Simulation of an irrigation system for peat soils

A cIa Agrophysica, 2002, 68, 97-108

SIMULATION OF AN IRRIGATION SYSTEM FOR PEAT SOILS'

.1 l 2 3

D. Kowa/ski , W. O/szla ' , H. Zaradny

ILublin Technical University, Nadbystrzycka 40 str., 20-618 Lublin, Poland

2Institute for Land Reclamation and Grassland Farm;ng, Głęboka 29 str., 20-612 Lublin, Poland 3Hydro-enginecring Institutc ofthe Polish Acadcmy af Scicnccs

Kościerska 7 str., 80-953 Gdańsk, Poland

A b s t r a c t. This artic1e presents the use ol' selected numeric.:al models which provide progno-ses orwaler changes in 50il profile, in steady stale and dynamie conditions. Thc ex~mples prcsented concem the estimation af Ihe capillary rise range (model PODSIAK), the dynnmics ot' moisture changc in the profile af irrigated 50il (model SWATREZ) and the economics

ar

irrigation (model UGWTPN - IRRDEC), in peal soils conditions. Using the mathematical model methods describcd, the authors indicate significant facilities in comparison with traditional methods of land rcclamation improvement and the process orregulalioll design.

K e y wo r d s: soil physics, water Illovement simulation, irrigation, peat soils INTRODUCTION

Land reclamation is a rather expensive undertaking; pro per design is very im-portant. Potential investors of reclamation systems want to obtain a good design in as short a period of time as possible with a favorable prognosis for its functioning in the future. Traditional design methods do not give such possibilities. Probably the only solution is the mathematical model method, whieh allows the simulation of reclamation system work, in various soil and plant conditions.

Processes connected with water flow on soil-atmosphere and soi l-plant boundaries are among the most difficult to solve. Their external lim itations cannot be precisely determined in most cases, except as "potentials". For this purpose, many computer simulation program S exist which try to approximate reality.

98 D. KOWALSKI el al.

This paper presents basic equations for the description of wat~r tlow in soi!,

taking into account water lIptake by plants and real bOllndary eonditions. The

equations were lIsed lo compute moislllre dynamics in a Iayered soil profile con

-taining variable drain-pipe spaeing (program, SWATREZ) [3,9), Ihe capillary rise

range of waler in soil profile, in steady stage conditions (program PODSIAK). A

simulation of the predietion of Ihe time and the amount of water for irrigation (the subrolltine IRRDEC of UGWTPN program) is also presented.

Ali ealeulations presented below were don e based on the conditions at the

Stawek-Stoki peat valley [4,5)located near the Cily of Lublin and on the

experi-menlal station of the Institute for Land Reelamation and Grassland Farming in

Sosnowica [6,7), located also on peat.

J\PPLlED MODELS

To solve the assumed design problem as outlined in Ihe inlroduetion, Ihe

allthors used the following compuler program s: EVAPOT (9) for estimaling

po-tential evapo-transpiration, PODSIAK (9) to estimate soil capillary rise properties

and SWATREZ for the simulation ofthe dynamics ofmoislllre change in the soil

profile. The model UGWTPN (subroutine IRRDEC) was used for the estimation

ofthe dry mass yield ofplant and irrigation scheduling. [6,7).

Water now in soil with plant - watcr "ptakc

The physical concept of the SW A TREZ model is presented in Fig. I.

Dynamie water now in soils with growing plants was simulated by the

SWA-TREZ (3) program based on Richard's equation, supplemented by the souree

fae-tor S(Ii), whieh reneets the waler uptake by plant rools:

ali

=

_

I

_~

[k(h)(ali

+

1

)

]

_

S(Ii)

a

t

ecli) ilz oz C(Ii) (I)

where: li - soil water potenlial (cm), t - time [d], CCh) - differential water

eapa-city, K(h) - water conductivity (cm d-I),

~

- vel1ical coordinate [cm], S(Ii) - souree

faetor (transpiration etc.) (cm d-I).

