Luận văn investigation of a lysimenter using the simulation tool siwapro dss and adaptation of this program to vietnamese requirements

SUMMARY The main objective of this thesis is to use SiWal'ro DSS to model and simulate the water flow process in the unsaturated zone with the available data from the lysime- ter number

TECHNISCHE

Gy UNIVERSITAT

DRESDEN

HANOI UNIVERSITY OF SCIENCE DRESDEN UNIVERSITY OF TECHNOLOGY

PHAM THI BICH NGOC

INVESTIGATION OF A LYSIMETER USING THE SIMULATION TOOL SiWaPro DSS

AND ADAPTATION OF THIS PROGRAM

TO VIETNAMESE REQUIREMENTS

MASTER THESIS

Supervisors:

Prof Dr Ing habil Peter Wolfgang Graeber

Dipl Ing Rene Blankenburg

Technical University Dresden

Institute for Waste Management and Contaminated Site Treatment

Hanoi - 2008

ACKNOWLEDGEMENTS

Two years have passed and marked a histarical pathway toward my Master degree

‘The two years were full of challenges, hopes, inspiration and wonderful support from many people I would like to thank you alll for a big variety of reasons:

‘My first greatest thanks go to my tutors Prof Dr Ing habil Peter Wolfgang Graeber and Dipl Ing Rene Blankenburg for having guided, supported and accom- panied me Unough the process of this Master thesis Thanks also for having greally contributed to the thesis with your vast experience and advice

Marry thanks to Prof Dr Bilitewski, Asse Prof Dr Bui Duy Cam and Asse Prof

Dr Nguyen Thi Diem Trang for making great efforls to establish and design the

training program frame for this master course and develop it, so I cam have a chance

to join this course

My acknowledgements go also to all teachers from Hanoi University of Sciences in

Vietnam and Institute for Waste Management and Contaminated Site Treatment in

Germany for giving me lots of valuable and interesting lectures and helping us to understand more clearly and have a thorough grasp of specific knowledge during

this master course

My grateful thanks to Dr rer nat Axel Fischer, Mr Christian and Mrs Hoang Phan

Mai tor helping and supporting during my time in Drosden and Pina, Germany

Thanks also to Pham Hai Minh for all administrative support during the Master

course time

J also would like to express my gratitude to:

© The Committee on Overseas Training Project, Ministry of Fducation and

‘Training for having granted the scholarship that supported this Mater thesis

« Hanoi University of Sciences and Institute for Waste Management and Con- laminated Site Treatment (TAA) for providing ail materials and equipments that [ used during the course

Vietam National University, Hanoi and Technical University Dresden and German Academie Exchange Sorvice (DAAD) for supporting this Master

training program in which | attended

Thanks to all the classmates for their nice and warm company for the encourage-

ment and support

And last but not least, special huge thanks to my family (sy parents mw law, my par-

ents, my husband, my son and my brothers and sisters) and all my friends (especial-

ly Mrs [la) and my relatives for thinking of me, helping me, and encouraging me in

my pathway to a Master degree

Ilove you alll

Hanoi, 10" December 2008

Pham Thi Bich Ngoc

SUMMARY

The main objective of this thesis is to use SiWal'ro DSS to model and simulate the water flow process in the unsaturated zone with the available data from the lysime- ter number 302 in Juelich, Germany

The unsaturated zone is the portion of the subsurface above the ground water table

Tt contains air as well as water in the pores This zone plays an important roll in

xuany especls of hydrology, such as infiltration, ex{ilualion, capillary rise, recharge, interflow, transpiration, runoff and erosion Interest in this zone has been increasing

inrecenl years because the movement of water along wilh conlaminanis ta this cone

have boon affecting the groundwater and the subsurface environment,

Water flow is concemed with movement of water in unsaturated porous media,

In order to handle water flow process under steady state or tiansient conditions in the unsaturated zone, a useful computer program is used to model and simulate this process This program combines the simulation module SiWalro for numerical amodeling of water flow and contaminant transport in variably saturated media with additional simulation and parameter estimation tools, data sources for the simula- tion and a graphical user interface

The computer-based decision support systers SiWaPro DSS sofiware is a program

for modeling and simulating the processes as water flow, solute transport, bio de-

gradation and sorption in variably saturated porous media

In SiWaPro DSS, the discretization of the modeling area is realized using finite

elements with the GALERKIN method SiWaPro DSS contains Ihe 2D triangular

mesh generator HasyMesh 1.4, he mesh generator allows the generation of meshes with varying element sizes and irregular mesh boundaries Currently, the generator allows flexible space quantization at modeling time given by the user

Ta validate SiWaPro DSS, the means of measurement data from a lysimeter expe-

riment are used Lysiuclers are đơvicos for mcusuring the characteristic properlics

of the soil water balance, amounts of seepage water and their quality In this thesis, lysimeter 302 looated in Juelich, Germany is used for calibrating model

