Author: Delio Tortosa, Taylor Scarr and Bill May
Company: ELIRIS Inc.
This pilot project was a cooperative effort between ELIRIS Inc. and the Ontario Ministry of Natural Resources to determine the effectiveness of using GPS/GIS technology in operational field conditions during two spray programs. Herbicide and insecticide spray aircraft were passively tracked using GPS receivers. GPS positional information was differentially corrected, translated and presented using desktop mapping software. Results of the pilot project were used to determine the effectiveness of the GPS Mapping System for monitoring and audit purposes.
GPS tracking of aerial sp ray aircraft was conducted by ELIRIS Inc. under contract with the Ministry of Natural Resource Provincial Silvicultural Unit on two aerial spray programs during 1995.
The purpose of the pilot project was to assess the capabilities of a low cost GPS tracking/mapping system in an operational setting. The GPS test were conducted in parallel with the standard method of assessment for aerial spray operations using an MNR audit aircraft.
The main objective of the pilot project was to determine whether a GPS tracking/mapping system could be used as a suitable audit tool for aerial spray programs. In addition, it was important to assess the capability and performance of such a system under typical field conditions.
In June 1995, GPS tracking was conducted for aircraft involved in an aerial spray program for the biological insecticide, B.t. The program was designed to protect jackpine foliage from defoliation by jackpine budworm in stands between Espanola and Sudbury. Spray operations were conducted from the Durban airstrip near E.B Eddy Camp 12 and from the Sudbury airport.
In August 1995, GPS tracking was conducted for aircraft involved in an aerial spray program using VISION to suppress competing vegetation in harvested blocks between Thessalon and Elliot Lake. Spray operations were conducted from the Aubrey airstrip near the MNR Peshu Lake operations base.
Tracking of the aerial spray aircraft was accomplished using the GPS Mapping System, developed by ELIRIS Inc. It consists of Garmin SRVY II GPS receivers and post-processing software, QuikMAP Desktop Mapping Software, and QuikELINX GPS-QuikMAP Translator.
The Garmin SRVY II GPS receivers have differential GPS capability, can track up to eight satellites and store up to 10 Mb of position data in memory (100,000 pseudorange positions). This amounts to about 10-12 hours of continuous data collection at one reading per second, and provided sufficient coverage for the range of operation of the aircraft. Differential GPS data correction was completed using the Garmin SRVY II post-processing software.
QuikMAP is a Windows-based, cartographic mapping software which is fast and relatively easy to learn. The system has spatial and attribute query capability which allows for additional GPS data processing, if necessary. The computer hardware and memory requirements are modest and runs well on a 486 DX33 microprocessor.
QuikELINX is a custom GPS data translator for the Garmin line of GPS receivers. The software converts GPS data into point, line and polygon data files which can be immediately displayed in QuikMAP. QuikELINX acts as a data bridge between the GPS receiver and QuikMAP and is bi-directional (i.e. routes and waypoints can be uploaded/downloaded).
For the B.t. aerial spray program, four digital 1:250,000 National Topographic Series (NTS) maps were used. The digital maps were subdivided into four major themes: Transportation Routes, Topographic Contours, Lakes and Streams, and Miscellaneous Features. The 1:250,000 NTS maps provided an appropriate scale for the size of the spray blocks. The Ontario Base Maps (OBM) at 1:20,000 would have provided better accuracy, but were not available for the project.
For the herbicide aerial spray program, two digital 1:250,000 NTS maps and the available 1:20,000 OBM were used. The NTS and OBM maps were also subdivided into major themes (as above).
Computer hardware consisted of a Notebook PC 486 DX50 and an colour inkjet printer. Both the Notebook PC and printer are highly portable and can be battery operated if required.
The GPS Mapping System is designed for use in a variety of resource-related applications and differs significantly from aerial tracking/guidance systems in that it is not customized specifically for aerial spraying and navigation. Total cost for the GPS Mapping System, excluding computer and peripherals, was approximately $10,000.
Since the objective of the pilot project was to accurately track the aerial spray aircraft, no attempt was made to use the system to aid aircraft navigation or to layout routes.
A prior study on the accuracy and precision of the Garmin SRVY GPS receiver (Tortosa and Beach, 1995) indicated that codecorellating receivers of this kind can achieve standard kinematic accuracies of better than +/- 30 metres, 80% of the time. With differential GPS, the kinematic accuracy increases to better than +/- 10 metres, 90% of the time (figure 1).
By averaging over a one hour period, the accuracy of standard GPS (i.e. with Selective Availability (S/A) on), is +/- 5 metres (figure 2). Using differential GPS and a 15 minute average, the accuracy increases to 0 +/- 1 metre. This is the limit of accuracy for GPS receivers which are not capable of sub-metre accuracy.