The now from or to the drain (ditch) is calculated by formula [1):

q 1-_

1>1 -</>2

T

SIMULATION OF AN IRIGATION SYSTEM FOR PEAT SOILS

...

_.

k

.

,

-f - --- t .-.- - - - i'-! !nfi!trat!on

~

li

evapotransp!ration:

"-'-

Q

I

\

~

H

--

1

Gl

1

Do

j

-

~~~/

-

~

-'--'

1

-

03

I

D ą

1

I

,.

.

-

l - °2

D'

T

g : :.

lllto~w~w~a~t~er8ic~on~djljUi!ct~. !i8aYiśe~r ~ł~' ąSSb~=~ąSS1lll+~qlS2~illIlSlill8ill&li~W'

ą₂

Fig. 1. Physic<ll cOl1cept ofthe SWATREZ model

99

where: 1>! - water level in drain (ditch) (cm), 1>2 - groundwater level in the middlc ofdrainage spacing (cm), T - drainage resistance (d), calculated from the Ernst [2] form ula;

(3) where: L - drainage sr,acing (m), Ki - saturated water conductivity of i,layer, for

horizontal flow (cm d' ), D; - thickness of i-Iayer (m), IV - radial resistance (d m-'l,

calculated by:

w

=_I_ln CDo nK,. U_z

(4)

where: Do - thickness ofsaturated soillayer, below the water table in drain [m], C - constant, for homogenous layer C= 1.0, Kr - saturated water conductivity for ra-dial flow (cm d-!), U

z - drain wetted profi le (m). The flow from or to the lower layer is calculated as:

100 D. KOWALSKI e! al.

where:

rp3

-

average wal er level in the subject territory (cm),

rp4

-

piezometric groundwater level in layer with low water conductivity (cm), Co - vertical resis

-tance ofthat layer (d).

The solution of the Eq.( I) needs the definition of the series of boundary condi-tions, such as rainfall, inliltration, potential evaporation and transpiration, height of waler level in the drain or ditch and the groundwater level. It is also necessary

to estimate the parameters of objective soil environment.

To determine the upper boundary conditions, the authors used meteorological

data, comprising temperature, relative humidity, wind velocity, rainfall and actual insulation, collected for the Stawek-Stoki valley. The potential evaporation and transpiration values were calculated using program EVAPOT [4] based on the Monteith, Rijtema and Ritche formulas:

(6)

ES= 0.0352·ó R . -0.39LAI

_<; nel e

u +y (7)

.'"1 _I _ "') _!

where: EV - evapotranspiration stream (g m - s ), ES - evaporation stream (g m - s ),

Lllp - latent heat of vaporization (hPa), A - constant value, connected with vari-ous units calculations, 0 - Ihe first differential offunction called the curve ofsatu-rated vapor pressure, by air temperature, Pa - air density (kg m-3), R"e!

--"

radiation net, calculated using actual insulation (W 111

-l

,

cI' - water specific heat

(J kg-I K-I), e_ll - actual pressure of water vapor (hPa), ed - pressure of saturated water vapor (hPa), r_ll - di ffusion resistance of water vapor (s m -2), y - psychrol11c

-ter constant, LA! - index of leaf surface.

The lower boundary condition, the grOlllldwater level, was determined based on the assumed water level in drainage, using resistance theory [3] included in program S W A TR EZ

Additionally the authors used the PODSIAK [9] program based on the Darcy law, which make possible the sil11ulation of capillary rise range "=" (cm) of water in soil profile, in conditions of steady state output:

z=-13

dh

qj 1

- -+ (8)

SIMULATION OF AN IRIGATION SYSTEM FOR PEAT SOILS [Ol

where: qj - elementary llow intensity (cm d-I).

This model doesn't concem the dynamics of moisture change, but facililates

the estimation of soil properties, degradation degree and also the reslllts of the

cal-culations obtained which can be used as the first approximation of lower boundary conditions for the SWATREZ model.

