Precision Agriculture and Site-Specific Management: Current Status and Emerging Policy Issues

CRS Report for Congress
Precision Agriculture and Site-Specific
Issues
August 7, 2000
Tadlock Cowan
Resources, Science, and Industry Division


Congressional Research Service ˜ The Library of Congress

Precision Agriculture and Site-Specific Management:
Current Status and Emerging Policy Issues
Summary
Precision agriculture (PA) is a suite of information technologies that can support
a farm-based and site-specific crop management system in agricultural production.
PA is not a single technology or farming system, but rather a cluster of different
techniques. PA uses advanced information technologies to identify and to evaluate
temporal and spatial variation in cropland. Rather than apply inputs of production,
e.g., fertilizers, pesticides, seed, or water, uniformly across a field, a grower using PA
techniques can apply inputs more efficiently based on the biophysical variability of
different areas of a field. Proponents argue that PA can increase the yield potential
of a farm while reducing the costs from inefficiently applying inputs. Further, by
applying fertilizers and pesticides in more efficient amounts, proponents argue that
farmers would reduce the environmental effects of over-application. Because PA is
in the early stages of adoption, however, these benefits have been difficult to measure
to date. Returns to investment in PA have also been mixed according to available
research. Questions persist about the appropriate scale of analysis and measurement
needed to validly assess the potential benefits of PA. Adopters of PA to date have
disproportionately been larger corn, soybean, and wheat producers, which may be
partly attributable to the early commercial availability of PA technology, such as yield
monitors for these crops; but no hard evidence has emerged that PA technologies are
size-biased or crop-specific. PA appears to offer the greatest future potential benefits
where a variety of inputs are used and input costs are high, sub-field spatial and/or
temporal variability is high, and adverse environmental effects, especially on water
resources, must be reduced.
Several bills have been introduced in recent Congresses to increase support for
public research on PA development. Several public and land-grant universities have
also initiated new programs in PA technology and development; additional research
and development are supported by the Agricultural Research Service. In contrast to
many previous U.S. technological innovations in agriculture, however, PA is currently
being developed and promoted largely by private companies. Further diffusion of PA
technologies could raise several public policy concerns that the Congress may wish
to address. These include questions about the role of public research, education, and
extension services; future effects on the organization of production; intellectual
property issues associated with ownership and control of the agronomic databases
created for individual farms; and other public or appropriate public-private roles to
help growers gain access to and achieve some of PA’s potential benefits.



Contents
Introduction ................................................ 1
Socioeconomic Factors in Adoption and Use of PA..................3
Technological Costs and Development Trends..................3
Socioeconomic Aspects in Adoption and Diffusion of PA..........5
Environmental Aspects of Precision Agriculture.....................9
Current Status of Precision Agriculture at the Federal and State Levels..12
Congressional Interest.......................................13
Role of Public Research and Education......................13
PA’s Effects on Farm Structure............................14
Data Ownership and Control..............................15
Other Public and Public-Private Roles in PA......................17
Conclusion ................................................ 18
Glossary ................................................. 20
Base Map.............................................20
Choropleth Map........................................20
Crop Scouting.........................................20
Differential Correction...................................20
Digital Elevation Models (DEMs)..........................20
Digital Soil Mapping....................................20
Expert System.........................................20
Geographical Information Systems (GIS).....................21
Geo-reference System...................................21
Global Positioning System (GPS)...........................21
Ground-based Sensors...................................21
Point Sampling.........................................21
Precision Agriculture (PA)................................21
Remote Sensing........................................21
Variable Rate Application Technologies (VRT)................21
Yield Monitoring Systems................................22
Selective Bibliography on Precision Agriculture for Further Reading....22
Overviews ............................................ 22
Environmental Issues....................................22
Socioeconomic Issues...................................22
Technological Development..............................23
Precision Agriculture Web sites............................23



Precision Agriculture and Site-Specific
Management: Current Status and Emerging
Policy Issues1
Introduction
Over the past decade, information technologies have fundamentally transformed
many industrial areas. These technologies, undergirding advanced
telecommunications, data acquisition and information processing, and commerce,
have become core elements for many contemporary production and marketing
processes. Not only have they permitted the modification of existing activities, they
have also become the catalyst for fundamentally redesigning production processes and
for creating original products and processes. The manufacturing and service sectors
already have seen significant changes to their production processes through
information technologies, while the agricultural sector has benefitted less.
Agriculture’s reliance on information technology, however, is beginning to change
rapidly, with many promised benefits according to proponents.
Rapid technological change always has been a central characteristic of U.S.
agriculture.2 Growing competitive pressures stemming from international agricultural
markets, improving yields, and new environmental concerns lead farmers to seek and
adopt technological innovations. The emergence of new technologies in conjunction
with new demands on the agriculture sector from a range of interest groups are
setting the broad context for further changes in U.S. agriculture. Along with changes
in the structure of agricultural production (e.g., increased contractual arrangements,
greater concentration in some sectors, changing government policies and regulations),
an increased acceptance and reliance on information technologies by farmers is also
altering U.S. agricultural production.
Expectations that farmers would rapidly adopt computers, for example, are now
beginning to be realized. In 1999, computers were owned or leased by approximately
40% of U.S. growers, up from 31% in 1997.3 The predominant use initially was for
financial applications, i.e, bookkeeping, business management. In 1999, 29% of U.S.
farms had Internet access compared to 13% in 1997. Farms with sales greater than
$250,000 per year, which are considered to be commercial or full-time farms, had


1 Prepared under supervision of Jeffery Zinn, Senior Analyst in Natural Resources Policy
2 Cochrane, W.W. Development of American Agriculture. Minneapolis: University of
Minnesota Press, 1979.
3 USDA. National Agricultural Statistics Service. “Farm computer usage: access, ownership,
and use by state and U.S., 1997 and 1999.

double the rate of computer ownership, business management use, and Internet access
relative to farms with sales of less than $250,000 per year. Farms with sales greater
than $250,000 also account for the great majority of total U.S. agricultural output,
so that the bulk of agricultural output today is likely from farms using computers in
some aspect of their operation. Growth in adoption of information technology for the
agricultural sector as a whole is now consistent with other sectors – approximately
3% per year. The largest impact of computers in agriculture until recently has been
in the support industries, but producers are increasingly finding potential benefits from
computer technologies, including precision agriculture.
Precision agriculture (PA) (also called site-specific crop production, prescription
farming, and soil-specific crop management ) refers to a suite of technologies that use
sensing and geo-referencing innovations to apply more precise inputs based on a
field’s biophysical variability. The objective of PA techniques is to quantify and
manage the spatial and temporal variability existing in virtually every field. This is
accomplished by remote sensing technology, by local sensors, or by using the
integrated data in a geographic information system (GIS) as the basis for identifying
particular management zones that may require different amounts or types of inputs.
How a farmer might use PA:
Growers might typically adopt a PA system using GIS software to permit growers to
capture, store, manipulate, analyze, and display a wide variety of data referenced to
the specific points on the Earth’s surface determined by the Global Positioning System
(GPS) or other remote sensing technology. Growers can then add data on other
attributes of a field’s variability, e.g., soil moisture, pest densities, weather, soil
compaction, to the GIS to create a multi-layered, field-specific information system for
guiding management decisions for more precise application of inputs by variable rate
technologies (VRT). Producers first determine the variability within their fields by
relating crop production and/or soil attribute data collected over several growing
seasons to geographic coordinates using a GPS monitor mounted on a combine.
While crop yield data are insufficient in and of themselves to understand the actual
source of field variability, when they are joined to digital soil maps, farmers have two
significant layers of information about field variability that can be integrated into the
GIS and supplemented by other layers of agronomic data as needed. Growers use the
GIS to guide their VRT to continually adjust input applications as they operate their
equipment based on the biophysical variability of the various zones they have
identified in their fields (See the glossary on pages 14 and 15 for definitions of
selected terms associated with PA)
The goals of PA are increased yields and reduced costs with fewer negative
environmental effects from inefficient input application. PA seems to offer the
greatest potential benefits where inputs are varied and cost is high, field variability is
high, high-valued crops are grown, and environmental effects, especially to water
resources, must be reduced. PA is foremost a sub-field management system that can
specify rules for guiding input application based on past, current, or expected
conditions. PA may potentially do for agriculture what just-in-time production and