The Juelich lysimeter 302 was established in August 2001, the monoliths were tak-

en out from Munich-Nenherberg in June 2001 and the installation of the measure- ment devices ocurred and the data logging started on December 10th 2001 This lysimeter is rưn by the Research Centre in Tuelich (FZJ) This lysimeter is a large undisturbed lysimeter with 2? in area and 2.4m in depth including 0.8m of refer- ence material The three suction cups are installed together with tensiometers, [DR and temperature sensors at 3-different depth layers distance from upper edge of the Iysineter im hơn aš 0,85m, 1,15m and 1,8m,

To model the water low of the lysimeter im SiWaPro DSS, the finite clement mesh

of the lysimeter is constructed with the cohunn of 1,6m in width and 1,6m in height (excluding 0,8m of reference material) The lower boundary condition is a first kind boundary condition that allows out{low only A second type boundary condition is applied at the upper boundary of the column of lysimeter, It is a transient boundary condition using time — variable boundary conditions to simulate precipitation in the tao Three soil water sampling device layers are applied as first kind boundary condition, and as the lower boundary condition, only outflow is allowed The col- umn of the lysimeter soil is divided into 5 layers: each of the soil layers is described

in ils hydraulics with 17 parameters

To calibrate model, two data sets of 11 soil hydraulic and van Genuchter parame-

ters with different initial pressure head and boundary condition of three suction cup layers as well as different amount of nodes and elements in the mesh are used Be- cause the time is short — besides, one model took from 25 hours to 50 hours for run- ning; some models took much more time, then they were stopped before they finish

So there are only 10 models were run After getting the result from simulation of each model, the simulation result was checked and analyzed and then the data set was changed or finite element mesh of the lysimeter was adjusted or the software

was reconsidered Ihe simulation results that were shown in diagrams in section 4.1 are the best model, but the results still show some difference of output between si-

mulation and measurement because input data which took from lysimeter station are

not well documented and some soil parameters which are estimated by the person

who operate the lysimeter are different from the fact The result shows that total inflow ard Lotal oullow of lystmeter are in balance That means the model and fi-

nite element mesh of the lysimeter is designed well Outflow of the suction cup

layer number 3 in the sinvulation is almost the same as measurement Cutflow of the

suotion cup layer number 1 and lower boundary condition in simulation are the same as measurement in the first year But in the second year, outflow of the suction

cup layer number 1 in simulation is higher than measurement; opposite to the out-

Now of the lower boundary condition the simulalion one is lower (ham mcasure-

ment Outflow at the suction cup layer number 2 is different increasing by time be-

tween simulation and measurement The differences come from the data mentioned

as above,

The SiWaPro DSS program have beon introducing to Federal Bnviromuentsl Bức reaus and Consulting Companies in Germany ‘these ureaus and Companies can use this software tool primarily for leachate forecasts with respect to the German soi] proisetion law, In Vietnam il also can be apply siuilar 10 Germany, but il takes

a bit time for Vietnamese to familiar with it, For Vietnamese to apply this software,

the GUI and help system were initially translated into Vietnamese

Therefore, i can he said that SiWaPro DSS is one of the useful tools for leachate

forecast However, it should be applied for a wide varicty of contaminants if the

software is revised to adapt with not only all available data but also a few available data, The lysimeter is good for calibrating the modet and will be betler if the dala is

documented well and frequency

21 Definition of soil and unsaturated zone seo

3.6.2.2 Mesh Generator -.isnniennninnuenninemninnnmnnnennn dd 3.6.3.3 Weather Cen€ratOF ác 4đ

3.6.2.4 12atabase LAYET, eo đÁÌ

3.6.2.6 Import and Export Interfaces - - Ag

3.0.3 Manual SiWaPro DSS Mesh enerator 50

3.6.3.1 Create a simple 2D mesh 41

3.6.3.3 Inserting a background image as construction basis 56

Appendix I: Precipitation using for simulation - - 80

Appendix 2: Brief of output af simulation for 784 days - 85

Appendix 3: Da from HiedSi0'6PNHE ào ¬—- 8D

Appendix 4: Data from simulation for the days equivalent with measurement

German Soil Protection Law

Classification and Regression Trees

Classification and Regression Trees

Decision Support System Equation

Finite Element

The Research Center in Juelich

Group Method of Daia Handling

The National Research Center for Environment and Health

Graphical User Interface

The North Rhine-Westphalia State Environment Agency

National Institute of Plani Projection Pedotransfer Function

Sickerwasserprognose / Leachate Forecast

Soil water sampling device layer 1 at 0,85m distance to upper edge of the lysimeter

