Beruflich Dokumente
Kultur Dokumente
IDV
by
.A
0
N0
78-WR-08
September 1978
Z
ACKNOWLEDGNEWNT
iii
LIST OF FIGURES
Fig. 1. Typical plot of the soil moisture retention curve based on Eq.
(3).
v
LIST OF TABLES
vii
Corey (1964) and Jeppson (1974) each used an analytical expression for
and Corey (1964, 1966) obtained fairly accurate predictions with their
curve at some negative value of the pressure head (this point is often
lems. It also appears that predictions based on the Brooks and Corey
equations are somewhat less accurate than those obtained with various
Recently Mualem (1976a) derived a new model for predicting the hydrau-
lic conductivity from knowledge of the soil moisture retention curve and the
sions. The theories of both Mualem and Burdine are used for this deriva-
dent parameters which may be obtained from the soil moisture retention
are discussed in this paper, a simple graphical method which enables one
2
MAT.EMATICAL DEVELOPMENT
ing the relative hydraulic conductivity (K.) from knowledge of the soil
K 1 dx / 1
(x) (1)
where h-h(G) is the pressure head, given here as a function of the dimen-
8ser
e -e (2)
s r
m
(3 M 1 (3)
[l+(ah)n]
4
6
- t0i
Ew
0..
LU
LU
U)
cc
co
0 .1 2 .3 .4 .5
MOISTURE CONTENT, 8 (cm3/cm3)
Fig. 1. Typical plot of the soil moisture retention curve based on
Eq. (3).
6
10
z
0
a:
0~~~n=.
> ~ ~ RESR
PRESSURE
0.0 (EA/cm)m
HEAD (cm)m
8
es either er or es . Note that the diffusivity becomes infinite when 6 ap-
mately between e-0.25 and e-0.45 in Fig. 3) does the diffusivity acquire
D(6) a BP (12)
(8 ~ 8 )q
for other integer values of k. For k-l, for example, the conductivity
becomes
While this particular model is not only more complicated than model (8),
Hence, (13) does not present an attractive alternative for (8), and will
Similar results as above for the Mualem theory may also be obtained
2 '1 d / 1
r h 2(x)
JQ~~ 0 h2(x)
10
The soil moisture diffusivity for this case is given by
most cases, in better agreement with experimental data than (19). Through
dictions based on his theory (i.e., based directly on Eq. (1) by means of
brief discussion of the equations derived by Brooks and Corey (1964) will
be given here, since their model of the soil moisture retention curve
this study.
Brooks and Corey (1964a 1966) concluded from comparisons with a large
number of experimental data that the soil moisture retention curve @(h)
(3), one sees that (3) reduces to (22) for large values of the pres-
12
.
-le~
-I0
0a 0%
6P )1
U U
2
'-&
&Rb
-I0 0
101I
0 .1 .2 .3 .4 .5
Fig. 4. Comparison of the proposed soil hydraulic functions (solid lines) with curves obtained by
applying either the Mualem theory (M; dashed lines) or the Burdine theory (BD dashed-dotted
lines) to the Brooks and Corey model of the soil moisture retention curve.
PARAMETER ESTIMATION
e-e +^ (es-er)
e - er + r (28)
[i +(h)4
Mualem model
m -1-1/n. (29)
able, the following procedure can then be used to obtain estimates of the
16
Combining (32), (33), and (34b) leads to the following expression for S
The best location on the 6(h) curve for evaluating the slope S is about
halfway between er and eS. Let P be the point on the soil moisture re-
tention curve for which (-' (see Fig. 5). From Eq. (2) and (31) it
ep (s 4 r)/2 (36a)
h - 1 ( 2 1/m 1) (36b)
n/(l-n)
SP(n) - 1.151 (n-1)(1-2 ). (37b)
18
.
0
1 2 3 4 5
-n
a Fig. 6. Plots of the dimensionless slopes Sp and SQ as functions of the parameters n and
m (=1-1/n).
m
SQ -[l~m] * (42)
Q e r+ (es -e r) [1 (43a)
(
h 1 m- (43b)
easy to determine this point accurately (even less so when the curve is
mate m from the slope of the curve. Substitution of (42) into (35) gives
SCm)
Q
r m]
2.303
-rn Ll- J4a (44a)
or, in terms of n,
2
-1/n
SQ(n) - 2.303 n .44b)
Figure 6 shows that Sp(n) and S Q(n) define approximately the same curve,
especially for the larger n-values. This is not surprising since the
points P and Q are generally very close together on the soil moisture re-
tention curve. Fig. 5, furthermore, shows that both points define approxi-
mately the same gradient. Hence the n-values obtained from the sketched
22
1 m-i
h m
a24 '0
- 10 2.43(0.49)- 0.0053.
