Variability of Soil Physical Properties (1)
1
Variability of Soil Physical Properties in a Clay-Loam Soil and Its Implication on Soil
2
Management Practices
3
4
Samuel I. Haruna and Nsalambi V. Nkongolo*
5
Center of Excellence for Geospatial Information Sciences, Department of Agriculture and
6
Environmental Science, Lincoln University, Jefferson City, MO 65102-0029
7
8
*Corresponding author: nkongolo@lincolnu.edu
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
1
Abstract
2
We assessed the spatial variability of soil physical properties in a clay- loam soil cropped to corn
3
and soybean. The study was conducted at Lincoln University in Jefferson City, Missouri. Soil
4
samples were taken at four depths: 0-10 cm, 10-20, 20-40 and 40-60 cm and were oven dried at
5
105o c for 72 hours. Bulk density (BDY), volumetric (VWC) and gravimetric (GWC) water
6
contents, volumetric air content (VAC), total pore space (TPS), air- filled (AFPS) and water-
7
filled (WFPS) pore space, the gas diffusion coefficient (Diff) and the pore tortuosity factor (Tort)
8
were calculated. Results showed that in comparison to Depth 1, means for AFPS, Diff, TPS, and
9
VAC decreased in Depth 2. In opposite, BDY, Tort, VWC and WFPS increased in depth2.
10
Semivariogram analysis showed that GWC, VWC, BDY and TPS in depth 2 fitted to an
11
exponential variogram model. The range of spatial variability (Ao) for BDY, TPS, VAC, WFPS,
12
AFPS, Diff and Tort was the same (25.77 m) in depth 1 and 4, suggesting that these soil
13
properties can be sampled together at a same distance. The analysis also showed the presence of
14
a strong (≤ 25%) to weak (>75%) spatial dependence for soil physical properties.
15
16
Key Words: Spatial variability, physical properties, semivariogram, depth, clay- loam soil
17
18
19
20
21
22
23
2
1
1. Introduction
2
Characterizing the spatial variability and distribution of soil properties is important in predicting
3
the rates of ecosystem processes with respect to natural and anthropoge nic factors [1], and in
4
understanding how ecosystems and their services work [2]. In agriculture, studies of the effects
5
of land management on soil properties have shown that cultivation generally increases the
6
potential for soil degradation due to the breakdown of soil aggregates and the reduction of soil
7
cohesion, and thus a decrease in soil nutrient content [3, 4]. Cultivation, especially when
8
accompanied by tillage, have been reported to have significant effects on topsoil structure and
9
thus the ability of soil to fulfill essential soil functions and services in relation to root growth, gas
10
and water transport and organic matter turnover [5, 6, 7]. The physical properties of the soil are
11
governed by many factors that change vertically with depth, laterally across fields and
12
temporally in response to climate and human activity [8]. A more permanent and important
13
factor is soil texture. Apart from soil texture, structure and porosity, a less permanent factor
14
affecting soil physical properties is human influence through agricultural practices of soil tillage,
15
crop rotation and cover cropping. Soil properties vary considerably under different crops, tillage
16
type and intensity, fertilizer types and application rates. Since this variability affects soil water
17
and nutrient content dynamics and other soil physical processes, knowledge of the spatial
18
variability of soil properties is therefore necessary.
19
To study the spatial distribution of soil properties, techniques such as classical statistics and
20
geostatistics that seek to show spatial and spatiotemporal variations have been widely applied [9,
21
10, 11]. Geostatistics provides the basis for interpolation and interpretation of the spatial
22
variability of soil properties [9, 12, 13, 14]. Information on soil spatial variability leads to better
23
management decisions aimed at correcting problems and at least maintaining prod uctivity and
3
1
sustainability of the soils, and thus increasing the precision of farming practices [1, 15]. A better
2
understanding of the spatial variability of soil properties would enable for refining agricultural
3
management practices by identifying sites where remediation and management is needed. This
4
promotes sustainable soil and land use, and also provides valuable base against which subsequent
5
future measurements can be proposed [14]. Despite its importance in agriculture, literature is
6
lacking on the variability of soil physical properties in central Missouri, especially at the lower
7
depths (30-60cm). The objective of this study, therefore, was to assess the spatial variability of
8
soil physical properties in a clay- loam soil field and determine how knowledge on this variability
9
can affect agricultural management practices.
10
11
2. Materials and methods
12
2.1 Experimental site
13
The study was conducted at Lincoln University‟s Freeman farm in Jefferson City, Missouri. The
14
geographic coordinates of the study site are 380 58‟16”N latitude and 920 10‟53”W longitude.
15
Prior to establishment in 2011, the farm was a 80.94 ha farm in the bottomland of the Missouri
16
river. It was mainly planted to corn and soybean, with conventional tillage (moldboard) for over
17
50 years. The soil type is mainly Waldron clay- loam (Fine, smectitic, calcareous, mesic Aeric
18
Fluvaquents). The study area is almost flat, with an average slope of 2%. The experimental field
19
(Fig. 1.) is a 4.05 ha field divided into three blocks (replicates) using complete randomized block
20
design. Each block of 1.35 ha contains 16 plots for a total of 48 plots. Each plot measured 12.2m
21
x 21.3m. Half (Twenty four) of the plots were planted to Corn (Zee mays) while the other half
22
was grown to Soybean (Glycine max). As part of the standard research protocol for the grant that
23
funded this study (acknowledged below), soybean and corn plots all received 26.31 kg/ha of
4
1
nitrogen, 67.25 kg/ha of phosphorus, and 89.67 kg/ha of potassium. Corn plots received 201.75
2
kg/ha of additional nitrogen in the form of urea.
3
4
2.2 Soil sampling
5
Cylindrical cores of 3.15 cm radius and 10 or 20 cm heights were used to collect soil samples in
6
the corn and soybean field at four depths; 0-10, 10-20, 20-40 and 40-60 cm, corresponding to
7
depth 1, 2, 3 and 4, respectively. The cylinders were 10 cm in height for samples collected at
8
depths 1 and 2 and 20 cm in height for sampling at depths 3 and 4. A total of 576 soil samples
9
were collected: 48 plots x 4 depths x 3 replicates (in each plots). To eliminate compaction that
10
may be caused by trafficking, samples were collected at the center of each plot where there is
11
very little-to-no traffic. Soil samples were taken after planting and seeds full emergences.
12
Collected samples were taken to the laboratory where they were weighed (fresh weight of
13
sample; FWS), then oven dried at 105o C for 72 hrs. The weight was taken after oven drying (Dry
14
weight of Soil; DWS). Soil physical properties were calculated as follows: Soil bulk density
15
(BDY, g.cm-3 ) = DWS/V, where DWS is the dry weight of soil and V the volume of cylinder
16
(total volume of soil); Volumetric water content (VWC, cm3 .cm-3 ) = (FWS-DWS)/V, with FWS
17
being the fresh weight of soil; Gravimetric water content (GWC, g.g-1 ) = [(FWS-DWS)/DWS]
18
where FWS is the fresh weight of soil; Total pore space (TPS, cm3 .cm-3 ) = 1-(BDY/PDY) where
19
PDY is the soil particle density (taken as 2.65 g cm-3 ); Volumetric air content (VAC, cm3 .cm-3 ) =
20
TPS-VWC; Water-filled pore space (WFPS, %) =100*(VWC/TPS); Air-filled pore space
21
(AFPS, %) = 100*(VAC/TPS); Relative gas diffusion coeffient (Diff., cm2 s-1 .cm-2 .s) =(VAC)2
22
and Pore space tortuosity (Tort., m.m-1 ) =(1/VAC) [16].
23
5
1
2.3 Statistical and geospatial analysis
2
After calculation, data on soil physical properties was first transferred to Statistix 9.0 to compute
3
summaries of simple statistics, then to GS+ (Geostatistics for environmental science) 7.0 for
4
semivariogram analysis. A semivariogram (a measure of the strength of statistical correlation as
5
a function of distance) is defined by the following equation [17]:
6
7
8
m(h)
γ (h) = 1/(2m(h)) Σ [z(xi + h) – z(xi)]2
(1)
i=1
9
where γ(h) is the experimental semivariogram value at a distance interval h, m(h) is number of
10
sample value pairs within the distance interval h, Z(Xi), Z(Xi + h) are sample values at two
11
points separated by the distance h. Exponential and spherical models were to the empirical
12
semivariograms. The stationary models, i.e., Exponential (Eq. (2)) and Spherical model (Eq. (3))
13
that fitted to experimental semivariograms were defined in the following equations [18]:
14
15
γ(h) = C0 + C1 [1 – exp {- (h/a)}
(2)
16
17
18
γ(h) = C0 + C1 [ (3h/2a) – (h3 /2a3 ) ] when h ≤ a
= C0 + C1 when h ≥ a
(3)
19
20
where Co is the nugget, C 1 is the partial sill, and a is the range of spatial dependence to reach the
21
sill (Co + C1 ). The ratio C o /(Co + C1 ) and the range are the parameters that characterize the spatial
22
structure of a soil property. The C o /(Co + C1 ) relation is the proportion in the dependence zone,
23
and the range defines the distance over which the soil property values are correlated with each
6
1
other [19]. A low value for the C o /(Co + C1 ) ratio and a high range generally indicates that high
2
precision of the property can be obtained by kriging [19]. The classification proposed by
3
Cambardella et al. [14], which considers the degree of spatial dependence (DSD = C o /(Co + C1 ) x
4
100) as strong when DSD ≤ 25%; moderate when 25 < DSD ≤ 75%; and weak when DSD > 75%
5
was used in this study to classify the degree of spatial dependence of each soil property.
