6.6   Equipment Enhancement to Support Variable-Rate Response Surface Development

Investigators

Scott A. Shearer, Biosystems and Agricultural Engineering, shearer@bae.uky.edu
Stephen F. Higgins, Biosystems and Agricultural Engineering, shiggins@bae.uky.edu
Tom Mueller, Agronomy, mueller@pop.uky.edu
Tim S. Stombaugh, Biosystems and Agricultural Engineering, tstomb@bae.uky.edu
Carl R. Dillon, Agricultural Economics, cdillon@ca.uky.edu
John P. Fulton, Biosystems and Agricultural Engineering, jfulton@bae.uky.edu

Introduction

            Much of the recent research effort in precision agriculture at land grant universities has been directed at developing criteria and evaluating the economics of variable-rate control of inputs that include seed, fertilizer and chemicals.  In general this work is being accomplished with field-scale equipment.  Inherent with the scale of the equipment is the need to cover large acreages quickly at planting, and again at harvest.  Many of the variable-rate fertilizer application studies have been conducted using spinner disc spreaders (e.g. Shearer et al., 2000).  Unfortunately, distribution patterns can be in error by as much as 30% when comparing the as “applied surface” to prescription, or the desired application maps (Fulton, et al., 2000).

Variable-rate studies require accurate seeding, and application of granular and liquid materials.  To this end the University has developed variable-rate seeding, granular fertilizer, agricultural lime and liquid (28% N) application capabilities.  While we have confidence in rate control of seeding and 28% N application, granular material application remains somewhat problematic.  Granular fertilizers are applied using a spinner disk applicator.  Unfortunately, the greatest single source of application error is spinner disk spreader mechanism.  Application errors are further compounded in consideration of the typical 50 percent pattern overlap.   Fulton et al. (2000) modeled the actual application surface, and found that nearly two-thirds of the field received an over- or under-application that differed by greater than 5% of the target value as determined from the prescription map.  Figure 1 illustrates the magnitude and distribution of errors generated as a result of the VRT spinner spreader application.  These differences were attributed to controller response speed, as well as distribution pattern shifts when changing from low to high application rates (Fulton et al. 1999), as illustrated in Figure 2.  While this application equipment may be typical of that used for the application of granular materials in many smaller fertilizer markets, such as those typical in much of Kentucky, these application errors obscure and confound the specification of crop yield response.

Figure 1.  Comparison of “As Applied” surface to the desired prescription for potash application, adapted from Fulton et al. (2000).

Figure 2.  Modeled rate change application surface from Fulton et al. (1999).

            Perhaps a more significant problem with many variable-rate investigations is the accuracy of crop yield response data.  Yield monitoring, perhaps the most significant driving force in precision agriculture, is a technology that continues to evolve.  Initially the yield monitor was designed as an aftermarket addition.   To this end the sensors are installed on combines where room is available.  Mass flow sensors, or force impetus devices, must be placed at the top of the clean grain elevator.  Positioning of this sensor too far above or below the trajectory of the flowing grain produces less than desirable results.  Unfortunately, combine designers to date have ignored mass flow sensors requirements, and the elevators continue to be designed for conveying grain only.  Therefore, yield monitor designers must live within the space constraints of existing clean grain elevator designs.

               A less publicized factor in variable-rate investigations is the nature of yield monitor calibration.  While some yield monitor manufacturers use multi-point calibration routines, others have elected to use single-point routines.  The multi-point routines have an advantage in that senor deviations at extremes are preserved.  While it be may be rationalized that a significant portion of yield data is collected within the linear response region of the sensor, it is the extremes, high and low yields, that are most important to grain producers.  These deviations are lost with single point calibration routines.  Further, most of the studies conducted to assess yield monitor accuracy, do so by comparing the accumulated mass flow, or total harvested mass of grain.  Unfortunately, integration of the mass flow rate also integrates out the errors.  Many popular press reports indicate yield monitors errors of less than 2% for accumulated mass, a number that has little bearing on the accuracy of yield estimates.      