For the aerial spray project, the base station survey control was obtained using three, one-hour standard GPS averages. If the three averages were all within 2-3 metres, then these were averaged to arrive at a final reference position. This method has been found to meet the horizontal position requirement for the 1:20,000 Ontario Base Maps (+/- 10 metres) (Tortosa & Beach, 1995). Digital NTS maps were primarily used for reference since the map accuracy is about +/- 50 metres.
Accuracy of the differentially corrected GPS position can be affected by the range of operation and distance between the base and field GPS units due to changes in the satellite geometry between the two locations. The maximum recommended distance for differential GPS is 300 kilometres; for OBM updates the recommended distance is 100 kilometres (to remain within +/- 10 metre accuracy). The range of operation for both spray programs was between 25 and 100 kilometres, the average range was 55 kilometres.
The base station GPS antenna was located on the roof of a building or trailer which acted as the operations base for the duration of the spray. A Garmin SRVY II was set to operate as a base station, collecting pseudorange data at a rate of 1 position/second. The base station survey session was started prior to the departure of the aircraft, and then ended when the aircraft arrived.
Dromodaire aircraft were used for both aerial spray programs. The Garmin SRVY II GPS antenna was fixed to an external bracket on the roof of the cockpit and connected to the GPS receiver with a coaxial cable. The GPS receiver was placed behind the pilot seat. The position of the antenna allowed for a maximum view of the sky, even with tight turns of the aircraft.
For the herbicide spray program, the aircraft operated independantly since the spray blocks were small. As such, GPS receivers were placed on each of the three spray aircraft. For the B.t. spray program, the aircraft operated in tandem and a single GPS receiver was placed on the lead aircraft.
The GPS receivers were assigned an identification for each aircraft. Each flight was identified by the date, a sequential number, and a description of the blocks to be sprayed. As each plane returned and refueled, a new tracking session with date, number, and description was added. Both the Base and Field GPS data was downloaded and differentially corrected at the end of each spray session. Differentially corrected data was then translated and plotted on the desktop mapping system. All GPS data files including any later processing of those files was logged on field notes in order to keep an archival record displaying each stage of the data processing.
For the B.t. spray program, the aircraft were tracked successfully without a significant loss of satellite signal reception. Differential correction of the field GPS units resulted in accurate tracking of the aircraft with more than 90% of the data resulting in 3D positioning. Since the aircraft were tracked throughout the flight path, the DGPS corrected data was filtered using the GPS time stamp in each data file to eliminate the travel path to and from each block. The resulting GPS positions were then presented as GPS point symbols and also converted into polylines. The size of each GPS point symbol was estimated to match the spray width (figure 3).
Results of the aerial spray tracks indicate gaps and cross-overs on earlier tracks. The booms on / booms off positions were not identified since the system was not customized for the aerial spraying application and no attempt was made to involve the pilot in the required procedure.
For the herbicide spray program, satellite signal loss was a more significant factor and resulted in the loss of DGPS positioning in some low lying areas which were surrounded by high relief. GPS signal were obtained at these locations but there were insufficient satellites in common with the base station to provide either a 2D or 3D position (figure 4).
The aerial spray tracks were plotted on 1:20,000 scale OBM's due to the small size of the spray blocks. Results indicate where the aircraft covered the spray block, but do not show the booms on - booms off position for reasons previously discussed.
The GPS tracking pilot project produced a number of favourable results:
1) Established a higher level of confidence in the aircraft spray track than what was previously available.For the B.t. spray program, reliance was placed on a navigator aircraft for guidance of the spray aircraft. Based on the results of this pilot project, improvements in spray accuracy can be achieved by eliminating overlaps and crossovers in the spray swath. For boreal forest spray applications involving large spray blocks, the level of navigational accuracy required does not necessitate the use of real-time DGPS. Real-time DGPS is fraught with problems, such as the limited telemetry range, the requirement of a base station near each spray block, and the potential high risk of DGPS signal loss. When applied to forest pest control programs such as the B.t. spray program which ranged over an area 14,000 square kilometres in size, the problems imposed by ground telemetry do not justify the potential increase in accuracy obtained.
Consideration should be given to the use of high quality standard GPS navigation and guidance. Code-corellating receivers such as the Garmin SRVY II can provide a navigator aircraft with an additional tool with which to improve guidance for the spray aircraft which can meet the basic objective of eliminating crossovers and overlaps. This approach also has a significant cost savings when compared to current GPS guidance systems and allows spray companies a lower-risk entry into this technology.
For governments involved in forest pest management, the results of this pilot project provide information on the realistic levels of accuracy and navigation which can be expected for aerial spray programs in boreal forest. The kinds of data generated during the pilot project can serve to assist in defining the expected levels of positional accuracy for a spray program, and result in a standard series of specifications for aerial spray contracts. Aerial spray programs could then be selectively audited (using DGPS) to determine how well the spray company was fullfilling its contractual obligations.
Tortosa, D. and Beach, P. 1995 (in prep.). Accuracy and precision tests using differential GPS for natural resource management; Northern Ontario Development Agreement, Northern Forestry Program.