6
7
3. Results and discussion
8
3.1. Summaries of statistics for soil physica l properties
9
Overall descriptive statistics for soil properties in this study showed moderate to high skewness
10
for some of the properties (Table 1). The highly ske wed soil parameters included soil bulk
11
density (BDY), diffusivity (DIFF), volumetric water content (VWC), whereas total pore space
12
(TPS) was moderately skewed. Air filled pore space (AFPS) had a low skewness. Highly skewed
13
parameters indicate that these elements have a local distribution, that is, high values were found
14
for these elements at some points, but most values were low [20]. The other soil parameters were
15
approximately normally distributed on the field. The underlying reason for soil parameters being
16
distributed normally or non- normally may be associated with differences in management
17
practices, land use, vegetation cover, and topographic effects on the variability of soil erosion
18
across the landscape of the field. Such factors can be the sources for a large or very small
19
concentration of soil properties in some of the samples that leads to the non-normal distribution
20
[21].
21
A wide range of spatial variability was observed for soil physical properties (table 1). For
22
instance, bulk density ranged from 1.01-1.23gcm-3 for depth 1, 1.15-1.46gcm-3 for depth 2, 0.96-
23
1.19gcm-3 and 1.04-1.18gcm-3 for depths 3 and 4 respectively. The mean value of AFPS was
7
1
significantly lower in the second depth (26.54%) than all the other 3 depths where it varied from
2
39.34% to 45.76%. Soil bulk density was also significantly higher in the second depth (1.47%)
3
than all the other 3 depths, where it varied between 1.18 g cm-3 and 1.24 g cm-3 . Soil pore
4
tortuosity factor (tort.) and water filled pore space were also significantly higher in the second
5
depth (12.46% and 73.46% respectively). However, diffusivity (DIFF), gravimetric water
6
content (GWC), total pore space (TPS) and volumetric air content (VAC) were significantly
7
lower in the second depth (0.02%, 0.21%, 0.42% and 0.12% respectively) (Table 1). Figure 3
8
shows the variability of bulk density, gravimetric water content, volumetric water content and
9
total pore space with depth.
10
All the variability in soil physical properties noticed in the second depth can be attributed to the
11
fact that Missouri has a smectite layer (clay-pan) in about 10-20cm deep in its soil corresponding
12
to the second sample depth. This layer of smectite is hard and compact, with very low pore
13
spaces, high mass-volume ratio (bulk density) and high water retention capability (because of
14
their large surface area). Correspondingly, the mean of water filled pore space (WFPS) was
15
slightly lower in the first depth (54%) than in all the four dep ths. This agrees with the fact that air
16
predominates the pore spaces in the first depth and also cultivation loosens the soil, thereby
17
allowing the water trapped in the pore spaces to evaporate. Higher GWC, VWC and TPS at the
18
lower depths (20-60cm) means crops (especially corn and soybean usually grown in the area)
19
will be able to access water and dissolved nutrients through their roots. Despite the clay-pan
20
layer (10-20cm), it has been reported by various researchers that crop roots were able to
21
penetrate into and through this layer of smectitic clay [22, 23, 24] and that root growth may
22
increase within the claypan layer [23] as a result of plant adaptation to water- limited soil layers.
8
1
In general, the use of coefficient of variation (CV) is a common procedure to assess variability in
2
soil properties since it allows comparison among properties with different units of measurement.
3
Overall, the coefficient of variation for all soil physical properties, in the four depths sampled,
4
ranged from 19.33% to 42.15% (AFPS), 5.57% to12.09% (BDY), 46.36% to 91.61% (Diff.),
5
10.43% to 22.48% (GWC), 4.83% to 16.40% (TPS), 22.69% to 85.91% (Tort.), 22.72% to
6
49.83% (VAC), 10.58% to 17.11% (VWC) and from 12.56% to 19.56% (WFPS) (Table 1). This
7
means that pore tortuosity factor (Tort.) showed the highest variation while soil bulk density
8
(BDY) showed the least variation. This classical statistics indicates a strong spatial variability of
9
the soil properties investigated. However, to have a better assessment of such spatial variability
10
across the entire field, the geostatistical procedure was used since it permits the dependence and
11
variability of a particular soil property.
12
13
3.2 Spatial variability of soil properties
14
Model fit was determined from the coefficient of determination (R2 ) values, which range from 0
15
(very poor model fit) to 1 (very good model fit). Table 2 shows soil physical properties which
16
mainly responded to exponential and linear variogram models, with exponential model providing
17
the better fit. In the 10-20 cm depth, exponential model provided the best fit for BDY (R2 =
18
0.93), with spherical models providing very poor model fit. Exponential model also provided the
19
better fit for pore tortuosity in the 20-40 cm depth (R2 = 0.57), although spherical model was
20
noticed. Linear and exponential models were observed in the 40-60cm depth for TPS (R2 = 0.46),
21
with linear model providing the better fit (Table 2). In general, for all depths, model fit was not
22
very strong with the exception of gravimetric water content and bulk density in the second depth.
9
1
Exponential model provided the best fit with about 65% of the physical parameters fitting this
2
model.
3
In geostatistical theory, the range of the semivariogram is the distance between correlated
4
measurements (the minimum lateral distance between two points before change in property is
5
noticed) and can be an effective criterion for the evaluation of sampling design and mapping of
6
soil properties. The value that the semivariogram model attains at the range (the value on the y-
7
axis) is called the sill. The partial sill is the sill minus the nugget [25, 26]. Theoretically, at zero
8
separation distance (lag = 0), the semivariogram value is zero. However, at an infinitesimally
9
small separation distance, the semivariogram often exhibits a nugget effect ( the apparent
10
discontinuity at the beginning of many semivariogram graphs), which is some value greater than
11
zero. The nugget effect can be attributed to measurement errors or spatial sources of variation at
12
distances smaller than the sampling interval (or both). Measurement error occurs because of the
13
error inherent in measuring devices. To eliminate this error, multiple sampled were taken from
14
each plot (Materials and Methods above). Natural phenomena can vary spatially over a range of
15
scales. Variation at microscales smaller than the sampling distances will appear as part of the
16
nugget effect. Table 2 shows that the spatial correlation (range) of soil properties widely varied
17
from 1m for volumetric water content (VWC) in depth 4 to 64m for gravimetric water content
18
(GWC) in depth 2. However, for the first and second depth (which are agriculturally more
19
important), the range of spatial correlation varied from 3m for volumetric air content (VAC) in
20
depth 2 to 64m for GWC in depth 2. Beyond these ranges, there is no spatial dependence
21
(autocorrelation). The spatial dependence can indicate the level of similarity or disturbance of the
22
soil condition. According to Lopez-Granados et al. [27] and Ayoubi et al. [17], a large range
23
indicates that the measured soil parameter value is influenced by natural and anthropogenic
10
1
factors over greater distances than parameters which have smaller ranges. Thus, a range of about
2
64m for GWC in this study indicates that the measured GWC values can be influenced in the soil
3
over greater distances as compared to the soil parameters having smaller range (Table 2). This
4
means that soil variables with smaller range such as VWC and VAC are good indicators of the
5
more disturbed soils (the more disturbed a soil is, the more variable some soil properties become.
6
The more variable properties have a shorter range of correlation). The different ranges of the
7
spatial dependence among the soil properties may be attributed to differences in response to the
8
erosion–deposition factors, land use-cover, parent material and human interferences in the study
9
area. The nugget, which is an indication of micro- variability, was significantly higher for water
10
filled pore space (WFPS) and air filled pore space (AFPS) when compared to the others. This
11
may be due to the fact that the selected sampling distance could not capture well their spatia l
12
dependence. The lowest nugget was for GWC (Table 2). This indicates that GWC had low
13
spatial variability within small distances. Knowledge of the range of influence for various soil
14
properties allows one to construct independent accurate datasets for similar areas in future soil
15
sampling design to perform statistical analysis [17]. This aids in determining where to resample
16
if necessary, and design future field experiments that avoid spatial dependence. Therefore, for
17
future studies aimed at characterizing the spatial dependency of soil properties in the study area
18
and/or a similar area, it is recommended that the soil properties are sampled at distances shorter
19
than the range found in this study.
20
Cambardella et al. [14] established the classification of degree of spatial dependence (DSD)
21
between adjacent observations of soil property > 75% to correspond to weak spatial structure. In
22
this study, the semivariograms indicated strong spatial dependence (DSD ≤ 25%) for soil
23
physical properties such as bulk density, gravimetric water content, volumetric water content,
11
1
total pore space and Diffusivity. The rest of the soil physical properties (water filled pore space,
2
air filled pore space, tortuosity) measured exhibited very weak spatial dependence (DSD > 75%)
3
(Table 2). The strong spatial dependence of the soil properties may be controlled by intrinsic
4
variations in soil characteristics such as texture and mineralogy whereas extrinsic variations such
5
as fertilizer application, tillage, soil and water conservatio n and other management practices may
6
control the variability of the weak spatially dependent parameters [14].
7
8
3.3 Spatial distribution of soil properties a cross the field.
9
Interpolated maps portraying the distribution of soil properties across the field are shown in Fig.
