26
S. C. Lessof! and P. Indelman
with ~ being a weighting constant to normalize the variable lengths of the depth
intervals. In Eq. (14) Cmn(S) is the model computed average concentration in depth
interval n at time m and C(tm, zn) is the measured average concentration at that
time and depth, the expected variance of the sample mean is given by:
-2
2 IN
CTmn = CTmn
nm
(15)
where d mn is the sample variance and N nm the sample size.
~ ~--A--------------------------------------------~
"
, I
, I
f \
: \
: \
I
\
k "
I
I
fr /'-i!t.,.
: l
\ \.
I :
\ "\
I :
\
"' ..
: :
\
"'''+
I :
"
.......
, .
'x '"
I:
',
........... .
! /
' . . . . . . . . . . _ · . . . ~f· .... ··-.... -__ . ___ ._ .... _ ....
o
' :
---~.
o
Depth (em)
II<
Fig. 3. Mean bromide concentration at two times t=6 days and 1=17 days (50 and 100 mm
applied water) after tracer application. Symbols field data; curves best fit analytic results;
x, --------- t=6 days; +,
t=17 days
Based on measurements made in the vicinity of the experimental site (Russo
and Bresler 1981; Russo and Bouton 1992), we estimate the mean 0,., ~ to be 0.1
and 0.4, respectively, and the coefficient of variation of Or to be CV Or = 0.2. The
Ks distribution was adopted from Russo et al. (1997) as described above. Based on
drying curves measured by Russo and Bouton (1992), we estimate P=1I3 in Eq.
(2). Because transport practically ceases by 2 days after water application, the
model is very weakly sensitive to K .. p, and ~ and more sensitive to Or. This time
behavior of moisture content in the model qualitatively agrees with the field
observation that within less than 2 days the field moisture content reaches a
constant field capacity of near 13%.
The distribution of the volume of infiltrating water W was estimated by fitting
model simulations of conservative tracer transport to experimentally measured
bromide concentrations. Assuming a lognormal distribution of W, the best fit is
achieved with the mean IJw = 2.2 cm and the coefficient of variation CV w = 1.0.
These values compare favorably to the values estimated for the mean and
S. C. Lessof! and P. Indelman
with ~ being a weighting constant to normalize the variable lengths of the depth
intervals. In Eq. (14) Cmn(S) is the model computed average concentration in depth
interval n at time m and C(tm, zn) is the measured average concentration at that
time and depth, the expected variance of the sample mean is given by:
-2
2 IN
CTmn = CTmn
nm
(15)
where d mn is the sample variance and N nm the sample size.
~ ~--A--------------------------------------------~
"
, I
, I
f \
: \
: \
I
\
k "
I
I
fr /'-i!t.,.
: l
\ \.
I :
\ "\
I :
\
"' ..
: :
\
"'''+
I :
"
.......
, .
'x '"
I:
',
........... .
! /
' . . . . . . . . . . _ · . . . ~f· .... ··-.... -__ . ___ ._ .... _ ....
o
' :
---~.
o
Depth (em)
II<
Fig. 3. Mean bromide concentration at two times t=6 days and 1=17 days (50 and 100 mm
applied water) after tracer application. Symbols field data; curves best fit analytic results;
x, --------- t=6 days; +,
t=17 days
Based on measurements made in the vicinity of the experimental site (Russo
and Bresler 1981; Russo and Bouton 1992), we estimate the mean 0,., ~ to be 0.1
and 0.4, respectively, and the coefficient of variation of Or to be CV Or = 0.2. The
Ks distribution was adopted from Russo et al. (1997) as described above. Based on
drying curves measured by Russo and Bouton (1992), we estimate P=1I3 in Eq.
(2). Because transport practically ceases by 2 days after water application, the
model is very weakly sensitive to K .. p, and ~ and more sensitive to Or. This time
behavior of moisture content in the model qualitatively agrees with the field
observation that within less than 2 days the field moisture content reaches a
constant field capacity of near 13%.
The distribution of the volume of infiltrating water W was estimated by fitting
model simulations of conservative tracer transport to experimentally measured
bromide concentrations. Assuming a lognormal distribution of W, the best fit is
achieved with the mean IJw = 2.2 cm and the coefficient of variation CV w = 1.0.
These values compare favorably to the values estimated for the mean and