shipment tracking technologies have essentially done for manufacturing: Managing
temporal and spatial variability in the production process for more efficient decisions.
Private companies have dominated PA’s technical development to date. In
contrast, basic agricultural technological research has predominantly been a public
undertaking. This difference may suggest a changing role between public and private
sectors in agricultural innovation. For example, the increasing importance for the role
of professional crop specialists in PA could presage change in the traditional role of
extension personnel; and the need for computer knowledge and skilled data analysis
suggests potential new demands on agricultural education and extension. Creating
site-specific GIS databases through service contracts with input suppliers also raises
issues about ownership and control of these databases. PA also may contribute to
new contractual arrangements through proprietary ownership of agricultural input
systems, e.g., PA data could become a new part of existing proprietary seed-herbicide
“packages.” This could have implications for current farm structure. Given PA’s
potential environmental benefits, there is precedent for a public role in research that
has little likely commercial benefits and is therefore less likely to be assumed by the
private sector, e.g., sensor development for environmental monitoring.
Socioeconomic Factors in Adoption and Use of PA
Technological Costs and Development Trends. Farmers face significant start-
up costs before they can exploit the potential of PA to achieve long-term advantages.
Much of the capital and labor costs of PA are in acquiring equipment and developing
a baseline data set. These costs include yield monitoring, crop scouting, and soil
testing over multiple growing seasons and learning how to use remote sensing
technology, GIS data, and VRT. Individuals may either buy this technology directly
and learn how to use it, or they may contract for various services with vendors,
including independent crop consultants and input suppliers. Most adopting farmers
will likely purchase some hardware and develop some expertise and contract for other
support services. Some illustrative equipment and service costs might include the
following:
!A yield monitor for corn, wheat, or soybeans could cost $4000-
$7000.
!Soil sampling has been estimated at $3-$7 per acre, considerably
more if samples are deeper or taken more frequently and in smaller
grids.
!A truck-mounted computer system and GIS software could add
$3000 to the cost of a PA system.
!Service costs including soil sampling, yield monitoring, crop
scouting, GPS receiver and, in some regions, a satellite signal
subscription;4 VRT controllers and fertilizer application (additional


4 The Department of Defense recently made its GPS signal accurate to 20 meters. Farmers
who do not need greater accuracy than this will no longer have to buy a differential correction.
A public provided national differential GPS network is also expanding and should make the
GPS signal more widely available in the future. As this network expands into new regions,
(continued...)

cost over uniform application rates) have been estimated at $13-$26
for a grower and $2.50-$14.50 for a dealer on a per acre basis.5
!Weather Services International, a private vendor, currently charges
$5-$7 per acre to subscribe to their basic weather service and $10-
$20 to obtain their integrated weather data packages.6
Digital soil maps based on grid sampling are generally expensive to acquire and
some soil attribute data may require periodic updating. Soil grid sampling could
become a high annual expense, but different sampling densities based on indicators
such as yield monitoring data or soil attributes can reduce the costs of sampling where
little variability exists or where information about variability of a particular area in the
field is of little benefit. For example, one of the most efficient uses for yield maps is
to identify those areas where yields are significantly lower than the field average.
Often the production problems in these areas can be readily identified and corrected,
leading to increased future revenue.
Other sensor-based applications, sometimes referred to as “on-the-go” sensing
or local sensors, are also becoming an important part of PA technology. A sensor
mounted on a machine or truck senses the field condition, e.g, soil pH, nitrogen
content, pest population, soil compaction; an on-board computer sends the data
through an algorithm; and a real-time signal goes to a VRT applicator, also mounted
on the machine, to deliver a precise application of fertilizer or pesticide. Soil nutrient
sensors, optical plant sensors, pest population sensors, and soil property sensors are
currently in use; research on other sensors is underway, e.g., moisture density sensors
and plant disease sensors. Accurately synchronizing sensor measurement with the
desired input application, however, has proved somewhat problematic to date. For
example, a bulk spreader moving at 25 mph passes over approximately 37 feet each
second, making optimal accuracy difficult to achieve.
Remote sensing technology is yet another technical development in PA. For
growers who need it, remote sensing, either from airplane or satellite platforms,
permits improved spatial, temporal, and spectral resolution compared to other sensing
devices such as on-the-go sensors or grid sampling. This technology could be very
beneficial for decisions regarding irrigation, pest and disease control, etc. that have
to be made during the growing season. Also the potential for this technology to
determine the cause, size and location of sub-field variation in plant vigor is increasing
with developments in the new hyper-spectral sensors. Some work is underway to
develop sensors that detect nitrogen deficiency and water stress. Such data will
improve the efficiency of professional crop scouts who can quickly find low


4 (...continued)
farmers who need more precise spatial accuracy may no longer have to purchase a satellite
subscription fee.
5 National Research Council. Precision Agriculture in the 21st Century: Geospatial and
Information Technologies in Crop Management. National Academy of Science Press.
Washington, D.C., April, 1997.
6 Weather is a critical and largely uncontrollable variable that can alter the outcomes of a PA
system under the best of circumstances. Weather data, nonetheless, can become an important
layer of information in a GIS.

vigor/stressed areas within large fields. Timely delivery of this information over the
Internet is now also possible.
Research continues to improve yield monitors and VRT. A yield monitor is
often the first piece of equipment PA growers acquire. Yield monitors allow
producers to map yield variability. These maps can then be used to assess changing
production practices. Accurate measurement can be difficult as calibration can be a
challenge. The fastest growing area of yield monitoring technology is currently in
grain-harvesting combines although there is research underway on monitors for other
crops such as cotton, root crops, and fruits. Lack of yield monitors for high-value
crops, e.g., fruits and vegetables, is possibly a barrier to adoption of PA techniques
for some operations. Given the limits of current sensing calibration and the absence
of a universal standard of yield measurement, the available grain sensing systems must
be calibrated for each type of grain and for each type of machine. Finally, a relatively
wide variety of commercially available VRT exists for liquid and dry chemical
application, for planting, and for manure spreading. Most of these systems rely on
electronically controlled hydraulic systems. Precision irrigation technologies are also
beginning to be developed in drip and center-pivot irrigation systems. Again,
accuracy of application can be lower than desirable, especially in systems that do not
have the capacity to anticipate rate changes as the VRT moves across a field or, in the
case of precision irrigation, to control water application involving a large number of
nozzles.
PA technologies are evolving rapidly. There is, however, little standardization
of measurement among manufacturers; companies are designing and developing their
own equipment and proprietary software systems. Ongoing research is likely to
establish what works better and at what cost in particular production areas and with
particular crop systems. Producers who adopt PA are likely to encounter a relatively
steep learning curve with many of the GIS software packages, especially in collecting
data that can be analyzed and translated into useful information for changing
management practices in the field, according to anecdotal information in the farm
press and on various Web sites. Learning costs can be expected to fall as adoption
of PA spreads, as equipment becomes more standardized, as equipment calibration
becomes more reliable, and as research supports the value of different techniques and
different data comprising particular GISs.
Socioeconomic Aspects in Adoption and Diffusion of PA. A grower’s
anticipated return on investment in PA technologies will most likely govern the
decision to adopt PA techniques. Increased yields and/or reduced input costs could
generate significant net revenue gains to an individual farmer. The more input-
intensive crops (e.g., sugar beets, vegetables, potatoes, cotton) generally are more
highly valued on a per acre basis. This suggests that growers planting these crops
might see the greatest incentives to adopt PA techniques, especially if they could
reduce the costs of expensive inputs. Most adopters to date, however, appear to
grow low-value, bulk commodities, e.g., corn, soybeans, and wheat, which may be
partly due to the early development of yield monitors for these crops.
Researchers are studying the profitability of PA. One of the more detailed
economic reviews of PA adoption to date examined phosphate and potassium