10
2. Gravimetric water content (GWC) showed a good spatial distribution across the field with the
11
highest values located around the southwestern portion of the field. Volumetric water content
12
(VWC) also showed good spatial distribution across the field with high values located in the
13
northern, central and south-western portions of the field. The other soil properties, however,
14
showed very poor spatial distribution in the field. This is most probably due to their poor sill
15
(Co +C), model fit and regression coefficient (r2 ). Even though the spatial variability was not
16
very pronounced, there were areas on the field that had slightly higher amount of these physical
17
properties than the rest of the field. In general, bulk density, total pore space, volumetric air
18
content, water filled pore space, air filled pore space, diffusivity, and tortuosity were very high in
19
the field even though they didn‟t exhibit very distinguishable variability. This lack of visible
20
spatial variability is supported by the fact that the sampling distance (range) is 26m for these
21
properties.
22
23
3.4. Implications of spatial variability of soil physica l properties on soil management
12
1
Results of this study indicated that the spatial variability of soil water content (GWC and VWC)
2
was high. This can be explained by soil type (clay- loam) which was able to hold more water. But
3
with intensive tillage, this soil water content could be adversely affected. Studies have shown
4
that tillage practices can alter soil physical properties and consequently the hydrological behavior
5
of agricultural fields, especially when a similar tillage system has been practiced for a long
6
period [15, 28, 29, 30, 31]. Tillage intensity has also considerable effects on spatial structure and
7
spatial variability of soil properties [15, 30]. Therefore, this study can help determine site-
8
specific soil management and decision making. To do so, spatial variability of the soil properties
9
developed through kriging will be an important tool. Different ranges of spatial dependence
10
were noticed in the field. The different ranges of the spatial dependence among the soil
11
properties may be attributed to differences in response to the erosion–deposition factors, land
12
use-cover, parent material and human interferences in the study area. The different ranges can
13
also be used in future studies to determine the sampling distance of different soil physical
14
properties on the field. Also, the sill (C o +C) can help determine where the variability or change
15
in soil property stops. This will be useful especially for irrigation purposes.
16
Generally, with farmers facing the decision of whether or not to till and the intensity of tillage, a
17
spatial variability study can help in this decision making. Maps produced in this study can also
18
be used for irrigation purposes as they can clearly indicate which portion of the field needs
19
irrigation (soil water content). Since different range of spatial dependence among soil properties
20
shows differences in response to human interferences and land use-cover, this will help reduce
21
human activities that increase soil bulk density and cause soil compaction like the use of heavy
22
equipments. It can also serve as a reference for the type of crop to be grown (cover crops for
23
erosion susceptible areas).
13
1
2
3
4. Conclusion
4
We conducted a study in central Missouri to test the variability of soil physical properties in a
5
clay-loam soil. Results show variability in soil physical properties with depth and across the
6
field. Soil physical properties either decreased or increased sharply in the second depth (due to
7
the presence of a smectite layer) before leveling up or dropping off, but without reaching the first
8
depth value in either case. In addition, depending on soil physical property, maps produced by
9
krigging showed either good or poor spatial distribution. The semivariogram analysis showed
10
the presence of a strong (≤ 25%) to weak (>75%) spatial dependence of soil properties. Our
11
understanding of the behavior of soil properties in this study provides new insights for soil site-
12
specific management in addressing issues such as „„where to place the proper interventions‟‟
13
(tillage, irrigation and crop type to be grown).
14
15
16
Acknowledge ment
17
This research is part of a regional collaborative project supported by the USDA-NIFA, Award
18
No. 2011-68002-30190, “Cropping Systems Coordinated Agricultural Project: Climate Change,
19
Mitigation,
20
http://sustainablecorn.org
and
Adaptation
in
Corn-based
21
22
23
14
Cropping Systems.” Project Web
site:
1
References
2
[1] Schimel D, Melillo J, Tian H, McGuire D, Kicklighter D, Kittel T, Rosenbloom N,
3
Running S, Thornton P, Ojima D, Parton W, Kelly R, Sykes M, Neilson R, Rizzo B.
4
2000. Contribution of increasing CO 2 and climate to carbon storage by ecosystems in the
5
United States. Science 287: 2004-2006.
6
[2] Kosmas C, Gerontidis S, Marathianou M. 200. The effects of land use change on soils and
7
vegetation over various lithological formations on Lesvos (Greece). Catena 40, 51-68.
8
[3] Igbal J. Thomasson JA, Jenkins JN, Owens PR, Whisler FD. 2005. Spatial variability analysis
9
of soil physical properties of alluvial soils. Soil Science Society of America Journal 69: 1-
10
11
14.
[4] Zhang S, Zhang X, Huffman T, Liu X, Yang J. 2011. Influence of topography and land
12
management on soil nutrients variability in Northeast China. Nutrient Cycle &
13
Agroecosystem 89: 427-438.
14
[5] Munkholm, L.J., A. Garbout, S.B. Hansen. 2013. Tillage effects on topsoil structural quality
15
assessed using X-ray CT, soil cores and visual soil evaluation. Soil & Tillage Research
16
128, 104–109.
17
18
19
[6] Franzluebbers, A.J., 2002. Soil organic matter stratification ratio as an indicator of soil
quality. Soil and Tillage Research 66, 95–106.
[7] Munkholm, L.J., Schjønning, P., Rasmussen, K.J., Tanderup, K., 2003. Spatial and temporal
20
effects of direct drilling on soil structure in the seedling environment. Soil and Tillage
21
Research 71, 163–173.
15
1
[8] Swarowsky, A., R.A. Dahlgren, K.W. Tate, J.W. Hopmans, A.T. O‟Geen. 2011 Catchment-
2
scale soil water dynamics in a Mediterranean-type oak woodland. Vadose Zone Journal,
3
10, 800-815.
4
5
6
7
8
9
[9] Webster R, Oliver MA. 1990. Statistical methods in soil and land Resource Survey. Oxford
University Press, Oxford University Press, Oxford, UK.
[10] Saldana A, Stein A, Zinck JA. 1998. Spatial variability of soil properties at different scales
within three terraces of the Henare River (Spain). Catena 33: 139-153.
[11] Wang ZM, Song KS, Zhang B, Liu DW, Li XY, Ren CY, Zhang SM, Lou L, Zhang CH.
2009. Spatial variability and affecting factors of soil nutrients in croplands of Northeast
10
China: a case study in Dehui county. P lant, Soil & Environment 55: 110-120.
11
[12] Webster R. 1985. Quantitative spatial analysis of soil in the field. In: Stewart, B.A. (Ed.),
12
Advances in Soil Science 3: 1-70.
13
[13] Pohlmann H. 1993. Geostatistical modeling of environmental data. Catena 20, 191-198.
14
[14] Cambardella CA, Moorman TB, Novak, JM, Parkin TB, Karlen DL, Turco RF, Konopka
15
AE. 1994. Field-scale variability of soil properties in central Iowa soils. Soil Science
16
Society of America Journal 58: 1501-1511.
17
18
19
[15] Ozgoz E. 2009. Long term conventional tillage effect on spatial variability of some soil
physical properties. Journal of Sustainable Agr iculture 33: 142-160.
[16] Nkongolo NV, Hatano R, and Kakembo V. 2010. Diffusivity models and greenhouse gases
20
fluxes from a forest, pasture, grassland and corn field in northern Hokkaido. Pedosphere
21
20(6):747-760. doi:10.1016/S1002-0160(10)60065-3.
16
1
[17] Ayoubi SH, Zamani SM, Khomali F. 2007. Spatial variability of some soil properties for
2
site-specific farming in northern Iran. International Journa l of Plant Production 2: 225-
3
236.
4
5
6
[18] Burgess TM, Webster R. 1980. Optimal Interpolation and isarithmic mapping of soil
properties: I. The variogram and punctual krigging . Journa l of Soil Science 31: 315-331.
[19] Parfitt JMB, Timm LC, Pauletto EA, Sousa RO, Castilhos DD, de Avila CL, Reckziegel
7
NL. 2009. Spatial variability of the chemical, physical and biological properties in
8
lowland cultivated with irrigated rice. Rev. Bra s. Cienc. Solo 33: 819-830.
9
10
[20] Grego CR, Vieira SR, Lourencao AL. 2006. Spatial distribution of Pseudaletia sequax
Franclemlont in triticale under no-till management. Science & Agr iculture 63: 321-327.
11
[21] Tesfahunegn GB, Tamane L, Vlek PLG. 2011. Catchment-scale spatial variability of soil
12
properties and implications on site-specific soil management in northern Ethiopia. Soil
13
& Tillage Research 117: 124-139.
14
15
16
[22] Grecu SJ, Kirkham MB, Kanemasu ET, Sweeney DW, Stone LR, Milliken GA. 1988. Root
growth in a claypan with a perennial-annual rotation. Soil Sci. Soc. Am. J. 52: 488-494
[23] Myers DB, Kitchen NR, Sudduth KA, Miles RJ, Sharp RE. 2007. Soybean root distribution
17
related to claypan soil properties and apparent soil electrical conductivity. Crop Sci. 47:
18
1498-1509
19
[24] Jiang P., Kitchen N.R., Anderson S.H., Sadler E.J., Sudduth K.A. 2008. Estimating palnt
20
available water using the simple inverse yield model for claypan landscapes. Agronomy
21
Journal. 100: 1-7.
