrocks with the dam foundation specified in terms of a no-flow
boundary. A cut-off wall below a dam (for reducing seepage flow) is also a no-flow boundary. Along the vertical
boundaries specifying the no-flow boundary conditions, qx=0 implies: , so , hence: hi+1=hi-1.
For the left boundary, the point referred to by indices (i+1,j) is inside the problem domain, but the point referred to
by indices (i-1,j) is outside. Therefore, we expand the finite difference problem domain by one additional column to the
left by putting in a column of so-called imaginary or fictitious nodes (yellow colour nodes in Fig. 4).

6 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

Fig. 4. Finite difference grid for no-flow boundary condition

2.6. Explicit 3D finite difference approximation

2.6.1. An Explicit 2D and 3D Solvers (Transient flow)
The finite difference method for the transient version of the well drawdown in spatial 2D is used as a test case to
develop a spreadsheet solver for (1) using an explicit finite difference scheme. For a grid aquifer of sides and , a
second order finite difference scheme discretises the continuous equation (1) in space and a first order finite difference
scheme discretises it in time, as follows [23]:

( ) (3.a)

Let x= y=a, (3a) may be solved for , as follows:

( ) ( )( ) (3.b)

The variable is unknown and evaluated in terms of the known (old) values of h (i.e. ) at the nodes
surrounding (i,j) at the time step of n. Again CFL condition is given by:

CFL Condition: ⁄( ) (3.c)

2.7. Implementation of the 2D Explicit Solver (Steady state flow)

A test case is devised in Fig. 5 in which a well is discharging at a constant rate of Q within a confined aquifer. The
boundaries are treated as constant head boundaries. Fig. 5 is implemented in a worksheet with the following tables: (i)
One table assigns the model and aquifer parameter values: = =a=200 m, T=350 m2/day and S=0.002. The well is
discharging at a constant rate of -2500 m3/day and R=-Q/( )=-2500/(2002)=-0.0625 m/day where R is the rate of
release from the aquifer. (ii) The computational table (the solver) is developed to estimate the drawdown throughout the
aquifer based on (3.e). (iii) Further tables are laid out to specify the distributed values of T, S and R. For 2D steady state
ground water flow we have:

( )
(3d)
( )

( )
(3e)

Volume 8 (2022), Issue 4 7

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

The user can also specify some boundary conditions or all of them as no-flow boundary conditions. The layout of
the grid is shown in Fig. 5 (left, right, top and down with no-flow boundary conditions) but the implementation of the
boundary conditions is discussed further. In this test case, the grid is divided into three types of cells: (i) the
computational cells are highlighted in grey; (ii) the cells representing no-flow boundary conditions are highlighted in
dark and (iii) yellow cells represent fictitious nodes.
Mathematical modelling of no-flow Drichlet boundary condition in Fig. 5 for nodes B, A and H are:

Node B: (3.f)

As mentioned previously, implementation of Eq. (3.f) needs for representation of fictitious nodes/cells. Another
way without need for fictitious nodes is to use Eq. (3.g) as the following. Hence with attention to (3.e), heads on
boundary cells (like node B) defined as:

( )
(3.g)

At corner nodes like node A, no-flow boundary condition (BC) are both in x and y direction. In x direction Eq. (3f)
satisfy no-flow BC but at y direction, Eq. (3r) must be define.

(3.r)

Again, another way without need for fictitious nodes is to use Eq. (3.n) as the following. Hence with attention to
(3.e), heads on corner boundary cells (like node A) defined as:

( )
Node A: (3.n)

Fig. 5. Excel spreadsheet for example of 2D explicit solver

Fig. 6 presents the simulated drawdown surface. As the aim is not here to explore the accuracy of the scheme, its
analytical solution is not derived. However, different aspects of this problem can be explored in group teaching.

