Identifying growth releases in dendrochronological studies of forest disturbance

• Published in: Canadian journal of forest research Vol. 35; no. 7; pp. 1648 - 1656
• Main Authors: Fraver, S, White, A.S
• Format: Journal Article
• Language: English
• Published: Ottawa, Canada NRC Research Press 01.07.2005; National Research Council of Canada
• Subjects: accuracy; Animal and plant ecology; Animal, plant and microbial ecology; Biologi; Biological and medical sciences; Biology; boreal forests; dendrochronology; Dendrometry. Forest inventory; forest growth; Forestry; Fundamental and applied biological sciences. Psychology; growth releases; NATURAL SCIENCES; NATURVETENSKAP; old-growth forests; regression analysis; statistical analysis; Synecology; temperate forests; Terrestrial ecosystems; tree growth; Maine; Forests; Growth rate; Identification; Species specificity; Dendrochronology; Time variation; Temperate zone; Dicotyledones; Angiospermae; Gymnospermae; Softwood forest tree; Response modality; Hardwood forest tree; Temperate forests; Species; Method study; Threshold; Ecology; Increase; Biomathematics; Perturbation; Percent-increase method; Historical statistics; Statistical regression; Statistical method; Response function; Logistic regression; Natural damage; Tree ring; Applied mathematics; Forestry; Absolute-increase method; Spermatophyta; Coniferales; Binary logistic regression; CANOPY RECRUITMENT; SPRUCE; AGE STRUCTURE; NEW-ZEALAND; HEMLOCK; VIRGIN FOREST; BEECH NOTHOFAGUS FORESTS; STAND; NEW-HAMPSHIRE; HISTORY
• ISSN: 0045-5067; 1208-6037; 1208-6037
• DOI: 10.1139/x05-092

Information on historical disturbances is vital to our understanding of current forest conditions. Dendrochronological methods provide one means of reconstructing disturbance histories in temperate and boreal forests. In particular, the dates of significant growth releases recorded on surviving trees provide strong inferential evidence of past disturbance events. The most common method of detecting releases (the percent-increase method) expresses the postevent growth increase as a percentage of the preevent rate. Despite its widespread use, the method is known to be overly sensitive at low rates of prior growth and overly stringent at high rates. We present an alternative method that directly follows the percent-increase method, but instead of dividing the postevent growth rate by the preevent rate, we simply subtract the two. If the difference exceeds a predetermined species-specific threshold, the event is considered a release. This absolute-increase method has convenient properties that remedy the shortcomings of the percent-increase method. We tested the validity of the absolute-increase thresholds by binary logistic regressions, and we compared the absolute- and percent-increase methods by various methods. We conclude that for the species evaluated in this study, the absolute-increase method represents an improvement over the standard percent-increase method.