My Account: Log In | Join | Renew
Table of Contents
Select All Chapters
Soil electrical conductivity (EC) sensing provides a means for rapidly mapping variations in soil properties such as salinity, moisture, and clay content. On claypan soils, EC measurements are related to topsoil depth above the claypan horizon, an important factor in spatial crop productivity differences on these soils. Soil EC data obtained with a noncontact sensor based on electromagnetic induction principles were compared with data from a direct contact, coulter-based sensor. Differences in EC readings were attributed to differences in sensing depth between the sensors and operating modes. The electromagnetic induction sensor generally provided better estimates of topsoil depth and correlations to crop productivity over the full range of topsoil depths encountered (up to 150 cm). The coulter-based sensor, however, performed well at shallower topsoil depths (up to 90 cm), and also would be useful for investigating soil differences in precision agriculture practice.
The development of a CAN (Controller Area Network) based control system for an air seeder system will be presented. The CAN bus is a high-speed serial communications system, which was developed by Bosch and is being used in automotive, truck and bus industries. It is also being adopted as an ISO standard (11783) for use on agricultural equipment. This paper will present the development of a controller-monitor for a variable drive air seeder. The variable drive air seeder allows changes in application rate on the go, making it well suited for precision farming applications. A discussion of the design process involving possible architectures of the system, expandability considerations, and compatibility of the electronics across various systems. The system architecture had to be a balance between cost, cable complexity, ECU complexity, and system flexibility.
The water application patterns of two center pivot irrigation systems in eastern Colorado were simulated over their entire area of operation and mapped for the 1997 growing season, using GIS, as part of a multidisciplinary precision fanning project. The system performance was evaluated over the entire field including locations where the operating characteristics were altered due to end gun operation and topography. Precipitation was measured at the center and around the perimeter of each field. Precipitation surfaces were created using four different interpolation methods and combined with the irrigation application depths to calculate a total water application uniformity coefficient. System mean application depth and uniformity coefficient varied with the topography and end gun settings. Total water application uniformity increased with the addition of the precipitation values and was insensitive to the interpolation method used.
Conceived in 1982, the Soil Doctor system is one of the earliest fundamental tools envisioned for Precision Agriculture. It senses soil and/or crop conditions on-the-go and immediately responds to the assays to provide real-time corrective chemicals and inputs to farm fields. The technology's suite of real-time sensing and controls produce high fidelity field treatments that optimize economic and environmental benefits. Alternatively, for other embodiments, the overall patent and product base also provide for a step-wise, GIS-based approach to field assessment and treatment. The system features both contacting and non-invasive sensor techniques embodying scientific principles drawn from electrochemistry, soil complex resistivity, and soil conductivity technical disciplines. Systems include equipment and methods for transient fertilizer and seeding rate control as well as data management, visualization, and interpretation.
Multi-year precision guidance of agricultural implements in the open field enables the exact placing of inputs, and to grow crops on predetermined spots. The on-line system proposed here, uses a digital map that contains all co-ordinates to describe the intended path of the implement in the field, a sensor to measure the actual position of the implement, a comparator to calculate the position error, a controller to generate a correction signal and an actuator, mounted between the tractor and the implement to side-shift the implement onto the intended path. On a test track a tractor was driven at a speed of 5.2 km h-1, what resulted in a repeatable sideways sway of the tractor plus and minus ±10 cm. The implement, mounted on the backside of the actuator, was side-shifted to a straight path programmed in the digital map. The true path of the implement was recorded and showed a deviation from the straight line of less than ±2 cm.
This paper reports the results of two studies comparing the efficiency of foam marker and submeter differential GPS (Trimble AgGPS132 with Parallel Swathing Option) guidance methods. Efficiency was determined by measuring overlap or skip, and by analyzing operational variables such as ground speed and offline distance. Both tests were conducted in ideal foam marking conditions. In the first study, an experienced foam marker and GPS guidance operator, found each method equivalent in terms of overall swathing accuracy. However, the operator drove 13% faster with GPS guidance than when using foam. In the second study, an inexperienced foam marker and GPS guidance operator proved more efficient when using submeter GPS guidance than when using foam. Additionally, average ground speed was 20% faster with GPS guidance than with foam.
The OmniSTAR Virtual Base Station system was originally developed in 1990 to provide users reliable differential GPS corrections for high accuracy real time applications. At the present time there are OmniSTAR systems covering all the World's major land masses (excluding the Poles and extreme Northern and Southern latitudes). The system delivers its corrections via geostationary satellite, thus providing seamless coverage over the entire service areas. The OmniSTAR system is unique in that corrections from the entire network are used in the user system solution to provide RTCM corrections for GPS that are at least as good as a dedicated base station within one kilometer of the users position. This paper will describe the OmniSTAR network and user system and address recent advances resulting in enhanced accuracy. It will show results accumulated over the last 4 yrs demonstrating how final position results are dependent on the quality of the GPS engine.
The site specific application of P, K, and Mg requires a highly accurate dosage system and specific features for the transport of fertilizer from the spreader tank to the outlets of the boom. The pneumatic spreader developed has the facility to regulate separately the application rates of its four 6 m sections via four different electrohydraulic systems. The unique design of the pneumatic system, dosage cells and outlets reduces variability within each section to 2%. A weighing system, which senses the weight of the tank, serves as a permanent control for the amount of fertilizer applied. This enables a continual recalibration of the system, reducing the negative effects of varying fertilizer qualities. The data of the working path and width, application rates and weighing sensors are stored by the on-board computer, providing a fully automated system with accurate documentation of the whole fertilizing process.
In addition to yield mapping in combines and in other harvesters, biomass distribution in growing plant populations is a possible parameter for surveying heterogeneity inside of fields. To develop the technique of acquisition of biomass data by mechanical scanning of plants with stems, it has been shown that definite, simple relationships between dynamic effect and biomass arise through strip scanning of plant populations. When a cylindrical body and a pendulous pivoted cylindrical body are moved horizontally through a plant population, the forces acting on the cylindrical body and also the angle of deviation are determined by the parameters of plant growth, mainly biomass. By measuring the forces on the cylindrical body and the angle of pendulum, conclusions can be drawn about the biomass of crops. With measurements are determined differences in still growing crops in order to optimise crop management and yields in Precision Agriculture.
Site-Specific Crop Management refers to a rapidly developing new direction in agriculture, promoting variable agricultural management practices within a field according to site conditions. Existing agriculture equipment for variable rate technology (VRT) for multiple input has a significant disadvantage - a constant size of compartments or bins for applying components. When one of the compartments is empty, the whole unit has to stop for a refill. Plus, it requires the presence of a tender's fleet on a field. This reduces productivity and increases the cost of operation significantly. To improve the performance and productivity of equipment for applying agriculture inputs it is necessary to have changeable size of compartments on VRT equipment, matching tenders, and automated systems for management.
Possible technical solutions suitable for fertilizer spreaders, liquid chemical sprayers, seed drills, and tenders are proposed.
In the rush to adopt the new technology that many people assume is the essence of precision agriculture, we often overlook the simple requirements of good system management. We have measured the poor distribution of ammonia across application tool bars. We have seen the results of using planters without checking calibration. Lawsuits can be based on sprayer drift or poor technique in loading or operating. The goal of this paper is to remind all of us of the importance of every step in the management of crops. Checking, calibration, understanding instructions (in other words basic good technique) are necessary to achieve the benefits that should be available through the use of precision agriculture.
The site-specific application of inputs such as seed, fertilizer and crop protection chemicals has the potential to reduce input costs, maximize yields, and benefit the environment. The economic returns currently received by the early adopters of precision farming methods need to be improved before wide-scale acceptance of this practice will occur. These improvements include cost-effective identification and management of the spatial variability of soil and nutrients, applying inputs based on each site's productive capacity, and correct decision-making using the available layers of information. Electrical conductivity (EC) measurements of soil have long been used to identify contrasting soil properties in the geological and environmental fields. The purpose of this paper is to discuss the applications where EC maps are proving useful in improving economic returns to precision farming.
Yield monitoring instrumentation may be divided into a measurement part and a user interface part. This paper addresses observations and discoveries relating mostly to the user interface of yield monitoring for conveyor harvested crops. Discoveries regarding the use of real time yield monitor information for operational decision making are discussed. Next, operator interface for ease of installation and system performance checking is addressed. Brief attention is given to additional sensors to augment the weight transducer technology to better quantify flow rates. The paper concludes with comments on the value of real time data transfer for management decision support.
Three independent yield-sensing systems were operated concurrently on a 1997 WIC Mini-Tank Sugarbeet Harvester. Data were collected on about 80 ha (200 acres) during the 1997 harvest season. One system was a set of load cells mounted near the end of the outlet conveyor. Another system was a second pair of load cells mounted on the scrub chain discharge conveyor. A third system was a torque sensor mounted in the scrub chain driveline. A HarvestMaster HM-500 yield monitor was used for the data collection. Data from each system were collected simultaneously to produce three parallel sets of yield data for each field. The systems were evaluated by comparing truckload error and standard deviation associated with each system and by comparing yield maps generated from the data produced by each system.