8 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

Fig. 6. Location of discharging well at cell H27 and head drawdown in other cells

2.8. Implicit 2D finite difference approximation

2.8.1. An Implicit Solver
Critical thinking in the student can be trained by learning the improvements in the solvers through using different
solvers, e.g. using implicit finite difference schemes for groundwater problems. In this approach, the continuous space
derivatives are discretised between t=n t and t=(n+1)t by weighting the average of the approximations at (n) and
(n+1). The weighting parameter is represented by , and it lies between 0.0 and 1, which normally set to 0.5 or 0.55.
The following implicit finite difference scheme is often employed (Wang and Anderson 1982):

( ) (4.a)

( ) (4.b)

( ) (4.c)

The substitution of (4.a)-(4.c) in (1) produces the following:

(4.d)

An example is presented to show how the solver described by (4.a)-(4.c) is transformed into a system of equations.
Consider an aquifer transformed into a grid of 33 in space. Each grid is governed by (4.d) and this leads to a system of
equations illustrated in Fig. 7.

Fig. 7. Illustration of discrete equations transformed into a system of equations

Volume 8 (2022), Issue 4 9

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

These equations at each cell may be transformed into an algorithmic format, as follows:

([( )[ ( )( ) ] (4.d)
) ]

2.8.2. Implementation
The implicit finite difference uses iterative spreadsheet solution technique. Thus, the formation of a system of
simultaneous equations is not required using matrix algebra. Fig. 8 formulates a test case in which the boundary cells
(the left, right and lower boundaries) are assumed to be constant head section (with h=100 m) and upper cells is
assumed with no-flow BC (this made mixed BC). In addition all initial value of cells selected h=100 m. The grid size is
selected at the spacing of x= y=a=100 m, t=0.001 days and =0.5 (Crank–Nicolson scheme). The aquifer
coefficients are T=300 m2/day and S=0.002. The well discharge is at a constant rate of 5 m3/day in cell K11. The test
case also assumes that a number of cells, namely, T4, T5 and S5 cells are recharged at a constant rate of 50 m3/day. This
implicit finite difference (Eq. 4d) is solved by using the spreadsheet and corresponding graphic head changes given in
Fig. 9.

Fig. 8. Excel spreadsheet for implicit 2D finite difference approximation

100.200
100.000
99.800
99.600
99.400
99.200
99.000
h (m) 98.800
98.600
98.400
98.200
1
4

10

13

16

19

98.200-98.400 98.400-98.600
98.600-98.800 98.800-99.000
99.000-99.200 99.200-99.400
99.400-99.600 99.600-99.800
99.800-100.000 100.000-100.200
Fig. 9. 3D representation of head variation by importing discharge well in K11 cell and recharge wells in T4, T5 and S5 cells

