Field-Measured Limits of Soil Water Availability as Related to Laboratory-Measured Properties1

Abstract

To make a unified and broad assessment of the accuracy of laboratory measurements for estimating field soil water, a comprehensive data base of field-measured upper and lower limits of the soil water reservoir was obtained and evaluated.-from Authors

Full text

Field-Measured Limits
of
Soil
Water Availability
as
Related
to
Laboratory-Measured
Properties
1
L.
F.
RATLIFF,
J.
T.
RITCHIE,
AND
D.
K.
CASSEL
2
ABSTRACT
Accurate
evaluation
of the
soil water reserves available
for
plant
use
is
vital
in
developing optimum water management
for
crop production
in
marginally
dry
regions. Laboratory estimates
of the
upper
and
lower
limits
of
soil water availability used
to
calculate
the
soil water reservoir
often
deviate significantly
from
the
limits measured
in the
field.
To
make
a
unified
and
broad assessment
of the
accuracy
of
laboratory
measure-
ments
for
estimating
field
soil water,
we
obtained
and
evaluated
a
com-
prehensive
data base
of
field-measured
upper
and
lower limits
of the
soil
water
reservoir.
The
field-measured
upper
limit
was
taken
as the
water
content
at
which
drainage from
a
prewetted
soil
had
practically
ceased.
The
lower
limit
was
taken
as the
water content
of the
soil
at
which
plants
were practically dead
or
dormant
as a
result
of the
soil
water
deficit.
These field-measured limits were compared
to
laboratory mea-
surements
at
—0.33
and
—15
bar
made
on
samples
removed
from
each
field
site.
A
total
of 401
observations
were
available
for the
comparisons
of —15 bar
measurements
to the
field-measured lower limits
and 282
observations
of
—0.33
bar
measurements were available
for
comparison
with
the
field-measured upper limit. Variation often existed
within
a
soil
series
at a
particular site
for the
field-measured
upper
and
lower limits.
However,
the
differences
between
the
field-measured
limits,
the
total
available
water reservoir,
were
relatively constant. Crop species caused
only
minor
differences
in the
lower limit water content
for the
upper
part
of the
soil
profile
where root length density
was
apparently above
some critical limit. However, some annuals extracted water
to
greater
depths
than others.
The
laboratory estimates
of the
upper limit obtained
by
—0.33
bar
water contents were significantly
less
than
the
field-mea-
sured drained upper
limit
for
sands, sandy loams,
and
sandy
clay
loams
and
were significantly more than
field
measurements
for
silt loams,
silty
clay
loams,
and
silty clays.
The
laboratory estimates
of the
lower limit
obtained
by
—15
bar
water content measurements
were
significantly less
than
field
lower
limit
measurements
for
sands, silt
loams,
and
sandy clay
loams
and
significantly more than
field
observations
for
loams,
silty
clays,
and
clays. Because
our
study included relatively
few
measurements
of
loamy sands, silts, sandy
clays,
and
clays,
it was
difficult
to
generalize
about
differences
in
field-measured
and
laboratory-estimated water lim-
its
for those textures. The results suggest that if absolute accuracy is
necessary
in
water balance
calculations,
laboratory-estimated soil water
limits
should
be
used with caution
and
field-measured
limits,
if
available,
would
be
preferred.
Additional
Index
Words:
soil water reservoir,
matric
potential,
in-
situ
water limits.
Ratliff,
L.F.,
J.T.
Ritchie,
and
O.K.
Cassel.
1983.
A
survey
of
field-
measured limits
of
soil water availability
and
related laboratory-mea-
sured properties. Soil
Sci.
Soc.
Am. J.
47:770-775.
A
GRICULTURAL
WATER
PROBLEMS
are
related
to
both
weather
and to the
reserves
of
soil water available
to
plants.
The
dynamics
of
water
in the
soil
are
related
to the
drainage
process,
the
capacity
of the
reservoir,
its
depletion and replenishment, and its
efficient
manage-
ment
for
agricultural production.
Accurate
calculation
of
the
soil water
balance
is
becoming increasingly
im-
1
Contribution
from
the
USDA,
SCS,
and
ARS
in
cooperation with
the
Texas
Agric.
Exp.
Stn.,
Texas
A&M
Univ. Received
22
June 1982.
Approved
14
Feb. 1983.
2
Soil
Scientist,
USDA-SCS,
and
Soil Scientist,
USDA-ARS,
Tem-
ple,
TX
76503;
and
Professor
of
Soil
Science,
North
Carolina
State
University, Raleigh,
NC
27650
(formerly Visiting
Scientist,
USDA-ARS,
Temple,
TX
76503).
portant
because
of the
need
to
manage water
as
effi-
ciently
as
possible.
Evaluation
of the
capacity
of the
soil water reservoir
requires knowledge
of its
upper
and
lower limits
in the
plant root zone.