22
23
[25] Utset A, Ruiz ME, Herrera J, Ponce de Leon D. 1998. A geostatistical method for soil
salinity sample site spacing. Geoderma 86: 143-151.
17
1
2
[26] Fu W, Tunney H, Zhang C. 2010. Spatial variation of soil nutrients in a dairy farm and its
implications for site-specific fertilizer application. Soil & Tillage Resea rch 106: 185-193.
3
[27] Lopez-Grandos F, Jurado-Exposito M, Atenciano S. Garcia-Ferrer A, De la Orden MS,
4
Garcia-Torres L. 2002. Spatial variability of Agricultural soil parameters in southern
5
Spain. Plant Soil 246: 97-105.
6
7
8
9
10
11
12
13
[28] Hill RL. 1990. Long-term conventional and no-tillage effects on selected soil physical
properties. Soil Science Society of America Journal 54: 161-166.
[29] Buschiazzo DE, Panigatti JL, Unger PW. 1998. Tillage effects on soil properties and crop
production in subhumid and semiarid Argentinean Pampas. Soil & Tillage Research 49:
105-116.
[30] Tsegaye T, Hill RL. 1998. Intensive tillage effects on spatial variab ility of soil physical
properties. Soil Science 163: 143-154.
[31] Gomez JA, Giraldez JV, Pastor M, Fereres E. 1999. Effects of tillage method on soil
14
physical properties, infiltration and yield in an olive orchard. Soil & Tillage Resea rch 52:
15
167-175.
16
17
18
19
20
18
1
2
Fig. 1. Study area (Lincoln University‟s Freeman farm) showing the plots
3
4
5
6
7
8
9
10
11
12
13
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Fig. 2. Spatial distribution of soil physical properties at four depths in a clay-loam soil
20
1
2
a) Soil Bulk Density
3
4
b) Gravimetric Water Content (GWC)
21
1
2
3
4
5
6
7
8
c) Total Pore Space (TPS)
d) Volumetric Water Content (VWC)
Fig. 3a, b, c, d. Variation of soil physical properties with depth
22
1
2
3
4
5
6
7
8
9
Table 1. Descriptive statistics for soil physical properties at four depths in a clay- loam soil
AFPS BDY
Diff.
GWC TPS
Tort
VAC
VWC
Depth 1, 0-10 cm
Mean
45.76
1.24
0.06
0.22
0.51
4.60
0.24
0.28
SD
10.61
0.11
0.03
0.03
0.04
1.50
0.07
0.04
C.V
23.18
8.99
56.29 15.25 8.36
32.60
29.64
15.26
Minimum
26.35
1.01
0.01
0.15
0.45
2.56
0.12
0.18
Median
45.71
1.24
0.06
0.22
0.52
4.10
0.24
0.28
Maximum
68.68
1.41
0.15
0.36
0.60
8.18
0.39
0.36
Skew
0.05
-0.24
0.78
1.03
0.22
0.89
0.25
-0.23
Kurtosis
-0.66 -0.76
0.04
3.56 -0.78 -0.06
-0.53
-0.60
Depth 2, 10-20 cm
Mean
26.54
1.47
0.02
0.21
0.42
12.46
0.12
0.31
SD
11.19
0.18
0.02
0.048 0.07
10.70
0.06
0.05
C.V
42.15 12.09 91.61 22.48 16.40 85.91
49.83
17.11
Minimum
7.73
1.15
0.00
0.08
0.19
4.11
0.02
0.17
Median
27.32
1.46
0.01
0.22
0.43
9.04
0.11
0.31
Maximum
50.01
2.07
0.06
0.32
0.55
60.85
0.24
0.43
Skew
0.10
1.21
0.88
-0.57 -1.20
2.89
0.30
-0.49
Kurtosis
-0.82
2.15
0.11
0.92
2.11
9.36
-0.87
0.83
Depth 3, 20-40 cm
Mean
42.35
1.20
0.06
0.26
0.53
4.72
0.23
0.30
SD
8.19
0.12
0.03
0.04
0.05
1.29
0.06
0.03
C.V
19.33
9.75
53.22 13.92 8.68
27.30
26.27
11.48
Minimum
24.69
0.96
0.02
0.15
0.41
2.71
0.12
0.23
Median
42.11
1.20
0.05
0.25
0.53
4.52
0.22
0.31
Maximum
61.59
1.51
0.14
0.37
0.62
8.01
0.37
0.42
Skew
0.24
0.30
1.38
0.20 -0.29
0.95
0.64
0.22
Kurtosis
0.29
0.08
1.69
2.20
0.04
0.75
0.53
1.82
Depth 4, 40-60 cm
Mean
39.34
1.18
0.05
0.28
0.54
4.94
0.21
0.32
SD
7.62
0.07
0.02
0.03
0.03
1.12
0.05
0.03
C.V
19.36
5.57
46.36 10.43 4.83
22.69
22.72
10.58
Minimum
25.25
1.04
0.02
0.21
0.48
3.11
0.13
0.23
Median
39.74
1.18
0.05
0.27
0.54
4.67
0.22
0.33
Maximum
57.56
1.32
0.10
0.36
0.59
7.53
0.32
0.40
Skew
0.382
0.27
0.76
0.25 -0.36
0.42
0.44
-0.41
Kurtosis
-0.44 -0.43 -0.10
0.84 -0.33 -0.56
-0.48
0.59
WFPS
54.24
10.61
19.56
31.32
54.29
73.65
-0.05
-0.66
73.46
11.19
15.23
49.99
72.68
92.27
-0.10
-0.82
57.65
8.19
14.20
38.41
57.90
75.31
-0.24
0.29
60.66
7.62
12.56
42.44
60.27
74.75
-0.38
-0.44
AFPS: Air filled pore space (%); BDY: Soil bulk density (gcm-3 ); Diff.: Relative gas diffusion coefficient (m2 s -1 ms); GWC: Gravimetric water content of soil (g.g -1 ); TPS: Total pore spaces (cm3 cm-3 ); Tort.: Pore tortuosity factor
(m.m-1 ); VAC: Vo lu metric air content (cm3 cm-3 ); VW C: Volu metric water content (cm3 cm-3 ); WFPS: Water filled
pore space (%).
2
23
1
2
Table 2. Variogram parameters for soil physical properties at four depths in a clay- loam soil
Depth(cm)
GWC
0-10
10-20
20-40
40-60
VWC
0-10
10-20
20-40
40-60
BDY
0-10
10-20
20-40
40-60
TPS
0-10
10-20
20-40
40-60
VAC
0-10
10-20
20-40
40-60
WFPS
0-10
10-20
20-40
40-60
AFPS
0-10
10-20
20-40
40-60
Diff.
0-10
10-20
20-40
40-60
Tort.
0-10
10-20
20-40
40-60
Model
Nugget(Co ) Sill(Co +C) Range(Ao )
R2
(C/Co+C)
DSD(%)
Exponential
Exponential
Spherical
Exponential
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.42
64.26
7.17
4.56
0.12
0.75
0.11
0.01
0.94
0.52
1.00
0.99
0.01
0.29
0.00
0.00
Exponential
Exponential
Exponential
Exponential
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.21
11.82
5.88
0.45
0.31
0.44
0.17
0.00
0.94
0.89
0.97
1.00
0.01
0.04
0.00
0.00
Linear
Exponential
Spherical
Linear
0.01
0.02
0.00
0.00
0.01
0.04
0.01
0.00
25.77
40.17
7.42
25.77
0.07
0.93
0.20
0.39
0.00
0.50
0.95
0.00
1.18
0.04
0.07
0.44
Linear
Exponential
Exponential
Linear
0.00
0.00
0.00
0.00
0.00
0.01
0.00
0.00
25.77
14.64
7.50
25.77
0.07
0.79
0.21
0.47
0.00
0.81
0.90
0.00
0.16
0.13
0.03
0.07
Linear
Exponential
Exponential
Linear
0.00
0.00
0.01
0.00
0.00
0.00
0.00
0.00
25.77
2.73
7.62
25.77
0.25
0.00
0.35
0.21
0.00
0.86
0.92
0.00
0.48
0.06
0.07
0.23
Linear
Spherical
Exponential
Linear
107.28
10.80
3.50
55.49
107.28
135.30
66.49
55.49
25.77
5.38
7.47
25.77
0.15
0.00
0.27
0.07
0.00
0.92
0.95
0.00
10727.60
1173.91
369.59
5549.30
Linear
Spherical
Exponential
Linear
107.28
10.80
3.50
55.49
107.28
135.30
66.49
55.49
25.77
5.38
7.47
25.77
0.15
0.00
0.27
0.07
0.00
0.92
0.95
0.00
10727.6
1173.91
369.588
5549.30
Linear
Exponential
Exponential
Linear
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
25.77
4.71
4.53
25.77
0.10
0.00
0.05
0.19
0.00
0.85
0.93
0.00
0.11
0.00
0.01
0.05
Linear
Exponential
Exponential
Linear
2.12
17.80
0.18
1.24
2.12
126.10
1.74
1.24
25.77
8.10
13.20
25.77
0.15
0.19
0.58
0.16
0.00
0.86
0.90
0.00
211.80
2072.20
19.44
123.68
24
1
2
3
DSD = Degree of spatial dependence: strong DSD (DSD ≤ 25%), moderate DSD (25 < DSD ≤ 75%),
weak DSD (DSD > 75%) according to Cambardella et al., (1994).