Soil water sampling device layer 2 at 1,15m distance to upper edge of the Iysimeter

Soil water sampling device layer 3 at 1,80m distance to upper edge of

LIST OF FIGURES

Figure 1: The unsaturated sone compares with the saturated zone

Figure 2: Division of soil fraction sizes, German (left) and American (right)

Figure 3: Dicretization / meshing of area to be madeled

Figure 4: Boundary conditions and discetization of a simple model for groundwater

flow (from Chris McDermott, 2003)

Figure $: Boundary conditions and discretization for a simple column model .26

Figure 6: Stress applied to the top of the rock colann causes deformation

Figure 7: Mesh in details

Figure 8: Pressing of the stainless steel bottom plate (left) and lifting of a readily

filled monolithic: lysimeter (right)

Figure 9: Lysimeter covered with grass (left), the round surface of the Lysimeter

(middle) and lysimeter cellar with complete instrument (right)

Figure 10: The lysimeter system at the Biel measurement site

Figure 11: Cross-section of a guideline lysimeter surrounded by a control plot 32 Figure 12: The lysimeter station in Munich-Neuherber,

Figure 13: The instrument for measuring the wind speed (right) and the rainfall

dF

Figure 14: Simplified sketch of the lysimeter and boundary conditions in the upper, lower and 3 suction cup layers at lysimeter 302 in Juelich „35 Figure 15: The schematic composition and the arrangement of measurement

evices

Figure 16: Structure of S¥aPro DSS

Figure 17: Graphical user interface (GUI) of SiWaPro DSS

Figure 18: SiWaPro DSS help sysiem

Figure 19: Search options for database access

Figure 20: GeODin interface form for data import

Figure 21: First start of the mesh generator

igure 22: Define the modeling domain

Generated mesh with internal curves

Convex internal curve (left) and concave internal curve (Vight) sec 56

Adjusting graphic

Construction with a background image

Boundary nades of the generaled mesh

Selected nodes for assigning material number

Selected nodes for assigning initial pressure head 61

Dialogue box of language options

10

LIST OF TABLES

Table 1: Nodal Coordinates

Table 2 Soi properties and van Genuchten Parameters using for Simulation Table 3: Switching surfaces for the assignment of the at the beginning of boundary

Table 4: Properties of the boundary conditions

Table 5:Submitted soil hydraulic parameters of the lysimeter at I"2 Juelic 64

Table 6: Soil layer list of the lysimeters

Table 7: Parameter limits and maximum allowable concentrations of pollutants in ground waier (according to Vietnam standard TCVN 5944-1995 and German

INTRODUCTION

‘The unsaturated zone (vadose zone) plays an important roll in many aspects of hydrology, such as infiltration (the movement of water from the soil surface

into the sail), exfiltration (water evaporation from the upper layers of the soil),

capillary rise (water movement from the saturated zone upward into the unsa- turated zone due to surface tension), recharge (the movement of percolating

waler (rom the unsaturaled zone to the subjacent saturated vame), interMow

(flow that moves down slope), transpiration (water is uptaken by plant roots) (Dingman SL, 2002, p 220), runoff (the movement of water/rain-water agross the surlace soil and enlering streams or other surface receiving waler) and erosion (wearing away of soil by the action of water, wind, glacial ive, ete

on the soil surface) (Simunek I et al, 1994, p 1) Interest in this zone has been inercasing in recent years because the movement of waler along with contaminants in this zone have been affecting the groundwater zone as well as

the subsurface environment One of the inlerested areas is Lo predict the water

movement and water quality in unsaturated zone that is recommended to use

computer models,

‘The past several decades have seen considerable progress in the conceptual understanding and mathematical description of water flow and solute transport processes in the unsaturated zone A variety of analytical and numerical mod- els are now available to predict water and/or solute transfer processes between the soil surface and the groundwater table These models are also helpful tools for extrapolating information from a linvited mi aber of ficld experiments to

different soil, crop and climatic conditions, as well as to different tillage and

waler management schemes (Sununek J ef al, 1994, p 1)

A useful computer model thal, allows predicting waler and solute transfer

processes in vadose zone is the computer-based decision support system Si-

WabPro DSS ‘his program combines the simulation module SiWaPro for nu-

merical modeling of water flow and contaminant transport in variably media with additional simulation and parameter estimation tools, data sources for the simulation and a graphical user interface

The ram objective of this thesis is 10 use SiWaPro DSS to model and simulate

the water flow process in the unsaturated zone with the available data from ly- simeter number 302 in Tuelich, Germany As mentioned above, the SiWaPra