The
most common procedure
for
esti-
mating
the
upper limit water content
is to
extract water
from
a
disturbed
or
undisturbed soil sample using
a
soil
water extraction apparatus
or
"pressure
chamber"
(Ri-
chards and Weaver, 1943). A matric potential of
—0.33
bar is
used
for
moderately coarse-
and
finer-textured
soils
whereas
a
—0.10
bar
potential
is
used
for
coarse-textured
soils
(Jamison
and
Kroth,
1958;
Colman,
1947).
The
lower
limit
water content
is
also estimated using
the
pressure
chamber
at a
matric
potential
of
—
15
bar.
The
soil
water
reservoir
for a
soil
profile
is
estimated
by
collecting soil
samples from
the
different
soil horizons
or
depths,
de-
termining
the
water content
at the
upper
and
lower limits
for
each horizon,
and
summing
the
differences over
the
entire rooting depth.
Laboratory methods for estimating the soil water res-
ervoir
have been criticized (Richards, 1960; Gardner,
1966;
Ritchie,
1981).
It has
been argued
that
some plants
remove water
from
the
soil
at
matric potentials
<
—
15
bar. Other plants
may not
remove water
to a
matric
po-
tential
of —15
bar.
Few
field-measured values
of the
matric potential
at the
lower limit have been reported.
For the
upper limit,
field
measurements often
do not
agree
well
with those values estimated using
the
—0.10
and
—0.33
bar
pressure apparatus
in the
laboratory.
Esti-
mates
of the
upper limit made
by
using
the
pressure
chamber
for
different
depths
of a
single soil
profile
may
overestimate
in-situ
measurements
at
some depths,
un-
derestimate
it at
others,
and be
nearly equal
to it at
still
others
(Cassel
and
Sweeney,
1974).
Because
of the
problems encountered
in
estimating
the
soil water
reservoir,
we
assembled
a
comprehensive
data
base
of
upper
and
lower limits
of
soil water availability
measured in the
field
for a broad range of soils through-
out
the
United
States.
Our
purposes were
to
make
a
broad
assessment
of the
value
of
laboratory measurements
for
estimating
field
soil water limits
and to
determine
if al-
ternative techniques might
be
needed
for the
accurate
evaluation of the soil water reservoir. In
this
paper we
summarize these field-measured upper
and
lower limits
and
compare them
to the
—0.33
and
—15
bar
laboratory
determinations.
In a
companion paper
we
report equa-
tions
for
estimating
the
potential upper
and
lower water
limits
of
soils based
on
routinely measured soil physical
and
chemical properties.
PROCEDURES
Soil
Selection
Process
To
develop
a
data
base
encompassing
a
broad
range
of
soils
with
respect
to
texture
and
other
chemical
and
physical
prop-
erties,
both
published
and
unpublished
data
meeting
certain
criteria
were
collected,
summarized,
and
tabulated.
Initially
a
literature
review
was
conducted
to
locate
published
data
on
upper
and
lower
limits
measured
in
situ.
The
literature
review
was
followed
by a
survey
of
about
250
researchers
who
were
conducting
or had
recently
conducted research that
in-
770
Published July, 1983

RATLIFF
ET
AL.:
FIELD-MEASURED LIMITS
OF
SOIL WATER AVAILABILITY
771
eluded
field
measurements
of
soil water content under various
crops.
Questionnaires were sent
to
researchers
identified
during
the
literature search
and
also
to
researchers
at
state
and
federal
institutions having research programs
in
soil physics
or
soil water
management. The questionnaire was designed to
identify
those
studies which
met the
following criteria:
(i)
the
crops growing
on
the soil in question had undergone severe water stress,
(ii)
the
soil water content
had
been measured throughout
the
root-
ing
zone periodically during
the
stress period,
and
(iii)
the
water
content measurement sites could
be
precisely located. Appli-
cable data were
found
from
28 respondents who agreed to con-
tribute
to the
survey.
After
identifying
the
soils
to be
included
in
the data base, the senior author visited all sites, discussed
the
in-situ
measured water content data
with
the researcher,
described
or
helped describe
the
soil
at the
site where
the
data
were
collected,
and
collected soil samples.
At one
location
the
soils
had
previously been described
and
sampled
by
individuals
experienced
in
soil classification. These soil samples
had
been
submitted to the same laboratory being used in this study and
the
resulting analyses were included in the data base. Eighteen
months
were required
to
assemble
the
data base. During
the
study, several other sets
of
water limit data were
identified.
However, the data and soil properties were either similar to
those
of
soils
already included
in the
data base
or the
cost
of
obtaining
a
single data
set
from
one
location
was
prohibitive.
Methods
for
Defining
the
Soil
Water
Limits
The
methods used
to
define
the
in-situ upper
and
lower limits
of
the
soil water reservoir available
to
plants were similar
to
those described
by
Franzmeier
et
al.
(1973)
and
Ritchie
(1981).
