|
|
Site-specific
farming is gaining in popularity in the corn belt of the U.S.
A concern expressed by many Kentucky producers is whether the same
technology is profitable on their land. Site-specific management differs from
traditional practices in that soil fertility levels are geographically
referenced, and the frequency of sampling is substantially increased.
Many service providers are recommending sampling grid sizes as small as
0.4 ha (1.0 ac.). Fertilizer and
lime application rates are then adjusted as these inputs are applied to match
the production goals and existing fertility for that region or cell of the
field. Field-average techniques,
the traditional approach, treat all areas within a field the same--without
regard to variability. With the
field-average approach a representative fertility level is established by
pulling multiple soil cores that are composited into one sample. Laboratory analysis of this single sample is used to make
fertilizer recommendations for an entire field.
Kentucky
grain producers and agricultural services providers see the potential for the
site-specific management approach and are
concerned about the profitability. This
new approach will require investments in hardware and increased management
skills. Therefore, the work reported in this manuscript was initiated to
address these concerns. The
primary objective of the work was to assess the potential economics of
site-specific fertility management in Central Kentucky.
Specific objectives were:
1.
Detect existing fertility levels and the degree of variability within
fields in the Outer Bluegrass Region via extensive grid sampling.
2.
Model corn, wheat and soybean yield potential reductions for areas with
low fertility levels.
3.
Compare the economics of field-average versus site-specific nutrient
management practices.
The
subject of soil sampling and soil parameter variation within fields has long
been an important topic in the agronomic community. For example, Reed and Rigney (1946) studied the sampling
procedures required to determine various soil properties to within a given
limit of accuracy. They concluded
that greater precision is used in laboratory analysis and therefore
field-sampling techniques are the limiting factor.
Beckett and Webster (1971) presented a review of lateral variability of
soil properties. They found up to
one-half of the variance within a field may be present in as little as one
square meter of land. Bonmati, et
al. (1991) investigated the variability of urase, phosphatase, casein-hydrolysing
activities, organic carbon and total nitrogen.
Strong correlations were found between total nitrogen and organic
carbon. Similar correlations were
found between phosphatase and total nitrogen, and with total carbon.
Recognition
of this variability and the advent of the Global Positioning Systems (GPS) has
prompted many researchers, service providers, and grain producers to consider
managing this variability. The
missing link of this scenario was the cost-effective availability of
positioning services such as GPS. Carr,
et al. (1991) suggested "farming soils, not fields."
Their study was initiated to measure crop yield differences between
contrasting soils within a field, and to compare the economics varying
nutrient application by contrasting soils with the traditional practice of
field -average application. Returns to farming soils were generally greater than when
farming fields. However, the
researchers noted that it was essential to establish appropriate crop yield
goals, conduct accurate soil tests, and utilize reliable fertilizer
recommendations to generate greater returns when managing nutrients.
Fixen
and Reetz (1995) used a model termed PKMAN, developed by the Potash and
Phosphate Institute of Norcross, Georgia, to estimate long-term profitability
of site-specific nutrient management. This
model included elements such as yield potential, crop price, nutrient uptake
by crop, crop removal, nutrient recovery, acceptable return on investment,
interest rate, and others. They
concluded from this work that site-specific interpretations should consider
both soil and producer characteristics, and that continued information
feedback and soil-test calibration will cause site-specific management to
continuously improve over time.
The subject of selecting an optimum grid or cell size to describe variability
is of interest to many potential users of this technology. The following examples address cell size selection from a
statistical perspective. Han,
et al. (1992a) used mean correlation distance (MCD) to specify the upper limit
of cell size for site-specific management of agricultural lands.
The MCD was obtained by fitting either a linear or spherical model to
semivariograms, a common geostatistical approach.
Han, et al. (1992b) also used the concept of data blocking procedures
to convert soil sample data into a best-fit data set.
A nonparametric statistical technique was used to select the best value
to represent the soil property of a particular cell.
They compared the results of this approach with the more traditional
approach of kriging, and found
their new approach to produce similar estimation errors.
However, the new procedure was much faster.
Four fields located in Shelby County, Kentucky were selected to represent
grain producing lands of the Outer Bluegrass physiographic region.
These fields were located in the northern portion of the county and
consisted primarily of Lowell-Nicholson soils.
The relief for this area is described as undulating ridges and rolling
side slopes. The Lowell-Nicholson
soils are deep and well drained to moderately drained.
Soil types include Lowell-C, Nicholson-B, Shelbyville-B and C, and
Nolin.
During the 1995 cropping season the fields were grid sampled to assess the
variability of soil parameters affecting inputs to crop production.