A continuous mass flow type yield—load monitor was developed for mapping spatial yield variability in processing tomatoes. The load monitoring system consisted of a load cell to measure weight of tomatoes over a section of a boom elevator and an angle transducer to measure the inclination of the boom elevator. A differential global positioning system [DGPS] was added to the load monitoring system to provide a yield monitoring system. The yield monitoring system was calibrated and validated using a GT weigh wagon during the early part of the 1997 harvesting season. It was found to be accurate within 2.5%. This yield monitor was used to map the variability in tomato yield. There were significant spatial variations in tomato yield. Typically lowest 20% of the yield was less than half the highest 20% of the yield.
During the 1997 harvest season, a peanut combine equipped with a load cell yield monitoring system was used to harvest 37.6 ha (119 ac) of irrigated land in Georgia. Yield data from two fields are presented to demonstrate the yield monitor's ability to characterize spatial yield variation as well as quantify cumulative yield. Data recorded from individual wagon loads averaged less than 5% error, and whole field errors were <1%. Analog and digital signal conditioning methods used in the combine instrumentation are discussed. Practical techniques for correcting erroneous data are also described. Interpolation methods used for map creation are identified and justified according to the spatial characteristics of the collected data. Improvements in the yield monitoring system over previous systems are discussed along with the system's limitations and suitability for commercial use.
The measurement of actual harvested area per unit time is an important component in the creation of accurate crop yield maps. For row crops, such as corn (zea mays L.), these measurements can be made manually on most conventional yield monitors. However, in drilled or broadcast crops a more accurate and automated method is required. In this study, a vector method is developed to determine actual combine harvest area at each time step of the harvest process from a global positioning system (GPS) trajectory. The algorithm was coded into a geographic information system (GIS) and modifications were made to increase computational efficiency. The method was compared with a previously reported raster method of swath width determination on data collected during the 1997 drilled soybean harvest. The vector method greatly improved yield accuracies over the assumption of constant swath width, and provided a number of distinct advantages over the raster method.
In 1997, commercial yield mapping was available for cotton pickers on a limited basis. The Zycom AgriPlan 600 system was installed and evaluated in Texas and Arizona. An experimental system based on weighing the harvested cotton mass was also tested in 1997 on a cotton stripper operated in the southern High Plains of Texas. Both systems were evaluated for their accuracy in predicting yield at individual points within a field. Two yield samples were manually harvested at each evaluation point. The sample variance was used to determine the confidence interval of the manual yield estimates, as a basis for judging the accuracy of the yield mapping systems. The relatively large sample variance resulted in a 95% confidence interval of approximately 0.8 bales of lint per hectare (1/3 bale a-1). The weighing system confidence interval was approximately equal to that of the manual samples, while the Zycom confidence interval was approximately 50% greater.
This paper describes two methods of yield mapping in Japanese paddy fields. One is the yield mapping of grain and straw by hand at the Takatsuki experimental farm of Kyoto University. The variety of rice was MINAMI-HIKARI. The field was divided into three kinds of the rectangular cells. Yield maps showed large variability in relation to position. The variability of grain and straw yield were 5.4 to 9.7t ha−1 and 6.2 to 18.0t ha−1, respectively. The moisture content of grain and straw varied 9.5 to 23.1% wb and 37.8 to 73.0% wb, respectively. The second method was conducted as a trial to measure the yield of grain and straw using a head-feeding combine. In this case, measurement of grain yield through straw yield also was investigated. We estimated the relation of yield between grain and straw and produced the grain yield map by this method.
The point-to-point accuracy of real-time grain yield mapping is decreased by a process of flow convolution as grain moves through a harvester. Yield data can be deconvolved with a transfer function. This paper outlines the possibility of using a strip of wetted grain to estimate the parameters of the transfer function. Some results are shown for a preliminary trial performed on a barley crop. The conclusion from the work is that the grain pulse was too wet to adequately represent the normal state of the harvester's threshing dynamics.
Digital maps are used to visually present the observed spatial variability in many soil and crop attributes. Real-time yield monitor data in particular provides observations which possess inherent measurement errors. These operational or measurement errors are additive and will impart a level of uncertainty to the yield data. The local spatial variation and the method chosen for predicting onto a regular grid for map construction will also contribute to the total uncertainty. This paper proposes that the uncertainty in any estimates should be quantified and that block kriging is an ideal prediction technique for this purpose. It is shown that the optimum block size (minimum map resolution) may be determined from the response of uncertainty to increasing block size. The information content of digital maps is also quantified, and when coupled with the optimum block size, can be used to determine the optimal spectral resolution for a map.
The objective of this study was to demonstrate an approach for determining the vertical accuracy of different global position satellite systems. Two hundred points were selected for comparison in a field with a gently rolling topography. The highest point was approximately 30 m higher than the lowest point. Elevation information from the differentially corrected GPS systems were compared with surveyed elevations. The regression equation between the local differentially corrected carrier phase receiver (y) and surveyed elevations (x) was: y = 0.01 + 0.99x; r = 0.99**. Based on this analysis, the carrier phase receiver was not biased and the vertical errors were <2 cm. The regression equation between the Satellite differentially corrected code phase receiver and surveyed elevations (x) on different days were: y = -1.35 + 1.28 x (r = 0.82**), and y = 1.79 + 0.70x (r = 0.78**). These results showed that the differentially corrected code phase receiver produced inconsistent biased vertical measurements.
Tractor hitch pins, instrumented by NASA Langley Research Center, are being used between tractors and cultivating equipment to log the spatial variability in the mechanical resistance or drag force required for cultivation. The INEEL Site-Specific Technologies for Agriculture (SST4Ag) research project, in collaboration with university, other federal agency and industry partners, is correlating soil physical spatial variability, measured as tillage tool drag force, with crop productivity.
Bulk soil electrical conductivity (EC) is influenced by soil water content, salts, and parent material. This study evaluated the ability of the EM-38 or Veris 3100 soil mapping system to describe soil parameters. The Geonics EM-38 is a widely used noninvasive soil electromagnetic induction meter that measures soil (EC) for the top 120 cm of soil (vertical configuration 30 cm from soil surface). Veris 3100 sensor cart is a direct contact soil EC meter that measures soil EC for the surface 33 and 100 cm of soil. A comparison was made between both systems at matched GPS location points. Additional information at these points were gravimetric water content and NO3-N concentration. The two measuring systems were highly correlated to each other and water content. The EM-38 and Veris deep reading were similar while the Veris shallow reading was lower and higher than the EM-38 in the dry and wet areas, respectively.
Livestock producers want to control manure application with more accuracy, for crop management and environment protection. The lack of appropriate equipment for solid manure has lead to the development of an integrated weighing platform, designed for box spreaders with bottom conveyor. After calibration, this sensing device allows measurements of the manure flow during spreading with an acceptable accuracy. In the near future, it could be incorporated into a regulation system with control on the conveyor velocity, to apply a constant rate or variable rates within a field, based on application maps.
Two commercial cotton yield monitoring systems (YMS) were marketed in the Fall of 1997. The systems were purchased and installed on three cotton pickers in south Georgia. Installation of both YMS involved much time, labor, and specialized tools. Field-scale accuracies of the systems were investigated by using the systems to map yields on three fields (113,19, and 7 ha) representing extremes in cotton production levels in Georgia. Results indicated one system's accuracies were within 3% of actual yield while the other system exhibited a considerably larger error. Instantaneous accuracies of the two systems were researched by harvesting small plots of varying cotton yield levels. The errors for both systems over 33 m (100 ft) plots were large but were not representative of actual system performance. The growers involved in the investigations were impressed that commercial yield monitors were available, but the YMS overall performance was not sufficiently rigorous or accurate to encourage their purchase. Improved versions of both systems are needed before widespread use will begin in Georgia.
Previous work has shown that force-time curves for impacting soybeans vary with soybean moisture content. This paper documents an effort to characterize the nature of the relationship between a parameter developed from the force-time curve and soybean moisture content. The underlying objective was to determine if impact force sensing could be used to reliably make real-time moisture content measurements for yield monitoring applications. A parameter, C, based on the peak force and duration of the force-time curve was calculated. C was regressed onto moisture content using a polynomial model. The model fit the data with R2 = 0.84 and 0.76 for the Conrad and Jack cultivars, respectively. The variability in C increased as moisture content decreased, and the strength of the relationship between C and moisture content decreased with increasing moisture content.
