Assessment
of Remote Sensing For Implementation of Precision Tillage
L.G. Wells, S.A. Shearer, J.P. Fulton and L.W. Murdock
L.G. Wells, S.A. Shearer, J.P. Fulton and L.W. Murdock. 2000. Assessment of Remote Sensing for Implementation of Precision Tillage. ASAE Paper No. 001084. Annual International Meeting, Midwest Express Center, Milwaukee, Wisconsin, July 9-12.
ABSTRACT
A yield map of a field in central Kentucky indicated the likelihood of
detrimental soil compaction. The
field was divided into 1-acre grid cells and 15 cone penetrometer readings were
recorded for each cell. Maximum
average cone index (CI) readings versus depth above an effective tillage limit
of 400 mm were then calculated. Average
reflectance of the 7-band LANDSAT LS-5 image was determined for each grid cell
in the field. Infrared reflectance
was compared to average maximum CI, average CI, and corn yield by linear
correlation.
INTRODUCTION
Yield maps of field are essential elements in the application of the crop production methodology known as precision agriculture. Such maps provide information concerning the variation of crop yield within a field that farmers interpret in conjunction with various data that has been collected which may include soil fertility, texture, depth, etc.
An important factor that may affect crop yield is the detrimental compaction of soil by equipment traffic and tillage operations. When a farmer suspects yield-limiting compaction, remediation by tillage is typically considered. However, deep tillage or subsoiling is energy intensive and expensive and is most effective when compaction is severe. Thus, the development of a methodology of applying site-specific deep tillage offers a potential means of increasing crop yield and minimizing tillage cost.
The
soil cone penetrometer (ASAE, 1999) has been widely used to characterize soil
strength. This device measures
penetration resistance as pressure versus depth and is relatively easy to use.
However, the resulting cone index (CI) is highly variable in the field
requiring several measurements to characterize a soil zone. Collection of
numerous measurements of soil cone index in field grid cells can be expensive
and time-consuming.
One way of limiting measurement of soil cone index would be to compare
areas of low yield to known high-traffic areas.
The use of remote sensing imagery may also provide important information
that can reduce or eliminate the need for measuring soil cone index.
The objectives of this study were to characterize the state of soil
compaction in a field using soil cone index and to determine if remotely sensed
images could identify compacted regions that require deep tillage.
Compaction of soil by machine traffic
often produces adverse changes in soil physical properties that result in
reduced crop yield. Soil cone index
(ASAE, 1999) is a measure of soil resistance to penetration by a small probe and
therefore indicative of both compactness and strength.
Numerous researchers have reported reduced crop yields associated with
high values of soil cone index (Carter et al. (1965), Taylor et al. (1964),
Douglas and McKeyes (1983), Al-Adawi and Reed (1996)).
Tillage is the mechanical disturbance of soil that reduces strength and
bulk density and thereby alleviates compaction.
Normal tillage operations do not disturb soil deeper than approximately
20-25 cm and, in the case of no-till crop production, there is generally no
disturbance. When traffic
compaction occurs below the normal depth of tillage, deep tillage or subsoiling
is required.
Al-Awadi and Reed (1996) reported reduction
in soil cone index from approximately 1.5 to 1.0 MPa at the 40-45 cm depth
attributable to subsoiling in a silty clay loam soil.
Chaudhary et al. (1985) found that subsoiling reduced penetration
resistance and increased grain yield in a loamy sand soil.
Blancher et al. (1978) reported that root growth decreased as cone index
increased above 1 MPa and virtually stopped at
2 Mpa. Foshee et al. (1998)
also suggested a cone index value of 2 MPa to characterize severe compaction.
Raper
et.al. (1998) reported that annual in-row subsoiling could circumvent the
detrimental effects of traffic in cotton production, allowing roots to reach
moisture in deep zones. A similar
concept of “precision tillage” was described by Carter et al. (1965) wherein
tillage depth was precisely specified to reach and disturb a compacted
“pan”. As the concept of
GPS-based precision agriculture has gained acceptance, the idea of precision
tillage has evolved to include real-time control of a “smart” tillage tool (Scarlett
et al. (1997)) and variable-depth deep tillage (Raper, 1999).
Precision deep tillage is attractive from the
standpoint of eliminating unnecessary tillage.
Evans et al. (1996) reported no improvement of corn yield resulting from
subsoiling and suggest that it be used only when compaction is evident.