10 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

3. Discussion

The analogy between the reproduction process of living thing and education as the reproduction of knowledge
provides some food for thought. The reproduction process is an encapsulation of the primitive past at lower complexity
preparing individuals for new generations. Likewise, knowledge is created and encapsulated in time but every
individual has to acquire it through education and lifelong learning, during which encapsulated knowledge has to be
uncovered for the users even though many parts of the learned knowledge will not be used directly during the lifetime
of the user. The paper argues that a great deal of knowledge on mathematical modelling has been encapsulated and the
ways of uncovering it for training and educational practices are being overlooked. Arguably, there are new
opportunities as outlined below.
Self-directed lifelong learning runs simultaneously with a programmed phase of formal education at a certain stage
to acquire basic skills necessary for professional practices by being taught the mainstream knowledge. In the past,
education of science/engineering did not include the uptake of new knowledge produced by the on-going R&D
activities as their uptakes were slow due to many barriers hampering their direct benefits for teaching. The barriers are
increasingly becoming irrelevant as learning no longer relies on textbooks alone due to the emergence of (i) multimedia
approaches to learning and teaching, the effectiveness of which is stressed, e.g. by Lindstrom [24], stating that
simultaneously seeing and hearing and doing result in better learning; (ii) game-based learning (for more information,
see Prensky [25], among others), (iii) the emergence of modelling animations of many complex scientific theories, (iii)
the emergence of student-centred learning (see Brandes and Ginnis [26]).
It is often the case that the graduate student joins professional organisations with a typical cold start as practices
are aligned with many drivers yet to penetrate higher educational practices. Arguably, the practising graduate engineer
is often shocked by the scale of new facets of knowledge to be picked up in professional practices. In civil engineering
related to water resources, these include a vast array of mathematical and spatial modelling tools and risk analysis. The
task is therefore to introduce the various drivers to teaching curricula that are parts-and-parcels of professional practices.
These drivers include research, policy, best practice and the various thinking sweeping industry, e.g. sustainable
thinking, system thinking, uncertainty thinking, jointed-up thinking, critical thinking, etc. Curriculum designs should
include appropriate levels of research-led, policy-led and best-practice-led teaching programs, as well as training the
student with appropriate levels of gaining an insight into the problems and exercising critical thinking, although
sustainable thinking has already penetrated into global cultures.
While researchers and students are aided in various ways in their learning as discussed above, this paper only
highlights the dimension of engaging them with developing different solvers to train their critical thinking. The use of
groundwater flow modelling in this paper is incidental and in particular, the state-of-the-art in this problem area has
reached its classical status with no further contribution from this paper on the subject. The paper suggests the following
procedure for learning and teaching practices on mathematical modelling:

1. A “living team” can be formed in universities to develop and peer reviews of new solvers by the students to
build up a library of modules of programs, where each model starts from the beginning as follows.
2. The derivation of the governing equations are integrated by their quality conditions normally comprising the
continuous partial differential equations satisfy the three conditions of existence, consistency and stability,
together with highlighting their assumptions;
3. These equations are discretized by using numerical schemes (often using finite difference, finite elements,
finite volume techniques), where their discretization satisfies the three more conditions of: convergence,
consistence (compatibility) and stability. However, the transformed equations must be shown to be “properly-
posed” by: (i) having the number of equations equal to the number of unknowns, (ii) the subsequent equations
are consistent, (iii) they are stable; (iv) they are not ill-conditioned;
4. The formulation of their inverse problems still has to satisfy the three conditions of identifiability, uniqueness
and stability.

Each of the above activities was topical in the past and now encapsulated within the expertise of a narrow section
of professionals. The users are likely to learn passively these concepts, but passive learning is arguably a weakness in
teaching any educational subject. The uptake of multimedia, game-based learning and model-based animations is
making learning easy and releasing proportionately time resources reusable for different active learning subjects. It is
suggested that the teachers can engage students with building a library of solvers based on the procedure suggested
above and create the conditions for accumulating the solvers developed by the students over the years. This way of
accumulating solvers is a bottom-up approach pooling together the efforts by students, which in turn can lead to the
emergence of open source modelling systems. In this way, universities can change the situation on mathematical
modelling and transform their capabilities into organically growing living systems and assure a better educational
program.
Groundwater models may be used to predict the effects of hydrological changes (like groundwater abstraction or
irrigation developments) on the aquifer by formulating their simulation problems. Nowadays the groundwater models
are used in various water management plans for urban areas. Spreadsheets may be applied to these problems or any