fertilizer applications on wheat, potatoes, and corn.7 The authors concluded that
when all costs were included (including education and training), PA was rarely
profitable for the farm operations studied. If other, more costly inputs (e.g.,
pesticides, seed) could be applied on a site-specific basis, however, the likelihood of
profitability could increase, although another study showed that profit actually
decreased with an increase in the variability of input requirements.8 If PA is adopted
by more producers, costs of training and education and the technologies themselves
are likely to decrease. Finally, at least one study has shown that variable rate
application may not be the key factor in PA profitability.9 Rather, this study showed
that it could be profitable in some cases to use PA to determine the optimal uniform
input application.
Information on the adoption of PA is fragmentary and unsystematic. Adoption,
however, is likely to vary across crops, farming systems, and geography. Because PA
is not a single technology or farming system, each combination of GIS data and
technologies has its own array of costs and benefits. In a comprehensive study of 950
corn growers in 16 states, PA adopters tended to be younger, full-time operators,
better educated, and currently using computers for business management than non-
adopters.10 Adopters also had less diverse operations, a higher debt-to-asset ratio,
and operated larger and more profitable farms. Total acres farmed, acres harvested,
asset values, return on equity, and net income each were between 1.5 to 3 times larger
for adopting farms. Over half the adopting farms had annual sales of $250,000 or
more; less than 20% of non-adopting farms were of that size. This suggests that PA
may be used in a higher percentage of total agricultural production than percentage
of farms data cited earlier might indicate. Over half the adopters were in Indiana,
Illinois, and Iowa, and overwhelmingly specialized in cash grain production. Two-
thirds of the non-adopters were located in the other 13 states in the study, suggesting
some geographic concentration in current adoption patterns.
Socioeconomic models of adoption patterns suggest that farmers positioned to
realize the greatest returns will more readily adopt new technologies. This might
mean that farmers who are currently not highly efficient in their management of inputs
would find early adoption of PA most beneficial. Yet, farmers who have adopted PA
or who might be expected to adopt it in the future appear to be those who already use
relatively sophisticated approaches to input application and who would, in comparison
to less efficiently managed operations, possibly realize fewer initial profit advantages


7 Lowenberg-DeBoer, J. and S.M. Swinton. “Economics of site-specific management in
agronomic crops.” Department of Agricultural Economics Staff Paper 95-14. Purdue
University, 1995.
8 Hennessey, D., B.A. Babcock, T.E. Fiez. “Effects of site-specific management on the
application of agricultural inputs.” Report 96-WP-156, Center for Agriculture and Rural
Development, Iowa State University, 1996.
9 Schnitkey, G., J. Hopkins, and L. Tweeten. “An economic evaluation of precision fertilizer
applications on corn and soybean fields.” Paper presented at the annual meetings of the
American Agricultural Economics Association, San Antonio, TX, July 1996.
10 Daberkow, S.G. and W.D. McBride. “Survey results: Adoption rate of site-specific crop
management technologies among U.S. corn growers.” Modern Agriculture 1 (7)(Fall) 1998.

by adopting PA. Several studies of technological adoption have also noted the
importance of the compatibility of an innovation with current practices, that is, the
innovation complements existing practices rather than wholly replaces them, at least
over the short term.
Adoption of PA for sub-field management is a refinement of good whole-field
management. Since 1977, agricultural research and extension services have promoted
Best Management Practices (BMPs) to growers for improved profitability and
improved environmental practices, e.g., soil conservation, pesticide record-keeping.11
Soil testing and crop scouting, both long-standing BMPs, are also central to the
success of PA. Adding small-area soil grid sampling and integrating other data into
a GIS, for example, are perhaps smaller incremental steps for growers already using
BMPs than for farmers who have not adopted BMPs or who use BMPs inconsistently.
Some research has indicated that large farms have adopted BMPs significantly more
often than small farms.12 Growers who have never soil tested or never crop scouted
might be expected to see in PA a uniquely attractive set of practices that could, in
time, significantly improve their operational efficiency over current practice.13 Yet,
these growers appear less likely to have adopted PA techniques to date and, given the
importance of prior BMP adoption, may only slowly adopt PA in the future. In a
statistical test of differences between PA adopters and non-adopters, with respect to
nutrient soil testing and crop scouting, PA adopters had significantly higher levels of
use with these BMPs than did non-adopters.14 Differences in management and
technical skills as well as initial costs of BMP adoption might help explain why some
growers have adopted BMPs and others have not. These indicators suggest that
adoption and diffusion of PA techniques may not be as straightforward as existing
adoption models of agricultural technology might predict, and further attest to the
site-specificity in growers’ decisions to adopt PA techniques for their cropland.
Would-be adopters desiring to contract for PA services have found that vendors
of these services are concentrated in potentially high-demand areas, i.e., areas where
large, commercial operations predominate. PA is also more likely to attract farmers
who are already comfortable with computers and are not daunted by the information
complexities of PA. As the technology matures and as the socioeconomic profile of
PA’s potential returns on investment become clearer, however, more growers are
likely to consider at least some of the PA technologies either through direct purchase
or, perhaps more likely, through contracting for PA services from private suppliers,
especially with growth in the private supply of PA expertise.15


11 BMPs were implemented as requirements of the 1977 amendments to the Clean Water Act.
12 Tweeten, L. “The structure of agriculture: Implications for soil and water conservation.”
Journal of Soil and Water Conservation 50(4), July-August, 1995.
13 Only about a third of acres planted to major field crops was soil tested for nutrients. Pest
scouting was practiced on slightly more than half of the acreage planted to major crops.
USDA. Cropping Practices Survey,1994.
14 Daberkow, S. and W. McBride. “Socioeconomic profiles of early adopters of precision
agriculture technologies.” Journal of Agribusiness 16(2), Fall, 1998.
15 Over the past five years there has been significant growth among agrochemical dealers in
(continued...)

Individual farm returns on PA investment will vary by the management skills of
operators, by the biophysical properties of each field, and by the interactive effects of
altering various inputs based on the site-specific database. The limited data on
economic returns showing mixed results is perhaps not unexpected given the difficulty
of maintaining any consistency of the system between comparative sites. Even side-
by-side farm comparisons reveal significant biophysical variability between fields; and
potential profitability of PA is highly correlated with the biophysical variability of the
setting to which it is applied. No research could be located that has examined the
economics of PA for an agro-ecological region, for a single production system, or
across all dimensions of a production system.16 The economic research available is
based disproportionately on large operations with relatively low-valued crops and
inexpensive inputs, variables that are perhaps less conducive to producing significant
profit advantages under a PA system. Research is currently underway on other
production systems (e.g., cotton, fruits and vegetables, forestry) that will clarify the
picture on returns on PA investment. Changes that might attract other growers to
invest in PA include cooperative arrangements to share technology, or training,
equipment leasing, and standardization of PA technology.
The rate of adoption does not appear to be uniform among crop systems either.
Farm press reports and regional agricultural meetings suggest that many adopters of
yield monitors have been corn, wheat, and soybean growers. Trade press reports
information reveal that adopters also include some Florida tomato growers, California
vegetable, grape, and orange growers, Arkansas cotton growers, Massachusetts
cranberry growers, and Northwestern foresters. The variety of adoption rates among
different crops may suggest the importance of examining PA adoption rates from a
crop-specific perspective or that the crop is less important than other variables in the
decision to adopt. For example, some PA practitioners have stated that producers are
using PA information to document yields on rented farmland which may be useful for
negotiating new leases and equitably dividing up crop production on crop sharing
arrangements. PA also may be used for loan applications, proving yield history for
crop insurance, and documenting losses for crop insurance purposes.
USDA data indicate that among the computer users noted above, about 64,000
farms (about 3.2% of total farms) also used computer-aided chemical application or
GPS technologies in 1995.17 More recent unpublished USDA data (1998) indicate
only a slight increase in acquisition of these PA technologies, perhaps suggesting a