4
25
Variability of Soil Physical Properties in a Clay-Loam Soil and Its Implication on Soil
2
Management Practices
3
4
Samuel I. Haruna and Nsalambi V. Nkongolo*
5
Center of Excellence for Geospatial Information Sciences, Department of Agriculture and
6
Environmental Science, Lincoln University, Jefferson City, MO 65102-0029
7
8
*Corresponding author: nkongolo@lincolnu.edu
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
1
Abstract
2
We assessed the spatial variability of soil physical properties in a clay- loam soil cropped to corn
3
and soybean. The study was conducted at Lincoln University in Jefferson City, Missouri. Soil
4
samples were taken at four depths: 0-10 cm, 10-20, 20-40 and 40-60 cm and were oven dried at
5
105o c for 72 hours. Bulk density (BDY), volumetric (VWC) and gravimetric (GWC) water
6
contents, volumetric air content (VAC), total pore space (TPS), air- filled (AFPS) and water-
7
filled (WFPS) pore space, the gas diffusion coefficient (Diff) and the pore tortuosity factor (Tort)
8
were calculated. Results showed that in comparison to Depth 1, means for AFPS, Diff, TPS, and
9
VAC decreased in Depth 2. In opposite, BDY, Tort, VWC and WFPS increased in depth2.
10
Semivariogram analysis showed that GWC, VWC, BDY and TPS in depth 2 fitted to an
11
exponential variogram model. The range of spatial variability (Ao) for BDY, TPS, VAC, WFPS,
12
AFPS, Diff and Tort was the same (25.77 m) in depth 1 and 4, suggesting that these soil
13
properties can be sampled together at a same distance. The analysis also showed the presence of
14
a strong (≤ 25%) to weak (>75%) spatial dependence for soil physical properties.
15
16
Key Words: Spatial variability, physical properties, semivariogram, depth, clay- loam soil
17
18
19
20
21
22
23
2
1
1. Introduction
2
Characterizing the spatial variability and distribution of soil properties is important in predicting
3
the rates of ecosystem processes with respect to natural and anthropoge nic factors [1], and in
4
understanding how ecosystems and their services work [2]. In agriculture, studies of the effects
5
of land management on soil properties have shown that cultivation generally increases the
6
potential for soil degradation due to the breakdown of soil aggregates and the reduction of soil
7
cohesion, and thus a decrease in soil nutrient content [3, 4]. Cultivation, especially when
8
accompanied by tillage, have been reported to have significant effects on topsoil structure and
9
thus the ability of soil to fulfill essential soil functions and services in relation to root growth, gas
10
and water transport and organic matter turnover [5, 6, 7]. The physical properties of the soil are
11
governed by many factors that change vertically with depth, laterally across fields and
12
temporally in response to climate and human activity [8]. A more permanent and important
13
factor is soil texture. Apart from soil texture, structure and porosity, a less permanent factor
14
affecting soil physical properties is human influence through agricultural practices of soil tillage,
15
crop rotation and cover cropping. Soil properties vary considerably under different crops, tillage
16
type and intensity, fertilizer types and application rates. Since this variability affects soil water
17
and nutrient content dynamics and other soil physical processes, knowledge of the spatial
18
variability of soil properties is therefore necessary.
19
To study the spatial distribution of soil properties, techniques such as classical statistics and
20
geostatistics that seek to show spatial and spatiotemporal variations have been widely applied [9,
21
10, 11]. Geostatistics provides the basis for interpolation and interpretation of the spatial
22
variability of soil properties [9, 12, 13, 14]. Information on soil spatial variability leads to better
23
management decisions aimed at correcting problems and at least maintaining prod uctivity and
3
1
sustainability of the soils, and thus increasing the precision of farming practices [1, 15]. A better
2
understanding of the spatial variability of soil properties would enable for refining agricultural
3
management practices by identifying sites where remediation and management is needed. This
4
promotes sustainable soil and land use, and also provides valuable base against which subsequent
5
future measurements can be proposed [14]. Despite its importance in agriculture, literature is
6
lacking on the variability of soil physical properties in central Missouri, especially at the lower
7
depths (30-60cm). The objective of this study, therefore, was to assess the spatial variability of
8
soil physical properties in a clay- loam soil field and determine how knowledge on this variability
9
can affect agricultural management practices.
10
11
2. Materials and methods
12
2.1 Experimental site
13
The study was conducted at Lincoln University‟s Freeman farm in Jefferson City, Missouri. The
14
geographic coordinates of the study site are 380 58‟16”N latitude and 920 10‟53”W longitude.
15
Prior to establishment in 2011, the farm was a 80.94 ha farm in the bottomland of the Missouri
16
river. It was mainly planted to corn and soybean, with conventional tillage (moldboard) for over
17
50 years. The soil type is mainly Waldron clay- loam (Fine, smectitic, calcareous, mesic Aeric
18
Fluvaquents). The study area is almost flat, with an average slope of 2%. The experimental field
19
(Fig. 1.) is a 4.05 ha field divided into three blocks (replicates) using complete randomized block
20
design. Each block of 1.35 ha contains 16 plots for a total of 48 plots. Each plot measured 12.2m
21
x 21.3m. Half (Twenty four) of the plots were planted to Corn (Zee mays) while the other half
22
was grown to Soybean (Glycine max). As part of the standard research protocol for the grant that
23
funded this study (acknowledged below), soybean and corn plots all received 26.31 kg/ha of
4
1
nitrogen, 67.25 kg/ha of phosphorus, and 89.67 kg/ha of potassium. Corn plots received 201.75
2
kg/ha of additional nitrogen in the form of urea.
3
4
2.2 Soil sampling
5
Cylindrical cores of 3.15 cm radius and 10 or 20 cm heights were used to collect soil samples in
6
the corn and soybean field at four depths; 0-10, 10-20, 20-40 and 40-60 cm, corresponding to
7
depth 1, 2, 3 and 4, respectively. The cylinders were 10 cm in height for samples collected at
8
depths 1 and 2 and 20 cm in height for sampling at depths 3 and 4. A total of 576 soil samples
9
were collected: 48 plots x 4 depths x 3 replicates (in each plots). To eliminate compaction that
10
may be caused by trafficking, samples were collected at the center of each plot where there is
11
very little-to-no traffic. Soil samples were taken after planting and seeds full emergences.
12
Collected samples were taken to the laboratory where they were weighed (fresh weight of
13
sample; FWS), then oven dried at 105o C for 72 hrs. The weight was taken after oven drying (Dry
14
weight of Soil; DWS). Soil physical properties were calculated as follows: Soil bulk density
15
(BDY, g.cm-3 ) = DWS/V, where DWS is the dry weight of soil and V the volume of cylinder
16
(total volume of soil); Volumetric water content (VWC, cm3 .cm-3 ) = (FWS-DWS)/V, with FWS
17
being the fresh weight of soil; Gravimetric water content (GWC, g.g-1 ) = [(FWS-DWS)/DWS]
18
where FWS is the fresh weight of soil; Total pore space (TPS, cm3 .cm-3 ) = 1-(BDY/PDY) where
19
PDY is the soil particle density (taken as 2.65 g cm-3 ); Volumetric air content (VAC, cm3 .cm-3 ) =
20
TPS-VWC; Water-filled pore space (WFPS, %) =100*(VWC/TPS); Air-filled pore space
21
(AFPS, %) = 100*(VAC/TPS); Relative gas diffusion coeffient (Diff., cm2 s-1 .cm-2 .s) =(VAC)2
22
and Pore space tortuosity (Tort., m.m-1 ) =(1/VAC) [16].
23
5
1
2.3 Statistical and geospatial analysis
2
After calculation, data on soil physical properties was first transferred to Statistix 9.0 to compute
3
summaries of simple statistics, then to GS+ (Geostatistics for environmental science) 7.0 for
4
semivariogram analysis. A semivariogram (a measure of the strength of statistical correlation as
5
a function of distance) is defined by the following equation [17]:
6
7
8
m(h)
γ (h) = 1/(2m(h)) Σ [z(xi + h) – z(xi)]2
(1)
i=1
9
where γ(h) is the experimental semivariogram value at a distance interval h, m(h) is number of
10
sample value pairs within the distance interval h, Z(Xi), Z(Xi + h) are sample values at two
11
points separated by the distance h. Exponential and spherical models were to the empirical
12
semivariograms. The stationary models, i.e., Exponential (Eq. (2)) and Spherical model (Eq. (3))
13
that fitted to experimental semivariograms were defined in the following equations [18]:
14
15
γ(h) = C0 + C1 [1 – exp {- (h/a)}
(2)
16
17
18
γ(h) = C0 + C1 [ (3h/2a) – (h3 /2a3 ) ] when h ≤ a
= C0 + C1 when h ≥ a
(3)
19
20
where Co is the nugget, C 1 is the partial sill, and a is the range of spatial dependence to reach the
21
sill (Co + C1 ). The ratio C o /(Co + C1 ) and the range are the parameters that characterize the spatial
22
structure of a soil property. The C o /(Co + C1 ) relation is the proportion in the dependence zone,
23
and the range defines the distance over which the soil property values are correlated with each
6
1
other [19]. A low value for the C o /(Co + C1 ) ratio and a high range generally indicates that high
2
precision of the property can be obtained by kriging [19]. The classification proposed by
3
Cambardella et al. [14], which considers the degree of spatial dependence (DSD = C o /(Co + C1 ) x
4
100) as strong when DSD ≤ 25%; moderate when 25 < DSD ≤ 75%; and weak when DSD > 75%
5
was used in this study to classify the degree of spatial dependence of each soil property.