Irrigation scheduling

Soil moisture, described as the function O(z,!) can change in time (I) and at dif-ferent depths (=), within a wide range of moisture, i.e., from critical moisture Ok, throllgh optimum moisture

0

0 to maXinllllTI water saturation of soil

O.w"

If the

value O is presented on the axis as below:

O

LI

_________

O

_

K

L

_ _ _ _ _ _ _ _ _

O

~

OI

L-

________

Os

~

a'l

then the prediction ar undertaking of a decision to carry out irrigation consists of maintaining, in the root zon e of the soil profile, the value of function O(z.t), in the range:

At point Ok> limited plant growth occurs. Thus this value will be the main crit e-rion for undertaking the decision to water. The value 00 is proposed as optimum

moisture, which corresponds to pF ; 2.0.

In improvement practices, the determination of the irrigation dosc (d_{n )} is re-duced in calcu lat ing the deficits between evaporation and rainwater supply and capillary rise, as well as soil water retention for a certain period of time. Then, knowing the ground water state and the moisture distribution in the soils profile, the amount of water necessary to carry out irrigation can be calculated, according to the simple scheme presented in Fig. 2.

fn conformity with the denotations in Fig. 28 the net dose (d_{n )} upon the area

unit is:

/I III

d_{n ;} 2:(0[; -0₂;)+ 2:(0.<0,-0₂₎

;=1 i=n+1

(9)

where: i - successive 5 cm thick layers, n - a num ber of 5 cm layers in the soil profile counted from the surface to a depth of hl' m - a num ber of layers in the soil profile counted from a depth ol' hl to _{171 (171} - the depth of water table in soil

[02 D. KOWALSKI el al. A B EVAPOTRANSPIRATION [TRRS (2JJ LAYEFl FLOW [1.11 Fl.OW (NU1( NU! FLOW( RCOT EXT.(2) TENPIIJ wc (I) NET FLOW II

Fig. 2. Physical concept ofthe soi I profile moi sten ing model UGWTPN

e

for e = (_2), e[ - optiTum ;vater content (cm3 cm-\ e₂ - water conte,nt at

~egin

­

ning ofirrigation (cm" cm-"),

e

_sat - water content under saturation (cm" cm-·'). The value of h[ is calculated on the basis of the function numerically

deter-mined [6]. The

e[

value is counted from the pF curves, whereas _{17 2} and 8₂, that is, the actual level of (he ground water and (he moistening of the soil profile being searched are given by model UGWTPN.

The whole process of computer simulation for the Stawek-Stoki peat valley (the PODSIAK and SWATREZ program s), had been based, except for the soil pa-rameters just presented, i.e., the upper and lower boundaries, on the following as-sumptions: the object will be equipped with a drainage system with a constant water level, irrigation water will be taken only from local sources (Stawek River)

and connected with it, irrigation can be done only once a week. Additionally, soil suction pressure, in the middle ofthe root strata (12.5 cm) should not fali below pF=2.7 for any period longer than 7 days. Simulations were done with climatic conditions noticed in 1994.

The simulations for the Sosnowica experimental object (sub-routine IRRDEC

ofUGWTPN program) takes into account the influence ofsoil moisIure variations

e(z,l) on plant growth. Another condition which decides for Ol' against irrigation is the economic evaluation ofthe effectiveness of irrigation consisting or calculating

the ratio of yield loss, caused by the drop in soil moisture, to the costs of Ihe pro-spective irrigation.

STMlJLATION OF AN TRIGATION SYSTEM FOR PEAT SOILS 103

RESULTS AND DISCUSSION

The simlllated investigation of the capillary rise of soil profiles demonstrates

