Skip to main content
SLU publication database (SLUpub)
Report, 2007

Erfarenhetr av GPS i riksskogstaxeringen 1996-2002

Bååth, Härje
Bååth, Härje (ed.)


GPS has been used in the Swedish National Forest Inventory (NFI) since 1996. During 1996-2000 GPS was used only to determine the coordinates of the sample plots. However, since 2001 (after the SA was stopped) the GPS was utilized even for navigation. The reason why GPS is used is that field data from the Swedish NFI are used together with satellite data. Furthermore it makes it easier to establish the plots and to find the permanent plots with GPS coordinates. From 1996-2002 a receiver with 6 channels was used to determine the coordinates of the plots. To get acceptable accuracy post processing was used with base data from SWEPOS. For navigation to the plots a cheap consumer GPS was adopted. Prior to the introduction of GPS a compass and measure tape was used. About 80 % of the plots had data that was possible to correct. The reason that not all of the data was available was due to, among other things, too few satellites available and too high PDOP, under 1996-1999. The positional accuracy of each location was calculated as the horizontal distance between the satellite-acquired position and the theoretical reference position. Comparison was carried out for data from temporary plots from 1996-1999, when a compass was used for navigation to the plots, and for data from 2001-20002, when GPS was used. The resulting accuracy was 45 m and 7 m respectively. Increasing the number of positions for averaging the mean for the coordinates of the plots resulted in increased accuracy. The greatest improvement was with GPS navigation, where the accuracy improved from 9,2 m for 20 positions to 6,5 m for >180 positions. The proportion of plots within 10 m from the theoretical plot centre was 83 % and 11 % respectively. The accuracy increased with decreasing density, expressed by tree height, basal area or volume, with GPS navigation. For tree height the accuracy ranged from 6,2 m to 11,3 m with tree height ranging from 0 m to >25 m. When navigating with compass the results were the opposite. Among the species, pinus sylvestris, bearch and pinus contorta the differences in accuracy was small, while for picea abies the accuracy was reduced by 1,5 m compared to the other species. With GPS navigation the average difference in accuracy between differentially corrected and uncorrected data from the 6 channel receiver and uncorrected data from the cheaper 12 channel receiver was very small. In all methods the accuracy increased with decreasing tree height, basal area and volume. The averaged differential correction was 2,8 m. The level was independent of if the sample plots were in an open area or in dense forest . When SA was applied the differentially correction was in average about 30 m. The number of positions hade a great effect on the correction. It varied from about 40 m with 20 positions to 21 m with >180 positions. When SA not was applied the differentially correction was 2,8 m. The accuracy increased with increasing number of positions. The differentially correction was rather independent of variables like tree height, basal area and volume. The results from the comparison of differentially corrected data from the 6 channel receiver and data from the cheaper 12 channel receiver showed that the differences were very small. That resulted in the use of the later since 2003 for both navigation and for determining the coordinates of the sample plots in the Swedish NFI

Published in

Arbetsrapport / Sveriges lantbruksuniversitet, Institutionen för skoglig resurshushållning
2007, number: 174
Publisher: Skoglig resurshushållning

    UKÄ Subject classification

    Forest Science

    Permanent link to this page (URI)