Volume 8 (2022), Issue 4 11

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

partial differential equation with specified initial and boundary conditions. These test cases can be extended to estimate
uplift pressure below a dam and for accounting for recharge or discharge from aquifers, as well as that for the variations
in the values of S and T. Boundaries were treated as constant values in these test runs but variable values along the
boundaries can also be accommodated by spreadsheet modelling, which may be the case along earth dams. Other
problems approachable by spreadsheet modelling include: simple isotropic aquifers, anisotropic aquifers, stratified
aquifers, leaky aquifers, artesian aquifers, seepage from an earth dam, canal bed.
The level of sophistication on the implementation of the solvers in this paper is deliberately kept low and rather
than using loops, the solver tables are repeated for each time step. However, within a reason, spreadsheet can be made
relatively more versatile. Another limitation of spreadsheet calculations is that finite difference grids can only be
rectangular or triangular since they are mapped onto spreadsheet cells, therefore excluding curved boundaries and
inclined layer geometries. The spreadsheet solution procedure is presented in a manner that is consistent with the
governing conservation statements. This, coupled with students’ pre-existing familiarity with the spreadsheet package,
enables students to focus on understanding of the engineering system rather than cumbersome computational
procedures. Obviously, learning the skills necessary to implement and analyze the behavior of selected numerical
algorithms is also an important aspect of the development and application of computational models.

4. Conclusions

This paper raises concern about the encapsulated body of knowledge contributed to the emergence and the
establishment of modelling software applications since 1980. This body of knowledge comprise a deeper understanding
of equations of often partial differential equations describing physical problems, as well as their numerical
transformation into systems of equations and their subsequent properly- and improperly posed systems of equation in
terms of their assumptions and quality conditions. The outcome is the emergence of a cookbook mentality among the
new breed of mathematical modelers without any critical thinking.
A solution is offered that stems from the attributes of new technology already changing culture and now is
penetrating theories of education. New technological solutions include multimedia, game-based learning and animation
of mathematical modelling results, which inevitably make learning and teaching easier than the past on the assumptions
that the students have been transformed into active learners. As already teenagers have demonstrated their capacity of
active learning in the market-pulled and technology-pushed Internet social networking, educationists have embraced
this capacity. Thus, significant time resources are releases that can be utilized by forming living teams in universities,
whereby students produce a library of modelling solvers, which also peer-reviewed by them. This provides the
opportunity for transforming the encapsulated knowledge on mathematical modelling into an active learning process.
This paper uses spreadsheet solution of groundwater equations to illustrate the preliminary steps.
The technique used in this paper on solving groundwater equations are classic knowledge, which are transformed
to solvers on commonly available spreadsheet platforms. The solvers employ explicit and implicit finite difference
schemes for groundwater problems, applicable to flow and seepage problems. The main advantage of such a modelling
capability is its ease of implementation and application to a wide variety of practical groundwater problems. From an
educational point of view, all the assumptions, equations, and calculation steps can be captured on the spreadsheets and
be clearly stated, formulated, and executed, which is rather uncommon for other nonlinear numerical techniques. The
primary advantage of this suggested solution is the possibility of training the user’s critical thinking.
Declarations
Funding: No funding was received
Conflicts of Interest: The authors declare no conflict of interest.
Data availability: Data are available on request due to privacy or other restrictions.

References

[1] Scriven, M., Paul, R., 2003. Defining Critical Thinking: A draft statement prepared for the National Council for Excellence in
Critical Thinking Instruction.
[2] Kyrpychenko O, Pushchyna I, Kichuk Y, Shevchenko N, Luchaninova O, Koval V, 2021. Communicative Competence
Development in Teaching Professional Discourse in Educational Establishments, International Journal of Modern Education
and Computer Science(IJMECS), Vol.13, No.4, pp. 16-27, DOI: 10.5815/ijmecs.2021.04.02
[3] Jewell, T.K., 2001. Teaching hydraulic design using equation solvers. J. Hydraul. Eng. 127(12), 1013–1021.
[4] Cheetancheri, K.G. and Cheng H.H. 2009. Spreadsheet-based interactive design and analysis of mechanisms using Excel and
Ch, Adv Eng Softw 40, 274-280. doi:10.1016/j.advengsoft.2007.08.003.
[5] Yudianto MRA, Agustin T, James RM, Rahma FI, Rahim A, Utami E, 2021. Rainfall Forecasting to Recommend Crops
Varieties Using Moving Average and Naive Bayes Methods, International Journal of Modern Education and Computer
Science(IJMECS), 13 (3), 23-33, DOI: 10.5815/ijmecs.2021.03.03
[6] El-Hajj, A., Kabalan, K.Y., Al-Husseini, M., 2003. Antenna array design using spreadsheets. IEEE Trans Edu 46(3), 319–24.
DOI: 10.1109/TE.2003.813518.
[7] Shoup, T.E., 2004. Using spreadsheet modules to augment tolerance dimensioning. In: Proceedings of the DETC. 2912, ASME,
36–44. DOI: 10.1115/DETC2004-57649.

