Key dates
- 23 July 1972The launch of ERTS-1, later Landsat 1, advances observation of the Earth’s surface.5
- 1987A historical review records commercial computer-controlled variable-rate fertilisation by Soil Teq.2
- 1992Al Myers’s yield monitor reaches commercial sale.8
- 27 April 1995GPS achieves full operational capability; the public announcement follows on 17 July.6
- May 2000The United States discontinues Selective Availability, the intentional degradation of civilian GPS.7
- 2004StarFire incorporates JPL software and reference-station data, one example of advances in navigation corrections.14
A different way of understanding the field
Farmers recognised differences between wet hollows, lighter soils and more fertile patches long before computers. The late twentieth-century change was to connect this variation with digital location records, measurements and machine control. Equipment for variable-rate application developed during the 1980s, while yield mapping became increasingly prominent in the following decade. There is no single date on which precision agriculture as a complete system was invented. Individual technologies moved from prototypes to commercial products in different countries and at different times.
The central task was managing variation within a field. One average for a large parcel could conceal areas with different production conditions. In 2019, the International Society of Precision Agriculture adopted a definition emphasising a management strategy, analysis of data and decisions responsive to estimated variability. This helps explain the earlier development: purchasing a satellite receiver did not create an entire management system. Reliable observations, agronomic interpretation and the ability to carry out the selected action were also needed.231
Satellites for observation and positioning
Two satellite technologies served different purposes. Landsat 1, launched on 23 July 1972 as ERTS-1, supplied images of the Earth’s surface. GPS enabled a receiver to determine its position. An image of a crop and the coordinates of a combine are therefore different kinds of information. Multispectral observation provided evidence about spatial conditions, while navigation made it possible to return to locations and attach measurements to them.
GPS achieved full operational capability on 27 April 1995, with the public announcement following on 17 July. In May 2000, the United States discontinued Selective Availability, the intentional degradation of the public signal. This assisted civilian applications but did not initiate precision agriculture. Mapping and methods using differential corrections already existed. Accurate machine guidance still required suitable receivers and corrections: removing deliberate degradation did not automatically give every device centimetre-level accuracy.56714
The combine records the geography of harvest
During the 1990s, measurements of grain flow were linked to moisture readings and the combine’s position. The resulting maps preserved differences between parts of a field that disappeared in a single harvest average. The Smithsonian documents Al Myers’s development of a monitor ready for commercial sale in 1992, together with equipment from subsequent years. This records the commercialisation of a specific design, rather than proving that no earlier yield-measurement experiments existed. The important change was that harvesting became a series of observations as well as the physical collection of a crop.
A yield map is a calculated result. Grain-flow signals, moisture, working width and distance travelled jointly produce an estimate of output per unit area. Comparison with independently weighed grain helps assess calibration. Grain also takes time to travel from the point of cutting to the sensor, so location processing must account for that delay. Without measurement checks, a compelling colour pattern might depict equipment or processing errors rather than an agronomic difference. Digital recording expanded the reach of observation while creating new demands for data-quality assessment.89104
Moving from a map to an explanation
Geographic information systems, or GIS, allowed layers to be compared: soil-test results, yields, terrain and previous operations. Each layer described a different aspect of a field. A soil sample concerned a particular location and depth, whereas a yield map recorded an outcome accumulated over a season. Combining the records could help identify management zones, but matching colours did not establish a causal explanation. A low-yielding area might be constrained by a factor that additional fertiliser would not remove.
Nitrogen-management research showed that expected yield did not always predict the most appropriate fertiliser rate well. Nutrient supply from soil, moisture and seasonal weather also mattered. A hollow could support crops through water availability in one year and suffer from excessive wetness in another. The development of mapping consequently increased the value of multi-year records and comparative field trials. Precision agriculture was not simply a process of dividing land into progressively smaller squares. Its usefulness depended on the quality of the rule connecting an observation with an action, and on whether the assumed relationship held under the conditions being managed.41117
Crop sensors and measurements made on the move
Aerial and satellite imagery made it possible to observe large areas, but the reflected signal had to be interpreted in relation to both soil and vegetation. Sensors positioned near the crop developed along a complementary path. Active instruments emitted their own light and measured its reflection, reducing some dependence on changing sunlight. Research into using these readings for nutrient management expanded particularly from the 1990s and investigated different crops and production environments.
A 2010 publication by Kitchen and colleagues investigated corn-canopy reflectance as a basis for varying nitrogen application. It illustrates a shift from a map prepared in advance towards systems capable of sensing and responding during a pass through the field. An optical signal is not a direct reading of every plant requirement. It needs a model and reference conditions, while growth stage and other stresses affect interpretation. The history of sensing therefore joins agronomy to electronics.12113
Applying a decision and guiding the machine
A prescription map could tell a controller how to change an implement’s operation across a field. The controller had to coordinate position, speed and the dosing mechanism. Ruixiu Sui’s 2019 assessment of a variable-rate fertiliser applicator distinguished a target rate from the quantity actually delivered. This separation is fundamental: a correct decision stored in a computer is insufficient when physical application is uneven or delayed. Evaluation therefore concerns the performance of the machine as well as the contents of a map.