the signiticant differences on the objective valley territory. The example ofthe in-vestigation reslllts are shown in Fig. 3. Profile 8 represents the peat zone at the Stawek River and the No. 10 - peat zone located about 300 m from that river.

,~, .1 \·10 o _I 120C 'Cv~

-.,

"

,

1 / /,1_~:_~.;:::,,~==--=-~Il

"

q [mm'ć[ \. q: o o 2. q" o:! 3 \1',,06 <\ q= l o 5. q: I 5 61;= 20 7 q" 3 o 8 (1= 4 o ~ tl"'5.0 10. (I'" 7.5 t 1,11'" 10 o , '1'11 ~s 1(1 ,~ : 0 ,~ 30 ,~ 40 .~ 50 55 ~o log Ih) leoO! I 1400 --j Il00

--j

1:)00

J

K

600

J

N

".

..

q (mili/d] / / 1.q=QO / ~.,"_ .... - 3 2. q= 03 / ~4 3.Q"'O.6 /I::_._,:.c;.:·~::';""",. 6 5 4 q" 1.0 /.Y/~- 75q;1.5 /j ... ___ .- - 8 6. q" 2.0 }'~;:- .• - ':"1~ 7 G:: 3.0 ,',f'" ~_____ 1\ 8. c.;= 4 O ,'- 9. q= 5 O 10 C;: 7.5 ',1. Q= 10 O ~---~--;r

'

('I~ 10 I~ 10 H lO ,~ <o ~~ 51) 5~ 5(1 log Ih)

Fig. 3. C;lpillary rise propcrlies lor Iwo different profiles ofSlawek-Sloki v;lllt!y, leli: prolile No. 8.

right: profile No. 10

The simulations performed show that holding the groundwBter level at a depth of 60 cm, satisfies water dem and for average, seasonal, "in vegetation"

evapo-transpiration. This level doesn't allow periodically higher evapo-transpiration

val-ues to be satisfied, which gives rise to significant moisture lack in soil profiles. Jt

is of the ut most importance to simulate the dynamics of moisture change in soil

profiles and take into consideration additional irrigation.

This simulation, for the 1994 vegetation season, was done using the

SWA-TREZ mode!. The investigation considered optimum drainage spacing. also the amount and height ol' additional periodical irrigation which stopped degrada-tion

processes and help plant growing. It assumed also that additional irrigation cannot

be more frequently than once a week.

Figure 4 shows an example ofsimulation results for profile No. 10, including

evapo-transpiration and rainfal!.

Figure 5 presents the comparison of simlllation results contained in soil water

potential changes at a depth of 12.5 cm, for additional irrigation, made for dra

104 D. KOWALSKI el al.

. .---r-,---~--_ ... ,

'?

.

Period of moisturo lacking

Fig. 4. Example ofsimulation results done by the SWATREZ program lor prolile No. 10

Based on the simulation performed for ali the Stawek-Stoki valley profi\es

in-vestigated, the authors could propose the following opti 111 um parameters to the

de-signers for a reclamation system: groundwater level in drainage - 80 cm depth,

drainage spacing - 24 m, height ofadditional irrigation - 30 mm.

The determination ofthe water demands connected with the realization ofthe

proposed water relations was done in three stages. At first, the irrigation height

was determined, connected with lifting the groundwater level to the optimum 80

cm. Next the water demand for a territory chosen earlier was calculated, as result

SIMULATION OF AN IRIGATION SYSTEM FOR rEAT SOILS

Ol

-500

~

irr. dosage 30 mm -1000

-~

-1500 I 100 120 140 160 180 ;:00 2;:0 240 260 280 300 320 . rrr----i" [l-

I

t i r I l O -~ -1000 -1-'

E

-2000-.!:!. -3000

'"

+:: -4000

~

-5000 l c: !li

8.

o

l

~ -5000 -l !li

~

·10000

_.~

;: l o -15000

J

J'\

r-\ ,\

:

\1

'J \

I

I ., I

\1

_.,

-\

\

,

\.,

, ...

en

-20000 - - - "

l --i

--

~-I O -~ _l __ ' \ ' -- \ -5000 j' \

! \

·10000 - ' - - J ; --I -15000 -

j

-20000

r---,-

--

-'

---\

_,

I _\ " ~-\/ " irr. dosage 25 mm .-....• irr. dosage 20 mm r--,-'~-'-\

-I