12 Volume 8 (2022), Issue 4

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

[8] Karahan, H. 2006. Implicit finite difference techniques for the advection–diffusion equation using spreadsheets. Adv Eng
Softw 37, 601-608. https://doi.org/10.1016/j.advengsoft.2006.01.003.
[9] Dehghan, M., & Mohammadi, V. (2017). Comparison between two meshless methods based on collocation technique for the
numerical solution of four-species tumor growth model. Communications in Nonlinear Science and Numerical Simulation, 44,
204–219. https://doi.org/10.1016/j.cnsns.2016.07.024
[10] Panigrahi, P. K., & Velusamy, K. (2019). A robust multiphase model to investigate molten material relocation during total flow
blockage in SFR fuel subassembly. Annals of Nuclear Energy, 133, 59–72. https://doi.org/10.1016/j.anucene.2019.05.007
[11] Macías-Díaz, J. E., & Villa-Morales, J. (2017). A deterministic model for the distribution of the stopping time in a stochastic
equation and its numerical solution. Journal of Computational and Applied Mathematics, 318, 93–106.
https://doi.org/10.1016/j.cam.2016.11.025
[12] Sattari MT., Mirabbasi R., Sushab RS, Abraham J. 2018. Prediction of Groundwater Level in Ardebil Plain Using Support
Vector Regression and M5 Tree Model. Groundwater 56(4):636-646. https://doi.org/10.1111/gwat.12620
[13] Cryer, C.W., 1970. On the approximate solution of free boundary problems using finite differences. Journal of the Association
for Computing Machinery 17(3), 397–411. doi:10.1145/321592.321593.
[14] Weiss, P.T., Gulliver, J.S., 2001. What do students need in hydraulic design projects? J. Hydraul. Eng. 127(12), 984–991. DOI:
10.1061/(ASCE)0733-9429(2001)127:12(984).
[15] Huddleston, D.H., 2002. Spreadsheet tools utilised to introduce computational field simulation concepts to undergraduate
engineering students. Comput. Educ. J. 12(1), 6–11.
[16] Huddlesston, D.H., Alarcon Vladimir J., Chen, W., 2004. Water distribution network analysis using Excel. Journal of hydraulic
engineering 103(10), 1033-1036.
[17] Pandit A. (2016). Water engineering with the spreadsheet: A workbook for water resources calculations using excel, Publisher:
Reston, Virginia: American Society of Civil Engineers, DOI: 10.1061/9780784414040.
[18] Niazkar M. and Afzali S.H. (2015). Application of Excel spreadsheet in engineering education. First International & Fourth
National Conference on Engineering Education Shiraz University, 10-12 November 2015. Pp 1-7.
[19] Divayana DGH, Ariawan IPW, Giri MKW, CIPP-SAW, 2021. Application as an Evaluation Tool of E-Learning Effectiveness,
International Journal of Modern Education and Computer Science (IJMECS), 13 (6), 42-59, DOI: 10.5815/ijmecs.2021.06
[20] Salmasi, F., Azamathulla, H. M., (2013). Determination of optimum relaxation coefficient using finite difference method for
groundwater flow, Arabian Journal of Geo-sciences, (6): 3409-3415, https://doi.org/10.1007/s12517-012-0591-9
[21] Bear, J., 1979. Hydraulics of Groundwater, Dover Books on Engineering.
[22] Hosseinzadeh Asl, R., Salmasi, F., Arvanaghi, H., (2020). Numerical investigation on geometric configurations affecting
seepage from unlined earthen channels and the comparison with field measurements. Engineering Applications of
Computational Fluid Mechanics, 14:1, 236-253, https://doi.org/10.1080/19942060.2019.1706639.
[23] Wang, H.F., Anderson, M.P., 1982. Introduction to Groundwater Modeling: Finite Difference and Finite Element Methods.
Academic Press, NY, ISBN-13: 978-0127345857, pp: 237.
[24] Lindstrom, R., 1994. The Business Week Guide to Multimedia Presentations: Create Dynamic Presentations That Inspire, New
York: McGraw-Hill.
[25] Prensky, M., 2007. Digital Game-Based Learning (Kindle Edition), Paragon House Publishing.
[26] Brandes, D., Ginnis, G., 1994. A Guide to Student-Centred Learning.

