ABSTRACT: We report on accuracy comparisons among a range of global positioning system (GPS) receivers and configurations when collecting data in the open and below northern forest canopies. We compared recreational receivers in Wide Area Augmentation System (WAAS) mode, and expensive receivers optimized for spatial data collection (GIS receivers) in autonomous, WAAS, real-time differential, and post-processed differential modes. Data were collected over accurately surveyed open and subcanopy locations. Individual position fixes were logged for varying time periods, and corrected using appropriate methods. Euclidian distance errors were calculated, and analysis of variance (ANOVA), Tukey's tests, and linear regression were used to identify significant factors and differences. There were significant differences in the mean positional error due to receiver type under forest canopies, but no statistically significant differences under open locations. There was no difference between differentially corrected and uncorrected data when using the GIS receivers. Recreational receiver accuracies were much less consistent than GIS receivers, with higher frequencies of large errors. Subcanopy tests indicate WAAS signals were available between 8 (moving) and 23 (stationary) % of the time for the recreational receivers, and between 22 (moving) and 33 (stationary) % of the time when using GIS receivers. North. J. Appl. For. 22(1):5-11.
Key Words: Clear sky, subcanopy, receiver.
The Global Positioning System (GPS) is becoming a common tool in forestry because it provides quick, accurate measurements of coordinate locations. GPS may be used to locate point, line, or area features such as cruise plots, property corners, property lines, or treatment boundaries. GPS have been shown to increase the accuracy and efficiency in field navigation, location, and area measurement (Bergstrom 1990, Gerlach 1991, Evans et al. 1992, Oderwald and Boucher 2003). This all-weather, 24-hour, robust utility is expected to be continuously available for the foreseeable future.
A number of studies have documented the accuracy of GPS receivers (Leick 1995, Lui and Brantigan 1995, Deckert and Bolstad 1996, Naesset 1999, Sawaguchi et al. 2003). The highest accuracies are obtained by survey grade or geodetic equipment. These receivers use carrier signals to obtain positions that may be accurate to within a centimeter or less, equivalent to a few tenths or hundredths of feet. The receivers work best with an unobstructed view of most or all of the sky, often require the receiver remain stationary for tens of minutes or longer for each coordinate measurement, and are typically among the largest and heaviest receivers available. These receivers are typically impractical for most forest management applications. Accuracies are usually lower for GPS receivers used in non-surveying measurements, in part because these receivers use the "Course Acquisition" code (C/A code), and in part because these C/A receivers provide acceptable accuracies for many forest management purposes (Gerlach 1991, Sigrist et al. 1999, Liu 2002, Oderwald and Boucher 2003). C/A code receivers are smaller, lighter, less expensive, and do not require open sky conditions, so they are appropriate for field sites below closed forest canopies. Accuracies have ranged from a few feet (submeter) to several tens of feet (5 to 10 m), depending on the type of receiver and collection methods used.
Assessments of realized accuracies are required if GPS receivers are to be used effectively. Manufacturers report average accuracies for most units. However, these accuracies may be optimistic, and may vary depending on field conditions or methods. Some units report an expected precision or figure of merit during data collection, and these are often wrongly interpreted as a direct measure of accuracy. Knowledge of realized accuracies should aid the adoption of appropriate equipment and methods.
Higher GPS accuracies are often obtained through three methods: (1) improvements in hardware and software (better antennas, tracking multiple frequencies, improved firmware to reject multi-path signals); (2) differential correction, which is the use of a stationary receiver at a known location to subtract systematic and stochastic errors; and (3) altering data collection methods, primarily by increasing the number of position fixes and restricting satellite geometry, or using carrier as well as code-phase processing (Leick 1995, Lui and Brantigan 1995. Naesset 2001. Hasegawa and Yoshimura 2003. Sawaguchi et al. 2003). More accurate C/A equipment costs more, while altering the collection methods to increase accuracy often requires more time for data collection and processing. For example, post-processed differential correction is reported to be most accurate, but it requires the combination of data from roving receivers with data from a fixed base station, which is a multi-step process (Leick 1995, Lui and Brantigan 1995. Deckert and Bolstad 1996, Sawaguchi et al. 2003). Real-time differential correction is quicker but often is not as accurate as post-processed differential correction. Receivers capable of real-time differential correction from surface or satellite base stations have been, until recently, more expensive than those that are not able to do differential correction because of the added cost of subsystems to receive and process the correction signal.
The Wide Area Augmentation System (WAAS) is a free differential correction utility that may substantially improve GPS data accuracy. WAAS was developed as a navigation and landing aid for civilian aviation, and consists of a network of correction stations and communications satellites. System and environmental errors are estimated from the network of correction stations, transmitted to the communications satellites, and re-transmitted to any handheld receiver in range. WAAS satellites are in a geosynchronous orbit, above the equator, and so are relatively low in the horizon at more northerly latitudes. An additional antenna is not required for WAAS reception because the WAAS signal is broadcast in the same frequency as a GPS signal. Many GPS receivers are WAAS compatible, including some of the least expensive receivers that are currently available. These receivers may provide substantially improved accuracy over unconnected positioning at low costs.
The realized accuracies of WAAS receivers under forest conditions arc not well known. Advertised WAAS accuracies typically range between 10 and 22 ft (3 to 7 m) under clear-sky conditions, although there are few published, independent tests of WAAS systems, and there are fewer side-by-side comparisons of WAAS, real-time differential, post-processed differential, and autonomous GPS data collection. Most data collection in northern forests occurs under obstructed conditions where trees or terrain mask a portion of the sky. Trees and hillsides may block signals from the GPS satellites and decrease accuracy, and they may also block transmission from WAAS satellites and render corrections impossible. All of these factors degrade positional accuracy to a currently unqualified extent.
We report an accuracy evaluation of two inexpensive, WAAS-capable receivers and for three commonly used, differential-capable receivers. Our main objective was to quantify the realized accuracy for these receivers, in particular, to establish the average accuracy of WAAS receivers relative to non-WAAS receivers in both open and forest conditions. Another objective was to establish the frequency distribution of errors for a single fix, for example, when a scries of individual points are collected when traversing a line or the edge of an area feature. We also aimed to quantify the relationship between the number of fixes averaged for a point and positional accuracy. Our final objective was to measure how often WAAS corrections are available under a forest canopy.
Methods