Slight modifications were required
to
accommodate
the
various
experimental approaches used by investigators throughout the
United
States.
Comparing
the
methods presented below
with
the
above references
will
show
the
differences.
To
maintain uniformity,
we
defined
the
water limits
to be
investigated before accumulating
the
data base
as: (i)
drained
upper limit
(DUL)—the
highest field-measured water content
of
a soil
after
it had been thoroughly wetted and allowed to
drain
until drainage became practically negligible; (ii) lower
limit
(LOL)—the
lowest field-measured water content
of a
soil
after
plants
had
stopped extracting water
and
were
at or
near
premature death
or
became dormant
as a
result
of
water stress;
(iii) potential
extractable
soil water
(PLEXW)—the
difference
in
water content between
DUL and
LOL. These three param-
eters—DUL,
LOL, and
PLEXW—are
expressed in percent by
volume.
The DUL for a
particular soil
was
derived
from
analysis
of
successive measurements
of
soil water content with depth after
the
soil
had
been thoroughly wetted
and
allowed
to
drain. Suc-
cessive measurements of such a thoroughly wetted soil exhibit
a
monotonic
decrease
in
soil water with time until
the
drainage
rate
becomes negligible.
The
soil
profile
was
considered
to
attain
a
negligible drainage
rate
and to
reach
the DUL
when
the
water
content
decrease
was
about
0.1 to
0.2% water content
per
day.
Soils
with
a
water table shallower than
200 cm at the
time
DUL was
measured were excluded. Some
soil
sites were covered
with
rainfall shelters
or
plastic sheeting which prevented evap-
oration losses or precipitation gains of water. Other plots were
uncovered
and
were subjected
to the
above gains
and
losses.
Typically,
2 to 12
d
were required
for
soils
to
reach
the
DUL.
Some
fine-textured
soils
and
soils with restrictive layers
re-
quired
up to 20 d of
drainage.
The LOL was
derived
from
successive measurements
of
soil
water content with depth during
a
period when
a
field
crop
was
subjected
to
severe water
stress.
Water content measurements
were continued until
the
plant
died,
nearly
died,
or
became
dormant.
Data
from
adequately fertilized
field
plots
in
which
plants
had
reached maximum vegetative growth before
undergoing severe water
stress
were preferentially
selected
over
data
from
plots inadequately fertilized
or
early season
stressed.
The
definitions
and
methods
of
selecting
the DUL and LOL
were
designed to
identify
the limits of the soil water reservoir
and do not
address water that
can be
taken
up by
plants while
drainage
is
occurring
(Ritchie,
1981).
In
addition, evaporative
losses of soil water
from
the soil surface or
from
near soil sur-
face
layers
of
uncovered plots result
in an
underestimation
of
DUL. Similarly, soil evaporation causes
an
underestimation
of
LOL for
layers near
the
soil surface. Also, there
is a
rooting
depth below which root density
is
inadequate
for
complete
ex-
traction
of
available soil water,
and
this
would
cause
the
water
content
at the LOL to be
overestimated.
The
above problems
were
recognized before compiling
the
data base; hence,
the
fol-
lowing
procedures were used
to
minimize underestimation
of
DUL and LOL and
overestimation
of
LOL.
All
possible values
of
LOL, DUL,
and
PLEXW were plotted
vs.
depth
for
each
soil
profile.
Possible
LOL and DUL
values near
the
soil surface
that appeared
to be
affected
by
soil evaporation
and
those
that
appeared
to
have inadequate root density
and
hence, incomplete
water
extraction, were
identified
and
omitted
in the
comparison
of
field-measured
and
laboratory-estimated water limits
and in
subsequent data analysis.
Two
procedures
for
measuring soil water content used
by the
various
investigators were gravimetric sampling
and
neutron
attenuation. Inherent
in the
data
set are
errors associated
with
the
variation
in
techniques used
by the
investigators providing
the
data.
This sampling error could
not be
removed
from
the
data.
Additional
Soil
Measurements
At
each location,
the
soil
was
described
and
sampled
as
close
as
possible
to the
point
at
which
the
soil water content
was
measured
when
DUL and LOL
were being determined. About
3 to 5 kg of
disturbed soil material
and
duplicate
5 cm
thick
and
7 cm
diameter undisturbed soil cores were collected
at
depth increments
that
coincided with
the
depth
of
water mea-
surement
and/or soil horizon.
All
samples were shipped
to the
National
Soil
Survey Laboratory, Lincoln,
Nebr.,
for
analysis
by
procedures described in Soil Survey Investigations Report
no.
1
(SCS,
1972).
Percent
sand, silt,
and
clay were determined
by
pipette analysis.
The
water content
at
—0.33
bar was de-
termined
with
the
pressure extractor, using
1-cm
thick slices
of
the
undisturbed soil
cores.
Disturbed soil samples were used
for
the
—15
bar
determination.