A cell size of 36.6 m (120 ft.) was selected in part because of the
18.3 m (60 ft.) spread width of commercial fertilizer application equipment.
A second limiting factor was the availability of resources. This cell size resulted in seven point samples per hectare (3
samples/ac). Gridding of each
field was accomplished by using a fiberglass tape, range poles and flags.
The origin of the grid was established by located a flag 18.3 m (120
ft.) from the field boundaries in one corner.
Composite
soil samples were collected at each grid point. Five soil cores were extracted to a depth of 17 cm (6.7 in.).
The five cores were collected equidistance round the periphery of a 9 m
(15 ft.) diameter circle, centered at the grid point.
This sampling technique minimized errors associated with true point
sampling practices by averaging soil parameters over an area that coincided
with grid point.
Soil samples were air dried and prepared for analysis. Organic matter content was determined using Carbon
Determinator methodology. Samples
were then submitted to the Kentucky Agricultural Experiment Station, Division
of Regulatory Services for analysis of pH, phosphorus and potassium.
Regulatory Services uses the Mehlich III extraction procedure to
determine phosphorus and potassium content.
A buffered pH was also determined using a modification of the
Shoemaker, McLean and Pratt method. The
buffered pH was reported for all samples having a water pH below 7.0 for the
Lexington laboratory.
The
justification for this work was to assess the potential for site-specific
management of croplands in Central Kentucky.
The variability that exists suggests that variable-rate application of
fertilizer and chemicals may result in reduced production costs and increased
yield. Quantification of
the cost-effectiveness of site-specific management is a major concern of
practitioners of this technology. Long-term
there must be productivity gains to offset increased management costs.
In the short-term producers will demand evidence of cost effectiveness. Therefore, the intent of this manuscript is to assess the
potential savings accrued by not applying nutrients or adjusting pH in those
regions of a field where levels are appropriate.
Soil
testing revealed current levels of available phosphorus and exchangeable
potassium, and the acidity/alkalinity or pH of the soil.
Fertility management practices for phosphate and potassium are
different from nitrogen in that the former nutrients are relatively stable,
and do not volatilize or move within the soil profile.
Changes in fertility levels for these nutrients are gradual.
Management practices must be followed for several years to change
levels within the soil. While pH
changes that result from liming may be rapid when using hydrated lime, for
limestone the reaction time may require up to four years.
Fertilizer and liming recommendations used in this investigation were obtained
from University of Kentucky (1996). For
corn, soybeans and wheat this publication recommends a pH of 6.4.
Table 1 is a summary of liming rates for agricultural limestone as a
function of water pH and buffered pH. These
rates are reduced 33% for hydrated lime.
Tables 2 and 3 provide recommended application rates for phosphorus and
potassium, respectively, for a corn and wheat with double-cropped soybean
rotation. Recommendations are made in kg/ha of P2O5
and K2O, on an oxide basis.
Yield reduction resulting from low nutrient levels is perhaps the most
difficult issue to address. The
mobility of nitrogen and potential in some soils for denitrofication to occur
encourages producers to apply nitrogen at the stage of growth when the plant
can most effectively utilize this nutrient.
Split application on corn and wheat is a practice that is growing in
acceptance. In addition P and K
levels have been found to limit crop growth.
OCES (1988) provides two figures that show the relationships between
relative yield and available P and exchangeable K, respectively.
The curves from these two figures are reproduced in Figures 1 and 2.
Jones, et al. (1991) provides a function for describing the effect of
soil-P buffering capacity on residual fertilizer-P effectiveness in which the
soil response curve is exponential,
|
|
where
f(x)
is the yield response,
a is the maximum expected yield, b
is the relative response to P, c
is the curvature coefficient, and x
represents the level of available P in the soil.
This equation was recast in a slightly different form and then used to
describe the data contained in Figures 1 and 2.
Available-P was expressed in units of kg/ha.
R-squared values for this relationship fit to the wheat, corn, and
soybeans data in Figure 1 were 0.9912, 0.9981 and 0.9881, respectively.
Applying the same relationship to K yielded similar results.
Here the R-squared value for corn and wheat was 0.9985, and for
soybeans was 0.9987. These
relationships were used to estimate yield reductions for those areas of the
field where one or the other nutrient level was limiting.
To
assess the economics of variable-rate fertilizer and lime application, the
following assumptions were used in the analysis:
Crop rotation for the study area is corn followed by wheat with double
cropped soybean for four growing seasons.
Grid sampling of the soil will be conducted once every four years.
Adjustment of soil pH will occur in the first year via the application
of agricultural lime. Application
rates will coincide with the recommendations of University of Kentucky (1996)
Phosphorus and potassium application rates will be determined using
University of Kentucky (1996). For
the double crop years the phosphate recommendation for small grains will be
used and the potash rates will come from the soybean recommendations.