6
7
3. Results and discussion
8
3.1. Summaries of statistics for soil physica l properties
9
Overall descriptive statistics for soil properties in this study showed moderate to high skewness
10
for some of the properties (Table 1). The highly ske wed soil parameters included soil bulk
11
density (BDY), diffusivity (DIFF), volumetric water content (VWC), whereas total pore space
12
(TPS) was moderately skewed. Air filled pore space (AFPS) had a low skewness. Highly skewed
13
parameters indicate that these elements have a local distribution, that is, high values were found
14
for these elements at some points, but most values were low [20]. The other soil parameters were
15
approximately normally distributed on the field. The underlying reason for soil parameters being
16
distributed normally or non- normally may be associated with differences in management
17
practices, land use, vegetation cover, and topographic effects on the variability of soil erosion
18
across the landscape of the field. Such factors can be the sources for a large or very small
19
concentration of soil properties in some of the samples that leads to the non-normal distribution
20
[21].
21
A wide range of spatial variability was observed for soil physical properties (table 1). For
22
instance, bulk density ranged from 1.01-1.23gcm-3 for depth 1, 1.15-1.46gcm-3 for depth 2, 0.96-
23
1.19gcm-3 and 1.04-1.18gcm-3 for depths 3 and 4 respectively. The mean value of AFPS was
7
1
significantly lower in the second depth (26.54%) than all the other 3 depths where it varied from
2
39.34% to 45.76%. Soil bulk density was also significantly higher in the second depth (1.47%)
3
than all the other 3 depths, where it varied between 1.18 g cm-3 and 1.24 g cm-3 . Soil pore
4
tortuosity factor (tort.) and water filled pore space were also significantly higher in the second
5
depth (12.46% and 73.46% respectively). However, diffusivity (DIFF), gravimetric water
6
content (GWC), total pore space (TPS) and volumetric air content (VAC) were significantly
7
lower in the second depth (0.02%, 0.21%, 0.42% and 0.12% respectively) (Table 1). Figure 3
8
shows the variability of bulk density, gravimetric water content, volumetric water content and
9
total pore space with depth.
10
All the variability in soil physical properties noticed in the second depth can be attributed to the
11
fact that Missouri has a smectite layer (clay-pan) in about 10-20cm deep in its soil corresponding
12
to the second sample depth. This layer of smectite is hard and compact, with very low pore
13
spaces, high mass-volume ratio (bulk density) and high water retention capability (because of
14
their large surface area). Correspondingly, the mean of water filled pore space (WFPS) was
15
slightly lower in the first depth (54%) than in all the four dep ths. This agrees with the fact that air
16
predominates the pore spaces in the first depth and also cultivation loosens the soil, thereby
17
allowing the water trapped in the pore spaces to evaporate. Higher GWC, VWC and TPS at the
18
lower depths (20-60cm) means crops (especially corn and soybean usually grown in the area)
19
will be able to access water and dissolved nutrients through their roots. Despite the clay-pan
20
layer (10-20cm), it has been reported by various researchers that crop roots were able to
21
penetrate into and through this layer of smectitic clay [22, 23, 24] and that root growth may
22
increase within the claypan layer [23] as a result of plant adaptation to water- limited soil layers.
8
1
In general, the use of coefficient of variation (CV) is a common procedure to assess variability in
2
soil properties since it allows comparison among properties with different units of measurement.
3
Overall, the coefficient of variation for all soil physical properties, in the four depths sampled,
4
ranged from 19.33% to 42.15% (AFPS), 5.57% to12.09% (BDY), 46.36% to 91.61% (Diff.),
5
10.43% to 22.48% (GWC), 4.83% to 16.40% (TPS), 22.69% to 85.91% (Tort.), 22.72% to
6
49.83% (VAC), 10.58% to 17.11% (VWC) and from 12.56% to 19.56% (WFPS) (Table 1). This
7
means that pore tortuosity factor (Tort.) showed the highest variation while soil bulk density
8
(BDY) showed the least variation. This classical statistics indicates a strong spatial variability of
9
the soil properties investigated. However, to have a better assessment of such spatial variability
10
across the entire field, the geostatistical procedure was used since it permits the dependence and
11
variability of a particular soil property.
12
13
3.2 Spatial variability of soil properties
14
Model fit was determined from the coefficient of determination (R2 ) values, which range from 0
15
(very poor model fit) to 1 (very good model fit). Table 2 shows soil physical properties which
16
mainly responded to exponential and linear variogram models, with exponential model providing
17
the better fit. In the 10-20 cm depth, exponential model provided the best fit for BDY (R2 =
18
0.93), with spherical models providing very poor model fit. Exponential model also provided the
19
better fit for pore tortuosity in the 20-40 cm depth (R2 = 0.57), although spherical model was
20
noticed. Linear and exponential models were observed in the 40-60cm depth for TPS (R2 = 0.46),
21
with linear model providing the better fit (Table 2). In general, for all depths, model fit was not
22
very strong with the exception of gravimetric water content and bulk density in the second depth.
9
1
Exponential model provided the best fit with about 65% of the physical parameters fitting this
2
model.
3
In geostatistical theory, the range of the semivariogram is the distance between correlated
4
measurements (the minimum lateral distance between two points before change in property is
5
noticed) and can be an effective criterion for the evaluation of sampling design and mapping of
6
soil properties. The value that the semivariogram model attains at the range (the value on the y-
7
axis) is called the sill. The partial sill is the sill minus the nugget [25, 26]. Theoretically, at zero
8
separation distance (lag = 0), the semivariogram value is zero. However, at an infinitesimally
9
small separation distance, the semivariogram often exhibits a nugget effect ( the apparent
10
discontinuity at the beginning of many semivariogram graphs), which is some value greater than
11
zero. The nugget effect can be attributed to measurement errors or spatial sources of variation at
12
distances smaller than the sampling interval (or both). Measurement error occurs because of the
13
error inherent in measuring devices. To eliminate this error, multiple sampled were taken from
14
each plot (Materials and Methods above). Natural phenomena can vary spatially over a range of
15
scales. Variation at microscales smaller than the sampling distances will appear as part of the
16
nugget effect. Table 2 shows that the spatial correlation (range) of soil properties widely varied
17
from 1m for volumetric water content (VWC) in depth 4 to 64m for gravimetric water content
18
(GWC) in depth 2. However, for the first and second depth (which are agriculturally more
19
important), the range of spatial correlation varied from 3m for volumetric air content (VAC) in
20
depth 2 to 64m for GWC in depth 2. Beyond these ranges, there is no spatial dependence
21
(autocorrelation). The spatial dependence can indicate the level of similarity or disturbance of the
22
soil condition. According to Lopez-Granados et al. [27] and Ayoubi et al. [17], a large range
23
indicates that the measured soil parameter value is influenced by natural and anthropogenic
10
1
factors over greater distances than parameters which have smaller ranges. Thus, a range of about
2
64m for GWC in this study indicates that the measured GWC values can be influenced in the soil
3
over greater distances as compared to the soil parameters having smaller range (Table 2). This
4
means that soil variables with smaller range such as VWC and VAC are good indicators of the
5
more disturbed soils (the more disturbed a soil is, the more variable some soil properties become.
6
The more variable properties have a shorter range of correlation). The different ranges of the
7
spatial dependence among the soil properties may be attributed to differences in response to the
8
erosion–deposition factors, land use-cover, parent material and human interferences in the study
9
area. The nugget, which is an indication of micro- variability, was significantly higher for water
10
filled pore space (WFPS) and air filled pore space (AFPS) when compared to the others. This
11
may be due to the fact that the selected sampling distance could not capture well their spatia l
12
dependence. The lowest nugget was for GWC (Table 2). This indicates that GWC had low
13
spatial variability within small distances. Knowledge of the range of influence for various soil
14
properties allows one to construct independent accurate datasets for similar areas in future soil
15
sampling design to perform statistical analysis [17]. This aids in determining where to resample
16
if necessary, and design future field experiments that avoid spatial dependence. Therefore, for
17
future studies aimed at characterizing the spatial dependency of soil properties in the study area
18
and/or a similar area, it is recommended that the soil properties are sampled at distances shorter
19
than the range found in this study.
20
Cambardella et al. [14] established the classification of degree of spatial dependence (DSD)
21
between adjacent observations of soil property > 75% to correspond to weak spatial structure. In
22
this study, the semivariograms indicated strong spatial dependence (DSD ≤ 25%) for soil
23
physical properties such as bulk density, gravimetric water content, volumetric water content,
11
1
total pore space and Diffusivity. The rest of the soil physical properties (water filled pore space,
2
air filled pore space, tortuosity) measured exhibited very weak spatial dependence (DSD > 75%)
3
(Table 2). The strong spatial dependence of the soil properties may be controlled by intrinsic
4
variations in soil characteristics such as texture and mineralogy whereas extrinsic variations such
5
as fertilizer application, tillage, soil and water conservatio n and other management practices may
6
control the variability of the weak spatially dependent parameters [14].
7
8
3.3 Spatial distribution of soil properties a cross the field.
9
Interpolated maps portraying the distribution of soil properties across the field are shown in Fig.