The
water contents obtained
at
—0.33
and
—15
bar
were expressed
in
percent
by
volume.
RESULTS
AND
DISCUSSION
The
geographical
distribution
and
number
of
soils
at
each
location
in the
data
base
are
shown
in
Fig.
1. The
data
will
eventually
be
published
in the
Soil
Survey
In-
vestigative
Report
series.
Seven
soil
orders
are
repre-
sented
in the
data
base,
but
over
60% of the
soils
are
Mollisols
or
Alfisols.
Histosols,
Oxisols,
and
Spodosols
are not
represented.
It
would
have
been
desirable
to in-
clude
more
soils
from
the
humid
temperate
midwest,
northeast,
and
southeast
regions
of the
United
States.
More
data
from
these
regions
were
not
included
because
additional
data
meeting
the
aforementioned
criteria
could
not
be
located.
The
fact
that
these
regions
experience
frequent
precipitation
events
during
the
crop
growing
season
precludes
the
casual
collection
of
field-measured
LOL
data.
Variation
of
DUL, LOL,
and
PLEXW
for
Morphologically
Similar
Soils
Before
analyzing
the
amassed
data
in its
entirety,
we
examined
the
uniformity
of
measured
DUL, LOL,
and
PLEXW
values
for a
given
soil
series.
Data
were
avail-

772
SOIL
SCI.
SOC.
AM.
J.,
VOL.
47,
1983
Fig.
1—Geographical
distribution
and
number
of
soils
at
each location
in the
data base.
able
from
several locations which could
be
used
for
this
task. For one of the locations,
DUL
and
LOL
were mea-
sured
in the
center
of
each
of 18
adjoining plots
on a
fine-loamy,
mixed,
hyperthermic
Typic
Camborthid
which
is
a
Variant
of the
Avondale
series. Graphs
of DUL and
LOL vs.
depth showed that
the
18
sets
of
soil water con-
tent measurements could
be
grouped into three closely
related,
but
different
soil water content
profiles.
In
Fig.
2 we
show DUL, LOL,
and
PLEXW
for one
represent-
ative
profile
from
each
of the
three groups. Field obser-
vation
of the
soils showed that they were morphologically
similar except
for
detectable
differences
in
clay content
with
depth. Percent clay
and
sand determined
in the
lab-
oratory
for the
three
profiles
are
shown
in
Table
1.
The
lowest
DUL and LOL values below the depth of 40 cm
were
for
profile
C
which
had the
lowest clay content.
Profiles
A and
B
had a
similar clay content between
0
u
20
40
E
o
1
60
X
£L
80
g
J
100
8
120
140
160
CAB
"
^
l!
-
•
*o
/I
//
-
••>
»e>
n
/i
-
•
CP
•
0*
1
'1
'I
In
i\
-
LOL
•
O
•
»o
•
DUL
U
/
M
-
.
o\
/
U
\
\
i\
\ \
•
O
•
"O
•
-
.Vuw
,
.
1
4
|
<
|
CAB
••O
a
F-
h
<l-
/
•9
_
/VPLEXW
•
&
r/,
"
0
10
20 30 40 0
10
30
WATER
CONTENT
-
VOLUME PERCENT
Fig.
2—Representative
field-measured
DUL,
LOL,
and
PLEXW
for
three
pedons
of the
same
soil
series.
and
150
cm,
thus suggesting that some factor other than
clay
content
influenced
DUL and
LOL. When PLEXW
was
computed,
the
three
profiles held nearly
identical
amounts
of
extractable
water. Thus,
for
this
field,
minor
variations
in
soil properties give rise
to
different
DUL
and LOL
values
but the
amount
of
extractable water
is
relatively
constant.
Crop
Effects
DUL
should
not
vary with
the
crop,
but LOL may be
crop dependent.
To
determine
the
effect
of
crop type upon
LOL, both crops would have to be grown on the same
soil
at the same time. Two crops may be grown in close
proximity,
but
because
of the
nonuniformity
that
exists
within
one
soil series
as
discussed above, such data must
be
carefully interpreted. Fortunately,
LOL
determina-
tions made
on the
same plots
for
different
crops
in
dif-
ferent
years allow
an
objective evaluation
of the
effect
of
crop type
on LOL and
PLEXW.
The DUL, LOL, and PLEXW values in Fig. 3 are for
wheat
(Triticum
aestivum
L.)
and
sunflower
(Helian-
thus
annuus
L.)
growing
on a
well-drained clayey Pull-
man
soil
(fine,
mixed, thermic
Torrertic
Paleustolls)
with
a
pronounced calcic horizon
at a
depth
of
112
cm.
Mea-
surements
were
made
on the
same
site
6
years
apart.
The
two
crops extracted about
the
same amount
of
water
to
Table
1—Percentages
of
clay
and
sand
for
three
pedons
of
the
same
soil
series.