The cost of soil sample collection and analysis is $7.00 per sample.
Agricultural lime will cost $14.30 per metric ton ($13.00 per ton).
Murate of potash (0-0-60) will serve as the source of potassium and
will cost $144.40 per metric ton ($131.00 per ton).
Diammonium phosphate (18-46-0) will serve as the source of phosphorus
and will cost $196.20 per metric ton ($178.00 per ton).
Constant-rate fertilizer and lime application will cost $8.72 per
hectare ($3.50 per acre).
Variable-rate fertilizer and lime application will cost $17.44 per
hectare ($7.00 per acre).
Field-average conditions will be determined by averaging all grid point
sample results together, and using the mean value to represent existing
fertility levels within a field.
Field-average application will be compared to variable-rate application
using a grid of 36.6m (120 ft.).
Yield reductions coincide with the limiting nutrient level at each grid
point (i.e., P or K).
The value of grain will be $165/Mg, $177/Mg, and $239/Mg ($4.50/bu, $4.50/bu, and $6.50/bu) for wheat, corn and soybeans, respectively.
Maximum yield potential for each field will be determined by selecting
the predominant soil type and slope of a particular field, and then reviewing
the yield potential as listed in the USDA-SCS Soil Survey of Shelby County,
Kentucky (USDA-SCS, 1990).
The maximum yield potential for soybeans as published in USDA-SCS (1980) is reduced by 33% to represent double-cropped soybean following wheat.
Yield reductions will decrease by 25% each year for the site-specific management approach as soil fertility increases in deficient areas by the guidelines of University of Kentucky (1996).
Soil
test summary results are presented in Table 4 for each field.
For a field-average management approach the test result averages
suggest that pH is not a major concern in any of the fields.
In all cases except Field 22, the pH is at or above the preferred 6.4.
For Field 22 lime application was determined using Table
1, a rate of
0.91 Mg/ha (1.0 ton/ac). The
available phosphorus levels were well above the 67 kg/ha (60 lb/ac) level,
thereby alleviating the concern for the need to add phosphate.
For Fields 23 and 25 the exchangeable
potassium was well below the high fertility level as suggested by
University of Kentucky (1996). Using Tables 2 and 3, potash application levels of 67 kg/ha
(60 lb/ac) and 45 kg/ha (40 lb/ac) were recommended for Fields 23 and 25,
respectively, for corn. For years
when wheat was followed by double cropped soybeans the application levels were
reduced to 56 kg/ha (50 lb/ac) and 34 kg/ha (30 lb/ac), respectively.
Fertilizer and lime recommendations made for site-specific management on a
0.134 ha (0.331 ac) cell basis revealed a slightly different need. Tables 5 and 6 were generated to summarize total lime and
fertilizer qualities under either management style. Entries in Table 5 are sparse when compared with
Table 6.
Table 6 indicates the need for adjustment of pH and phosphorus and
potassium levels in all fields. Fields 23, 25 and 26 will require pH adjustments with the
sight-specific approach. Similarly
all four fields will require the application of phosphate.
Although, it might be argued that application of 0.34 Mg (0.38 tons) to
Field 22 is so minor that it is not economically feasible.
In general the quantities of fertilizer to be applied are about the
same under either management approach, if application is warranted under the
field-average approach. Most
notable is the site-specific approach for those fields where the field-average
management approach would suggest no additional nutrients.
For example, the potash recommendation for Field 22 calls for no
fertilizer application for the four years.
The site-specific approach suggests applying a total of 9.8 Mg (10.82
tons) over four years. The
overall difference between the management approaches for all four fields
results in the additional application of 32.3 Mg (35.6 tons) of lime, 11.3 Mg
(12.5 tons) of phosphate, and 14.3 Mg (15.8 tons) of potash.
Perhaps the most important question is if the return on investment in
site-specific management practices is justification for increased management
costs. Table 7 summarizes the
returns for both management practices. The
site-specific management approach can be expected to return $13,218.20 more in
the way of increased yield over four years.
Notably the increased return for Field 22 and 23 at $4675.53 and
$5702.58, respectively. Contributing
to the returns for Field 22 was the increase in potash application whereas
most of the return on Field 23 came from the application of phosphate.
Under the field-average management approach elevated phosphorus levels
in the back of the field masked deficiencies in the front.
Approximately one-half to two-thirds of this field had low phosphorus
levels.
When looking at the overall economics a much different conclusion may be drawn
(Table 8). To implement site-specific
management using a 36.7m (120ft) grid will cost the grain producer $2760.14 in
potential profit over the 4-year production cycle.
