Exploiting irregular measurement intervals for the analysis of growth trends of stand basal area increments: A composite model approach

A method was developed to extract environmentally driven growth trends from data of periodically repeated measurements of experimental stands. The method was applied to 470 long-term Norway spruce (Picea abies [L.] Karst.) experiment stands located in southwest Germany. The detection of trends in stand basal area increment was based on a composite estimator consisting of a tree-level and a stand-level growth model. Expected increment was predicted in annual time steps and depended on stand age, site quality, stand density, and thinning regime. With respect to the later, tree growth response to thinning was modeled distinguishing between target crop-trees and other trees. The long-term growth variation component was finally depicted using generalized additive models.Basal area growth displayed only minor fluctuations around expected growth during the first decades of the 20th century. After a distinct growth depression in the 1940s basal area growth increased sharply relative to the expected growth until the early 1970s and continued to slightly increase until the late 1980s. Since the early 1990s growth acceleration slowed down and observed basal area growth converged to the expected growth in most recent years (last vegetation period included in the study: 2007). Splitting the experimental stands in two age groups (younger vs. older than 60 years) or stratification according to two regions of different regional climate characteristics (mountainous vs. sub-mountainous) did not result in major differences of this general growth trend pattern. However, temporal variability of ratios of measured and expected stand basal area growth of younger stands (stand age 60 years) was higher than that of older stands. Maximum growth ratios in mountainous regions exceeded those in sub-mountainous regions and the recent decrease in basal area growth ratios appeared more pronounced in sub-mountainous regions.

► We use irregular repeated stand measurements to investigate environmentally driven growth trends. ► The developed methodology allows extraction of the environmental growth component. ► The expected growth is based on a composite estimator of tree- and stand-level growth model. ► GAMs were adapted to smooth the long-term environmental-driven growth trends. ► After increasing from the 1950 to 1990s, the growth trend in Norway spruce reversed.