10
2. Gravimetric water content (GWC) showed a good spatial distribution across the field with the
11
highest values located around the southwestern portion of the field. Volumetric water content
12
(VWC) also showed good spatial distribution across the field with high values located in the
13
northern, central and south-western portions of the field. The other soil properties, however,
14
showed very poor spatial distribution in the field. This is most probably due to their poor sill
15
(Co +C), model fit and regression coefficient (r2 ). Even though the spatial variability was not
16
very pronounced, there were areas on the field that had slightly higher amount of these physical
17
properties than the rest of the field. In general, bulk density, total pore space, volumetric air
18
content, water filled pore space, air filled pore space, diffusivity, and tortuosity were very high in
19
the field even though they didn‟t exhibit very distinguishable variability. This lack of visible
20
spatial variability is supported by the fact that the sampling distance (range) is 26m for these
21
properties.
22
23
3.4. Implications of spatial variability of soil physica l properties on soil management
12
1
Results of this study indicated that the spatial variability of soil water content (GWC and VWC)
2
was high. This can be explained by soil type (clay- loam) which was able to hold more water. But
3
with intensive tillage, this soil water content could be adversely affected. Studies have shown
4
that tillage practices can alter soil physical properties and consequently the hydrological behavior
5
of agricultural fields, especially when a similar tillage system has been practiced for a long
6
period [15, 28, 29, 30, 31]. Tillage intensity has also considerable effects on spatial structure and
7
spatial variability of soil properties [15, 30]. Therefore, this study can help determine site-
8
specific soil management and decision making. To do so, spatial variability of the soil properties
9
developed through kriging will be an important tool. Different ranges of spatial dependence
10
were noticed in the field. The different ranges of the spatial dependence among the soil
11
properties may be attributed to differences in response to the erosion–deposition factors, land
12
use-cover, parent material and human interferences in the study area. The different ranges can
13
also be used in future studies to determine the sampling distance of different soil physical
14
properties on the field. Also, the sill (C o +C) can help determine where the variability or change
15
in soil property stops. This will be useful especially for irrigation purposes.
16
Generally, with farmers facing the decision of whether or not to till and the intensity of tillage, a
17
spatial variability study can help in this decision making. Maps produced in this study can also
18
be used for irrigation purposes as they can clearly indicate which portion of the field needs
19
irrigation (soil water content). Since different range of spatial dependence among soil properties
20
shows differences in response to human interferences and land use-cover, this will help reduce
21
human activities that increase soil bulk density and cause soil compaction like the use of heavy
22
equipments. It can also serve as a reference for the type of crop to be grown (cover crops for
23
erosion susceptible areas).
13
1
2
3
4. Conclusion
4
We conducted a study in central Missouri to test the variability of soil physical properties in a
5
clay-loam soil. Results show variability in soil physical properties with depth and across the
6
field. Soil physical properties either decreased or increased sharply in the second depth (due to
7
the presence of a smectite layer) before leveling up or dropping off, but without reaching the first
8
depth value in either case. In addition, depending on soil physical property, maps produced by
9
krigging showed either good or poor spatial distribution. The semivariogram analysis showed
10
the presence of a strong (≤ 25%) to weak (>75%) spatial dependence of soil properties. Our
11
understanding of the behavior of soil properties in this study provides new insights for soil site-
12
specific management in addressing issues such as „„where to place the proper interventions‟‟
13
(tillage, irrigation and crop type to be grown).
14
15
16
Acknowledge ment
17
This research is part of a regional collaborative project supported by the USDA-NIFA, Award
18
No. 2011-68002-30190, “Cropping Systems Coordinated Agricultural Project: Climate Change,
19
Mitigation,
20
http://sustainablecorn.org
and
Adaptation
in
Corn-based
21
22
23
14
Cropping Systems.” Project Web
site:
1
References
2
[1] Schimel D, Melillo J, Tian H, McGuire D, Kicklighter D, Kittel T, Rosenbloom N,
3
Running S, Thornton P, Ojima D, Parton W, Kelly R, Sykes M, Neilson R, Rizzo B.
4
2000. Contribution of increasing CO 2 and climate to carbon storage by ecosystems in the
5
United States. Science 287: 2004-2006.
6
[2] Kosmas C, Gerontidis S, Marathianou M. 200. The effects of land use change on soils and
7
vegetation over various lithological formations on Lesvos (Greece). Catena 40, 51-68.
8
[3] Igbal J. Thomasson JA, Jenkins JN, Owens PR, Whisler FD. 2005. Spatial variability analysis
9
of soil physical properties of alluvial soils. Soil Science Society of America Journal 69: 1-
10
11
14.
[4] Zhang S, Zhang X, Huffman T, Liu X, Yang J. 2011. Influence of topography and land
12
management on soil nutrients variability in Northeast China. Nutrient Cycle &
13
Agroecosystem 89: 427-438.
14
[5] Munkholm, L.J., A. Garbout, S.B. Hansen. 2013. Tillage effects on topsoil structural quality
15
assessed using X-ray CT, soil cores and visual soil evaluation. Soil & Tillage Research
16
128, 104–109.
17
18
19
[6] Franzluebbers, A.J., 2002. Soil organic matter stratification ratio as an indicator of soil
quality. Soil and Tillage Research 66, 95–106.
[7] Munkholm, L.J., Schjønning, P., Rasmussen, K.J., Tanderup, K., 2003. Spatial and temporal
20
effects of direct drilling on soil structure in the seedling environment. Soil and Tillage
21
Research 71, 163–173.
15
1
[8] Swarowsky, A., R.A. Dahlgren, K.W. Tate, J.W. Hopmans, A.T. O‟Geen. 2011 Catchment-
2
scale soil water dynamics in a Mediterranean-type oak woodland. Vadose Zone Journal,
3
10, 800-815.
4
5
6
7
8
9
[9] Webster R, Oliver MA. 1990. Statistical methods in soil and land Resource Survey. Oxford
University Press, Oxford University Press, Oxford, UK.
[10] Saldana A, Stein A, Zinck JA. 1998. Spatial variability of soil properties at different scales
within three terraces of the Henare River (Spain). Catena 33: 139-153.
[11] Wang ZM, Song KS, Zhang B, Liu DW, Li XY, Ren CY, Zhang SM, Lou L, Zhang CH.
2009. Spatial variability and affecting factors of soil nutrients in croplands of Northeast
10
China: a case study in Dehui county. P lant, Soil & Environment 55: 110-120.
11
[12] Webster R. 1985. Quantitative spatial analysis of soil in the field. In: Stewart, B.A. (Ed.),
12
Advances in Soil Science 3: 1-70.
13
[13] Pohlmann H. 1993. Geostatistical modeling of environmental data. Catena 20, 191-198.
14
[14] Cambardella CA, Moorman TB, Novak, JM, Parkin TB, Karlen DL, Turco RF, Konopka
15
AE. 1994. Field-scale variability of soil properties in central Iowa soils. Soil Science
16
Society of America Journal 58: 1501-1511.
17
18
19
[15] Ozgoz E. 2009. Long term conventional tillage effect on spatial variability of some soil
physical properties. Journal of Sustainable Agr iculture 33: 142-160.
[16] Nkongolo NV, Hatano R, and Kakembo V. 2010. Diffusivity models and greenhouse gases
20
fluxes from a forest, pasture, grassland and corn field in northern Hokkaido. Pedosphere
21
20(6):747-760. doi:10.1016/S1002-0160(10)60065-3.
16
1
[17] Ayoubi SH, Zamani SM, Khomali F. 2007. Spatial variability of some soil properties for
2
site-specific farming in northern Iran. International Journa l of Plant Production 2: 225-
3
236.
4
5
6
[18] Burgess TM, Webster R. 1980. Optimal Interpolation and isarithmic mapping of soil
properties: I. The variogram and punctual krigging . Journa l of Soil Science 31: 315-331.
[19] Parfitt JMB, Timm LC, Pauletto EA, Sousa RO, Castilhos DD, de Avila CL, Reckziegel
7
NL. 2009. Spatial variability of the chemical, physical and biological properties in
8
lowland cultivated with irrigated rice. Rev. Bra s. Cienc. Solo 33: 819-830.
9
10
[20] Grego CR, Vieira SR, Lourencao AL. 2006. Spatial distribution of Pseudaletia sequax
Franclemlont in triticale under no-till management. Science & Agr iculture 63: 321-327.
11
[21] Tesfahunegn GB, Tamane L, Vlek PLG. 2011. Catchment-scale spatial variability of soil
12
properties and implications on site-specific soil management in northern Ethiopia. Soil
13
& Tillage Research 117: 124-139.
14
15
16
[22] Grecu SJ, Kirkham MB, Kanemasu ET, Sweeney DW, Stone LR, Milliken GA. 1988. Root
growth in a claypan with a perennial-annual rotation. Soil Sci. Soc. Am. J. 52: 488-494
[23] Myers DB, Kitchen NR, Sudduth KA, Miles RJ, Sharp RE. 2007. Soybean root distribution
17
related to claypan soil properties and apparent soil electrical conductivity. Crop Sci. 47:
18
1498-1509
19
[24] Jiang P., Kitchen N.R., Anderson S.H., Sadler E.J., Sudduth K.A. 2008. Estimating palnt
20
available water using the simple inverse yield model for claypan landscapes. Agronomy
21
Journal. 100: 1-7.