Nowadays precision agriculture requires new equipment and systems that are mounted in agricultural tractors and implements. These systems usually have distributed architecture and are composed of several devices like sensors, actuators, control elements and supervision and control units, all of them intercommunicating in real time. This application requires robustness, flexibility and expansion possibility, involving devices of different manufactures. In that sense, several standards are being proposed in order to help to achieve these goals. The ISO 11783 standard specifies a serial data network for communication and control in tractors and implements, standardizing the method and format of the data interchange between control elements, actuators, computers, sensors and other intelligent devices connected in a system. It is based on the CAN specification (Controller Area Network), used nowadays in other applications. This paper discusses the main characteristics of the ISO 11783 standard, and presents its adoption in a planter monitor with a GPS receiver, used in precision agriculture in order to generate a planting map. This monitor has distributed architecture, and is composed of a main module in the tractor and a sensor module in the planter. This implementation allows for the use of new intelligent sensors and modules, and is compared with other usual implementations.
The design of a cotton Yield Monitoring System, using electro-optical devices, was tested in field and laboratory conditions with better then expected results. As the cotton passes through the chute, it is illuminated by the emitters. The detectors that are positioned across the chute detect the light pulses. The detection circuit then generates pulse train proportional to the length of time the light detected at the detector was blocked. The pulse count is relative to the mass of the cotton passing through the conduit. The sensor is constructed of an array of infra red emitters and detectors assembled in self cleaning brackets. The optical components are kept clean by the application of continuous clean airflow under pressure. Light test is being conducted during operation to verify performance and compensate for inoperative or partially operating light emitter - detector pairs. A sequence of calculations is being performed taking into account the sensor performance, the yield level, moisture of the cotton, the variety, and other parameters. The cotton weight and the yield is displayed on the cab indicator and also stored on a recording media, together with the GPS position of the picker. The data can then be transferred to the office PC, processed and displayed as a yield map.
Over the last decades much effort has been invested in the development of complex simulation models, incorporating and integrating the current understanding of soil–water–plant interactions. The use of such models in precision agriculture has been shown to have great potential. This research presents a methodology to derive basic units for precision agriculture, referred to as management units. Their main purpose is to reduce the theoretically infinite variability of growth conditions in the field to a limited set, which can be evaluated using mechanistic models. A quantitative criterion is applied to ensure that management units accurately represent local variation with respect to growth conditions. Using representative soil profiles for each management unit, real-time simulations can provide insight in crop performance and the nutrient status of the soil. This information can be used to optimise farm management, maintaining crop performance while reducing environmental impacts.
Past efforts to correlate yield from small field plots to soil type, elevation, fertility, and other factors have been only partially successful for characterizing spatial variability in corn (Zea mays L.) yield. Furthermore, methods to determine optimum N rate in grids across fields depend upon the ability to accurately predict yield variability and corn response to N. In this paper, we developed a technique to use the CERES-Maize crop growth model to characterize corn yield variability. The model was calibrated using 3 yrs of data from 224 grids in a 16 ha field near Boone, I A. The model gave excellent predictions of yield trends along transects in the field, explaining approximately 57% of the yield variability. Once the model was calibrated for each grid cell, optimum N rate to maximize net return was computed for each location using 22 yrs of historical weather data. These results were used to evaluate the economic benefit associated with variable rate N prescriptions.
Precision farming revolves around the use of information-based technologies. The biological information link between the crops and these technologies are obtained from scouting reports and plant mapping. However, these activities are time consuming and expensive. Crop simulation model running on real time can provide many of the soil-plant-weather information at a fraction of the cost. Initial performance evaluation of the ICEMM-cotton simulation model to a precision farming study at the King Ranch in Kingsville, TX, shows promising use of the models. A 40-ha field divided into 20 blocks was used for the study. Agronomic practices, daily weather data and soil physical and chemical properties (collected at the start of the planting season) were used as inputs to the model. The date of occurrences of developmental events were predicted within 2 d of actual occurrences. The final plant height and node numbers were close to the observed values in the 20 blocks. Although the overall average yield data approximated the actual yield, individual block yield comparisons were inconsistent partly due to incorrect soil hydrologic parameters.
Crop simulation models have been used historically to predict average field crop development and yield under alternative management and weather scenarios. The objective of this paper was to evaluate and test a new version of the CERES-Maize model that was modified to improve the simulation of site-specific crop development and yield. Seven sites within a field located in central Missouri were selected based on landscape position, elevation, depth to the claypan horizon, and past yield history. Detailed monitoring of crop development and soil moisture during the 1997 season provided data for calibration and evaluation of the model performance at each site. Mid-season water stress caused a large variation in measured yield with values ranging from 3.0 Mg ha-1 in the eroded side-slope areas to 11.7 Mg ha-1 in the deeper soils located in the low areas of the field. The results obtained demonstrated that the modifications improved the ability of the model to simulate site-specific crop development. Areas of potential model improvement and further investigation were discussed.
The CERES-Maize model was used to estimate the spatial variability in corn (Zea mays L.) yield for 1995 and 1996 using data measured on soil profiles located on a 30.5 m grid within a 3.9 ha field in Michigan. The model was calibrated for one grid profile for the 1995 and then used to simulate corn yield for all grid points for the 2 yrs. For the calibration for 1995, the model predicted corn yield within 2%. For 1995, the model predicted yield variability very well (r2 = 0.85), producing similar yield maps with differences generally within ± 300 kg ha-1. For 1996, the model predicted low grain yields (1167 kg ha-1) compared with measured (8928 kg ha-1) because the model does not account for horizontal water movement within the landscape or water contributions from a water table. Under nonlimiting water conditions, the model performed well (average of 8717 vs. 8948 kg ha-1) but under-estimated the measured yield variability.
Two machine learning methods based on the induction of regression trees and Bayesian networks from data were used to predict wheat yield in 1996 from fourteen soil, remotely sensed, and other variables for a field located near Wyalkatchem, Western Australia. The regression tree model explained 39.7% of the variation in the data, while the Bayesian network model explained 69.1% of the variation. Both models predicted similar spatial trends in yield, and both models predicted yield to within ±0.25 t ha-1 of measured yield >50% of the area of the field. In comparing model types, both models suffered from an inability to predict yields >0.33 t ha-1 or yields >1.89 t ha-1. We suggest that Bayesian network models are more suitable frameworks for site specific management research than regression trees, or other machine learning methods similar to regression trees.
A robust new approach for describing and segmenting landforms that is directly applicable to precision farming has been developed in Alberta. The model uses derivatives computed from DEMs and a fuzzy rule base to identify up to 15 morphologically defined landform facets. The procedure adds several measures of relative landform position to the widely used classification of Pennock et al., (1987). The original 15 facets can be grouped to reflect differences in complexity of the area or scale of application. Research testing suggests that a consolidation from 15 to 3–4 units provides practical, relevant separations at a farm field scale. These units are related to movement and accumulation of water in the landscape and are significantly different in terms of soil characteristics and crop yields. The units provide a base for benchmark soil testing, for applying biological models and for developing agronomic prescriptions and management options.
Winter wheat-fallow is the conventional cropping system in the semiarid central Great Plains. Fallow weed control relies on tillage, but this practice leads to erosion and loss of organic matter. Cropping systems studies in this region, conducted since the early 1980s, are demonstrating that annualized yield can be doubled and soil quality improved with continuous no-till cropping. However, with only four research sites in the central Great Plains, producers question how far experimental results may apply across this region, given the extreme variation in soil properties and climatic characteristics. Therefore, we identified agroecozones of similar soil and climatic characteristics in relation to the four cropping systems sites. The winter wheat-fallow region can be divided into two agroecozones, with a zone of uncertainty identified. Within an agroecozone, soil differences may have a greater impact on land productivity than climatic factors.
Adequate information systems are one of the main needs of precision agriculture. Although many systems have been proposed as research tools and some others have been available in the market, there are many issues concerning their completeness, compatibility, cost and user-friendliness, to mention a few. Aiming at contributing to the understanding and to the solution for the availability problem of more adequate systems a model for such a class of systems was developed. It was based on software engineering fundamentals such as domain analysis concepts and object oriented modeling methods. This paper discusses the need for a broader understanding of the role of information systems in precision agriculture in view of the quick changes in the technology. The modeling method is described and some of the results are presented.
Well water N03 --N concentrations in the central part of the San Luis Valley had been found as high as 72 mg N03--N L-1. Precision management practices according to differences in soil type found under a center pivot irrigation sprinkler may have the potential to improve N-use efficiency. No Nitrogen Leaching Economic Analysis Package (NLEAP) model simulation of effects of different soil types and crops on residual soil N03--N (RSN) has been conducted under a similar irrigation system. Lettuce (Lactuca sativa L. ), potato (Solanum tuberosum L ), winter cover rye (Seca/e cereale L.), and winter wheat (Triticum aestivum L.) were grown on sandy loam and loamy sand areas of a center-pivot. A new NLEAP 1.2 version simulated soil type and crop effects on RSN (P < 0.001). NLEAP is a potential technology transfer tool that can be used to protect water quality by evaluating the effects of crops and soil types on RSN and can potentially be used with precision agriculture management practices.