Objectives

            The primary objective of this equipment proposal is to develop within the University the capabilities to conduct controlled fertilizer response investigations, and then to have confidence in the yield response data collected from these investigations (seeding, lime, nitrogen, and granular fertilizer application).  Specifically, the objectives are:

1.       To improve the accuracy of variable-rate granular fertilizer application by acquiring a pneumatic fertilizer metering and distribution applicator.

2.       To improve the quality and efficiency of yield response data collection from large scale field investigations by acquiring  

Enhancement of Variable-Rate Investigation Capabilities

Current precision agriculture research projects being conducted at the UK College of Agriculture involve the collection, manipulation and utilization of spatially referenced data for the purpose of managing crop production.  The goal of this research is aimed at aiding producers in the selection and adoption of appropriate precision agriculture practices for optimizing inputs to crop production systems for maximum economic yield, and to improve environmental quality.  Specific areas that we are concentrating on include, the quantification of crop yield variability within a field and the determination of the causes of this variation, and the development of criteria and decision aids for managing the variability that exists within crop production units.  With spatially referenced information, we could develop methodologies to assess the economics of the adoption and use of precision agricultural crop production practices, and develop and evaluate methodologies to improve the environment including utilization of animal waste resources.

            A compliment of machinery is envisioned this would allow researcher to layout and conduct replicated, field investigations to study the interaction of multiple variables, for example, the interaction of seeding and nitrogen application rates in corn.  This investigation might be approached using a randomized block approach with existing equipment.  However, if we are to consider how the interaction of population and nitrogen application affect profitability across typical soil landscapes, the experiment must be modified to include consideration of variations of the soil landscape.  Restating this concept, should we expect the most profitable combination of seeding rate and nitrogen application to change with landscape features such a topsoil depth or slope?  When answering these questions prior to the development of GPS, researchers were limited by their ability to layout and conduct intensive field investigations.  However, with GPS we were able to conducted a 30-acre investigation with over 300 treatments (seeding rate, nitrogen rate, and landscape position) in Shelby County last year.  Existing seeding and nitrogen application equipment allowed researchers to plant and side-dress nitrogen for this investigation in a timely manner.  However, at this point in time similar investigations with granular fertilizers are not possible because of equipment scale and marginal application accuracy. 

At harvest, it is desirable to have adequate time and flexibility to generate yield data that are accurate, and on a scale where large replicated plots are possible.  Although we have had of the privilege of working with numerous enthusiastic and excellent cooperators who accommodate many of our field research needs, we also feel the imposition on these gracious hosts can have a significant negative impact on the timeliness of their own field operations.  To solve either situation simultaneously we propose to purchase and modify a small (Class IV), used, combine such as a Gleaner R42 or John Deere 9400.  Specifically, we seek to acquire a late model machine that can be fitted with a four-row corn head and 15 feet wide small grain platform.  The small wheel base and lower gross machine weight of a Class IV machine will make it possible for us to transport the machine to field research locations around the state.  Further, and in conjunction with information learned from field investigations and the yield monitor test facility (Shearer et al., 1997; and Burks et al., 2000), we will refine and enhance the clean grain elevator geometry to improve the quality of yield data at both the upper and lower limits of elevator capacity.

The acquisition of a combine will also increase our capacity to layout and conduct field investigations that seek to develop multivariate response surfaces.  An existing university-owned plot combine has ground speed (less that 2.0 mph), header width (2 row corn head, 5 feet wide small grain platform) and grain tank capacity (25 bushel) limitations.  The reality is that actual corn harvest rates rarely exceed 5.0 acres per day under ideal conditions.  Obviously, acquisition of a Class IV machine will enhance our productivity.

To resolve concerns regarding the application accuracy of granular fertilizers with a spinner disk spreader, we propose to purchase a cost-effective air metering unit for granular materials, provide for variable-rate control of this unit, and fabricate a field applicator that is capable of uniformly distributing granular fertilizers to within 4% of the target application rate.  To this end a Gandy 62 Series Orbit Air (Model No. OA6250BS24C) pneumatic applicator will be purchased along with a Rawson Control Systems, Inc. Accu-Rate drive for fertilizer and planters.  Our experiences to date with the Rawson Controller have been excellent.  By coupling this rate control device with a systems that accurately meters and conveys granular materials to discrete locations across the applicator bar, we can insure more uniform delivery of materials to the soil surface. This combination of components will enable accurate granular fertilizer metering across application widths up to 20 feet.