22
23
[25] Utset A, Ruiz ME, Herrera J, Ponce de Leon D. 1998. A geostatistical method for soil
salinity sample site spacing. Geoderma 86: 143-151.
17
1
2
[26] Fu W, Tunney H, Zhang C. 2010. Spatial variation of soil nutrients in a dairy farm and its
implications for site-specific fertilizer application. Soil & Tillage Resea rch 106: 185-193.
3
[27] Lopez-Grandos F, Jurado-Exposito M, Atenciano S. Garcia-Ferrer A, De la Orden MS,
4
Garcia-Torres L. 2002. Spatial variability of Agricultural soil parameters in southern
5
Spain. Plant Soil 246: 97-105.
6
7
8
9
10
11
12
13
[28] Hill RL. 1990. Long-term conventional and no-tillage effects on selected soil physical
properties. Soil Science Society of America Journal 54: 161-166.
[29] Buschiazzo DE, Panigatti JL, Unger PW. 1998. Tillage effects on soil properties and crop
production in subhumid and semiarid Argentinean Pampas. Soil & Tillage Research 49:
105-116.
[30] Tsegaye T, Hill RL. 1998. Intensive tillage effects on spatial variab ility of soil physical
properties. Soil Science 163: 143-154.
[31] Gomez JA, Giraldez JV, Pastor M, Fereres E. 1999. Effects of tillage method on soil
14
physical properties, infiltration and yield in an olive orchard. Soil & Tillage Resea rch 52:
15
167-175.
16
17
18
19
20
18
1
2
Fig. 1. Study area (Lincoln University‟s Freeman farm) showing the plots
3
4
5
6
7
8
9
10
11
12
13
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Fig. 2. Spatial distribution of soil physical properties at four depths in a clay-loam soil
20
1
2
a) Soil Bulk Density
3
4
b) Gravimetric Water Content (GWC)
21
1
2
3
4
5
6
7
8
c) Total Pore Space (TPS)
d) Volumetric Water Content (VWC)
Fig. 3a, b, c, d. Variation of soil physical properties with depth
22
1
2
3
4
5
6
7
8
9
Table 1. Descriptive statistics for soil physical properties at four depths in a clay- loam soil
AFPS BDY
Diff.
GWC TPS
Tort
VAC
VWC
Depth 1, 0-10 cm
Mean
45.76
1.24
0.06
0.22
0.51
4.60
0.24
0.28
SD
10.61
0.11
0.03
0.03
0.04
1.50
0.07
0.04
C.V
23.18
8.99
56.29 15.25 8.36
32.60
29.64
15.26
Minimum
26.35
1.01
0.01
0.15
0.45
2.56
0.12
0.18
Median
45.71
1.24
0.06
0.22
0.52
4.10
0.24
0.28
Maximum
68.68
1.41
0.15
0.36
0.60
8.18
0.39
0.36
Skew
0.05
-0.24
0.78
1.03
0.22
0.89
0.25
-0.23
Kurtosis
-0.66 -0.76
0.04
3.56 -0.78 -0.06
-0.53
-0.60
Depth 2, 10-20 cm
Mean
26.54
1.47
0.02
0.21
0.42
12.46
0.12
0.31
SD
11.19
0.18
0.02
0.048 0.07
10.70
0.06
0.05
C.V
42.15 12.09 91.61 22.48 16.40 85.91
49.83
17.11
Minimum
7.73
1.15
0.00
0.08
0.19
4.11
0.02
0.17
Median
27.32
1.46
0.01
0.22
0.43
9.04
0.11
0.31
Maximum
50.01
2.07
0.06
0.32
0.55
60.85
0.24
0.43
Skew
0.10
1.21
0.88
-0.57 -1.20
2.89
0.30
-0.49
Kurtosis
-0.82
2.15
0.11
0.92
2.11
9.36
-0.87
0.83
Depth 3, 20-40 cm
Mean
42.35
1.20
0.06
0.26
0.53
4.72
0.23
0.30
SD
8.19
0.12
0.03
0.04
0.05
1.29
0.06
0.03
C.V
19.33
9.75
53.22 13.92 8.68
27.30
26.27
11.48
Minimum
24.69
0.96
0.02
0.15
0.41
2.71
0.12
0.23
Median
42.11
1.20
0.05
0.25
0.53
4.52
0.22
0.31
Maximum
61.59
1.51
0.14
0.37
0.62
8.01
0.37
0.42
Skew
0.24
0.30
1.38
0.20 -0.29
0.95
0.64
0.22
Kurtosis
0.29
0.08
1.69
2.20
0.04
0.75
0.53
1.82
Depth 4, 40-60 cm
Mean
39.34
1.18
0.05
0.28
0.54
4.94
0.21
0.32
SD
7.62
0.07
0.02
0.03
0.03
1.12
0.05
0.03
C.V
19.36
5.57
46.36 10.43 4.83
22.69
22.72
10.58
Minimum
25.25
1.04
0.02
0.21
0.48
3.11
0.13
0.23
Median
39.74
1.18
0.05
0.27
0.54
4.67
0.22
0.33
Maximum
57.56
1.32
0.10
0.36
0.59
7.53
0.32
0.40
Skew
0.382
0.27
0.76
0.25 -0.36
0.42
0.44
-0.41
Kurtosis
-0.44 -0.43 -0.10
0.84 -0.33 -0.56
-0.48
0.59
WFPS
54.24
10.61
19.56
31.32
54.29
73.65
-0.05
-0.66
73.46
11.19
15.23
49.99
72.68
92.27
-0.10
-0.82
57.65
8.19
14.20
38.41
57.90
75.31
-0.24
0.29
60.66
7.62
12.56
42.44
60.27
74.75
-0.38
-0.44
AFPS: Air filled pore space (%); BDY: Soil bulk density (gcm-3 ); Diff.: Relative gas diffusion coefficient (m2 s -1 ms); GWC: Gravimetric water content of soil (g.g -1 ); TPS: Total pore spaces (cm3 cm-3 ); Tort.: Pore tortuosity factor
(m.m-1 ); VAC: Vo lu metric air content (cm3 cm-3 ); VW C: Volu metric water content (cm3 cm-3 ); WFPS: Water filled
pore space (%).
2
23
1
2
Table 2. Variogram parameters for soil physical properties at four depths in a clay- loam soil
Depth(cm)
GWC
0-10
10-20
20-40
40-60
VWC
0-10
10-20
20-40
40-60
BDY
0-10
10-20
20-40
40-60
TPS
0-10
10-20
20-40
40-60
VAC
0-10
10-20
20-40
40-60
WFPS
0-10
10-20
20-40
40-60
AFPS
0-10
10-20
20-40
40-60
Diff.
0-10
10-20
20-40
40-60
Tort.
0-10
10-20
20-40
40-60
Model
Nugget(Co ) Sill(Co +C) Range(Ao )
R2
(C/Co+C)
DSD(%)
Exponential
Exponential
Spherical
Exponential
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.42
64.26
7.17
4.56
0.12
0.75
0.11
0.01
0.94
0.52
1.00
0.99
0.01
0.29
0.00
0.00
Exponential
Exponential
Exponential
Exponential
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.21
11.82
5.88
0.45
0.31
0.44
0.17
0.00
0.94
0.89
0.97
1.00
0.01
0.04
0.00
0.00
Linear
Exponential
Spherical
Linear
0.01
0.02
0.00
0.00
0.01
0.04
0.01
0.00
25.77
40.17
7.42
25.77
0.07
0.93
0.20
0.39
0.00
0.50
0.95
0.00
1.18
0.04
0.07
0.44
Linear
Exponential
Exponential
Linear
0.00
0.00
0.00
0.00
0.00
0.01
0.00
0.00
25.77
14.64
7.50
25.77
0.07
0.79
0.21
0.47
0.00
0.81
0.90
0.00
0.16
0.13
0.03
0.07
Linear
Exponential
Exponential
Linear
0.00
0.00
0.01
0.00
0.00
0.00
0.00
0.00
25.77
2.73
7.62
25.77
0.25
0.00
0.35
0.21
0.00
0.86
0.92
0.00
0.48
0.06
0.07
0.23
Linear
Spherical
Exponential
Linear
107.28
10.80
3.50
55.49
107.28
135.30
66.49
55.49
25.77
5.38
7.47
25.77
0.15
0.00
0.27
0.07
0.00
0.92
0.95
0.00
10727.60
1173.91
369.59
5549.30
Linear
Spherical
Exponential
Linear
107.28
10.80
3.50
55.49
107.28
135.30
66.49
55.49
25.77
5.38
7.47
25.77
0.15
0.00
0.27
0.07
0.00
0.92
0.95
0.00
10727.6
1173.91
369.588
5549.30
Linear
Exponential
Exponential
Linear
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
25.77
4.71
4.53
25.77
0.10
0.00
0.05
0.19
0.00
0.85
0.93
0.00
0.11
0.00
0.01
0.05
Linear
Exponential
Exponential
Linear
2.12
17.80
0.18
1.24
2.12
126.10
1.74
1.24
25.77
8.10
13.20
25.77
0.15
0.19
0.58
0.16
0.00
0.86
0.90
0.00
211.80
2072.20
19.44
123.68
24
1
2
3
DSD = Degree of spatial dependence: strong DSD (DSD ≤ 25%), moderate DSD (25 < DSD ≤ 75%),
weak DSD (DSD > 75%) according to Cambardella et al., (1994).
4
25