Threadgill (1982) showed that the loosening effect of subsoiling was
temporary, suggesting that regular tillage would be required to achieve
beneficial results as indicated by Raper et al. (1998).
In
order to apply site-specific deep tillage cost-effectively, a method is needed
to minimize cost associated with identifying compacted zones in fields.
Soil cone index is perhaps the least complicated and easiest measurement
of soil strength and density to acquire. However,
the collection, recording and analysis of 10, 15 or more cone penetrometer
readings per field grid cell using equipment such as described by Raper et al.
(1999) would involve substantial effort and expense.
The use of supplementary information, such as remotely sensed images, in
conjunction with yield maps and farmers’ knowledge of fields could lead to
more efficient and cost-effective implementation of precision tillage.
Remote sensing by instruments orbiting the earth is
recognized as the only practical method for gathering spatially distributed data
for watershed analysis (Engman, 1995). Schmugge (1983) described the use of
microwave reflectance to sense soil moisture near the surface.
Entekhabi et al. (1994) described a method whereby remotely sensed
observations of multispectral irradiance could be used to determine soil
moisture and temperature profiles. Synthetic
aperture radar (SAR) provides a remote sensing tool to measure soil moisture
that is not compromised by cloud cover.
Even though remote sensing methods are generally imprecise in quantifying soil moisture content, they may be useful in identifying areas of poor drainage that may indicate compacted soil. When this information is combined with crop yield maps and known field traffic patterns, areas of suspected soil compaction can be identified and, if necessary, verified by a cone penetrometer before applying precision deep tillage.
MATERIAL AND METHODS
Figure 1 shows a 1999 corn yield map from Field no. 31 that is part of the Worth and Dee Ellis Farms in Shelby County, Kentucky. The average yield in this field was approximately 50 bu/ac when yield in other nearby fields was 90-100 bu/ac. The field was divided into 0.41 ha grid cells for site-specific application of nutrients and for monitoring crop yield. Prior to being planted in corn, this field was in forage production for several years and was subjected to concentrated traffic by heavy equipment. This knowledge of field history, along with indication of yield suppression, led to the assessment of soil compaction using a cone penetrometer.
Figure 2 shows a multiple-probe, tractor-mounted soil cone penetrometer that was fabricated to assess compaction in fields for the purpose of implementing precision deep tillage. The device consists of five standard penetrometer shafts and cone as specified by ASAE Std. S313.3 (ASAE, 1999). The five penetrometers are forced in the soil profile simultaneously by a single hydraulic cylinder with a stroke length of 75 cm. Allowing for clearance for the penetrometer tips in the extracted position, the effective depth of penetrometer readings was approximately 68 cm. Hydraulic flow rate was adjusted to attain the recommended probe penetration rate of 183 cm/min.
Each shaft is mounted to a 4.45 kN
load cell, and the five load cells are monitored by a Campbell Scientific Data
Logger. Millivolt output of the
five cells is recorded at 4 Hz and
recorded in memory. The
penetrometer is equipped with a programmable logic controller and a bypass valve
that prevents overloading and damage to the load cells when any load cell
reading exceeds 2.25 kN.
Triplicate measurements were recorded in each grid cell using the tractor mounted penetrometer equipped with GPS navigation. The measurements were taken as the tractor was driven across each grid cell, resulting in a total of 15 penetrometer measurements per grid cell. The system can record penetration resistance in depth increments of approximately 0.75 cm.
A LANDSAT-5 image of Ellis Field 31 was acquired to illustrate the use of remote sensing to identify areas of possible soil compaction. The image was acquired on May, 1999 and consisted of seven bands of measured reflectance from the surface, from visible to near infrared. The image was recorded approximately one week after 3.5 cm of rainfall was recorded near the field, thus, the image should be indicative of variation of surface soil moisture content caused by slower drainage corresponding to soil compaction. The field was planted in wheat that was to be harvested in early June. Reflectance data from the LANDSAT-5 image was mapped onto Ellis Field 31 using ARCINFO.
RESULTS
AND DISCUSSION
Figure
3 shows the maximum average cone index measured in each 0.4 ha field grid cell
at or above a depth of 45 cm, which was considered to be the maximum depth of
subsoiling. The values range from
1.07 to 1.97 MPa, with the field average being 1.51 MPa.
These maxima occurred as shallow as 7 cm and as deep as 45 cm. These
results indicate the potential applicability of applying tillage at variable
depths over such a field. A key determination would be the threshold value of
maximum CI below which tillage would not be applied.