15 (...continued)
qualifying to become Certified Crop Consultants, a program administered by the American
Society of Agronomy. Private consultants are likely to become an important means by which
PA knowledge is transmitted.
16 Some studies of Western sugar beet production have indicated that PA holds substantial
promise in improving the efficiency of nitrogen application. Adding more nitrogen increases
yield but lowers beet quality. PA potentially offers the technical capacity to optimize the
application for yield and quality.
17 Sommer, J.E., R.A. Hoppe, R.C. Green, and P.J. Korb. “Structural and Financial
Characteristics of U.S. Farms, 1995: 20th Annual Family Farm Report to Congress, AIB-

746, ERS-USDA, December, 1998.



continuation of the initial phase of the adoption process.18 It is, however, reasonable
to regard PA as an extension of mechanization and an elaboration of existing
chemical-based agricultural practices, two technological systems that researchers
regard as encompassing significant scale economies and that have contributed to
changes in farm size and the industrialization of the farm sector.19 In that sense, PA
may be regarded more as an evolutionary technology reinforcing certain existing
agricultural practices rather than as a revolutionary one that holds the potential to
radically alter existing production practices.
Given its character as a site-specific set of practices, PA, in contrast to many
earlier agricultural technologies, seems likely to develop as customized systems
specific to each producer’s operation even as the basic technologies become more
standardized. The particular PA package for each farm will be based on the
biophysical variables that are being addressed. The biophysical variability of cropland
makes it likely that different farms will find different combinations of data in a GIS
most useful. A soil map based on small area sampling, for example, is likely to
provide more useful information to a grower whose soil variability is high throughout
a field than it would to a grower whose soils are more uniform. Similarly, a farm with
a more varied topography may find digital elevation modeling (DEM) very useful
while for another grower with level fields, such information would add little or no
benefit to crop management decisions.
Environmental Aspects of Precision Agriculture
PA is used within field boundaries, but environmental issues in agriculture
transcend these boundaries. Ecological and health implications of using chemical
fertilizers, soil fumigants, herbicides, and insecticides in large amounts are central
targets of environmental reform. Farmers increasingly find themselves under pressure
to improve their output efficiency while reducing the adverse environmental effects
of necessary inputs. Proponents of PA have strongly emphasized its potential to
significantly reduce these effects. Others, however, doubt that PA techniques can
address many of the key factors that produce negative environmental outcomes.20
There is little convincing evidence to date for either position.
Field-level studies do suggest that more accurate calibration of inputs can reduce
over-fertilization and excessive use of pesticides at the farm level. To the extent that
such reduction improves the environmental performance of individual farms, PA
would seem to offer some improvement over current practices. Research at the


18 Classic studies of adoption and diffusion of technology have identified a characteristic S-
shaped curve that models the socioeconomic process of adoption: a slow takeoff followed by
a rapid phase of adoption followed by a slow and steady diffusion across a production sector.th
See Rogers, Everett M. Diffusion of Innovation, 4 Edition. New York: Free Press, 1995.
19 Wolf, S.A. and F.H. Buttel. “The political economy of precision farming.” American
Journal of Agricultural Economics, 78(December),1269-1274. 1996.
20 Groffman, P.M. “Ecological constraints on the ability of precision agriculture to improve
the environmental performance of agricultural production systems.” Pp. 52-64 in J.V. Lake,
G.R. Bock, and J.A. Goode, eds., Precision Agriculture: Spatial and Temporal Variability
of Environmental Quality, New York: Wiley Press, 1997.

watershed or regional scales may produce better evidence than field-level studies
about the environmental effects of PA, but these studies will have to wait for wide-
spread adoption of PA. Moreover, because environmental problems are not evenly
distributed, where PA is widely adopted becomes central to its potential success for
reducing them. It is possible that PA could produce less environmental improvement
in commercial-scale agriculture than individual field tests indicate. Growers who find
they can reduce certain inputs could be offset by other growers who learn they are
under-fertilizing or using sub-optimal levels of pesticides and decide to increase their
inputs. In assessing PA’s future potential for improving environmental quality, the
scale at which one examines PA becomes significant: Producers are motivated
primarily by considerations of their own actions and their own costs; those concerned
primarily with what PA potentially offers in terms of environmental improvements are
concerned with agricultural producers as an aggregate.
A reduction in agriculture’s impact on water resources through more efficient
application of fertilizers and improved drainage management could potentially become
a significant environmental benefit of PA. Controlling non-point runoff from U.S.
farms is considered to be the most significant water quality need, according to EPA’s
nationwide survey.21 Water quality impairment from cropland sediment alone has
been estimated to cause between $2 and $8 billion per year in damages across the
country.22 Farms and ranches, however, have been largely exempt from most
regulations governing water, air, and land that other industrial sectors must adhere to.
The Clean Water Act of 1972 (§303) required states to establish ambient water
standards for surface water. Few states, however, have set the total maximum daily
loads (TMDL) of pollutants to improve the quality of remaining pollutant-impaired
water bodies as required by the Act. A number of lawsuits since 1992 led the EPA
to issue a requirement in 1997 that states must develop specific schedules for TMDLs
of impaired waters, including those impaired by non-point source pollution, within the
next 8-13 years. A recent agreement between EPA and USDA has extended this time23
to 15 years. There is some uncertainty whether agricultural operations other than
large, confined animal enterprises legally fall under the TMDL requirement. If it is
determined that they do, PA could potentially play a significant role in helping
implement the TMDL program requirement through development of a site-specific24
management program to control non-point sources of pollution. Recent legislation


21 EPA. National Water Quality Inventory: 1996 Report to Congress.
22 ERS-USDA. Agricultural Resources and Environmental Indicators: Agricultural
Handbook 712. Washington, DC, 1997.
23 EPA. Office of Water. Joint Statement of the Department of Agriculture and the
Environmental Protection Agency Addressing Agriculture and Siliviculture Issues within
EPA Revisions to TMDL and NPDES Rulings. May, 2000.
24 Most agricultural pollution is classified as “non-point” and is not subject to controls under
the Clean Water Act. The TMDL program is designed to target waters that are impaired even
after point-source control measures have been applied. If agriculture is determined to fall
legally under the TMDL program, as a recent federal district court case in California
(continued...)

(H.R. 4502), however, was proposed to prohibit EPA from imposing new rules
regulating TMDLs from non-point sources; and S. 2417 would leave power to
regulate TMDLs in the states while increasing state funding for non-point source
control. On May 22, 2000, H.R. 4502 was referred to the Subcommittee on Water
Resources and Environment. S. 2417 was referred May 18, 2000 to the Committee
on Environment and Public Works Subcommittee on Fisheries, Wildlife, and Drinking
Water. Hearings were held; and on July 11, 2000 EPA issued final regulations to
revise the TMDL program.
Agricultural policies to reduce the environmental effects of crop production
usually have been based on the combination of voluntary participation and incentives
to attract participants including various technical assistance and cost-sharing
arrangements. Idling agricultural acreage has also been a policy measure for reducing
adverse environmental effects while limiting supplies of commodities. However,
pesticide registration, drainage for wetland modification, protection of endangered
species, and pollution from large confined-animal operations are four areas where
some agricultural practices have been regulated.25 PA has potential value for reducing
the need for regulation in some situations by documenting/monitoring pesticide and
fertilizer use, and by identifying environmentally sensitive areas where reduced inputs
may be appropriate. Each of these measures could play a significant role in the
development of site-specific control programs where run-off is shown to be a source
of non-compliance under the TMDL requirement noted above. With heightened
environmental awareness and increasing urban/suburban development also
encroaching onto rural agricultural areas, more farmers are likely to find themselves
under public pressure to reduce the environmental effects of agricultural production,
especially those that adversely affect the quality of their neighbors’ lives and its impact
on water resources.
The National Research Council also has recently recommended development of
a national strategy to combat nitrogen and phosphorus pollution in coastal waters
after its research showed that synthetic fertilizers have contributed to more than 50%26
of the doubling of nitrogen and phosphorus pollution from 1960-1990. There is
evidence based on contingency-valuation methods suggesting that, in shifting from a
logic of maximizing output to one of reducing inputs, farmers would accept some
reduction in output per acre in return for avoidance of a moderate environmental