Depth
0-15t
15-50
50-130
130-150
150-170
Clay
20.5
22.9
23.5
31.5
39.8
Sand
37.2
34.3
36.0
39.0
33.4
Clay
20.5
20.5
22.9
31.5
29.4
Sand
'o
——————
38.9
38.5
32.8
33.3
23.9
Clay
21.0
20.3
18.6
17.5
19.1
Sand
38.0
37.4
44.5
54.3
50.6
t
Upper
and
lower boundary layers
are
within
± 4 cm
of
depths
listed.

RATLIFF
ET
AL.:
FIELD-MEASURED LIMITS
OF
SOIL WATER
AVAILABILITY
773
v
20
40
e
60
o
1
80
X
a.
100
Ul
Q
_j
120
0
w
140
160
180
•
i
i i
t
X
\
:
/
/
I/
/
•
:
/
/
'
LOL
<B
*OUL
-
0
•
•
x
\
/
\\
/
0
•
•
-
SUNFLOWER->il
/
]
.<*
•
1
!
i
i
•O
I/
?/"<-WHEAT
-
1
/
i
1
I
-
-
-
-
• 0
/
Af-SUNFLOWER
/ /
—
• O
'
ll
PLEXW
*>
I I
0
10
20 30 400
10
20
WATER
CONTENT
-
VOLUME
PERCENT
Fig.
3—
Field-measured
values
of
DUL,
LOL,
and
PLEXW
for
crops
grown
on the
same
soil.
-
3<
two
a
depth
of
about
110
cm.
Below this depth
the
wheat
root density
was
apparently inadequate
for
complete water
extraction,
whereas
the
sunflower
roots were able
to
pen-
etrate
the
calcic horizon
and
obtain more water
to a
depth
>
180 cm.
Grain sorghum [Sorghum
bicolor
(L.)
Moench]
grown
on a
nearby plot
with
almost identical
soil properties extracted about
1.5
volume percent less
water
than wheat
for all
depth increments. Total water
removed
by
evapotranspiration
from
a
210-cm
deep soil
profile
was
24.6, 22.9,
and
20.1
cm for
sunflower,
wheat,
and
grain sorghum, respectively.
An
evaluation
of all
data
collected
in
this study sug-
gests that
the
effect
of
crop type
on LOL and
PLEXW
for
the
same soil
is not
large among many annual crops,
particularly
in the
upper portion
of the
soil
profile
where
root density
is
high.
The
major
difference
observed
is the
apparent ability
of
some annuals
to
extract water
at
greater
depths.
Although
evidence
is not
conclusive,
the
observed trend
is for
annual taproot systems
to
extract
water
from deeper
in the
soil than
fibrous
root systems
and for
perennials
to
extract water deeper than annuals.
Evidence also exists
to
support
the
idea suggested
by
Franzmeier
et
al.
(1973) that perennials
may
extract
slightly
more water than annuals
at all
depths.
The ap-
parent differences observed
for the
effect
of
crop type
on
the
water limits
may be
related
to
genetic, climatic,
or
soil
factors or to experimental errors associated
with
the
soil
water content measurements.
Texture Effects
on
DUL, LOL,
and
PLEXW
The
field-measured values
of
DUL, LOL,
and
PLEXW
for
four
soils representing
a
wide range
in
soil texture
are
presented
in
Fig.
4. The
four
soils were deep,
well
drained
or
excessively drained,
and had no
root restrict-
ing
layers
in the
upper
1
m.
Soil
A was
classified
as a
fine,
mixed, thermic
Torrertic
Paleustoll.
Texture ranged
from
silty
clay
to
silty
clay loam
in the
upper
64 cm and
was
clay loam from
64 to 200 cm. A
pronounced calcic
horizon
that
might partially restrict rooting occurred
at
112
cm.
Soil
B, a
fine-loamy, mixed,
hyperthermic
Typic
Camborthid,
had
loam texture
to 127 cm and
clay loam
to
2 m. Soil
C
had silt loam texture
from
0 to 2 m and
\j
20
40
o
,
60
X
!L
80
Ul
o
_i
100
o
(O
120
140
160
_
•
-
•
-i
-«
-i
j
1
•
1
1
i
D
C B A
•D
•
•
O
•
n
\
/
?
\
VI
\
/
1
\
*O
•
•
0
•
F!
\l
I
I
»O
to O
•
!
1!
i
/
»
o
•
o
•
/
/
/
/
1
I
/
lO
••
O
•
\
1)
\
1
\
1;
l
/
>
0
••
O
•
1
v
»
io
*>
K
i\
/i
l
•
•
•
m o
•
o
t
•
D
C
AB
• *
•>
.
•
BO
I//
L
•
mo
-I
<
-
-
i
I!
too
y
'
/I
1
•
mo
III
'
tt
PLEXW
U
,
-1
10
!>0
30 400
10
20
WATER
CONTENT
-
VOLUME PERCENT
30
Fig.