DSS can be used also for modeling and simulating the water New process in the saturated zone and the solute transport process (including bio degradation

and sorption) in the unsaturated and saturated zone, but this thesis does not

consider these processes bevause of Lime limitation,

Refore focusing on the main objective (discussed in the chapler 3 and 4), the fundamentals of soil hydrology will be discussed with the basics of soil phys-

ics and soil water of the unsaturated zone that are relative to the model (see

chapter 2)

The Tuekch lysimeter and lysimeter station descriplion are also mentioned as

an overview to understand more about the model (sce chapter 3.3)

Furthermore, the demands by law (thresholds for contaminants in groundwa- ter), the graphical user interface and help system of SiWaPro DSS should be translated into Vietnamese and adapled to Vietnamese requirements (sve chap- ter 4.2),

Hopefully, initial achievement of the study in this thesis will prepare the ground for an application SiWaPro DSS inta leachate forecasting in Vietnam

14

2 FUNDAMENTALS OF SOIL ITYDROLOGY

2.1 Definition of soil and unsaturated zane

There are several definitions of soil and the unsaturated zone in some science books and websites, bul wilhin the scope of this [hesis only a shorl compilation

of important terminology concerning, soil and unsaturated zoue which will be used in the following chapters as well as relevant to content of the thesis is

considered

Soil:

Soil is an extraordinarily complex medium, made up of a heterogeneons mix- ture of solid, liquid, and gaseous material, as well as a diverse community of living organisms (Jury W & [orton R., 2004, p 1)

Soil is a rather thin layer over the earth’s surface consisting of porous material

wilh properlies varying widely Tt can be seer: as a sand-sill-clay matrix, com

taining inorganic products of weathered rook or transported material together with organic living and dead matter (biomass and necromass) of the flora and

fauna (lanhaler C., 2004, p.13)

Unsaturated zone:

The zone between the earth's surface and the groundwater surface is to speak

of the unsaturated zone, also called zone of aeration (Lanthaler C., 2004, p.14, quoted from Ward R.C., 1975)

The unsaturated zone is the portion of the subsurface above the pround water

table Tl conlains air as well as waler in Lhe pores (see Figure 1) Tis thickness

can range from zero meters, as when a lake or marsh is at the surface, to hun-

dreds of meters, as is common in arid regions (Unsaturated zone flow project,

2001)

2.2

The unsaturated zone is the subsurface zone in which the geological material

contains both water and air in pore spaces It is different from the saturated

zone, in which all pores

in the aquifer are filled

with water (see Figure 1)

Figure 1: The unsatu- rated zone compares

with the saturated zone

(Unsaturated zone flow

ble to retrieve a pollutant from the unsaturated zone A pollutant that enters the

topsoil is transferred by the water movement through the big reactor, and if it

does not decompose, or become consumed by vegetation, or attached to the

soil material, it will finally reach the aquifer and contaminate groundwater

supplies

Soil hydraulic parameters

Determine water and solute transport with numerical modeling needs informa-

tion about soil hydraulic parameters Before go to the SiWaPro DSS for mod-

eling and simulating water flow in vadose zone, getting more knowledge about soil hydraulic properties is important This section will talk about some soil

hydraulic properties that are related to the model

16

Schlute ee

20 HH; ae

Figure 2: Division of soil

fraction sizes, German (left)

and American (right) nomen-

clanna Where Hloecke is

Block; Stee ix Stone; Kies

is Gravel, Schiff is Silt and

(from SCHEFFER 2002, p 157)

Ton is Clay

- the clay fraction < 2 pm in diameter,

has been formed as a secondary prod- uct from the weathering of rocks (pri- mary minerals) or from transported deposits,

- the norrelay fraction > 2 pan, can be

divided into the subclasses: silt, sand, and gravel (Marshall TJ et al., 1996,

p.4)

Size limits can differ between the German and the American classifications; therefore, limits are not natural but defined by man Vigure 2 show the 2 classification systems

of German and American The system of American coming from the United State Department of Agriculture uses 50 ym as the limiting size between silt and sand; the syslem of Gorman lakes lintils of 63 jum between silt and sand,

According to (Lanthaler C., 2004, p.15) another size dependent classification: coarse soil has a size of > 2 mm and fine soil < 2 mun, This is based on a suggestion by Aller berg (1912) to use the mumber 2 as a limit

between fractions.

where Mis mass af mineral grvins

Viyis volume of mineral grains Bulk density:

Bulk density, py, is the dry density of the soil

Volumetric water content:

Volumetric water content or simply water content in soil, @, is the ratio of wa-

ter volume to soil volume:

2.3 Soil water balance

Soil as an important storage medium can also be explained systematically in

the following soil water balance, where 4H’, the change of the amount of wa-

ter stored in a certain period, is according to (Marshall T.J et al, 1996, p

248) composed of

Precipitation (P) and irrigation (7) are balanced against (he amounts of losses

of surface runoff (A), underground drainage (D), and evapotranspiration (£) doring a given period Usually, quantities are given in mm A can be negative when water runs [rom soil lo the surface and D is nogative when (ground) wa-

ter gets to the root zone

Precipitation (P)