24 (...continued)
(Prosolino, et al. v. EPA, No. C. 99-01828WHA, March 30, 2000), suggests it may, control
of non-point pollution will be site-specific. For further information on agriculture and water
quality see CRS Report RL-30437, Water Quality Initiatives and Agriculture and CRS
Report 97-831 ENR, Clean Water Act and Total Maximum Daily Loads (TMDLs) of
Pollutants.
25 EPA was empowered to review each pesticide to ensure it met minimum risk criteria. These
criteria were altered by the Food Quality Protection Act of 1996 and EPA has not completed
details of their implementation of the new criteria.
26 National Research Council. Clean Coastal Waters: Understanding and Reducing the
Effects of Nutrient Pollution. National Academy of Science Press. Washington, DC. April,

2000.



risk.27 Also, some suggest that cost savings can sometimes outweigh lost income.
Databases on chemical use generated by PA could also help farmers meet any future
environmental standards as well as provide a long-term record of input applications,
which may potentially help future land sales.
Current Status of Precision Agriculture at the Federal and State
Levels
Although the emerging innovations of PA are being developed and promoted
largely by private companies, there has also been a public presence in PA
development. Recent Congresses considered several bills to promote agricultural
research and extension targeting PA and PA technologies. Three bills were
introduced in the 104th Congress to amend the Competitive, Special, and Facilities
Research Grant Act and one to amend the research categories in the Fund for Rural
America. None of these bills was enacted.
Three relevant bills were introduced or re-introduced in the 105th Congress. An
original bill (S. 1150) to address high priority national agricultural concerns and to
reform, extend and eliminate certain agricultural research programs was enacted as
P.L.105-185, The Agricultural Research, Extension, and Education Reform Act of
1998. Section 403 of P.L.105-185 sets general research priorities and authorizes
grants to study and promote components of PA technologies and to integrate
research, education and extension efforts in PA. Section 403 (b) (2) limits funding
under this section to those projects that the Secretary determines are unlikely to be
privately financed; and §403 (c) (6) specifically targets research on the applicability
of PA for small and medium-size farms. In further defining these grant priorities,
§403 (d) (3) appears to limit funding for certain projects that would investigate how
PA might enhance environmental quality stemming from a reduction in the use of
inputs. Section 403 (d) (3) gives priority to research that demonstrates the efficient
use of agricultural inputs rather than uniform reduction in the use of inputs. As noted
above, however, growers may be willing to accept some trade-offs in total output in
return for a reduction in adverse environmental impact. Research on this relation and
on PA’s contribution to an overall reduction in input use would seem to be foreclosed
by this subsection. Congress has not appropriated funding under this authority to date.
Section 401 of P.L.105-185 authorizes a new 5-year Initiative for Future Agriculture
and Food Systems research that has PA as one of six priority mission areas (coupled
with natural resource management). About $113 million is available for research
grants for the Initiative and this has been appropriated only for FY2000.
At least eight public or land-grant universities have established PA research
institutes or laboratories: Purdue, University of Georgia, Oregon State, University of
Minnesota, University of Arizona, Texas A&M, University of Missouri, and North
Carolina State. In 1996, the Cooperative State Research, Education, and Extension
Service (CSREES) reported funding about $9 million to PA-related research at land-
grant universities. The Coastal Plain Experiment Station in Georgia received $3.4
million between 1991 and 1993 to construct the National Environmentally Sound


27 Lohr, L., T. Parker, and L. Higley. Farmers risk assessment for voluntary insecticide
reduction.” Ecological Economics 30(1)(July), 1999:121-130.

Production Agriculture Laboratory where PA is a major element of the research
program. The Agricultural Research Service (ARS) also invested $4.4 million in PA-
related research in 15 locations in 1995. According to the ARS Budget Office, the
agency spent $20.4 million on PA from 1996 through 1999. For 2000, it has
allocated $7.7 million for PA-related research. In addition, the ARS Subtropical
Research Laboratory in Weslaco, Texas is the site of RESOURCE 21, a project
designed to launch four solar-powered satellites for remote sensing that will support
PA. This public-private venture is being developed under a USDA Cooperative
Research and Development Agreement (CRADA) involving six partners and the ARS.
Other Federal agencies conducting research and development with PA
applications include the Department of Energy (e.g., Idaho Engineering Laboratory)
and the National Aeronautical and Space Administration (e.g., LandSat, Commercial
Remote Sensing Program at Stennis Space Center). It also should be noted that the
U.S. Department of Defense has, quite unintentionally, probably had the largest
federal investment in PA with its launching and maintenance of the 24 Earth-orbiting
satellites that comprise the GPS.
Congressional Interest
Congressional interest in PA involves several issues, including the role of public
research and education, PA’s potential effects on U.S. farm structure, and questions
of data ownership and control.
Role of Public Research and Education. Inasmuch as PA is evolving largely
under private auspices, potential roles for the land-grant universities are unclear. If
PA is truly a new paradigm for agricultural production, as some observers believe, the
traditional role of extension agronomy, for example, may not be as well-suited for the
managerial demands that PA places on the grower as extension was for many previous
agricultural innovations. In addition, some observers believe that the land-grant
institutions are not well-equipped to provide the highly skilled data analysis that will
be needed by farmers. Several land-grant universities, however, have begun research
programs in PA (See the list of Websites in the Bibliography) and the University of
California-Davis now offers a minor in PA in its Department of Biological and
Agricultural Engineering. An increase in the number of private crop consultants in
PA, either as independent contractors or as employees of agro-input firms, also may
suggest changes in the traditional role of land-grant university research and extension.
More private crop consultants could make public education and extension programs
for PA redundant. On the other hand, some farmers may perceive crop consultants
with ties to agro-input firms to be less than objective in some of their managerial
recommendations and may consider the perceived neutrality of land-grant extension
agents and university researchers to be a valuable asset. Land-grant researchers could
help guide future PA developments. Regionally specific socioeconomic and
environmental analyses will be necessary to gauge the potential of PA as it becomes
more widely used. These usually have been publicly-supported areas of research by
land-grant scientists and extension personnel.
Traditional models of agricultural research and extension services may not be
adequate to some of the challenges of PA, which is based on site-specific