Manipulating agricultural systems for ecological, economic and agricultural gains is becoming increasingly important to agricultural policymakers and planners in both developed and developing countries and to the national agricultural commodity and trade industry. Systems tools such as crop growth simulation models are an important component in satisfying the above requirement. Crop growth duration is among the major determinants of yield. Existing crop growth simulation models have reliably predicted effects of temperature and photoperiod on crop duration. However, these models generally do not consider the effects of extremely high or low temperatures, drought stress, and nutrient deficiencies on crop duration. Drought stress and deficiencies of N and P during the vegetative phase have resulted in delayed tassel initiation and silking. Similarly, stresses during the ripening phase have resulted in early senescence and maturity. Thus, harvesting, yield forecasting, and planting of the following crop in a sequence, livestock rearing or fisheries would be influenced. A modified model that simulates the effect of N deficiency on phyllochron and phenological stages is presented. Simulation results are compared with field trials from Nigeria, Hawaii and Florida for tropical maize. Simulation studies showing importance of timely N management on yield and risk avoidance are presented.
Modeling is an important tool for the decision-making process for site-specific management. An integrated model such as the Environment-Policy Integrated Climate (EPIC) can be used to assess the potential economic and environmental consequences of site-specific nitrogen fertilization. In this paper we test this application of EPIC to four small watersheds (280 to 380 ha) in southern Ontario, under intensive cropping dominated by soybeans, corn and winter wheat, and various tillage systems. In order to minimize data requirements, a simple area-weighting average will be used to produce watershed values from soil, crop and tillage-specific model runs. We will analyze the ability of the model, under this pooling strategy, to represent current conditions and two hypothetical 20-yr scenarios - N-fertilization using recommended rates or using site-specific N requirements. The analysis will focus on corn and winter wheat grain yields, and on total N discharge at the watershed outlet.
Weed species occur in non-uniform Patches across agricultural fields with the amount of patchiness differing among weed species and field. This patchiness complicates herbicide recommendations; however, herbicide applications can be targeted to specific areas by identifying the locations and weed species in the field. The size, shape, and location of weed infestations can be determined by intensive field scouting, but this approach is expensive. Remote sensing information, combined with ground-truth data, may provide a useful method to solve this problem. A study was conducted to determine the feasibility of using remote sensing techniques to detect weed populations at several stages of corn growth. A charge-coupled device (CCD) with four spectral filters mounted in an airplane was used to obtain several near digital images with 1 m * 1 m resolution of a 65 ha no-till corn field from May through September 1997. The CCD sensor contained four spectral filters sensitive in the blue, green, red, and near infrared (NIR) wavelengths. Latitude and longitude coordinates of the field perimeter were integrated into a geographical information system so that coordinate of anomalous areas of the field could be identified. Using the coordinates of the anomalous areas, ground scouting from the aerial images was conducted to define the nature of the anomaly. Comparing the aerial images to the ground-truthed data indicated that the NIR wavelength showed the greatest differences between corn and weeds. A normalized difference vegetative index was generated using the image corresponding with maximum green canopy cover and was highly correlated with corn yield, indicating that future yield predictions may be possible using remote sensing, however, field scouting was necessary to distinguish weed species and densities. Remote sensing combined with ground scouting provided an excellent method to determine the location of weed infestations over an entire field, to create a database for site specific herbicide management, and to monitor changes in weed species density over time.
Major changes have occurred and continue to occur in remote sensing technology in recent years. Significant changes related to advances in spatial, spectral and temporal resolution. For precision farming, this means that remote sensing can provide (i) much greater detail of an individual field, (ii) much more precisely defined colors or delineations of variations of the vegetation, residues or surface soils, and (iii) repeat viewing of the same scene every 2 to 7 days. In effect, the farmer can detect missing plants or stress damage, assess the causality of a stress by looking more closely at the spectral responses, and receive data—information in a timely fashion so that corrective action may be taken. Within the next 10 yrs there will be more than 50 new land viewing satellites. The impact on precision farming means more data available, varying types of data such as wavelength bands that are specific to stress conditions, more timely information and costs which should be reasonable to the farmer because of the competition.
Studies during the past 25 yrs have shown that measurements of surface reflectance and temperature (termed optical remote sensing) are useful for monitoring crop and soil conditions. Far less attention has been given to the use of radar imagery, even though Synthetic Aperture Radar (SAR) systems have the advantages of cloud penetration, all-weather coverage, high spatial resolution, day-night acquisitions, and signal independence of the solar illumination angle. In this study, we obtained coincident optical and SAR images of an agricultural area to investigate the use of SAR imagery for precision farm management. Results showed that SAR imagery was sensitive to variations in field tillage, surface soil moisture and vegetation density. The coincident optical images proved useful in interpretation of the response of SAR backscatter to soil and plant conditions.
Remote sensing provides an opportunity to collect information about crop growth during the growing season. The ability to relate remotely sensed data to physical anomalies of the growing crop or soil will permit producers an opportunity to correct problems in a timely manner. Relationships between remotely sensed data and field anomalies are very limited. In this study, three production fields (16 to 24 ha), two in corn and one in soybeans, were used to compare ground and remotely sensed data. These data were collected four times during the 1997-growing season by low-altitude fixed-wing aircraft with instrumentation having a resolution of one meter. The wavelengths of the remotely sensed data were red, blue, green, infrared and near-infrared. Geospatial crop and soil data were collected on the ground throughout the growing season. At the times of remote sensing, intensive ground soil and crop data were collected. Some of the anomalies observed in these production fields were caused by white mold infestation in soybeans, weed infestations, wheel traffic, planter and sprayer malfunctions, tillage, topography and soil morphology. The results of these anomalies were observed in the remotely sensed data.
Conventional recommendations of N fertilizer are based on composited soil samples taken over the entire field. This may result in either underfertilization or overfertilization due to the neglected spatial variability in the field. This paper shows a methodology to estimate in-season soil N at various growth stages of irrigated corn. Soil N was estimated from plant N assessed through remote sensing. The nitrogen reflectance index (NRI) was employed to estimate plant N. Regression analysis between soil N and plant N showed good relationships at various growth stages. Geographic information system (GIS) mapping of measured and estimated soil N showed an agreement except in locations where hot spots were measured.
A significant problem with the site-specific management of crops is the amount of data collection required to confidently map crop conditions and related parameters. In Taylor et al (1997) high resolution remote sensing data using Airborne Digital Photography (ADP) was used successfully to map the within-field spatial distribution of crop parameters and yield potential for winter barley in the England. The calibration was based on the regression equation derived from field sampling vs. NDVI measurements. Our first approach used a large number of field observations to calibrate the ADP data, which inhibits the approach from being adopted for practical use. This paper explores the possibility of reducing field sampling and demonstrates a practical methodology for field calibration of ADP data as a source of management information for precision farming in wheat and barley. Data taken from 90 field quadrat observations is compared statistically to a reduced sample size (24 quadrats) that was extracted from the same data. Using parallel lines analysis the reduced sample is shown not to be significantly different from the larger sample. The reduced sample technique was put into practice in four fields growing cereals. In each field seven sample sites were selected representing the range of tiller variation. The crop was sampled at each measurement site, at two times during tillering, using three 50□50 cm quadrat samples, arranged in a triangular subset. The coefficients of determination were very high, e.g., 0.98, 0.95, 0.94, 0.78, with probabilities <0.002; in one case an apparent background soil effect reduced the r2 to 0.63. The ability to map crop parameters and yield potential during a growing season provides a valuable source of management information to the farmer and agronomist. The development of a rapid methodology for field-based calibration will allow the use of ADP for the monitoring of crop parameters, bringing this technology closer to being used for monitoring both crop condition and the effects of managing variable inputs throughout, a growing season.
A case study is presented using 2 yrs of yield data in conjunction with high spectral resolution data, HYDICE and AVIRIS, and high spatial resolution imagery, simulated IKONOS. Procedures for integrating yield monitor data and remotely sensed data are discussed. There are many errors inherently associated with yield monitor data that need to be considered when using the data for scientific research. Correction methods which best suited this case study are presented. Statistical correlation between individual bands of hyperspectral imagery and yield data is analyzed on the basis of the entire agricultural scene, crop type, varieties, bare soil, vegetation cover, and sensor type. Averaging filters are used to study the effect of differing spatial resolutions on hyperspectral data correlation to yield. These results were then used to assess the performance of classifications to identify yield patterns prior to harvest. Traditional methods of detecting vegetation patterns are discussed, such as filters, linear stretches, tasseled cap, principal components, and classifications. An evaluation of hyperspectral bands with relation to yield for specific anomalous regions is given. Statistical correlation between specific anomalies identified in the field to the corresponding yield information was studied. It lays the groundwork for further research related to spectral responses of specific anomalies and their effects on yield. These results will be applied in other continuing studies to develop yield evaluation and anomaly detection models using varying spectral and spatial resolutions.
