A dense network of differential global positioning system reference stations for precision farming

Scope and method of study. The purpose of this study was to assess the feasibility of a differential global positioning system having a dense network of reference receivers to enable advanced precision farming applications. The objectives of the study were; (1) to validate previously published research that atmospheric effects increase as distance between receivers increase; (2) to identify economic justification for a Dense Network of Differential GPS (DNGPS) for individual farm precision applications; (3) to determine the resolution, repeatability and accuracy obtainable with DNDGPS corrections derived from a dense network of multiple reference receivers; and, (4) to develop a farm-based DNDGPS prototype and develop a plan for Oklahoma Mesonetwork DNDGPS prototype that identifies necessary resources required for implementation. During 1995-1997, GPS data were collected from seven National Geodetic Survey (NGS) High Accuracy Reference Network (HARN) monuments in five North Central Oklahoma Counties adjacent to the vicinity of Stillwater. Data analyses utilized EXCEL Analysis Tool Pack and Statistical Analysis System (SAS) for the calculation of the positional error statistics for each one second interval in all files and correction of a designated receiver file using error values from combinations of one, two, three, four (and five) other files. EXCEL and SAS were used to evaluate the distribution of raw, corrected, and error data via means, standard deviations, and ranges; these tools provided plots of raw data, corrected data, and error data.; Finding and conclusions. It was found that a low-cost, DNDGPS of reference receivers (RR) was feasible and met or exceeded a high percentage of precision farming operations. All of the objectives were accomplished. The research found that averaging DGPS error data derived from a multiple reference receiver network dramatically improved corrected position accuracy for an autonomous receiver at an unknown location, as compared with single reference correction. Improvement was a function of network topography, density, and node (monument) position preciseness. A simulated network density on the order of the Oklahoma Mesonet (19 miles between RRs) produced consistent accuracy within the 0.75 meter requirement defined for most precision farming operations; precision improved proportionally with additional reference stations. Based on the major findings, it was concluded that the DNDGPS network concept provided locational accuracy required for precision farming. It was recommended that a DNDGPS prototype network be implemented, capable of supporting the 0.75 meter accuracy requirements of precision farming.