The only natural input in this system is precipitation and its appearance can be divided inta a liquid (drizzle, rain, dew) and a solid type (snow, glaze, frost, Tule) The geographical variations, the regional pattern of precipitalion and ils distribution during a yearänonth with different variability (regime) are the

most important aspects for hydrology and soil hydrology Rainfall intensity

(amount of precipitation divided by duration) is relevant in catchments areas

of rivers/streams susceptible to floods Whenever precipitation is collected with any type of rain gauge, uncertainties about the amounts ocour due to wind influcnee (especially in mountain areas), the topography and site around the gauge, rain drop size, the material and condition of the gauge itself or

splash and gauge errors (Ward R.C., 1975, p 16-34)

Irrigation (2)

While some areas have more than enough rainfall, agricultural land in other arcas has lo be irrigated Nol oriy arid and semi-arid regions arc irrigated but also sub humid areas where itrigation supplements natural reinfall Litigation

aims to recharge soil to the field capacity in the layer from which roots absorb water The amount of water applied depends on weather, soil, plant, and eco- nomic conditions Insufficient water supply leads to a decrease of yield but too much inigation will increase losses of percolation (and can cause a higher wa- ter table and salinization of soil) and evapotranspiration, see below (Marshall

TI et al, 1996, p 268-271)

Surface Runoff or Overland Flow (4)

Tn case ual the nvinfall rate exceeds the infillration vate, the surplus waler tra- vels over the ground surface without infiltration to reach a stream chanel and finally the outlet of the drainage basin On most soils covered with vegetation this is a rather rare phenomenon The following conditions are relevant for overland flow and the infiltration capacity, respectively: saturation of soil/topsoil, agricultural practices, freezing of the ground surface or when soils show a hydrophobic nature (Marshall TI el al., 1996, p 261-264, Ward RC,

1975, p 240)

Underground Drainage (D}

The amount of water percolating through soil to the water table and recharging groundwater is to be considered as the underground drainage Water flows downward to the groundwaicr tible, and drainage soil water content decreases after infiltration have stopped (Ward R.C., 1975, p 193)

(Kutilek M & Nielsen D.R., 1994, p 133, Ward B.C., 1975, p 166) defined

infiltration as a process of water (precipitation) entering soil through the sur-

face (Kutilek M & Nielsen D.R., 1994, p 133) denoted the flux density of

water across a topographical soil surface as the infiltration rate (formerly de- scribed as infiltration capacity, infiltration velocity and infiltrability) The rate

determines the maximum waler amounl tfiltraling sot under spscified coudi-

tions in a given time, not limited by the rate of supply Soil surface condition

substantially affects infiltration (Marshall TI et al., 1996, p 134)

20

According to (Kutilek M & Nielsen D.R., 1994, p, 176-178) two cases are

important: soil water redistribution oceurs when water percolates from wetted

topsoil to the drier subsoil, secondly, the process of drainage to the groundwa-

ter level when wetting front is not far from groundwater level or reaches it,

water flows at/near steady state conditions Excess water is able to move di-

rectly from the Lopsoil to the groundwater afler nfiltraiun has ceased

Evapotranspiration (ET): Evaporation and Transpiration

Evaporation (F) is the water loss from bare suil or a free water surface la the atmosphere and is not the same for these two kinds of surfaces because their properties are different, for example the surface roughness, the area of air- waler imlerfaee, (he heat capacity and heal conductance leading to different surface temperatures, Water extracted from soil by roots to the dry organic matter of plants and then transported to the atmosphere is called transpiration (TR) These two pracesses often cannot be separated and are then unified in

£ + TR, Vurthermore, a distinction has to

the term evapotranspiration 1

be made between the actual and potential evaporation/evapotranspiration, the actual FE or ET (ETa) rellevis the real amount of evaporation resulting from given meteorological conditions of a surface providing limited quantity of wa- ter for soil and plants: it is highly dependent on the water and energy supply

In contrast, the potential E or ET (ETp) describes (he maximal amount of evaporation that is possible under given meteorological conditions Maximal evaporation will occur when enough water is supplied, for example above areas of surface water (Kutilek M & Niclscn D.R., 1994, p 182-218, Ward R.C, 1975, p 95-124)

(Marshall 'l.J et, al., 1996, p 393-395) provides another balance, which is the water balance of a lysimeter:

2.4

AW can be determined when the container is weighable When / is known due

to recording and P is measured by rain gauges, Z can be determined by ba-

lancing input versus output variables

Soil water flow

This section deals with the movement of water in unsaturated porous media,

focusing on infiltration, which is the movement of water from the soil surface into the soil and redistribution, which is the subsequent movement of infil-

tated waler in the unsaturated zone of a soil

Infiltration, Percolation

In

ction 2.3, infiliration water was slrcady mentioned (Kutilek M & Nick

sen DR, 1994, p 133, Ward R.C., 1975, p 166) define infilwation as a

process of water (precipitation) entering soil through the surface ‘I'he term

percolation is used when the downward Now/movement of water through the

unsaturated zone is to be explained, (Kutilek M & Nielsen D.R 1994, p 133)

denote the flux density of water across a topographical soil surface as the infil-

tration rate (formerly described as infiltration capacity, infiltration velocity

and infiltrability)

Redistribution

Redistribution can involve cxfiltration (evaporation from the upper layers of the soil), capillary rise (movement from the saturated zone upward into the un- saturated vone duc to surface tension}, recharge (the movement, of percolating

water from the unsaturated zone to the subjacent saturated zone), interflow

(flow that moves downslope) and uptake by plant roots (transpiration) (Ding- man ST 2002, p 220)

3.1

MATERIAL AND METIIODS

Theoretical approaches and methodology

To model or simulate the water flow process, a water flow model of Juclich

Lysimeter number 302 is developed based on the finite element mesh method Tho principle behind the application of the finile clement technique is thai eve- rything is broken down into matrices representing the governing equations,

which are then solved for the unknown values (McDermott C.1., 2003, p.8) It

means that the modeled area is divided into smaller elements linked in a mesh

‘The shape of the element used in this program is triangle Nodes at the comers

of the elements define the boundaries of the element The nodes are numbered

and assigned walural coordinates of the area im question The elements are

numbered and the nodes assigned to each element recorded (see section 3.2)

The necessary simulation data are collected, documented and verified (sce sce- tion 3.7) The model input data, which are related to evapotranspiration, preci- pilalion and hydraulic soil, were Laken from ficld monitoring station Guclich

lysimeter station)

‘Theoretically, the parameters for model ae estimated either from the function

O(w) according to (van Genuchten M Th, 1980) or from the continuous fune-

tion k(y), relauion [rom (Mualem, 1976) and (van Genuchten M Th, 1980)

according to (Wésten J.H.M et al., 2001), where @(y) is water content as a

function of matrix potential and k(v) is unsaturated hydraulic conductivity as a

function of the matrix potential Rut in this thesis, ihe parameters for simula-

tion are taken from the report of (Puetz I et al., 2004) ‘hese parameters are

estimated and checked

For calibrating the model, the simulated values of outflow are compared with

the amount of water leaching from the lysimeler as well as the water conten

measurements at different depths of a soil profile

3.2 Finite element method

In SiWaPro DSS, the discretization of the modeling area is realized using fi- nite elements with the GALERKIN method, so understanding about the finite clement method is necessary before going into SiWaPro DSS description This section will discuss about the finite element method-detailing object to be modeled to finite element mesh and principle behind finite element calcuta-

tions

Object to be modeled to finite element mesh

The object / arca (from now on area) lo be modeled is divided imto smaller

elements linked in a finite elements mesh

The shape of the element is variable, bars, triangle, squares, tetrahedral and

cubes are most commonly used The boundaries of the element are defined by

nodes usually at the carners of the elements, bul sometimes also along the

boundary of the elements and within the element ‘The nodes are muambered and

assigned nalural co-ordinates of the area in question The elements are num-

bered and the nodes assigned to cach clement recorded More clements are

generated in places of special interest, or where there are expected to be higher

than normal changes in the parameters being ineluded in the model This process, known as discretization or meshing is illustrated in figure 3 above

Once the area has been discretized the construction of a mathematical model to describe the processes being investigated is undertaken This mathematical model is unique to the process being simulated, similar processes having simi- Tar expressions An example is looking for the head (measure of water pres- sure) distribution in an area, where only boundary values of the head are known Ilusirated in Figure 4, or a rock under applied stress in Figure 5

34

In areas of interest where

large differences in pa- rameters are expected the

mesh is made finer

Mesh made up of several finite elements

described by the nodes and the coordinates of

‘the nodes in the natural coordinate system

Figure 3: Dicretization /meshing of area to be modeled

(Reproduce from McDermott C.I., 2003)

Sea level 20m above sea level

all other heads are not known

Figure 4: Boundary conditions and discetization of a simple model for

groundwater flow (from McDermott C.I., 2003)

Figure 5: Boundary conditions and discretization for a simple column model

(from McDermott CI, 2003)