Precision farming involves crop management in parcels smaller than field size. Yield prediction models based on early growth stage parameters are one desired goal to enable precision farming approaches to improve production. To accomplish this goal, spatial data at a suitable scale describing the variability of yield, crop condition at certain growth stages, soil nutrient status, agronomic factors, moisture status, and weed—pest pressures are required.
This paper discusses the potential application of aerial imaging to monitor and predict the potential yield for corn and soybean at various growth stages in the season. Included in the analyses were aerial images, yield monitor data and soil grid sampling. The relationship between remotely sensed Normalized Difference Vegetation Index (NDVI) and yield was best at 9 m spatial resolution. Preliminary results indicate that it is possible to use NDVI to estimate the potential yield for soybean and corn when canopy reaches full cover.
Yield of processing tomato was measured in the Sacramento Valley, California, in three furrow-irrigated farm fields of 32 to 44 ha in size with a prototype weighing yield monitor mounted on a conventional single-row harvester. Normalized difference vegetation index derived from false-color infrared aerial photographs taken at full bloom was not related to fruit yield. However, fruit yield and vine dry weight just before harvest were closely related, suggesting that late-season factors caused yield variability. In one field, fruit yield was lowest both in the poorest-drained and the best-drained areas. In the best-drained area, soil texture was coarser, but yields were low, possibly because of inadequate lateral flow of water from the furrow to bed center. Yields were also low in the poorest-drained areas of the field due to prolonged saturation of soil following irrigation.
This study uses GPS-GIS-RS techniques to analyze cranberry (Vaccinium macrocarpon Ait.) crop health and yield. Extensive field sampling has been used in the past as a means of estimating potential bed yields. The major problem for predicting yield appears to be to high intra-bed spatial variability. For this study, color-IR photography from commercial cranberry beds (May 1996) was rectified to earth coordinates using GPS technology. An unsupervised multi-spectral classification and an NDVI were done to statistically group pixels in the image. Results indicate that a number of features within cranberry beds can be identified, including variations of vegetative cover, irrigation and drainage systems, and areas of beds damaged by insects and fungal disease (Phytophthora cinnamomi). In the future, remotely sensed imagery will be linked to ground based data to gain further insight into the spatial variation of factors affecting crop yield and health.
The number of suppliers of remote sensing images taken from airborne platforms and space vehicles will increase in future and thus extend the limited availability of this information tool. Remote sensing enables a fast and economic data acquisition for the efficient investigation of soil and crop features. This paper presents existing and future tools for the collection of remotely sensed data, alternatives of data processing and examples of how this information can be efficiently used for the coordination and optimization of soil and plant sampling strategies. Special emphasis is put on the potential misinterpretation of vegetation indices, which leads to an inappropriate evaluation of the plant nitrogen status of the crop.
One of the most important irrigation management decisions in California Cotton Production is the timing of the last irrigation. We analyzed data obtained by scanning a sequence of false color infrared aerial photographs of an experiment at two field sites. The experiment involved a comparison of different dates on which the last irrigation was applied. Our preliminary results, based on one year of data, indicate a strong relationship between yield and vegetation index. They also indicate that remote sensing can detect differences in crop development due to soil texture differences within the field. Thus, remote sensing may be of value in assisting in timing the final irrigation and in determining whether within-field differences in soil texture are sufficient to economically justify a switch from furrow irrigation to other more precisely controlled systems.
Spectral reflectance sensors have been developed that can detect the presence of plant material against a background of bare soil. These systems are able to identify areas of weed infestation in fallow fields or in the area between the rows of a row crop and control the application of herbicides. In typical U.S. corn production, a number of options are available for controlling weeds in between the rows, e.g., broadcast herbicide application or mechanical cultivation. In-row weed control might be possible with nonselective herbicides if detected plant material areas can be segmented into crop and weed subsets. Accomplishing this task in real time would allow application of the herbicide to only the weeds, which could significantly reduce the amount of herbicide used. This study reports on the use of an industry-developed spectral reflectance sensor to identify and locate plant material and the experimental C-language software that was developed to segment the corn plants and weeds.
A sensor is proposed here which distinguishes between crop and weed based on their different spectral reflectances. The sensor is built upon an imaging spectrograph. We chose this spectral reflectance sensor because of the fast spectral imaging and possible high spatial and spectral resolution. Parameters like the angle-of-view and the quality of the optics were optimized for maximal performance within reasonable cost.
Classification success rates depend not only on spatial and spectral filtering, both characteristics of the device, but also on the number of wavelengths and the crop itself. Under controlled conditions, corn and sugar beet can be separated from weed with a success rate of at most 90, respectively 80%. Herbicide savings which depend on weed density, the nozzle activation frequency and the spray resolution (width), are maximal with the MLNN classifier.
Remotely sensed spectral data were used to assess the incidence of Sclerotinia stem rot of soybean caused by the fungus Sclerotinia sclerotiorum and to determine its effect on variability of soybean yields. Multispectral data were obtained with an ATLAS sensor (Airborne Terrestrial Applications Sensor), yields were mapped with a combine-mounted yield monitor, and field disease assessments made both visually and by means of spectral reflectance observations obtained with a handheld radiometer. Limitations in data obtained during the ground truth survey prevented use of multispectral data for disease assessment. However, our results indicate that disease incidence and crop yield can be estimated from spectral reflectance data, that plant disease can explain a high percentage of yield variability in a production soybean field, and that diseased areas can be mapped using precision agricultural techniques. This information will enable growers to use variable rate technologies to control Sclerotinia stem rot.
Some nutrient levels are often related to topography in the Northern Great Plains, particularly N. The use of topography to direct sampling is attractive because it may require fewer samples to reveal fertility patterns compared with dense grid-sampling methods. However, the use of topography-based sampling requires an additional measure of agronomic knowledge from the sampler. Determining where to establish soil-sampling zones may be difficult using only a topography map. Yield mapping, satellite imagery, and soil conductivity may be useful in determining sampling zone boundaries. When similar patterns to topography are produced using yield mapping, satellite imagery or soil conductivity sensors, soil sampling zone boundaries may be more confidently established or refined.
Development of new methods for weed control is one of the biggest scientific and technological challenges in the future. Sensor steered sprayers and prototypes of robots steered by digital video cameras have been developed for non-chemical weed control. In 1997, a collaboration project between the Danish National Railways Agency, the Danish Institute of Agricultural Sciences and Hardi International was initiated. The project aimed at developing a digital camera system (WeedEye) and a spot sprayer to reduce herbicide usage to locations with weeds. The results of the first experiments show that WeedEye detects small weed seedlings at a speed of 45 km h−1. Further, the results show that WeedEye provides a reasonable estimate of the percentage leaf area m−2 at this speed. The perspectives of using sensors and digital video cameras in agriculture are discussed.
This paper presents a system for the detection of weed amongst crop based on the extraction of structural field information. Data is gathered online with a sensor built upon an imaging spectrograph optimized for this purpose. The optical sensor splits the light from a line on the ground parallel with the spray boom in its spectral components that are projected on a camera. One of the main advantages of this sensor type is the high spatial resolution (up to a few mm). The weed detection algorithm uses the reflectance differences between the red and the near-infrared bands to make the distinction between plants and soil. Detection of weed amongst crop is based on the detection of weed patches and the extraction of the crop rows. A mathematical model for the herbicide reduction and the hit rate is presented that shows the relations between the different system parameters such as the spray resolution and the performance of the classifier.
Sugarbeets stand to benefit from site-specific management to a greater extent than many other crops. Yield maps for sugar production have been unavailable as no effective method of spatially quantifying sugar content has been available. An experiment was conducted in an attempt to relate spectral properties of beet canopy with sugar content and quality at harvest. Detailed spectral measurements, taken in August, September, and October of 1997, were related to beet quality. Models were developed using canopy reflectance at three and four spectral bands combined in indices and related to beet quality. The best model involved spectral bands at 500, 550, and 830 nm and was able to account for just >50% of the variation in sugar content of the sampled beets. The model was also able to predict one half or more of the variation in sodium, amino N, and recoverable sugar per ton.
With the future launch of high spectral and spatial resolution satellites, we foresee that that it will be possible to use the information from these satellites in a timely fashion for the detection of stress, the prediction of yield and as a diagnostic tool for recurring low productivity areas for numerous crops. Preliminary results indicate a correlation between the vegetation fraction of airborne hyperspectral pixels and ground truth validation measurements. The advantages of the pixel-unmixing algorithm for within-field crop properties mapping will be demonstrated.