Figure 4 shows the average reflectance in each grid cell of Band 7 of the LANDSAT-5 image measured over the entire profile depth in each grid cell. A visual comparison of figs. 2 and 4 appear to indicate a correspondence of infrared reflectance and yield. A similar qualitative correspondence appears to exist between yield and average CI as indicated by comparing figs. 2 and 3.
Table 1 presents the results of linear correlation determinations between the various indices associated with the cone index measurements and the LANDSAT reflectance data as recorded in the grid cells of Ellis Field 31. The strongest correlation was between yield and average CI measured in the total profile and in the tillable zone. Correlation between the reflectance data and the CI indices were disappointing. While much better correlation must be established before remote sensing measurements can be used to with confidence to identify areas of likely compaction, other strategies and types of images should be investigated due to the substantial cost savings that would result.
Additional strategies should be explored in a search for more useful remote imagery for identifying areas of potential soil compaction. The most opportune time to acquire such images would be during late autumn and winter when fields are without vegetation. Such conditions should be optimal for remote measurement of soil moisture. If such a measurement could be acquired within a short time after a substantial rainfall event, differences in soil moisture may indicate areas of restricted drainage that can correspond to soil compaction.
Other types of imagery such as passive microwave reflectance and synthetic aperture radar (SAR) should be investigated. The large area coverage of remote satellite images may cooperative use by multiple farmer feasible. Other strategies such as acquiring images with remotely guided miniature aircraft may also be possible.
REFERENCES
Al-Adawi, S.S. and R.C.
Reed. 1996. Compaction and subsoiling effects on corn and soybean yields and
soil physical properties. Trans. of ASAE 39(5): 1641-1649.
ASAE Standard. 1999. ASAE
S313.3: Soil cone penetrometer, St. Joseph, MI.
Blancher, R.W., C.R.
Edmonds and J.M. Bradford. 1978. Root growth in cores formed from fragipan and
B2 horizons of Hobson soil. Soil Sci. Soc. Amer. J. 42:437-440.
Carter, L.M. and J.R.
Tavernetti. 1968. Influence of precision tillage and soil compaction on cotton
yields. Trans. of ASAE 11: 65-67,73.
Chaudhary, M.R., P.R.
Gajri, S.S. Prihar and R. Khera. 1985. Effect of deep tillage on soil physical
properties and maize yields on coarse textured soils. Soil Tillage Res. 6:31-44.
Douglas, E. and E. McKyes.
1983. Tillage practices related to limiting plant growth factors and crop
yields. Canadian Agr. Eng. 25 (1):47-55.
Engman, E.T. 1995. The use
of remote sensing data in watershed research. J. Soil and Water Conservation
50(5): 438-441.
Evans, S.D., M.J.
Lindstrom, W.B. Voorhees, J.F. Moncrief and G.A. Nelson. 1996. Effect of
subsoiling and subsequent tillage on soil bulk density, soil moisture, and corn
yield. Soil Tillage Research 38: 35-46.
Foshee, W.G., R.L.Raper,
W.D. Goff and M.G. Patterson. 1997. Orchard floor practices affect soil
compaction around young pecan trees. Amer. Soc. Hort. Sci. 32(5): 871-873.
Raper, R.L. 199.
Subsoiling considerations for site-specific tillage. ASAE paper no. 991058, St.
Joseph, MI.
Raper, R.L., D.W. Reeves
and E.C. Burt. 1998. Using in-row subsoiling to minimize compaction caused by
traffic. J. Cotton Sci. (The Cotton Foundation) 2(3):130-135.
Raper, R.L., B.H. Washington and J.D. Jarrell. 1999. A tractor-mounted multiple-probe soil cone penetrometer. Applied Engr. in Agr.
Scarlett, A.J., J.C. Lowe
and D.A. Semple. 1997. Precision tillage: in-field real-time control of seedbed
quality. Procs. of 1st European Conf. on Precision Agr. Pp 503-510.
Schmugge, T.J. 1983. Remote sensing of soil moisture: recent advances. IEEE Trans. on Geosci. And Remote Sensing GE-21:336-344.
Taylor, H.M., L.F. Locke
and J.E. Box, Jr. 1964. Pans in Southern Great Plains soils. III. Their effect
on yield of cotton and grain sorghum. Agron. J. 56:542-545.
Threadgill, E.D. 1982.
Residual tillage effects as determined by cone index. Trans. of ASAE 25(4): 859-
863.