Authors’ Profiles

Dr. Farzin Salmasi, Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Tabriz, Iran.
He is a Professor in Department of Water Engineering, University of Tabriz, Iran. He was born in Tehran, Iran,
1970. He received the B.Sc. degree in Irrigation and drainage engineering from the University of Tabriz, Iran, in
1992 and M.Sc. degree in Hydraulic structures engineering specializing in mathematical modeling of ground water
from Shahid Chamran University (SCU), Ahvaz, Iran, in 1996 and He received Ph.D. degree in Hydraulic
structures engineering specializing in physical modeling of stepped spillways from Shahid Chamran University
(SCU), Ahvaz, Iran, in 2005. His doctoral thesis was on “Hydraulic investigation and physical modeling on stepped spillways”. His
main scientific interests are: physical modeling of weirs in irrigation canals, dam spillways, seepage analysis under diversion
dams and through earth dams, pump stations, design of earth and gravity dams and finite element analysis with commercial software
including Geo-studio (Seep/w) and ANSYS-Fluent. He has authored and co-authored about 150 national and international journal
articles and participated in 20 national and international conferences. He regularly serves as a reviewer for 10 high impact-factor
journals in engineering and material science, among them: Soil and Water (University of Tabriz), Civil Engineering (University of
Tabriz), Journal of Irrigation and Drainage Engineering (ASCE), Journal of Hydraulic Engineering (ASCE), Flow Measurement
and Instrumentation, ISH Journal of Hydraulic Engineering and Journal of Hydrology. In addition to these, he has performed five
scientific projects supported from “University of Tabriz -Iran”.

Halit Apaydin is professor at Ankara University Faculty of Agriculture. He received his MSc. and PhD. degrees
at the same university in 1995 and 2002, respectively. His main research interests include hydrology, geographic
information systems and artificial intelligence.

Volume 8 (2022), Issue 4 13

 Mathematical Based Implicit and Explicit Finite Difference Techniques for Solving the Ground Water Flow Equations Using
Spreadsheets

Mohammad Taghi Sattari is an associate professor of water resources management in the department of water
engineering, the University of Tabriz in Iran. He completed his BSc education at the University of Tabriz, MSc
education at Isfahan University of Technology and Ph.D at Ankara University in Feb 2009. More than 100 articles
in the field of water engineering have been published in national and international journals and scientific meetings.
Mohammad does research in Water Resources Management, Environmental Engineering, Civil Engineering and
Application of Data Mining in water engineering fields. Mohammad is currently a guest lecturer at Ankara
University.

How to cite this paper: Farzin Salmasi, Mohammad Taghi Sattari, Halit Apaydin, " Mathematical Based Implicit and Explicit Finite
Difference Techniques for Solving the Ground Water Flow Equations Using Spreadsheets", International Journal of Mathematical
Sciences and Computing(IJMSC), Vol.8, No.4, pp. 1-14, 2022. DOI: 10.5815/ijmsc.2022.04.01

14 Volume 8 (2022), Issue 4