.L irr. dosage 15 mm i .. _ - watcr Icval 60 cm water level 80 cm i: ,-80 100 120 140 160 1,-80 200 220 240 260 280 300 320 Simulation day

[

IV--,,

__

J~~r~11I _' IX

-C

x

-105

Fig. S. The comparison

or

simulation resu[ts contained the soil ""[Iter potential changes at a deplh ol'

12.5 cm, for various addilional irrigation levels, made IQI' drainngc sp"cing 8 m. The assumed water

Icvel in drainage 60 and 80 cm depth. Example ofprofilt: No. 10

Solution for long-Iasting drought periods and irrigation prcdiction

The calculalion results oblained using Ihe IRRDEC model for Ihe Sosnowica object are presented in Fig. 6A for the drought period. The following parameters

are given: a - rainfall distribulion, b - dynamics of soi I moistllre lension al layers 5-10,25-30, and 55-60 cm depth -lines 1,2,3 (respeclively) for Ihe initial stale of

106 D. KOWALSKI el ol.

for the state of the ground water table at H = 60 cm (line 4), c - dynamics of the

groundwater states.

Significant differences between the suction of the upper layer (5-10 cm) and

the deeper ones we re observed. II proves that capillary rise cannot prevent the

vio-lent drying of the root zone and therefore, suction at a depth of 5-10 cm reached

600 cm of the water column for the initial H = 60 cm (Iine 4) on July 5, that is af-ter 27 days of rainless weather, whereas that for the depth of H = 80 cm was, in

the same day, 1500 cm of the water column. The differences are equal to about

1000 cm. and lasted till the end of the simulation period, i.e., until August 5. The

difference ofthe initial (instantaneous) states ofground wat er (part c in Fig. 6B) at

the beginning ofthe simulation process reached 20 cm, while at the end

ofsimula-tion is decreased to zero.

The results of calculations during the 24 h growth of the dry mass of hay and

quantities of water used for transpiration, obtained from simulation, are presented

in Fig. 6d. At the initial state of H = 80 (curve I), the mean value of the 24 h

in-crease of the dry mass of hay (APG) reached about 100 (kg ha-I d-I) at the end of

the calculation period August 3rd.

Similarlyas for the former solution, the calculation was carried out with appli-cation of capillary rise irrigation (Fig. 6B). 1t can be seen that for the ground water state at H = 80 cm five irrigation were carried out, whereas for the H = 60 only four were carried out during simulation. II should also be stressed that for the H = 80 cm the first irrigation was done on June 28, and that the H=60 cm on July 23 (i.e., 25 days later).

The results of the calculations on grass growth are presented in Fig. 6Bd,

where line I presents the 24 h increase ofthe dry mass ofhay with the H = 80 cm.,

the dashed line represents the mean value of the 24 h increase for the second and

the third swathe obtained from the experimental field of the Institute of

Meliora-tion and Grassland Farming at Sosnowica, from the climatic condition noticed in

1983.

The prognosis of water relations for long-Iasting periods of drought (in this

case as an inward experiment optimally determined) allows the economie solution

of many important practical problem s eonnected wit h irrigation.

CONCLUSJONS

The paper presents an example of computer simulation methods for the esti-mation of improvements and regulations in water relations compared to traditional

SIMULATION OF AN lR1GAT10N SYSTEM FOR PEAT SOILS 107 A MONTH B MONTH

"

..

10 ) D

."

,on

"

""

,on

'''''

- -StMULATION "o. ___ MEASUR.ED ,o.

"

170 20 ~

non \.YiELoIOrH~8~m ---"-"f-~'I-\

4· Y1ELO lor H260 cm 1"1-. EWlPOTAANSPIAAlIOtl for H=8G r.m'" 4l)O~ ~ ~, EVAPOTRANSl'lAATION 10( H=60 cm