4—
Field-measured
values
of
DUL, LOL,
and
PLEXW
for
four
soils
having
different
textures.
DUL and LOL
values
are
plotted
on
the
left side
of
the
figure.
DUL
values
are
identified
as
soils
A, B, C,
and
D.
For a
given
set of
symbols,
the
line
on the
left
represents
the
LOL.
was
classified
as a
fine-silty,
mixed
Argic
Cryoboroll.
Soil
D
was a
thermic,
coated
Typic
Quartzipsamment
with
a
texture
of
fine
sand
or
sand throughout
the
2-m
depth.
Corn
(Zea
mays
L.) was
cropped
on
soil
D
whereas wheat
was
cropped
on the
others.
The
soil water content
at the DUL for the
four soils
was
highly correlated with soil texture. Soil
A had the
highest
DUL and the
highest average clay content
throughout
the
profile;
its
clay content reached
a
maxi-
mum
at a
depth
of 38 cm and
then gradually decreased
with
depth. Soil B had an intermediate clay content
that
remained
nearly constant
to
about
120 cm and
then
in-
creased with depth.
The DUL for
soil
B was
less than
that
for
soil
A
except
at the
135-cm
depth. Soil
C had
a
lower clay, but a higher silt content than soils A and
B.
The
clay
content
of
soil
C
reached
a
maximum
at 25
cm and gradually decreased with depth. This gradual
decrease
in
clay with depth
is
reflected
in the
gradually
decreasing DUL. Soil
D,
which
had a
relatively
uniform
and
low
(<
5%)
clay content,
had a low
DUL.
The
soil water content
at the LOL for the
four
soils
was
also highly correlated with texture.
For
soils
A, B,
and
C, the
respective
DUL and LOL
curves
are
nearly
parallel
from
the 20- to
120-cm
depth (Fig.
4). The DUL
and
LOL
curves
for
soil
D are
nearly parallel
to the
1
50-
cm
depth.
For
soil
B the LOL
curve below
120 cm in-
creases sharply compared to DUL, thus indicating in-
adequate extraction
of
soil water
to the
true LOL. This
behavior
exists for soils A and C also, but to a lesser
degree.
The
differences
in
depth
of
water extraction
to
the
true
LOL
between soil
D and the
others cannot
be
conclusively
explained.
Extractable
water
for the
four
soils
is
also shown
in
Fig.
4. The
patterns
and
amounts
of
water extracted
by
wheat
from
soils A, B, and C are very similar. For ex-
ample, between depths
of 30 to 120 cm an
average
of
12.8,
14.8,
and
1
1.9
volume percent water
was
extracted
from
soils
A, B, and C,
respectively.
In
contrast,
an av-
erage of
7.1%
was
extracted
from
soil D over the
same
depth.

774
SOIL
SCI.
SOC.
AM.
J.,
VOL.
47,
1983
100
o
fr-
ill
g
20
40 60
PERCENT
SAND
80
100
Fig.
5—Textural
distribution
of the 401
observations
in the
data base.
In
Fig.
5 we
show
the
textural distribution
of the 401
observations available
for
comparison
of the
laboratory-
estimated water limits with
the
field-measured limits.
Some samples that
had
nearly identical textures appear
as
single points
on the
graph.
The
number
of
samples
and the
observed range
of
sand, silt,
and
clay
for
each
textural
class
are
presented
in
Table
2. All
textural classes
were
well
represented except
for
sandy clay, silt,
and
clay.
The
mean
and
standard deviation
for
DUL,
LOL,
and
PLEXW
for
each textural class
are
also shown
in
Table
2.
Values
of the DUL
range
from
a
minimum
of
11.8
±
4.9%
for
sand
to a
maximum
of
35.0
±
6.2%
for
silty
clay.
The LOL
ranged
from
a
minimum
of 3.8 ±
2.2%
for
sand
to a
maximum
of
21.9
± 1.0 for
clay (based
on
only
three
observations).
The
mean
and
standard deviation
for the
—0.33
and
—
15
bar
determinations
for
each textural class
are in-
cluded
in
Table
2. The
—0.33
bar
determination over-
estimates
by
2.0%
or
more
the
DULs
for
silt loams, clay
loams,
silty clays,
and
clays; underestimates
by
2.0%
or
more
the
DULs
for
sands, loamy sands, sandy loams,
and
sandy
clay loams;
and is
within
±
2.0%
for
loams
and
silty
clay loams. It is recognized that a better laboratory
procedure
to
estimate
the DUL for
sands
and
loamy sands
would
be
—0.10
bar;
however, incomplete data
for the
—0.10
bar
value precluded such
a
comparison.
The
—15
bar
determination overestimates
by
1.0%
or
more
the
LOL for
loams, silty clays,
and
clays; underestimates
by
1.0%
or
more
the LOL for
loamy sands, silt loams,
and
sandy
clay loams;
and
estimates within
±
1.0%
the LOL
for
sands, sandy loams, silty clay loams,
and
clay loams.