This is compared to a $511.68 loss using the field average technique.
For Fields 22, 25, and 26, neither management approach results in a
substantial increase in profit. For
Field 23 the situation is much different.
In this instance there is the potential to increase profits by $797.94
and $1980.48, respectively for field-average and site-specific management
approaches.
Results of
this investigation suggest that site-specific management offers potential in
situations where field-average soil testing indicates the need for adjustment
of soil fertility, and where there is a high degree of variability for the
soil parameters of interest.
Figures Available Soon!
Figure 1:
Available phosphorus levels in Field 23 at the Worth and Dee Ellis Farm
in Shelby County, Kentucky during 1995.
Figure 2:
Exchangeable potassium levels in Field 23 at the Worth and Dee Ellis Farm
in Shelby County, Kentucky during 1995.
Figure 3:
Wheat, corn and soybean response to available phosphorus (adapted from
Miller, et al., 1988).
Figure 4:
Wheat, corn and soybean response to exchangeable potassium (adapted from
Miller, et al., 1988).
Figure 5:
Corn yield potential for Field 23 at the Worth and Dee Ellis Farm in
Shelby County, Kentucky during 1995.
Figure 6:
Wheat yield potential for Field 23 at the Worth and Dee Ellis Farm in
Shelby County, Kentucky during 1995.
Conclusion
1.
Site-specific management of soil fertility has the potential to be
profitable in the Outer Bluegrass Region where field-average soil sampling
indicates the need for fertility level adjustments.
2.
A high degree of variability for some soil parameters may mask
potential production deficiencies where high levels skew the mean for that
parameter. In these instances
site-specific management is more appropriate.
3.
Grid sampling at higher resolutions than current practices dictate may
be practical and profitable.
Conclusions reached in this manuscript were based on the assumption that
phosphorus or potassium was the limiting nutrient, and that yield reductions
were consistent with the data presented by Miller et.al. (1988). Yield monitoring has been implemented for Fields 22, 23, 25,
and 26. This data, along with
verification of VRT application of lime and fertilizer, will be used to adjust
the response curves for yield potential to better represent the soils of the
Outer Bluegrass Region.
Bonmati, M., B. Ceccanti, and P. Nanniperi.
1991. Spatial variability
of phosphateas, urease, protease, organic carbon and total nitrogen in soil.
Soil Biology and Biochemistry.
Pergamon Press. 23(4): 391-396.
Carr, P. M., G. R. Carlson, J. S. Jacobsen, G. A. Nielsen, and E. O.
Skogley. 1991.
Farming soils, not fields: a strategy for increasing fertilizer
profitability. Journal
of Production Agriculture.
4(1): 57-61.
Fixen, P.E. and H.F. Reetz, Jr.
1995. Site-specific soil
test interpretation incorporating soil and farmer characteristics.
Appearing in the Proceedings of Site-Specific Management for Agricultural Systems.
Edited by P.C. Robert, R.H. Rust, and W.E. Larson.
Minneapolis, Minnesota. 27-30
March.
Han S., C.E. Goering, J.W. Hummel, and M.D. Cahn.
1992a. Selection of cell
size for site-specific crop management. An
ASAE Meeting Presentation No. 927007. St.
Joseph, Michigan.
Han S., C. E. Goering, J. W. Hummel, and M. D. Cahn.
1992b. Blocking of spatial
soil data for site-specific crop management.
An ASAE Meeting Presentation No. 927008.
St. Joseph, Michigan.
Jones, C.A., Sharpley, A.N., and J.R. Williams. 1991.
Modeling phosphorus dynamics in the soil-plant systems.
Modeling Plant and Soil Systems
No. 31. Edited by
J. Hanks and J.T. Ritchie. ASA-CSSA-SSSA:
Madison, Wisconsin.
Miller, F., J. Beuerlein, J. Johnson, M. Loux, J. Street, W. Schmidt
and J. Underwood. 1988.
Ohio Agronomy Guide, 12th Edition.
Ohio Cooperative Extension Service, Bulletin 472, Ohio State
University.
Reed, J.F., and J.A. Rigney. Soil
sampling from fields of uniform and nonuniform appearance and soil types.
Journal of the American Society of Agronomy.
pp. 26-40.
University of Kentucky. 1996.
1996-1997 Lime and Fertilizer Recommendations.
Cooperative Extension Service, University of Kentucky, College of
Agriculture.
USDA/SCS. 1980.
Soil Survey of Shelby County,
Kentucky. United States
Department of Agriculture Soil Conservation Service in cooperation with the
Kentucky Agricultural Experiment Station and the Department for Natural
Resources and Environmental Protection.