An improved understanding of linkages between plant crop stress level, actual canopy vegetation, and vegetation index is needed to properly evaluate the relationship between false color infrared aerial photographs and yield. In Part II of this paper, we identify some of the key plant and soil based parameters useful in the evaluation of vegetative indices and yield prediction for irrigated cotton. The accumulation of crop water stress over time is partially related to climatic conditions and soil characteristics such as water holding capacity and rooting depth. These soil water storage parameters were found to be highly effective in predicting an ideal irrigation termination date for cotton and associated canopy development. Our preliminary analysis of one-year results suggests a strong correlation between plant water stress and crop canopy size in the late-season for fields having highly contrasting soil moisture regimes.
A Critical issue for Crop Management is to get the variability at the subfield level. Ground measurements and yield monitoring techniques are one method to achieve these goals. But remote sensing also can be a very efficient way to reach these objectives with the use of hyperspectral (superspectral) data. MATRA MARCONI SPACE (MMS) one of the largest space companies worldwide has been involved since 3 yrs in the definition of a new commercial satellite system able to generate such superspectral remote sensing data with a high revisit frequency and fast delivery. Working with the National French Agronomic Research Institute and using agronomic reflectance models it was possible to generate from the superspectral images biophysical indexes very important to detect stress (water, disease) or to create application maps (fertilizer). Since the last 2 yrs MMS has performed a lot of commercial campaigns in the USA and in Europe to validate this concept and test on different types of crops. The robustness of the agronomic algorithms and demonstrate the accuracy of the method.
Now that remotely sensed images of agricultural fields are commercially available, questions arise about proper use of this technology and interpretation of the data. A ground-truthing guide for remotely sensed fields is in progress. One of the two methods studied for this guide includes interpretation of weekly images to guide sampling to characterize soil variability and crop stresses. This is in contrast with pre-season determination of sampling sites based on GIS guided classification that utilizes available image data. The focus on the latter is a nontraditional approach to ground truthing stresses based on GPS guidance to predetermined points. The methodology and timeliness of sampling for soil variability and crop stresses are of interest to the users of remotely sensed products.
With the large number of herbicide options available, deciding which herbicide is best for a given situation may be overwhelming resulting in poor or incorrect decisions. GWM, General Weed Management, a bioeconomic model, has been proposed as an approach to simplify herbicide selection. Estimated net return is based on predicted yield and commodity price minus the loss from weed competition and control cost. Weed species and density in the evaluation areas are used as input parameters and a list of herbicide options based on predicted net return and efficacy is generated. The experiment evaluated different weed management strategies at different landscape positions. Five different strategies were tested in a RCBD at three locations in two fields in 1996 and 1997. Locations within fields were selected based on landscape position and previously mapped weed populations. Treatments included a weedy control, the weed management strategy selected by the producer, and three GWM generated treatments (preemerge, postemerge and preemerge plus post emerge). Preliminary results suggest that GWM may perform as good or better than the producer blanket herbicide treatment. Model generated decisions with regards to economics may play a role in the future of weed management as cropping practices move to more intensive management practices with precision farming techniques, and therefore need to be further evaluated. Specific results will be presented later.
Plant parameters such as leaf area, plant N status, and crop coefficients for crop Et have been successfully estimated from canopy reflectance measured in discrete, narrow spectral wavebands. Site specific water and N management for irrigated corn will require at least weekly inputs from spectral data to assist with management decisions. Unfortunately, weekly acquisition over large fields with narrow-band spectral data is still not feasible. Therefore, the objective of this paper is to determine the usefulness of 35-mm broad-band spectral data for estimating the plant N status of irrigated corn. Thirty-five mm color plus black and white infrared as well as false color infrared aerial photography was acquired at several growth stages over corn plots with established N treatments. The slide film was developed and scanned at five resolutions and written to CD ROM media by a commercial source. Ground-based canopy reflectance was measured with Exotech four band radiometers filtered in the blue, green, red, and near infrared portions of the electromagnetic spectrum. SPAD chlorophyll meter and leaf area measurements as well as plant samples for N analysis were also taken. All of these measurements were made on the same day of the over flights.
This study was conducted to evaluate the economics of site-specific N management for irrigated corn in central Kansas. Spatially detailed yield mapping and soil sampling were used to impose spatially variable N rates from which economic comparisons of variable and uniform N management was conducted. Less N was applied using variable-rate rather than uniform N management in every year studied. Variable-rate management was determined to be economically viable in some years while not in others. Additional research is necessary in order to generalize these results for application at other sites.
Adopting variable-rate technology (VRT) by agricultural producers depends on several factors including spatial variation in yield potential. Many farmers currently use consulting firms to implement VRT. A farmer breaks even using the technology when there exists a minimum yield variability such that the VRT custom cost is offset by additional returns. A farmer with greater spatial variability than the break-even variability will employ VRT. In this analysis, we have first presented a methodology for determining this spatial break-even variability and then applied it to a hypothetical 30-ac corn field consisting of two land qualities. Break-even variability was determined for a range of nitrogen and corn prices by varying land quality proportions. For the response functions assumed and a custom VRT rate of $4.67/ac, VRT was economically feasible within range of 15 to 70%of the field in good quality land.
This analysis works out the economic implications of the corn (zea mays L.) population response curves by yield potential category estimated by Pioneer Hi-Bred agronomists. Variable rate seeding examples are developed for various mixes of low, medium and high yield potential soil, as well as for a range of seed costs and variable rate equipment costs. The strategies analyzed are: variable rate planting using agronomic recommendations for each yield potential zone, variable rate planting using an economic decision rule, and two information strategies that set a uniform planting rate based site specific yield potential information. The general conclusion is that variable rate seeding by yield potential zone has profit potential only for farmers with some low yield potential land (<100 bu a-1). Farmers with mix of medium and high potential land are better off with uniform rate seeding. The surprise is that variable rate seeding is potentially profitable when the proportion of low yield land is small. In the example, the farm with 10% low yield potential soil shows positive returns to variable rate planting. The results are not particularly sensitive to seed cost, corn price or variable rate equipment investment cost.
In Indiana, variable rate application of lime is often considered a good place to start site-specific management (SSM). This is because pH is one of the most variable of manageable soil characteristics and because spreaders can be retrofitted relatively inexpensively to do variable rate application (VRA). The objective of this study is to evaluate the profitability of VRA for lime as a standalone activity. The methodology involves a spreadsheet model using corn and soybean pH response functions estimated with small plot data. The overall results indicate increased annual returns to corn and soybean production with site-specific pH management strategies. On average, SSM with agronomic recommendations provides an increased annual return of $2.93 acre-1 (+1.78%). SSM with the economic decision rule provides an average increase in annual return of $7.91 acre-1 (+4.82%).
Several studies have evaluated the profitability of variable rate application of one type of input, but few have investigated the profitability of variable rate application of two or more types of inputs. In addition, most studies ignore the impact of variable rate input application on the environment. This study evaluates the profitability and environmental outcomes associated with spatial variation of nitrogen fertilizer and irrigation water in seed potato production. Seed potato yields and nitrogen losses are simulated for four different areas of a 63 ha field using the EPIC (Environmental Policy Integrated Climate) crop growth model. A dynamic optimization model is used to determine optimal levels of N fertilizer for each area of the field. Average nitrogen losses and economic returns are evaluated for both uniform and variable rate application of nitrogen and water. The results indicate greater economic and environmental benefits may be achieved from varying water application than from varying nitrogen application across the field.
A conceptual model is developed to measure the value of information from in-field soil sensing technologies as compared with grid and other soil sampling methods. Soil sensing offers greater spatial accuracy and the potential to apply inputs such as nitrogen fertilizer immediately, avoiding changes in nutrient status that occur with delays between soil sampling and fertilizer application. By contrast, soil sampling offers greater measurement accuracy, because it does not rely on proxy variables such as electrical conductivity to infer nutrient status. The average profitability and relative riskiness of soil sensing versus sampling depend upon (i) the trade-off between, on the one hand, the spatial and temporal accuracy of sensing and, on the other hand, the measurement accuracy of sampling, (ii) the cost of data collection, and (iii) input and product prices. Similar trade-offs govern the relative riskiness of sensing versus sampling.
Objective: To offer farmers an effective analysis tool to evaluate profitability of a field or farm using yield data. While there are numerous options in global information system (GIS) software for the presentation of visualized data, the actual tools to analyze are more limited and certainly less used. One aspect of the analysis of data is most often overlooked by commercial providers is the analysis of profitability. This analysis will require the confidence and cooperation of the producer to be accurate and useful as a management tool. Algebraic formulas that convert yield data to gross and net profit are not that difficult to write, but most yield oriented GIS packages lack the ability to apply them. In our presentation we will demonstrate the practicality and necessity of such financial analysis. We will make a practical application that a Precision Farmer might use in analyzing items such as cash rent, input costs, drainage cost effectiveness and tillage practices. Results: By applying algebraic formulas to yield data we can represent yield as revenue per acre as well as a profitability by acre, hybrid or soil type. We believe that better managers will use this data to decide whether to continue to rent some farms and use the data to re-negotiate other cash rents. Conclusions: Farmers who use yield data for a variety of decision making processes will be the farmers of the next century. Yield analysis comparing varying crops is another area that could be like comparing apples and oranges. By taking yields to an index value based on yearly averages we believe farmers can make financial decisions more readily. Financial analysis will also be performed on this data, demonstrating the practical application of yield data in the determination of profitability for various farms and fields.