characteristics. Most agricultural research, by contrast, generalizes from a
representative sample (or experimental data) to a larger population. Extension
personnel apply the validated research results to individual cases. That model may be
less useful to PA analysis than it was for earlier agricultural innovations. A grower
needs to interpret the data from a PA system so that the information can be used to
make field-level management decisions. Growers need to know how to integrate the
information for precision application decisions when a series of field characteristics,
e.g., soil quality, disease, pests, weather assumes a particular configuration.
Traditional research and extension systems do not provide for this level of precision
interpretation. Some land-grant education and training programs may opt out of
providing PA extension education, letting private crop consultants take over this role,
whereas others might opt to change their education and training programs to better
reflect the specificity of PA-related knowledge. Several Midwestern 2-year colleges
have new PA education and training programs that could help provide publicly
supported extension services to PA adopters. Graduates of these and other PA
programs at 2-year colleges in collaboration with traditional extension services might
become a significant public source of regional expertise in PA.
PA’s Effects on Farm Structure. Some observers argue that smaller farms
could gain from PA because they are often operated under tighter labor constraints
where automation might enhance field monitoring capabilities, especially if the
operator is a part-time producer. Other observers argue that net benefits of PA are
greater for large farms and that the comparative advantages of small farms for
focusing on individual, high-value plants could be lost when larger producers gain an
ability to do the same. As discussed above, there is no evidence that PA exhibits bias
favoring larger producers, although a profile of early adopters does suggest that the
majority are large, well-financed commercial operations.
If a tendency toward scale bias does exist, it could be like many technical
innovations in agriculture that research has shown to have contributed to increasing
the gap between large commercial operations and smaller operations. Given existing
differences in computer and information management skills between typical large
operators and smaller ones noted above, PA’s data complexities could further widen
this divide. Absence of viable public involvement in PA could have the effect of
making it more likely that large farms benefit more from PA. The private sector is
more likely to put its major resources, e.g., data analysis services, VRT services, in
those areas where their returns are likely to be the highest. That will probably mean,
for the near term, in areas such as the Northern Central corn belt where nitrogen,
herbicides, and phosphorus applications can be most significantly affected.
Other effects on farm structure are also possible. Creation of proprietary seed-
chemical-PA “packages,” for example, also could conceivably increase costs to
farmers or create conditions where contract farming arrangements become more
common.28 While larger, better capitalized growers may find such contractual


28 Monsanto, a corporation currently developing PA technologies, charges a technology fee
for the use of its Roundup Ready soybeans and Bt corn, a proprietary genetically engineered
seed. This is a new way of profiting from agricultural technology. Proprietary ownership of
(continued...)

arrangements attractive, smaller farm operations that do not enter into these
arrangements could be at a disadvantage. Some research discussed above also
suggests that older or less educated farmers may find the information volume and
computer-use skills that PA requires a significant barrier. This does not mean that
PA necessarily forces older, less educated farmers out of farming, only that it may be
an additional factor either separating them further from more financially successful
growers or expediting their decision to exit farming. Limited-resource farmers,
including minority farmers, most with annual gross sales under $25,000, may also find
the costs to adopt PA techniques prohibitive.
Previous technological changes have contributed to changes in the structure of
U.S. agriculture. Socioeconomic research on PA will provide a clearer picture of its
potential for accelerating such change. Should new contracting arrangements develop
between farmers and proprietary owners of farm inputs, for example, research on the
dynamics of this process could better inform policy makers who wish to respond.
Expanded efforts in public education for developing computer skills, including
bringing educational opportunities to farms, could enhance the possibilities of smaller
farmers adopting PA techniques. Existing funding mechanisms through CSREES and
ARS may provide a means for supporting regional research and extension should the
demand for PA arise in areas unlikely to be serviced by private companies.
PA is developing in the context of a number of significant changes affecting
agriculture and the rural sector more generally.29 Given the adoption patterns
observed for PA to date, smaller operations, especially those located in areas not
dominated by larger, commercial operations, may find that their opportunities to
participate in PA are limited. The implications of under-serving certain areas with
PA services might become part of a broader examination of changes in the structure
of agriculture and the role of public policies in responding to these changes. If, for
example, the private sector’s domination of PA development is part of a more
fundamental change in the way future agricultural innovation might progress, how
public policy might respond could become a significant issue before Congress.
Data Ownership and Control. Arguably the most valuable commodity
generated by PA is the data acquired to create site-specific recommendations about
particular fields. With the large amount of data that will be collected, intellectual
property issues may emerge about data ownership, storage, access, disposal, and
protection. Currently, public institutions have more fully developed systems of
privacy protection and disclosure than does the private sector. Increased contract
production, with proprietary control over production inputs and outputs, and potential
private sector domination of PA data systems, could create conflict over who owns
farm-generated data and who may have or grant access to it. Growers’ groups have
expressed some concern that without appropriate safeguards, PA data could provide
the means for monitoring a farmer’s operation either by state agencies for compliance


28 (...continued)
PA databases could confer similar advantages to a firm.
29 Welsh, Rick. The Industrial Re-Organization of U.S. Agriculture: An Overview and
Background Report. Policy Studies Report #6, Henry A.. Wallace Institute for Alternative
Agriculture, 1996.

or by private firms for the creation of proprietary products. Moreover, depending on
how data controls develop, information collected by a contractor could be
copyrighted or protected under trade secrets and sold to others, even without the
farmer’s permission. Farmers have expressed concern about being able to exert a
right over the intellectual property in some form.
Databases such as those created for PA presumably could be protected under
copyright law (17 USC, §102). Protection is granted by copyright law irrespective
of the form or medium in which the data are embodied. If the database exhibits
original, creative expression through the selection, arrangement, or coordination of
the data elements, copyright protection subsists in these facets of the database.30
International law on intellectual property rights in databases is emerging that could
conceivably impinge on issues of ownership and use of databases created for PA.31
The traditional land-grant approach to research involved developing basic
knowledge and technology at university and agricultural experiment stations and then
disseminating it to farmers through the Extension Services. With PA, the farm itself
becomes the source of valuable knowledge. Because this knowledge is valuable, there
is an incentive to convert it to intellectual property, especially if the creator of the
knowledge is a private firm. The land-grant system’s orientation was toward the
creation of generalized knowledge; PA is directed at the creation of information that
is “context-specific/decision-focused knowledge.”32 With PA technologies, farmers
can acquire data about their own operations, bypassing the need to adapt agricultural
experiment station research to their own operations. This shift in how agricultural
information is generated and used, and the potential for someone other than the
farmer owning and using the site-specific knowledge may create intellectual property
issues for agriculture that are similar to ones that the manufacturing and service
sectors have addressed.
The 106th Congress is currently considering legislation on intellectual property
concerns in databases. H.R. 354, Collections of Information Antipiracy Act, would
amend Title 17 of the U.S. Code to provide broad copyright protection to databases
and the facts that generate them. Last reported action on the bill was its discharge by
the Committee on Commerce and placement on the Union Calendar (Calendar No.
212) on October 8, 1999. Although this bill does not directly address the site-specific
databases that would be created through PA techniques, it would seem to offer
additional protection to the creator of the database. If the creator of the database is,
for example, an employee of any agricultural business and not the grower, it is unclear


30 For more information on the legal basis of database protection, See CRS Report 98-902,
Intellectual Property Protection for Non-Creative Databases.
31 Databases are protected as a “content industry” within the information industry and are
defined in Article 2 as “collections of (other) materials such as texts, sounds, images,
numbers, facts or data ....” World Intellectual Property Organization, Draft Treaty on
Intellectual Property in Respect of Databases, Diplomatic Conference on Certain Copyright
and Neighboring Rights Question, December, 1996.
32 Boehlje, M. “Information and technology transfer in agriculture: the role of the public
and private sectors.” Pp. 23-38 in Steven A. Wolf, ed. Privatization of Information and
Agricultural Industrialization. Soil and Water Conservation Society, CRC Press, 1998.