26

Principle behind finite element calculations

Tho principle behind the application of the FR (Finite Element) technique is that everything is broken down into matrices representing the governing equa- lions, which are then solved for the unknown values The principle equation

given is

Ku=f (Eq 8) Where u contains some variable related to f through the matrix K In ground- water K would be the Conductance Matrix, u would be the heads and f would

be the flux How to get the values for K is discussed later, but in principle that

can see from above that

so if Cis known, u can be solved, or if u is known, f can be solved The difli-

culties start in understanding how K, u and f are constructed For this stage let

us take a look at Figure 6 Let u be the displacement, arxl f be the forces in-

volved, The finite clement mesh approximation is shown in Figure 7 In the

mesh, there are 8 nodes, from 0 to 7, and in this case 7 elements

‘Applied

TÐưếc displacement >

Figure 7: Mesh in details (from McDermott C.F, 2003)

Assigning a simple coordinate system, where the units of x and y are meters

and the nede 3 is ai the posilion 0,0 Then defining all the nodes of this mesh:

and the elements as illustrated in table 1

Table 1: Nodal Coordinates

28

2, by assuming a unit thickness of z the necessity to include variations in

the z direction is removed

The force vector f and the movement vector u are then composed of an x and

ay component for every nade:

3.3 Lysimeter

3.3.1 General information about lysimeter

‘The terra lysimeter is a combination of the Greek words “tusis” = solution

and “metron” = measure and the original aim of lysimeter is to measure soil

leaching (Lanthaler C., 2004, p 37, quoted from Muller J.C., 1996, p 9), Ac-

cording to (Lanthaler C., 2004, p 39), lysimeters are also used for determining actual evapotranspiration and proundwater recharge and therefore for setting

up a water balance Dus lo an increase of pollution and conlaminaiion of

groundwater, the original sonse of lysimetors gained more and more impor- tance in the last decades and not only quantitative but also qualitative aspects

predommate

According lo (Kutileck M & Nielsen D.R., 1994, p 215) the explanation and

the use of lysimeters are extended as following,

- soil is hydrologically isolated from the surrounding soil,

- lysimeters are containers filled with disturbed (= artificially filled) or un-

disturbed bare soil or soil covered with natural or cultivated vegetation,

- seepage water is measured directly, vertical water movement is also to be

determined,

- porcolating water is collected cither gravimetrically (_ gravilalion lysime- ter) or through suotion cups‘a suction plate with a negative soil water pressure head, identical to that in the field next to the lysimeter (= suction lysimeter),

- anartificial groundwater level can be simulated,

- lysimeters are either weighable or non-weighable; weighable lysimeters provide information about the change of water storage W for any time pe-

riod; non-weighable lysimelers collec ouly the water percolating from

the soil column

According to (Lanthaler C., 2004, p.40) there are several criteria thai can be

applied to classify the lysimeter as following:

30

size: small (< 0.5 mê), standard (0.5-1 m?), large (> 1 m2)

weighability: weighable, non-weighable

soil filling method: disturbed (backfilled) or undisturbed (mono- lith/monolithic lysimeter)

if groundwater occurs: groundwater lysimeter with a variable or invari- able groundwater level; lysimeter without groundwater contact and with

or without applied vacuum

vegetation: bare soil, grassland, arable land, forest

soil fractions: sandy, silty, clayey soil

Figure 8, Figure 9, Figure 10 and Figure 11 below show an overview about lysimeters

Figure 8: Pressing of the stainless steel bottom plate (left) and lifting of a rea-

dily filled monolithic lysimeter (right)

Figure 9: Lysimeter covered with grass (left), the round surface of the Lysime-

ter (middle) and lysimeter cellar with complete instrument (right)

(from http:/lysimeter.com)

Figure 11: Cross-section of a guideline lysimeter surrounded by a control plot

(from http:/Avww fe-juelich.de/icg/icg-4/index php? index=155 )

3.3.2 Juelich lysimeter station description

Juelich lysimeter station was established in 1980 It is located in Juelich, near

Cologne, Germany There are several types of lysimeter, more than 20 in total

10 lysimeters were used for the research project of German Ministry for Edu-

cation and Research, All lysimeters are weighable Their surfaces consist of

bare soil, grassland and arable land.

Figure 12: The lysimeter station in Munich-Neuherberg (from http:/Avww fe-juelich.de/icg/icg-4/index php? index=155 )

At the lysimeter station, the weather parameters are measured such as rainfall,

wind speed, temperature Figure 13 show the instruments for measuring the

wind speed and rainfall at the lysimeter station

Figure 13: The instrument for measuring the wind speed (right) and the rain-

{fall (left)at lysimeter station

3.3.3 Description of the Juelich lysimeter number 302

The Juclich lysimeter number 302 wa

This lysimeter is run by the Rescarch Centre in Juclich (FZJ)