The profitability associated with site-specific nutrient management will be impacted by the spatial scale that is used for the interpolation of management decisions. A 12 ha research site near Windom, MN was established to quantify the impact of sampling scale on potential increased return. The field was soil sampled on a 0.06 ha grid for routine soil test values (P, K, pH, etc) to a depth of 15 cm and to a depth of 60 cm for nitrate-N. Three replications of five N rate treatments (0, 67, 112, 156, and 202 kg ha"1) and three P rate treatments (0, 56, and 112 kg P205 ha"1) were applied as constant rate strips, randomized within a split block arrangement with P rate as the main block. Grain yields were obtained every 15 m along each treatment transect. Multiple regression techniques, using a split-block-in-space approach were used to determine yield response and economic optimum rates of N and P fertilization at different locations within the field (minimum resolution was 0.18 ha). The field was subdivided into several different management areas based upon grid cell size (0.36, 0.52, 1.04, 2.08, and 12 ha) and potential management areas based on soil test values, organic matter and topographic features. Each management area was described according to the appropriate soil testing procedure for making fertilizer recommendations. The impact of that fertilizer application was evaluated economically utilizing the response equations determined for that sub-region of the field.
Spatial variability of weeds in a field was used as input information for a bio-economic weed control model to generate preemergence, pre- plus postemergence, and postemergence herbicide strategies at three field locations. Weed control effectiveness, crop production, and profitability were estimated and compared with a producer's blanket herbicide application at each site. The recommendations from the bio-economic model were less expensive than the producer's treatment and resulted in similar or better weed control, yield, and net return. Using site-specific herbicide application and placement would optimize economic returns and environmental safety, benefiting the producer and society.
In a 2-yr field experiment the influence of sewage sludge application to some soil properties was investigated. The soil was a Typic Xerochrept and it was cultivated with cotton after amendment with sewage sludge rates ranging from 0 to 300 ton ha-1 yr-1 in a completely randomized blocks experimental design. The results showed that sewage sludge application significantly increased cotton yield. Its implication on soil properties was as follows: Organic matter content was significantly increased while soil pH was slightly decreased. Available phosphorus was highly increased in the surface layer. Nitrate N was highly increased in the whole soil depth. Phosphorus was strongly correlated with organic matter content.
Soil properties that affect the fate and transport of herbicides and weed density were studied across a soil-landscape in Blue Earth County, MN. Soil properties such as organic matter, texture, pH, and adsorption coefficients of herbicides and weed density varied spatially. Adsorption coefficient (Kd) of imazethapyr was strongly correlated with soil pH while Kd of alachlor was strongly correlated with organic matter. Distribution of broad leaf weeds were related to soil-landscape characteristic. Preliminary results of this research suggest that site-specific application of herbicide (pre- or post-emergence) based on soil properties and weed density can reduce herbicide use.
In a 2 yr field experiment the influence of sewage sludge application on the total and available forms of heavy metals, i.e., Zn, Mn, Cu, Fe, Pb, Cd, and Ni was studied. The soil was a Typic Xerochrept and it was cultivated with cotton after amendment with sewage sludge rates ranging from 0 to 30-ton ha-1 yr-1. The experimental design was completely randomized blocks with five treatments each replicated four times. After 2 yrs of sewage sludge application, soil samples were taken from all the plots from a depth 0 to 50 cm and the concentrations of the total and available forms (DTPA extractable) of the above mentioned heavy metals were determined. The results showed that sewage sludge application significantly increased the total concentration of Cu, Zn, and Pb in the surface layer. DTPA exctractable forms were increased significantly in all the metals studied except Mn. Total concentration of Cu, Zn, and Cd were strongly correlated with organic matter content. DTPA extractable Cu, Zn, Cd, and Fe were also strongly correlated with organic matter content positively and negatively with soil pH. For all the metals studied except Cu and Fe, there was no evidence of leaching in the deeper layers.
In recent years much concern has been given to toxic heavy metals that enter the human food chain. Inorganic fertilizers are considered among the potential avenues of such entry. In this work, we report the analyses of 77 samples of commercial fertilizers, marketed in the Kingdom of Saudi Arabia, for their heavy metal concentrations. Fertilizer samples included 20 samples of phosphatic fertilizers (MAP,DAP,TSP), 11 samples of liquid fertilizers, 34 samples of water soluble multiple nutrient fertilizers (WSMF), and 12 samples of solid multiple nutrient fertilizers (SMNF). Concentrations of heavy metals varied according to the type of fertilizer and the tested metal (Cr levels were the highest and Co were the lowest). The data revealed that Cd ranged from 36.8 to <1 mgkg−1. Nevertheless, the average Cd content was 32.2 mgkg−1 for the phosphatic fertilizers, 13.4 mgkg−1 for the liquid fertilizers, 18.4 for the (SMNF), and 4.5 mgkg−1 for the (WSMF). Concentrations of Pb, Ni, Co, and Cr in the phosphatic fertilizers averaged 17.8, 72.3, 12.9, and 276.8 mgkg−1 , respectively. However, the corresponding average values of these elements, in the liquid fertilizers, were 13.3, 19.4, 12.5, and 85.1 mgkg−1. In the (SMNF) were 14.5, 44.7, 11.7, and 162.0 mgkg−1 and in the (WSMF) samples were 10.0, 7.8, 7.4, and 12.5 mgkg−1. Data showed that Cd, Co, and Ni concentrations were lower than the tolerance limits for heavy metal addition, and apart from Cr metal, concentrations of the other heavy metals were comparable to those recorded worldwide.
A survey of agricultural service providers was conducted to help identify adoption rates and the status of precision farming techniques in Missouri. Results of the survey provide useful information to agribusinesses designing precision farming business strategies and to University Extension educators assisting farmers in making decisions about precision agriculture. The survey revealed that precision farming activities have developed in clusters around agricultural service providers. Established precision farming service providers offer technical support, services, and educational opportunities for farmers in the community. The survey also revealed that agricultural service providers add precision farming services because they predict a growing demand for the services and determine that they must be equipped to meet that demand to remain competitive, especially in fertilizer sales. Guidelines for estimating the costs of precision agriculture services were developed. Although many similarities exist in pricing structures for precision agriculture services, prices were more similar among service providers within geographic locale such as the bootheel or the West Central regions of Missouri. While the return on the investment in precision agriculture tools and services depends on the influence of many variables, the farm level costs can be predicted with confidence.
Long-term research in the SE Coastal Plain shows that soil variability is widespread. Areas of low-yielding soils within fields often significantly reduce yield below that expected for the typical soils within the field. Farmers, though qualitatively aware of both variability and its effect on yield, appear to perceive that purported economic and environmental benefits of variable-rate technology do not justify the initial cost. They need data on economic effects of field-scale variability to allow rational strategic decisions. A multi-agency project was funded to both document existing variability in on-farm yields and to communicate the significance of the problem. Yield monitors were installed on three combines in Duplin and Sampson Counties, NC. These monitors collected data during 1997, totaling 900 ha of wheat and 120 ha of rye, followed by approximately 1500 ha of corn and similar area of soybean. Preliminary data were processed in vendor's yield mapping software. For further analyses and presentation, they were aggregated into ARC/Info GIS. Dramatic variability was documented both within and among fields, operators, and soil types.
Computers, global positioning systems (GPS), and geographical information system (GIS) technology are enabling farmers and agricultural professionals to enter a new era in farm management, often referred to as precision farming. The objectives for this case are to evaluate the different approaches to precision farming of a farmer and fertilizer consultant and determine how each might improve their utilization of the new technology. The case will follow a fertilizer consultant's decisions on how best to position himself and his business to provide precision farming services. We will examine the interaction of the consultant and the farmer as they work to develop a precision farming program. Through evaluating the management approaches followed by these two businesses, others beginning to adopt precision farming technologies can follow a more effective decision making process.
Precision Agriculture will be an integrative part of agricultural farm management practices in future and the exploitation of this tool for fertilizer use will highly depend on the decision making strategies used for the calculation of variable rate fertilizer applications. Fertilization with basic nutrients such as P, K, or Mg traditionally relies on soil analysis. Geocoded grid sampling or directed sampling procedures provide information about spatial variability of nutrient contents in soil. The local fertilizer requirement can easily be determined employing the BOLIDES software which sets geocoded soil analysis data in relation to geocoded yield data. Geocoded yield maps offer the possibility to calculate the nutrient demand on the basis of nutrient removal of each crop and whole crop rotations, respectively and thus follow the concept of balanced fertilization. The optimization of the fertilizer rate for highly mobile nutrients such as nitrogen preferably goes along with the spatial variability of key variables as the available nitrogen content in the soil is extremely variable in space and time. The possible benefits of variable rate fertilization include a locally higher productivity, improvement of yield stability, decreased risk of lodging, improved crop quality, reduced nutrient losses to the environment and economic savings.