where ownership and use rights might lie. Because the source of the facts presumably
protected is a farmer’s cropland, it is unclear whether the protection would lie with
the facts in the database or the facts associated with the grower’s land on which the
database is developed.
H.R. 1858, the Consumer and Investor Access to Information Act, is designed
to promote electronic commerce by improving consumer access to electronic
databases. Last reported action on the bill was its discharge by the Committee on
Judiciary and placement on the Union Calendar (Calendar No. 213) on October 8,
1999. It is a narrower bill than H.R. 354 in that it would extend copyright protection
to the databases but not to the facts in them. As in H.R. 354, the kinds of databases
created under PA are not specifically identified in the bill, and there may be
differences between PA databases and those that these bills are designed to protect.
If Congress wishes to specify that farmers own and control site-specific PA databases,
regardless of whether they are created by the farmer or by someone contracted by the
farmer, bill language might be added.
As noted above, public institutions generally have stronger privacy rules than the
private sector. Central storage and control over the data in a public institution, e.g.,
data warehousing at a land-grant institution, for example, might provide a mechanism
to increase farmer control. With appropriate protections (e.g., no owner
identification), such an arrangement might also provide a publicly accessible database
of regional adopters of PA techniques that would be useful to land-grant and other
researchers as well as to extension personnel.
Other Public and Public-Private Roles in PA
Enhancing potential society-wide benefits such as improved environmental
quality may be an appropriate public sector responsibility. Most sensor research in
the private sector, for example, is currently targeted toward crop production and
quality. Sensor development targeted to environmental measuring (e.g., pesticide
residue, pollution loading, indicator species detection) holds potential benefits for
society more generally. If promoting adoption and diffusion of PA for its
environmental potential becomes desirable, different policies for agricultural
production may give farmers different incentive to adopt.
Farmers currently have little incentive to deal with most environmental
externalities of agricultural production; thus new technologies like PA that hold some
promise for ameliorating environmental effects may not be valued by producers as
highly as having changes to policy that increase that value. However, policy
incentives encouraging growers to raise its importance in their production decisions,
as is occurring for other industries, may enhance their interest in adopting PA to
reduce some of the adverse environmental impacts of contemporary agricultural
production. For example, if buying and selling pollution credits, already adopted in
a few areas of environmental regulation, was authorized for agriculture, that policy
measure could rely on PA data to document crop herbicide and fertilizer applications.
In addition, the Environmental Quality Incentives Program (EQIP) might be used for
cost-sharing the adoption of PA techniques where significant environmental
improvement was possible.



Rural areas generally have less of the advanced information technology
infrastructure that PA requires, and incentives to meet these needs may be insufficient
for private companies. The cost of making differential correction of the GPS signals
varies depending on the location of reception towers. Some areas of the US receive
this correction from public sources such as the U.S. Coast Guard system; individuals
in other areas must purchase the differential correction from private vendors.
Recently, however, the Department of Defense ended its selective availability (SA)
of the GPS signal (which made the signal less accurate for civilian use) and made it
accurate for civilian use to approximately 20 meters. Additionally, most rural areas
lag in access to the effective communication systems and high-speed data connectivity
that will become increasingly significant for adoption of PA. These technologies
include an advanced telephone system called a digital subscriber line (DSL), cable
modem technology, satellite technology and land-based wireless technologies.33 The
private sector may be less likely to invest in high-speed data connectivity in rural,
dispersed markets than in more urban or suburban areas where the cost to serve each
user would be less. The lack of broadband connectivity may also act as a deterrent
to some PA service providers. Conversely, the availability of broadband connectivity
may make those locations more attractive to some private providers of PA services.
Changing pubic attitudes and an environmentally oriented marketplace may
increase pressure on the agricultural sector to improve its environmental performance.
This has been happening in confined-animal feeding operations, but pressure to
redesign some other agricultural systems is growing. A number of policy questions
could arise if this pressure continued to grow. For example: Could taxation be linked
with environmental performance in revenue neutral ways for the agricultural sector?
Could PA techniques play a role in reducing environmental regulation and in
switching to performance-based measures? Could research on coordinating PA with
environmental policy facilitate this switch? Would increased public funding for basic
research on PA-related subjects be appropriate? Would public/private collaboration
on pre-competitive PA research and development, e.g., environmental sensors, on
new environmental performance incentives, and on other market-based mechanisms
be appropriate? For example, might advances in remote sensing technology become
useful tools in regulatory compliance for environmental quality?
Conclusion
Research in the public and private sectors is currently developing and assessing
PA technologies and documenting the benefits and costs of adopting PA in particular
areas and for particular crops. Research on PA’s economic returns, on its potential
to reduce adverse environmental effects, and on its adoption profile is hardly
conclusive. As results from PA-related research are evaluated over the next few years,
proponents will probably be in a much stronger position to advocate PA’s benefits.
A small but active presence for PA in several land-grant universities and within ARS
may help to better define potential benefits of PA and clarify when, where, and to
what degree these benefits may be achieved.


33 From more information see CRS Issue Brief 10045, Broadband Internet Access:
Background and Issues.

The private sector is likely to continue its research into many PA-related areas,
especially those that offer market and profit potential, including: improving sensors
to more precisely monitor crop production and crop quality characteristics;
standardizing and refining the calibration of VRT; designing and developing new PA
equipment for higher-value crops; simplifying GIS data acquisition and interpretation;
and refining remote sensing technologies. The National Research Council report on
PA (See Bibliography) also noted the lack of what it called “decision support
systems” research to quantify the complex relationship between the quantity of an
input and yields within a changing bio-physical environment in a grower’s field. The
basic knowledge to link precise data and precise recommendations is a significant gap.
While traditional research and extension information models do not appear to provide
well for PA interpretation at this point, the area of “decision support research” may
not be attractive to the private sector, suggesting a role for public research and
extension.
Private sector firms seem convinced that larger commercial growers in bulk
commodities will continue to find adoption of some PA techniques or contracting for
some PA services valuable. Growers in other crop systems, as well as perhaps in
livestock and range-land management systems, will likely draw from this experience
to help make their own decisions on adoption, especially if the technologies become
less expensive, if training and support are provided, if early on-farm data collection
and analysis suggests that an operation is a good candidate for PA, and if producers
see their neighbors and friends adopting PA.
Over the longer term, agronomy experts generally believe that PA seems likely
to develop a viable presence in agriculture, concentrated in some regions and for some
crops. PA seems likely to take many forms with some growers adopting individual
techniques of PA and others adopting highly sophisticated PA systems. Large
commercial operators, younger on average, with strong computer and technical skills,
seem likely to remain the modal PA adopters. As further research on the economic
and agronomic benefits emerge, some farmers will continue to adopt and others will
decide against it. Database ownership, control, and privacy concerns may cause some
producers to bypass PA until these questions can be satisfactorily answered. Several
broad issues could become the focus of congressional review, including, (1)
intellectual property issues in PA databases; (2) PA’s role in moving from a34
regulation-based to a performance-based environmental policy for agriculture; and
(3) PA’s role in the larger context of the changing industrial structure of U.S.
agriculture.
PA will increasingly become part of the technological fabric of changes and
pressures facing farmers, distributing its costs and benefits varyingly across the
agricultural sector. PA seems unlikely, however, to be a single cause of any


34 The emphasis in performance-based policies, e.g., market-based controls, information
management systems, is on achieving measurable environmental results with flexible,
innovative policy tools versus a prescriptive regulatory policy approach that tells a company
or community how to manage a pollution problem. See Howes, J., D. John, and R.A. Minard,
Jr., Resolving the paradox of environmental protection.” Issues in Science and Technology

14(4), summer, 1998.



significant shift in agricultural production’s dominant trajectories. It may produce
significant economic and environmental benefits at acceptable costs on some farms,
in some regions, and perhaps for entire production systems; but it may also offer only
marginal benefits to others.
Glossary
Base Map. A binary digital map that is the outline of a field with its proper
coordinates. Data collected within the field by a yield monitor can be defined in
location by the base map.
Choropleth Map. A thematic map such as a soil image where spatial data is
displayed through the use of shading or color variations.
Crop Scouting. Precise assessments of the spatial distribution of pest
populations and crop performance. The data from crop scouting, collected weekly
in some cases, can be tied to specific coordinates on the grower’s base map permitting
the targeting of pesticides where and when they are most needed.
Differential Correction. Correction of the GPS signal to make it more
accurate. The uncorrected signal is currently about 20 meters. Correction of the
signal is done from a second GPS receiver at a known fixed location. A corrected
signal can be accurate to about 1-3 meters. The Coast Guard, FAA, Corps of
Engineers, and the Department of Transportation provide corrected GPS signals. In
many areas of the country where these publicly provided corrections are not available,
commercially available services sell corrected GPS signals to subscribers who need
greater accuracy than 20 meters.
Digital Elevation Models (DEMs). Digital representation of topographic data.
These models can enhance the validity of soil survey maps by factoring in the
slope/gradients of a sub-field area. Topographic variation in a field can influence soil
quality, seed germination time, or moisture holding capacity of otherwise similar soil
types.
Digital Soil Mapping. Fields are divided into grid cells defined by a GPS
receiver. Cells are approximately 2-3 acres. Larger scale resolution is also possible
and, depending on the fields biophysical variability, perhaps necessary for more
precise input application. Soil sample data from each cell are transferred to a digital
map that is then used for variable rate application.
Expert System. Often regarded as a branch of artificial intelligence (AI), an
expert system is designed for solving particular problems in a particular area, e.g.,
agriculture or medicine. Using the stored knowledge base developed through the
GIS, an expert system would define rules for timing pesticide application aimed at
specific pests, or when to begin tilling, or when to apply fertilizers for optimum
performance. These decision rules are modified by soil, pest, weather conditions, and
other data integrated into the GIS.