In
general,
the
standard deviations
for the
—0.33
and
—
15
bar
determinations
are
less than those
for the
cor-
responding
DUL and LOL
determinations.
The
higher
standard deviations
for the
field-measured values
are
thought
to be
attributable primarily
to
errors associated
with
the
field
measurements
of
water obtained
by
differ-
ent
techniques
and
different
personnel.
The
mean
and
standard deviation
for
PLEXW,
which
is
equal to DUL minus LOL are also
shown
in
Table
2
and
are
plotted
as a
function
of
soil textural class
in
Fig.
6. The
values range
from
a
minimum
of 8.0 ±
3.1%
for
sands
to
14.8%
for
just
one
observation
for the
silt.
The
second highest value
is
14.3
±
3.3%
for
silt loam.
The
sand,
as
expected,
has the
least
PLEXW
because
the
large pores
in
sandy soils drain easily
and
rapidly under
field
conditions; moreover,
the
particle surface
area
is
low,
resulting
in the
presence
of
little
adsorbed water
at
the
LOL.
The mean
PLEXW
values for the remaining
textural classes are relatively constant with a range of
only
11.0
to
14.8%.
The
associated standard deviations
range
from
2.1 to
3.6%.
The
values support
the
com-
monly
held concept that plant available water increases
with
fineness
of
texture
up to
silt loam
but
suggests that
the
amount
of
increase
is not
large.
The
water retention
difference
(WRD)
defined
in Ta-
ble 2 as
—0.33
bar
minus
—15
bar,
ranges
from
a
min-
imum
of 5.6 ± 1.9% for sand to a maximum of
18.6
±
3.1%
for
silt loam.
Silt
has
been omitted
from
the
dis-
cussion
because only
one
observation
was
available. Com-
parison
of WRD
with
PLEXW
reveals that
WRD
over-
estimated
by
1.0%
or
more
the
observed
PLEXW
for
silt
loams, silty clay loams,
and
clay loams; underestimated
by
1.0%
or
more
PLEXW
for
sands, loamy sands, sandy
loams,
and
loams;
and
estimated
PLEXW
to
within
±1.0%
for
sandy clay loams, silty clays,
and
clays.
For
each of the textural
classes
except silt loam and silt, the
mean
WRD was
within
one
standard deviation
of the
mean
PLEXW.
Table
2
—
Texture
and
water retention data
by
textural class
for
the 401
observations.
Tex-
ture
s
Is
si
1
sil
si
sicl
cl
scl
sc
sic
c
No.
samples
76
7
31
51
83
1
53
41
24
0
31
3
Soil separate
Sand
————
Weif
87.4-97.5
73.7-88.3
53.1-83.3
29.0-49.4
0.9-25.4
2.2
0.9-18.8
20.0-44.6
47.4-72.7
1.2-15.1
5.8-20.0
Silt
*ht
percent
<
0.8-
8.5
3.4-23.5
2.8-30.7
29.7-47.1
53.6-84.8
86.4
44.0-71.8
25.3-46.2
6.6-26.5
40.7-55.2
38.9-39.8
Clay
2mm
————
1.2-
7.7
2.8-12.6
4.4-19.3
8.9-26.9
13.1-27.0
11.4
27.0-39.9
27.2-38.3
20.7-30.7
40.2-52.1
41.1-54.4
Upper
limit
DUL
-0.33
bar
Lower
limit
LOL
-15
bar
(DUL-LOL)
WRD
(-0.33bar-
-15
bar)
—————————————————————
Volume
percent
—————————————————————
11.8
± 4.9
18.9
± 6.0
23.7
± 5.4
25.0
± 5.1
29.0
± 7.0
32.3
33.8
±3.5
30.9
± 4.5
29.0
± 3.6
35.0
± 6.2
34.8
± 2.9
8.9 ± 2.2
16.0
± 5.3
21.4
± 5.5
25.2
± 3.9
31.6
± 4.1
36.1
34.9
± 2.8
33.0
± 4.4
26.3
± 3.3
37.3
± 3.3
39.3
±1.0
3.8 ± 2.2
5.9 ± 4.0
10.5
± 5.2
11.4
± 4.5
14.7
± 5.9
17.5
20.8
± 3.4
18.4
± 4.9
18.0
± 5.2
21.5
± 6.8
21.9
± 1.0
3.3
± 1.3
4.4
± 2.3
9.9
± 2.0
13.8
± 4.0
13.0
± 2.3
6.9
20.8
± 2.6
19.2
± 3.8
15.0
± 2.7
24.1
± 5.4
27.0
±
1.0
8.0
±
3.1
12.9
±
3.6
13.2
±
2.2
13.6
±
3.0
14.3
±
3.3
14.8
13.0
± 2.1
12.5
±
3.2
11.0
±
3.5
13.4
± 3.0
12.9
± 3.6
5.6
±
1.9
11.6
± 3.3
11.5
± 3.9
11.4
±
3.3
18.6
± 3.1
25.4
14.1
± 3.6
13.8
±
4.2
11.3
± 2.4
13.2
± 3.4
12.3
± 1.3

RATLIFF
ET
AL.:
FIELD-MEASURED
LIMITS
OF
SOIL
WATER
AVAILABILITY
775
eu
H
z
ID
£
15
S
111
1
I0
o
>-ti
I
*
.