1 1 9
Fir
DO 0.50 12
22.000.
Krr() - 9 L -( -- ] (based on S ) (46a)
Kr (9)- e
½ r2.o4
Ll-(1-4 )
0.49
] 2
(based on S ). (46b)
Equation (46a) exactly reproduces the conductivity equation one would have
obtained if the original data shown in Fig. 5 were used in Eq. (8). Equa-
tions (46a) and (46b) generate nearly the same curve when plotted versus
or versus h. Minor differences between the curves occur only at the extreme
dry side of the curves, and are caused by the fact that the same slope was
used to calculate both Sp and S (in reality, SQ should have been measured
retention data which are plotted on a normal e versus h scale. The pro-
cedure for finding the two parameters is similar to that used before.
Equation (37) still holds provided, however, that S is calculated with Eq.
(33) and (34). These two equations show that now estimates of both h
and the slope, de/dh, are necessary for evaluating S. Equations (43)
and (44), on the other hand, have to be modified because the inflection
point of the e(h)-curve does not coincide with the inflection point of the
24
INFLUENCE OF THE RESIDUAL MOISTURE CONTENT
is defined as the moisture content for which the gradient (da/dh) becomes
zero (excluding the region near 8s which has also a zero gradient). Also
content at some large negative value of the pressure head, e.g., at -10 6
In fact, they would be inconsistent with the general shape of the 8(h)-
predictions.
The open circles represent data points of the curve, and were taken from
26
Three different values for 8 were chosen rather arbitrarily (0.05, 0.10,
was assumed to be the same for all three cases (step 6 in Table 1), a
r r r
curve c (based on 6 ) fits the data points somewhat better at the dry
r
end of the curve than the other two. On the other hand, this curve also
28
100
0. 0 \
00
z
-J
10 0
10
30
I
HYGIENE SANDSTONE
U
E
U
a e .0079
n * 10.4
O x .153
E - *S o f PREDICTED-.l
U
E
U
a .00505 104_
no 7.09
.1_ r.I90l
0 I I a I *
h (cm) h (cm)
Fig. 10. Observed (open circles) and calculated curves (solid lines) of the soil hydraulic properties
of Touchet Silt Loam G.E.3. The relative hydraulic conductivity was predicted from knowledge
of the curve-fitted soil moisture retention curve.
program also provides for a correlation matrix between the different
with n but much less than with a, and that a and n are nearly inde-
pendent of each other. Some of these effects are also noticeable from
Beit Netofa Clay (Rawitz, 1965), however, were found to be much less
accurate (Fig. 12). The higher conductivity values are seriously under-
predicted, and also the general shape of the predicted curve is consider-
ably different from the observed one. It seems that much of the poor
the observed soil moisture retention data. For example, the residual
since clay soils have generally higher e.-values than coarser soils
Limited data at the lower moisture contents further increases doubt about
curve shows that the gradient of the curve changes fairly suddenly at
inconsistent with the general shape of curves based on (28). With some
curve at a pressure head of about -2,000 cm. The observed curve should
have become flatter from that point on if equation (28) were to describe
e the break in the slope of the curve at h-10,000 cm, and the presence
36
of an inflection point at h--2,000 cm, an attempt was made to improve the
the dry side of the curve. Fig. 13 shows that the soil moisture retention
curve is now much better described (with the obvious exception of the
last four data points). Also the description of the conductivity curve
the right of the observed one. The example shows that by deleting only
four points at the dry end of the curve a completing different value of
Results for Guelph Loam (Elrick and Bowman, 1964) are given in
in the soil moisture retention curve. The observed data of this example
were taken directly from the original study (Figs. 2 and 3 of Elrick and
0.434 for the moisture content was used, being the highest measured value.
Also the wetting branch of the hydraulic conductivity curve was matched
to the highest value of Xr measured during wetting (Fig. 14). The value
of 6r, furthermore, was assumed to be the same for drying and wetting,
and was obtained from the drying branch of the curve. Both the drying
and wetting branches of the soil moisture retention curve are adequately
described, even though the predicted curves are slightly below the observed
38
GUELPH LOAM
Kr 0
O I .0115
* 2.03 ~WETTING
2
I0
Ci ~~~~~~~~~~~~~~~DRYING
Li nu 2.76
cb. ~~~~~~~~~~4
Eo
*32 A' . 3 4.
9, .218 1
0 0
-10O -io2 -1o0. .3 .4 .5
h (cm) 0 (cm 3 /cm3 )
Fig. 14. Observed (circles) and calculated curves (solid lines) of the soil hydraulic properties of
Guelph Loam. The drying and wetting branches of the relative hydraulic conductivity curve were
predicted from knowledge of the curve-fitted branches of the soil moisture retention curve.