Guidance systems developed alongside variable-rate control. Stanford records precision-farming demonstrations in the mid-1990s, while NASA describes the transfer of GPS-correction methods into agricultural machinery, including cooperation between JPL and industry. In 2004, StarFire receivers incorporated software and reference-station information associated with JPL. This is one example within a wider technological development. Following a specified track and varying an application rate remain separate functions: the former organises movement, while the latter changes the intensity of an operation.131514
Why adoption followed different paths
The purchase price was only one part of adoption costs. Training, data preparation, maintenance and interpretation also required resources. A 2016 USDA study used information on American corn production in 2010 to investigate relationships between adoption and farm returns. Estimated benefits were positive but small, and depended on the technology and characteristics examined. Such findings cannot be converted into a promised return for any farmer.
A subsequent USDA report covering observations from 1996–2019 documented substantial differences among crops and technologies. Automated guidance, soil mapping, yield maps and variable-rate application did not follow one adoption curve. Farm scale and access to advisory services mattered too. Comparisons must distinguish the percentage of farms from the percentage of planted acreage: large operations can manage many hectares while accounting for a minority of producers. Evidence from one country describes that setting and cannot automatically represent the world. Recording a technology as available also differs from establishing that farmers use its outputs in their decisions.16172
Information becomes part of farm management
Precision agriculture expanded the role of machinery: equipment could collect information for later decisions while carrying out a physical task. Observation, interpretation, planning, application and assessment formed a feedback cycle. Every stage did not have to be automated, and the judgement of operators and agronomists remained consequential. ISPA’s definition also emphasised variation through time. The term therefore addressed a broader management problem than steering a tractor by coordinates.
The FAO’s 2022 automation report, summarised by the European Commission, also highlighted conditions of access: finance, skills, connectivity and electricity. Digitalisation did not remove these constraints and could deepen disparities when tools remained available mainly to some producers. The historical shift of the 1980s and 1990s joined spatial measurement with production management. Its assessment depends on which decisions actually improved, under what conditions and for whom. The number of sensors or the visual detail of a coloured map is not, by itself, a measure of a farming system’s success.1183
Questions and answers
Did precision agriculture begin in 2000?
No. Discontinuing Selective Availability in May 2000 improved civilian GPS, but yield mapping and variable-rate applications had developed earlier.72
Sources
- International Society of Precision Agriculture (2019). ISPA Newsletter 7(7): Official Definition of Precision Agriculture
- James Lowenberg-DeBoer and Bruce Erickson (2019). Setting the Record Straight on Precision Agriculture Adoption
- Chenghai Yang and Won Suk Lee (2013). Precision agricultural systems — USDA ARS publication record and abstract
- National Research Council (1997). Precision Agriculture in the 21st Century: Geospatial and Information Technologies in Crop Management
- U.S. Geological Survey (n.d.; accessed 2026). Landsat 1
- U.S. Department of Defense (2020). Global Positioning System Standard Positioning Service Performance Standard, 5th edition, Appendix C
- The White House, Office of Science and Technology Policy (2000). Improving the Civilian Global Positioning System (GPS)
- Smithsonian National Museum of American History (n.d.; accessed 2026). Precision Farming — American Enterprise exhibition
- R.L. Nielsen, Purdue University (2018). Yield Monitor Calibration: Garbage In, Garbage Out
- Levi Powell, Matt Darr and Ryan W. Bergman, Iowa State University (2020). Tips for Calibrating Your Combine’s Yield Monitor
- P.C. Scharf, J.P. Schmidt, N.R. Kitchen, K.A. Sudduth, S.Y. Hong, J.A. Lory and J.G. Davis (2002). Remote sensing for nitrogen management
- N.R. Kitchen and colleagues (2010; online 2009). Ground-Based Canopy Reflectance Sensing for Variable-Rate Nitrogen Corn Fertilization
- Ruixiu Sui (2019). Performance Assessment of a Variable-Rate Fertilizer Applicator
- Loura Hall, NASA (2018). How NASA and John Deere Helped Tractors Drive Themselves
- Stanford University, GPS Laboratory (n.d.; accessed 2026). Precision Farming/Agriculture
- David Schimmelpfennig, USDA Economic Research Service (2016). Farm Profits and Adoption of Precision Agriculture, ERR-217
- Jonathan McFadden, Eric Njuki and Terry Griffin, USDA Economic Research Service (2023). Precision Agriculture in the Digital Era: Recent Adoption on U.S. Farms, EIB-248
- European Commission, Knowledge4Policy; FAO (2022). The State of Food and Agriculture Report 2022 — official summary of automation findings
Approximate dates (c.) describe the nature of the evidence: archaeology dates settlements, layers and remains, not single moments of invention.