The Accuracy of Forest Damage Assessments – Experiences from Sweden
- Cite this article as:
- Wulff, S. Environ Monit Assess (2002) 74: 295. doi:10.1023/A:1014287228807
- 43 Downloads
In this article the consistency of forest damageassessments on conifer trees is analysed, by usingdifferent methods in estimating the accuracy inassessments of defoliation and discolouration. The dataoriginate from control surveys in the Swedish nationalforest damage survey, as well as from national andinternational training courses. Standard deviation ofdifferences in the assessment of defoliation on singletrees is found to be about 10% units for Norway spruceand 8.5% for Scots pine. Problems in correctly assessingdamaged Scots pine trees and discoloured Norway sprucetrees are revealed by measures of agreement (Kappastatistic). Results from several years of nationaltraining courses indicate that, on an average, theobserver teams do not significantly differ from a nationalstandard, but significant differences between observerteams are found. The presented estimates indicate asubstantial within observer error compared to the betweenobserver error. The results indicate that the long-termdevelopment of forest damage, rather than short-termfluctuations, is the most important information fromthese kinds of inventories.