"

40 20

Fig. 6. Rninfall (n), dynamie of soil moisture lensian (b), ground water depth (e) [Ind dl)' mass

ar

yicJd and transpiration (d): A (len) - for a long-Iasting period of drought, B (right) - for a

long-bst-ing period ol' drought unde!" irrigation (N)

methods. The determination of the optimum parameters of a reelamation system

are done by simulating the efTeet or their exploitation in assumed meteorologieal and

plant conditions. Before the simulation process it was necessary to estimate the

parame-ters oflhe soil's environment. From Ihat point ofview, the necessity ofjoinl

eollabora-tion between praetiee designers and seientists is elearly visible.

Of eourse it is very neeessary to contillUe the investigation in order to ereate

and verify new, more satisfied models. The authors of the paper are now working

on field experimental objects for verification of the present models. The next

108 I. 3. 4. 5. 6. 7. 8. 9. D. KOWALSKI el al. REFERENCES

Bclmans

c., 'Vc

sscling J.G., Feddcs R.A.: Simulatioll model ol' Ihl! wale!' balance ol' cropped

soil providing difTerent types of boundary conditions (SWATRE). Nota 1257 leW, Wngeningen,

Holl,nd. 1981.

Ernst L.F.: Grondwatcrstromingen in de ,'cr.wdige zone en hun berckcning bij aanwczighcid

van horizonlale open leidingcn. Versl. Landbouwk. Onderz., 67.15. PUDOC, 1962.

Feddcs RA., Kowalik p" Zaradny H.: Simulation of field wate!' lIse and erop yield. John

Wiley and sons, New York·Toronto, 1978.

Kowalski O.: Physico-water properties of organie soils dried by groundwnter inlake in

Wier-zchowiska. Zesz. Prob1. Post. Nauk RoJn .. 419. 53-58, 1995.

Olszta \V.: Characteristics of soil eQver of the "Stawck-Stoki" o~ject considcring their actual

degradation dcgrce and rccommcndations for their futllrc usc (in Polish). UnpubJished study,

Lublin. 1994.

Olsztu W.: Predicling irrigation using the rnathemalical modeling mcthod. Zesz. Probl. Post.

N.uk Roln .• 281, 163-176. 1982.

Olszt:ł W.: The investigation ol' soil moisture dynamics, grass growth and irrigation prediction, by mathematical Illodeling method (in Polish). Wyd. Falenly, 1981.

Z:,rmlny H.: A method for dimensioning subsurface drainnge in he:wy soils cOllsidering the

reduction in potentinl transpiration. ProG. lnl. Sern. on Land Orainage. 258-266. '-Il'lsinki,

1986.

Zaradny H.: Mathematical modeling ol' water and pollution transport in saturated (lnd

unsalu-rated soils for irrigation necessities (in Polish). Internal report !BW PAN for

CPBR.1O.8. 7.I.B.12.03. projecl, Gd.ńsk, 1990.

SYMULACJA SYSTEMÓW NAWADNIAJĄCYCH W WARUNKACH GLEB TORFOWYCH

l I ., 3

D. Kowa/ski, IV. O/s=la '-, H Zaradl1y

!Politechnika Lubelska. ul. Nadhystrzycb 40, 20-618 Lublin, Polska 21nstytut Melioracji i Użytków Zielonych, ul. Glęboka 29, 20-612 Lublin, Polsko

31nstytut Budownictwa Wodnego PAN, ul. Kościcrska 7. 80-953 GdaJlsk, Polska

S t r e s z c z e n i e. W artykule zaprczcntow<lno zastosow:lnic wybr<lnych modeli numery-cznych umożliwiających prognozowanie stosunków wodnych w profilu glebowym, Z<1.równo w usl<I-lanych. jak i niclIstalonych warunkach. PI-Ledslawionc przykłady obcjmowały wyznaczenic zasięgu wzniosu knpilarnego (model PODSJAK), dynamiki zminn uwilgotnienia nawadnianego profilu

gle-bowego (model SWATREZ) oraz ekonomiki nawodnień (model UGWTPN - IRROEC), w

warunkach gleb torfowych. Autorzy. wykorzystując opisnne metody modelowe, wskazują na

znaczne ułatwienie prac projektowych związanych 7. melioracją obszarów torfowych w stosunku do

metod tradycyjnych.