REFERENCES
Brooks, R.H., and A.T. Corey. 1964. Hydraulic properties of porous media.
Hydrology Paper No. 3, Civil Engineering Dept., Colorado State
University, Fort Collins, Colorado.
Brooks, R.H., and A.T. Corey. 1966. Properties of porous media affecting
fluid flow. J. Irrig. Drain. Div., Am. Soc. Civil Eng. 92(IR2):
61-88.
Daniel, C., and F.S. Wood. 1973. Fitting equations to data. Wiley-Inter-
science, New York. 350 pp.
Endelman, F.J., G.E.P. Box, J.R. Boyle, R.R. Hughes, D.R. Keeney, M.L. Northup,
and P.G. Saffigna. 1974. The mathematical modeling of soil-water-
nitrogen phenomena. Oak Ridge National Laboratory. EDFB-IBP-74-8.
66 pp.
Elrick, D.E., and D.H. Bowman. 1964. Note on an improved apparatus for
soil moisture flow measurements. Soil Sci. Soc. Am. Proc. 28(3):
450-453.
42
APPRENDIX A
SOHYP:
44
program only calculates best-fit values of a and n, and assumes that er
is known beforehand. The value of er is now given as an input variable
Eq. (Al) and (A2) (i.e. the Mualem theory still applies). If MODE
equals three, the computer model again calculates best-fit values of the
three parameters (dr a, and n), but it is now assumed that the Burdine
theory applies. Hence Eq. (Al) and (A3) are now used in the program. In
each case the computer program provides for a table of the hydraulic
properties of the soil (see Table A4), consistent with the value of MODE
selected.
46
TABLE Al (CONTINUED):
VARIABLE DEFINZTION
STOPCR.
be positive).
48
Table A3. Input data for example 3 (Silt Loam G.E.3).
1 2 3 4 5
Column: 12345678901234567890123456789012345678901234567890
Card
1 1
2 SILT LOAM G.E.3
3 1 3 13 0 0.18 0.396 4.96
4 0.180 0.002 2.3
5 WCR ALPHA N
6 10.0 0.396
7 20.0 0.394
8 43.0 0.390
9 60.0 0.3855
10 80.0 0.379
U1 111.0 0.370
12 190.0 0.340
13 285.0 0.300
.14 400.0 0.260
15 600.0 0.220
16 800.0 0.200
17 900.0 0.194
18 1000.0 0.190
50
.. .. . .
*
* NCN-LINEAR LEAST SQUARES ANALYSIS
* SILT LOAN G.E.3 *
*
*********.***********************.* **************.*************..***.**.**..*.**..
INPUT PARAMETERS
MODEL NXE..................
NUMBER OF COEFFICIENTS........................... 3
NUMBER OF OBSERVATIONS..*.....****.*******....... 13
RESIDUAL MOISTURE CONTENT IFCR MODEL 24....... 0.1800
SATURATED MOISTURE CCNTENT.................... 0.3960
SATURATED HYDRAULIC CONDUCTIVITY.............. 4.9600
OBSERVED CAI
L'
OBS. NO. PRESSURE HEAD MOISTURE CONTENT
1 10.cO 0.3960
2 20.00 0.3940
3 43.00 0.3900
4 60.00 0.3855
5 60.00 0.3790
6 111.00 0.3700
7 190.00 0.3400
8 285.00 0.3000
9 400.00 0.2600
10 60O.0O 0.2200
11 800.00 0.2000
12 900.00 0. 1940
13 1000.00 0.1900
*. .