x
5
ui
_l
a.
o
_
-
i
~
<
[/'
I
1
1
)•——
<
1
1
—
•
I
1^^
"N
i
i
l^^
\
1
—
<
_
-
"
Table
3—Results
of
t-test
for
paired
comparison
between
field-measured
and
laboratory-estimated
soil
water
limits
for
each
textural
class.
Is
si
I
sil
si
sicl
cl
scl
sc
sic
c
SOIL
TEXTURE
Fig.
6—Field-measured
PLEXW
as a
function
of
soil
textural
class.
To
determine
if the
field-measured limits were
signif-
icantly
different
from
the
laboratory-estimated limits,
a
t
statistic
was
calculated
for the
following
comparisons
for
each textural class:
DUL
vs.
—0.33
bar;
LOL
vs.
-15
bar;
and
PLEXW
vs.
WRD. Results
of
these anal-
yses
are
shown
in
Table
3.
Examination
of the
table shows
that
one or
more comparisons were
significantly
different
at the
0.10
level, usually
at the
0.05 level,
for all
textural
classes
except loamy
sands
and
clay
loams. However,
the
PLEXW
and WRD
values were
significantly
different
only
for
sands, loams, silt loams,
and
silty
clay loams.
The
mean
DUL and
—0.33
bar
values reported
in
Table
2
suggests that better agreement between the
field-mea-
sured
and
laboratory-estimated upper water limits
can
be
expected
by
using
matric
potentials
>
—0.33
bar for
soils
with
sandy textures
and
matric potentials
<
—0.33
bar for
soils
with
silty textures. Similarly, better agree-
ment
between
the
field-measured
and
laboratory-esti-
mated
lower water limits
can be
expected
by
using matric
potentials
>
—15
bar for
sands, silt loams,
and
sandy
clay
loams
and
matric potentials
<
—15
bar for
loams,
silty
clays,
and
clays. From
our
data,
it is not
possible
to
determine what alternative potentials would
be
needed
to
calculate
more
appropriate
water
content
limits
for
various
soil textures.
Soils
in the
data base
we
assembled were mostly deep
and
moderately well
or
better drained. Soils having root
restrictive layers were included
in the
data base,
but
since
root
density
in the
restrictive layers
was
generally
.in-
adequate
for
complete water removal,
the
values were
excluded
from
the
data
reported
herein.
We
also
recog-
nize
that some
of the
variation
in the
field-measured soil
water
data results
from
variations
in
techniques used
by
the
investigators providing
the
data
and
from
natural
within-site
soil variations. Assuming
the
errors
due to
•variation
in
measuring technique
and
soil heterogeneity
are
random,
our
comparisons between field-measured
limits
and
laboratory-estimated limits should
be
valid.
Texture
s
Is
si
1
sil
si
sicl
cl
scl
sic
c
DUL
vs.
-0.33
bar
*
NS
*
NS
*
_
*
NS
*
t
NS
LOL
vs.
-15
bars
*
NS
NS
*
*
-
NS
NS
*
*
*
PLEXW
vs. WRD
*
NS
NS
*
*
..
*
NS
NS
NS
NS
*
and t
Indicate
significant
differences
at the
0.05
and
0.10
levels,
re-
spectively.
NS
indicates
not
significant
at the
0.10
level.
The
results
suggest
that
if
absolute
accuracy
is
necessary
in
soil water balance calculations, laboratory estimates
of
limits
of the
soil water reservoir should
be
used with
caution.
Field-measured limits
are
usually
a
more
ac-
curate alternative
if
they
are
available.
ACKNOWLEDGMENTS
The
assistance
of
scientists
who
provided information
for
this
study
is
gratefully acknowledged. Appreciation
is
also extended
to the scientists and technicians who assisted in describing, sam-
pling,
processing,
and
reviewing
the
soils
data.
The
authors especially thank
R.D.
Jackson,
USDA-ARS,
Phoenix,
for the
data
shown
in
Fig.
2 and
soil
B
of
Fig.
4;
W.C.
Johnson,
USDA-ARS
(retired),
Bushland,
Tex.,
for the
data
shown
in
Fig.
3 and
soil
A of
Fig.
4;
P.L.
Brown,
USDA-ARS,
Bozeman,
Mont.,
for the
data
shown
for
soil
C of
Fig.
4; and
L.C.
Hammond,
University of Florida,
Gainesville,
for the data
shown
in
soil
D
of
Fig.
4.