Based on the classification of lysimeter types mentioned above (section 3.3.1),

Juelich lysimeter number 302 is a large undisturbed lysimeter with 2m’ in area

and 2,4m in depth In Figure 14, a simplified sketch of Juelich lysimeter num- ber 302 is presented together with boundary conditions of upper and lower edges as well as 3 layers of suction cups The three suction cups are installed

together with tensiometers, TDR and temperature sensors at 3 different depth

layers distance from upper edge of the lysimeter in turn as 0.85m; 1,15m and 18m

After the drying of the applied soil, the reference material soil was imple-

meted with a thickness of 47 em Above the reference maternal, a coarse sand

layer (33 cm) was implomented in ticrs of 10 to 15 om (compacted) until the

upper edge of the lysimeter On the coarse sand a further water tracer was ap-

phed (65% concentralcd deuterium oxide 1,0) and then sprinkled wilh Milh-

pore-water (Puetz ‘I et al., 2004) ‘The schematic composition and the ar-

rangement of measurement devices are presented in Figure 15

In order to match the amount of precipitation at the reference location Mu- nich-Neuherberg, the lysimeter was sprinkled with equivalent amounts of Mil- lipore-water at given times At the measurement layers the outflow or ex- tracted seepage water was sampled for a month and then analyzed

The lysimeter wasn’t saturated with water at the beginning of the experiment, Thorclore, a parl of the infiltrating water was slored for an up-saluration of the

soil matrix

3

Figure 14: Simplified sketch of the lysimeter and boundary conditions in the

upper, lower and 3 suction cup layers at lysimeter 302 in Juelich

(Puezt T et.al., 2004)

Figure 15: The schematic composition and the arrangement of measurement

devices (Puezt T et.al., 2004)

The scenario-based lysimeter set-up allows observation of its component's

behaviour under most natural conditions, Precipitation infiltrates the column through a covering layer of inert material (quartz gravel) in order to achieve

36

3.4

uniform permeation into the reference material At the bottom of the contami-

nated layer, suction cups are mounted, which extract pore water by vacuum

for the determination of the source term Additional suction cups at different,

depths of the transport zone can be added for the observation of concentration

fronts progressing to the bottom (Puetz T et al., 2004) An inert sand filter

with pradualed grain sizes helps 1o avoid negative pressure thal is otherwise

generated by a capillary fringe Trickling off the sand filter, the seepage water

is collected by a tank at the battom before it is analysed

Meteorological data are recorded to determine the amount of infiltrating water

and Lo compare with the amounl measured al the outlet, thus acquiring infor-

mation on the flix regime inside the lysimeter Additionally, lysimeters are weighable for the same purpose Any oover of vegetation on top of the lysime-

tor is continuously removed

‘Water flow model

The mathematical model behind SiWaPro, the modeling module of SiWaPro DSS, is based on the mathematical model used in SWMS_2D (Simunek J et al., 1994)

According to the (Blankenburg R et al., 2005, p 1) the flow mode! desoribing l-dimnensional vertical water flow in the unsaturated zone is given by the

where the independent variables are time t and spatial coordinate z The de-

pendent variables of equation 11 and 12 are the water pressure head hp and the

water content @ wy is the sink/source term.

A, the residual air content B, the scaling factor a and the slope parameter m

‘The unsaturated hydraulic conductivity k(0) depends on the water content in the soil The function of unsalurated hydraulic conductivily was movleled by

(Luckner L, et al., 1989) with:

The parameters of equation 14 are the hydraulic conductivily ke{@} at a known

degree of water mobility Sq =(8-A)/(@-B), the parameter and the transfor- mation parameter m

The parameters «, ky and @ must be estimated in advance using lab and/or

ficld Iesls The parameter 4 in the model may range botweon O<A<I, bul il is kept fixed at 2=0,5

Description of the finite element mesh of the lysimeter

As mentioned in section 3.3.3, the height of the Juelich lysimeter 302 is 2,4m including 0,8m of reference material In the water flow model, the reference

‘The lower boundary condition is a first kind boundary condition of equation (11), only outflow is allowed in this boundary The flix difference between

precipitation and evapotranspiration is applied as a second type boundary con-

dition at the upper boundary of the column of lysimeter ‘Ihe transient boun- dary condition with using time — variable boundary conditions is used in the model Three soil waler sampling device layers (SKE 1, SKE 2, SKE 3) are applied as first kind boundary condition and, as the lower boundary condition,

only outflow is allowed

The column of the lysimeter soil is divided into 5 layers; each of the soil layers

is described in its hydraulics with 6 paramoters and in $ van Genuchten para- metes ‘Their values were taken from the field soil description of (Puetz T et

al., 2004) and are listed in table 2

Table 2: Soi properties and van Genuchten Parameters using for Simulation

Conductivity (ko) 0,0143 | 00135 9.0135 [0.0135] 00135 Residual water content § 9 0 0 0