I 'I
56
MAIN
IF(NOVE.NE.2) WRITE(6,10261 NlT,(lJ,8b(2),Bf3),SSQMOOE
C
C -- BEGIN OF ITERATION -
34 NITTNIT+l
GAwO. *GA
00 38 J1,NP
TEMPPTH(J)
THilJul.Ol*TH(J)
Q(Jj=O
CALL MODEL(THDELZ(1,JlNOBXWCStODENPWCRI
DO 36 K-lpNG9
DELZ( IJlxELZ(lJ)-F(I3
36 g(J)-QlJ)+OELz(IJ)*R(IJ
-QJJ3lOO**C(J)/TH(J)
C
C -- STEEPEST CESCENT
38 TH(J)=TEMP
DO 44 I*INP
00 42 ju1,1
00 40 KslN08
40 SUP-SUM+OELZ(XI)*0ELZ(KJJ
D(IJ.Il0000.*SUM/(TH(I)*TH(J)l
42 O(JIJ-D(IJ-
C
C 0 a MOMENT MATRIX -
44 E(I)SQRT(O(IIl))
50 00 52 Iul,NP
00 52 J1,NP
52 A(IJ)-O(IJl/(E(I3*EIJ3)
C
C - A IS THE SCALED MOMENT MATRIX -
DO 54 Iu1,NP
P(IJ-Q(IJ/ECIII
PHI(It)P(l)
54 A(II)=AItl)+G
CALL 14ATINV(ANPPI
C
C -- P/E IS THE CORRECTION VECTOR
STEPl1.O
56 DO 58 Iu1,NP
58 T8(ImP(I)*STEP/E(I).TH(I3
DO 62 IJ1,NP
tF(TH(II T8(1J )66,66,6Z
62 CONTINUE
SUMBuO.O
CALL MOOEL(TB,FNO8,XWCSIODEtNPGCRI
00 64 Iu,NCB
R1I-Jn(IJ-F(II
64 SUMeUSUMB+RhI)*R(II
66 SUMlmOmO
SUM2aO.0
SUP3u0.O
00 68 IZlNP
58
rw
MAIN
ISEC-TVAR*SECOEF
TMCOQTHI I)-TSEC
TPCCE*TH( ZITSEC
K02*I
J-K-1
108 WRITE(6910581 StlJ) 5IlK),TH(I 1ISECOEF.TVALUEtTMCDE.TPCOE
C
C --- PREPARE FINAL OUTPUT --
LSCRT( 11-I
00 116 J=2#NCB
TEMP=RtJ)
K=J-1
DO Ill LalK
LULLSORT(L)
IF(TEMP-R(LL)i 1129112oll
111 CCNTINUE
LS"RT(J)=J
GO TO 116
112 KK-J
113 KK-KK-1
LSCRTCKK+IJ=LSORT(KKI
IF(KK-L) 115,L15,113
115 LSCRT(L)uJ
116 CONTINUE
WRITE(6 1066)
00 118 1=19NOB
JaLS0RT( NOB+1-I)
118 WRITE(61068) IXCI)tY(I13F(tJR(I),JJX(JIY(J),F(J)IR(J)
C
C ---- WRITE SOIL HYDRAULIC PROPERTIES -
WR!TE(69 1069)
PRESSa1. 18850
RN I-O 0*
RKLN-.10
WR!TE(6,lC72) RNLWCSRKLNeSArK
00 140 1a1,75
IF(RKLN.LT.(-16.)) GO T0 142
PRESS- 1. 1885o*PRESS
IF(MOOE-21 1209 122,120
120 hCR-TH(I)
ALPHA.TH( 2)
RN*TH13)
GO TQ 124
122 ALPHAwTH(11
RN*TH(2)
124 RHMl.-l./RN
IF(MODE.EQ.3) RMul.-2./RN
RN 1zRM*RN
RWC-1./( 1.+(ALPI4A*PRESSI**RN)**RM
WCuwCR+giCS-hCR)*Rwc
TERM*U.-RWC*(ALPHLSPRESS)**RNI
IF((TERM.LT.S.S-O5).OR.(RWC.LT.0.061) TERM RM*RWC**(.I/RM)
TF(MQUE.EC.3) RKwRWC*RWC*TERM
IFIMoO0.NE.*3 RKaSGRT(RWCI)*TEM*TERN
60
MATINV
SUBROUTINE MATINV(AiNPBi
DIMENSION A(3,3),8(3lt1NDEXl3t2J
00 2 J-1,4
2 INDEX(J,1)u0
ItO
4 AMAX--1.O
DO 10 JmINP
IF(INDEXIJ9 II 10,o,10
6 DO 10 KuwlNP
IF(INOEX(KIJI i0td810
J PuABSlAlJiK)I
IF(P.LE.AMAXJ GO TO 10
IR=J
IC-K
AMAX=P
10 CONTINUE
IF(AKAX) 30,30,14
14 INDEX(ICtlJlIR
IF(IR.EQ.IC) GO TO 18
DO 16 Lm1NP
P=uAIR,LJ
A(IRtL)0AlICtL)
16 A(IC#L)=P
P=BIRJ
B(IRl-B11C)
81IC)UP
INDEXII,2)1IC
18 P-1./AllCqIC)
AC ICtIC)L.0
DO 20 L*19NP
20 A(ICtLJUA(ICLI*P
l11C)nB(lC)'P
O0 24 KAlNP
IFlK.EQoIC) GO TO 24
P-AIK9 IC
A(KvIC)m0.0
00 22 LauINP
22 A(KtLJOA(K#LI-A( ICL)SP
B(K)nB(Ki-d(IC)*P
24 CONTINUE
GO TO 4
26 ICwINOEXlIt2J
IR-INDEXI IC.l)
00 28 KuLNP
P-AK, IRI
A(KitRIA(KtIC)
28 AIK9IC)aP
1-1-1
30 IFli) 26,32.26
32 RETURN
END
62