Anticipated Equipment Use

This equipment request and related justification for the acquisition of these capabilities is focused on extending our field research capabilities for projects funded under Phases I and II, and in anticipation of projects to be funded under Phase III of the USDA/CSREES Special Grants Program.  Similarly, this equipment will be utilized in conjunction with other existing projects, and several new projects that are being initiated at the request and with support from equipment manufacturers (i.e, John Deere, Case IH-New Holland, AGCO, and Caterpillar).  With regard to precision agriculture research, the newly acquired capabilities will be utilized to support the following projects:

USDA/CSREES Phase I Funding

Assessment of Grain Yield Monitoring Accuracy.  Investigators: Scott Shearer, Richard Barnhisel, Sam McNeill, Tom Mueller, Larry Wells and Steve Higgins.

Field Demonstration of Variable-Rate P, K and Lime Application.  Investigators: Scott Shearer, Tom Mueller, Richard Barnhisel, Sam McNeill, Lloyd Murdock, Steve Issacs, Carl Dillion and Steve Higgins.

Investigation of Machinery and Controls Limitations on Input Management Resolution.  Investigators: Scott Shearer, Sam McNeill, Tom Mueller, Richard Barnhisel, Larry Wells, and Steve Higgins.

USDA/CSREES Phase II Funding

Dynamic Testing of Force-Impetus Yield Monitors Under Rough Terrain Conditions.  Principal Investigators: Thomas F. Burks, Scott A. Shearer, Larry G. Wells, Sam G. McNeill, John Fulton, and Steve Higgins.

USDA/CSREES Phase III Funding (Anticipated)

Voice Recognition for Concurrent Field Scouting and Machine Operation.  Principal Investigators: John Fulton, Scott Shearer, Tom Mueller, and Sam McNeill.

Sensors and Variable Rate Management.  Principle Investigators: Tom Mueller, Tim Stombaugh, Scott Shearer, Richard Barnhisel, Carl Dillon, Lloyd Murdock, Haluk Cetin, Moris Bitzer, Mike Collins, Grant Thomas, and John Grove

Other Projects

Swine and Dairy Waste Management Using Precision Agriculture.  Senate Bill 271, General Assembly, Commonwealth of Kentucky.  Investigator: Scott A. Shearer, Richard I. Barnhisel, Joseph L. Taraba and Stephen F. Higgins.

Water and Crop-Yield Management Improvement with Data from Remote and Ground-Level Sensors (A8223009). Idaho National Energy and Environmental Laboratory, U.S. Department of Energy. Investigators: J.Alex Thomasson, Scott A. Shearer and Dean A. Pennington

Site-Specific Nutrient and Biosolids Management for Agricultural Lands.  USEPA Nonpoint Source 319(h) Program.  Investigators: Scott A. Shearer, R.I. Barnhisel, Doug G. Overhults and John H. Grove.

Expected Benefits and Deliverables

The primary benefit, with respect to the development of enhanced variable-rate application and yield response sensing capabilities, will be the development and specification of multivariate response surfaces.   These response surfaces are necessary and essential if Kentucky producers are to realize profits from variable-rate technologies, the very foundation of precision agriculture.  True benefits can only be realized if engineers, agronomists and economists work together to develop equipment systems and variable-rate management protocols that work.  While the effort described within this document represents the establishment of research capabilities, one must recognize that it is use of these capabilities in conjunction with other current and proposed field investigations that will result in the generation of information that is of value to Kentucky producers.  Perhaps just as significant is the development of field platforms from which to launch additional research in the area of precision agriculture.  Most notably is the extension of the laboratory yield monitor testing as summarized in Burks et al. (2000), and work proposed by Fulton et al. in the project entitled Voice Recognition for Concurrent Field Scouting and Machine Operation.   The identifiable deliverables are aligned with the individual projects, as noted above, that will benefit from the acquisition of these capabilities.