The DSS4Ag is an expert system under development at the INEEL through the Site-Specific Technologies for Agriculture (SST4Ag) precision farming research project. The system uses artificial intelligence and other computer information technologies to assist in making spatial, site-specific management decisions. Using this decision support system, we generated a variable-rate fertilizer recommendation recipe for a 135 acre wheat (Triticum aestivum L.) field with the goal of optimum economic return, not maximum yield. The field was split into blocks, alternately fertilized with the variable-rate recipe and with the uniform application method used by the farmer. The DSS4Ag fertilizer recipe reduced fertilizer costs 39.7% and yields 3.3%, which resulted in a net economic gain of 2.8% as compared with the uniform application used by the farmer.
A characteristic of precision agriculture research is the cooperation between different research disciplines. Especially when research is conducted at farm scale level, the data, collected by various measurement methods, has to be available in a well described format to the participants. Based on an information model described by Goense et al. (1996), a database structure was implemented to facilitate on farm research. Part of the database incorporates storage of farm management actions with their respective spatial measurements linked to geographical primitives like locations and areas. Functionality has been extended to use crop growth simulation models to optimize prescription rates. The implementation in a relational database management system offers the opportunity to use standard GIS software as a graphical front end. Specific tasks like data exchange with field equipment has been implemented by a structured query language (SQL) interface of the database management system.
Corn (zea mays L.) producers are the largest users of cropland and agri-chemicals in U.S. agriculture and represent a major market for precision agriculture technologies. Based on a USDA survey of corn producing farms in 16 states, about 9% used some aspect of precision agriculture for corn production in 1996, representing nearly one-fifth of 1996 harvested corn acreage. Among precision agriculture adopters, 70% used some form of grid soil-sampling, 32 percent used VRT for lime or fertilizer application, and 54% used a yield monitor. A logit analysis indicated that farmers were more likely to adopt precision technologies if they farmed a large number of corn acres, earned a sizable farm income, and had high expected corn yields. The probability of adoption was also higher for farm operators using a computerized record system, who were <50 yrs of age, and who relied on crop consultants for precision agriculture information.
There have been many developments in technology, such as precision agriculture, as well as an evolution in the relationship between the farm and its environment (administration, advisers, suppliers, etc.). All this now makes a review of the Information System of the farm necessary. In order to better understand the new environment of the farm and to better meet the new information demand from end-users, we have tried to define an up-dated global Information System for farming. This global Information System take into account the needs of Precision Farming.
The changing demographics of American higher education are placing new demands on institutions. In 1994, for example, an estimated five million working adults were enrolled part-time in U.S. colleges and universities. That number masks an even larger adult population who want to pursue a college education or take a course but cannot attend a traditional college because of campus inaccessibility, inconvenient class hours, or family responsibility. Colleges and universities are required to respond to these paradigm shifts in education by reengineering courses and curricula to incorporate emerging technologies that break the constraints of time and place of learning. Realizing the potentials and promise of emerging technologies in facilitating change, colleges and universities are navigating toward networked, student-centered learning environments. This paper describes major components of an authentic, Web-based learning environment in precision agriculture that incorporates an interactive multimedia instructional/resource materials, virtual field trips, K-12 activities, glossary and links, and other resource materials that enhance learning.
Economic profitability is a key issue in the adoption of any new technology that requires high investment. Most of the published work on precision farming profitability is being done assuming extremely well behaved crop response curves for major inputs. Tools like crop simulation modeling and new wave optimum search procedures can be used to overcome these limitations. This paper presents the results of a case study on the profitability of N site specific management in a rainfed corn field. A set of 35 yrs of weather data is used along with CERES-Maize crop model to compare the conventional N application schedule with the averaged optimized site specific one. The benefits of N site-specific management are extremely variable with soil type and weather year.
In this paper what has been developed is a mathematical model with the objective of determining stages for developing operational procedures such as harvest, handling, transportation, industrialization, and distribution that, helped by a business management system, gave room for the creation of not only an instrument of making decisions, but also an instrument of administration.
Within-season management decision aids are important for precision N03 management because the final fate of the N in soil—plant systems largely depends on within-season events and management. Uncertainty in future weather challenges N models for within-season management. MOM uses weather forecasts to estimate rainfall in the near future and simulates other components in the soil—plant systems. In addition to its management-oriented optimization, MOM-guided within-season management has the advantages of (i) High efficiency in predicting timely information. Users are advised of the probable N status of soil—plant systems in advance of sensors and soil tests, (ii) Low cost to implement. No within-season soil or tissue sampling and testing are required except an initial soil test, (iii) Transparency of the systems' status. Daily descriptions of the N cycle in soil—plant systems during the cropping season graphically advise users how to control the fate of N. MOM also presents within-season estimates of leachate N03 and mineralized N, which are not provided by standard soil tests. Within-season observed data on precipitation and crop growth update MOM-guided management with current events, which improve the precision with which MOM traces the N cycle in soil—plant systems. MOM-guided within-season management was not designed to match future events exactly, but to dynamically adjust probable consequences of management strategies to fit changing conditions within a cropping season. Two scenarios illustrate how MOM can help in precision N management for maximizing profits and yields while minimizing N03 leaching by updating management of irrigation and fertilization within-season.
Since the founding of an agricultural experiment station in Crookston more than 100 years ago, the University of Minnesota has recognized the importance of agriculture to the region. Over the years, University services have evolved to meet the needs of farmers and agri-businesses. Today, the University of Minnesota, Crookston, the Northwest Experiment Station, and University of Minnesota Extension Service provide essential support and leadership to this dynamic and increasingly capital intensive, technology-driven industry. The University of Minnesota, Crookston (UMC), has issued laptop computers to all full-time students for 4 years, and according to Microsoft, was one of the first to understand that today's college graduates need highly developed competence in technology. Since about one-half of UMC's students are enrolled in agricultural programs, an institutional commitment was made to serve the emerging precision agriculture industry and to provide students education in new agricultural technologies. In late 1996 and early 1997, UMC faculty and a team of industry advisors crafted the first learner outcomes in precision agriculture. A course was developed, and 26 students took the first class in the fall of 1997. In addition, UMC faculty and its Continuing Education team collaborated with leading precision agriculture vendors and practitioners to offer an intensive day-long training workshop for professionals in agricultural extension, agri-business, and production. This presentation describes UMC's experience in precision agriculture course design and implementation. It reviews the use of technology enhanced teaching in combination with hands-on, experiential learning to achieve desired learner outcomes. UMC's collaborative course management and delivery strategy will be highlighted. The experience gained to date may benefit those interested in developing similar educational initiatives, and will contribute to the advancement of productive dialogue between educators and other professionals with similar and complementary interests.
Private industry and government agriculture departments have collaborated sine 1994 to provide workshops on precision agriculture in Alberta. The 2-day workshops have been completed by 424 participants, 79% of whom are farmers. The goal of the workshops is to enhance understanding of land resources and precision-farming technology. The participants use enlargements of aerial photography to map field variability. The variability maps are used, along with soil survey information, to develop management zones and soil-sampling strategies. The components of precision farming technology are discussed. The farmers work with agribusiness personnel and consultants to develop a site-specific management plan most suitable for their land base, economic situation, and management scheme. Course evaluations indicate the most valuable features of the workshops include the practical mapping exercises, the explanation of soil and landscape variability, the discussion of strategies for site-specific management, in particular soil sampling, and the building of awareness for implementing precision farming.
A Precision Agriculture 2-yr Associate Degree, focusing on the use of emerging technologies in agriculture, has been implemented at Hawkeye Community College as a result of a Phase INSF/ATE grant. Valication and implementation of this curriculum at other community colleges and its use by other types of educational institutions is the focus of a Phase II NSF/ATE grant. Creating a Midwest regional network of educational institutions and industry will result in a coordinated effort to develop curriculum material that will provide basic math and science skill applied to agriculture, specific technology skills needed by industry and articulation with universities. Specific components of this project include: educational opportunities for current teachers or preservice teachers; instructional-curriculum material for use as a complete course, integrated into current courses or laboratories to give hands-on experience in the technology; and working with other teachers to develop curriculum materials.
Sixteen workgroups comprised of a combination of conference participants based on affiliations - industry/agribusiness, producer, researcher, and other, convened in separate meeting rooms on Tuesday from 10:40 AM to 12:00 with the task of recommending curriculum in precision agriculture for technical, college undergraduate and graduate, and professional levels. This is a summary of comments and recommendations made by participants assembled from reports and flip charts.