Geographical Information Systems (GIS). A combination of computer
hardware, software, and geographic data designed to capture, store, manipulate,
analyze, and display data that is referenced to specific points on the earth’s surface.
The capacity to perform many sophisticated spatial operations on these data
differentiates GIS from simple mapping software.
Geo-reference System. A coordinate system tracking specific points on the
Earth’s surface. An example of such a system is the Universal Transverse Mercator
system (UTM), a commonly used map projection that uses a set of transverse
Mercator positions for the globe that are divided into 60 zones, each covering 6
degrees longitude with an origin of the central meridian and latitude of 0 degrees.
Global Positioning System (GPS). Most PA will make use of the U.S.
Department of Defense’s 24 earth-orbiting satellite system to assign map coordinates.
These satellites emit radio signals at precise intervals accurate to a billionth of a
second. Through triangulation, a ground-based receiver translates the time lag
between emission and reception of the signals into precise geographic coordinates on
the targeted cropland.
Ground-based Sensors. Sensor-based application of inputs implies that sensors
on the VRT can measure information in real-time and make the necessary precise
adjustments. Ground-based sensors, also called real-time or local sensors, have been
developed for measuring soil nutrient status, soil physical properties, and plant
qualities. Research into other sensors is underway.
Point Sampling. A method of soil grid sampling where a sample is taken in a

10-30 foot radius at the center point of each grid cell (See digital soil mapping).


Precision Agriculture (PA). An emerging set of farming techniques whose
fundamental feature is assigning geographic coordinates to the biophysical variability
of cropland for improving efficiency of inputs. PA’s technology consists of four
major components: a Global Positioning System (GPS) receiver, a yield monitor, a
digital soil map, and variable rate application technologies (VRT). Data from these
components are integrated through a geographical information system (GIS). The
concept of PA focuses on the use of advanced information technologies in assessing
and managing temporal and spatial variation in cropland.
Remote Sensing. Data from light reflectance collected by instruments in
airplanes or orbiting satellites. The data can estimate vegetation characteristics, e.g.,
efficiency of nutrient uptake, on small areas within a field. High resolution (e.g., 1-5
meters) satellite images are currently available to producers from private vendors
(e.g., Space Imaging, Inc. will sell high-resolution satellite images from the satellite
IKONOS during the 2000 crop season).
Variable Rate Application Technologies (VRT). Computer-controlled
equipment continually readjusts the input application. Soil grid data provide the
prescription for the particular fertilizers or pesticides to be applied to each grid. A
GPS receiver in the truck enables the computer to recognize where it is in the field.
Computer-controlled nozzles can then vary the types and amounts of inputs according



to the soil fertility or pest population maps. Similar equipment is also available to
vary seeding rates for some crops.
Yield Monitoring Systems. Area-specific yields in the field are measured using
combine-mounted sensors or volume meters. A display of the yield weights and
moisture content of the grain every few seconds is translated to bushels/acre. When
used in conjunction with a GPS receiver mounted on the combine, estimates of yield
can be assigned to areas of a field to create a yield map. Proper calibration of yield
monitors can give a load accuracy of from within 5% of true yield under normal
operating conditions.
Selective Bibliography on Precision Agriculture for Further Reading
Overviews.
National Research Council. 1997. Precision Agriculture in the 21st Century:
Geospatial and Information Technologies in Crop Management, Washington, DC,
National Academy of Science Press.
Robert, P.C., R.H. Rust, and W.E. Larsen, eds. 1995. Proceedings, Third
International Conference on Precision Agriculture. American Society of Agronomy.
Madison, WI.
Stafford, J.V., ed. 1997. Precision Agriculture ‘97. BIOS Scientific Publishers, Ltd.
Oxford, UK.
Pierce, F.J. and P. Nowak. 1999. “Aspects of precision agriculture.” Advances in
Agronomy, 67: 1-85
Vanden Heuval, Richard. 1996. “The promise of precision agriculture.” Journal of
Soil and Water Conservation, 51:38-40.
Environmental Issues.
Khanna, M. and D. Zilberman. 1997. “Incentives, precision technology, and
environmental protection.” Ecological Economics 23: 25-43
Rejesus, R.M. and R.H. Hornbecker. 1999. “Economic and environmental evaluation
of alternative pollution-reducing nitrogen management practices in central Illinois.”
Agriculture, Ecosystems, and Environment 75(1): 41-53
Socioeconomic Issues.
Batte, M.Y. 2000. “Factors influencing the profitability of precision farming
systems.” Journal of Soil and Water Conservation, 55(1): 12-18.
Babcock, B.A. and G.R. Pautsch. 1998 “Moving from uniform to variable fertilizer
rates on Iowa corn: effects on rates and returns.” Journal of Agricultural and
Resource Economics, 23(2):385-400.



Lowenberg-DeBoer, J. and S.M. Swinton. 1997. “Economics of site-specific
management in agronomic crops.” Pp. 369-396, In F.J. Pierce and E.J. Sadler, eds
The State of Site-Specific Management for Agricultural Systems, American Society
of Agronomy Miscellaneous Publication, American Society of Agronomy. Madison,
WI,
Swinton, S.M. and J. Lowenberg-DeBoer. 1998. “Evaluating the profitability of site-
specific farming.” Journal of Production Agriculture, 11(4):439-446.
Technological Development.
Stombaugh, T.S. and S. Shearer. 2000. “Equipment technologies for precision
agriculture.” Journal of Soil and Water Conservation, 55(1): 6-11.
Modern Agriculture: The Journal of Site-Specific Crop Management. Quarterly
Sudduth, K.A. , J.W. Hummel, and S.J. Birrell. 1997. “Sensors for site-specific
management.” Pp. 69-79 in F.J. Pierce and E.J. Sadler, eds., The State of Site-Specific
Management for Agriculture, American Society of Agronomy Miscellaneous
Publication, American Society of Agronomy, Madison, WI.
Precision Agriculture Web sites.
Cranfield University, UK: [http://www.silsoe.cranfield.ac.uk/cpf/]
John Deere: [http://www.johndeere.com/greenstar]
North Carolina State University:
[http://bae.ncsu.edu/programs/extension/agmachine/precision]
Ontario, Canada: [http://www.gov.on.ca/OMAFRA/english/environment/precision]
Oregon State University: [http://www.orst.edu/dept.hort/precag/]
Pioneer Hibred: [http://www.pioneer.com/usa/technology]
Purdue University: [http://dynamo.ecn.purdue.edu/~biehl/Site Farming/]
Rutgers University: [http://www.crssa.rutgers.edu]
Texan A&M University: [http://www.agcen.tamu.edu/txprecag]
University of Arizona: [http://www.ag.arizona.edu/precisionag/]
University of Georgia: [http://nespal.cpes.peachnet.edu/]
University of Minnesota: [http://precision.agri.umn.edu/]
University of Missouri: [http://www.fse.missouri.edu/mpac]
University of Sydney, Australia: [http://www.usyd.edu.au/agric/acpa/]
University of Wisconsin-Madison: [http://www.ersc.edu/ersc/Projects/VIP/Precag]