Growth trends in basal area increments: The underlying problem, consequences for research and best practices

It is generally recognized that time series of tree-ring widths must be detrended, i.e. the age- or size-related trend must be removed before any climate-growth analysis can be performed. However, in contrast to ring-width (RW) analyses, there are no best practice rules on how to use basal area increment (BAI). BAI as an important variable in forestry has become also an increasingly used variable in dendrochronology. BAI is considered a better representative of the overall tree growth than RW and it is supposedly less affected by age trends than RW.

In this paper, we provide guidance by investigating the common rhetoric surrounding the use of BAI in the literature, and by highlighting possible problems and consequences from apparent misconceptions. We performed a literature review for the period 2018–2023, whether age trends in BAI time series are considered in the analysis. We also conducted additional analyses with a published tree-ring dataset of Klesse et al. (2018) to investigate whether BAI or RW time series exhibit stronger trends. Both the literature review and the data analysis provide no support for the frequently used claim that BAI time series have smaller age trends which would preclude detrending. In fact, in 58 % of the investigated time series, age trends in BAI are stronger than in RW when the trees are younger than 50 years old, and even between the ages of 51 and 100 years, we found stronger trends in BAI in 44 % of the cases. We also provide an overview of three general methods for calculating the BAI together with their advantages and disadvantages. We discuss whether the inside-out method, the outside-in method, or the Bakker (2005) method are useful to convert RW to BAI when the pith offset estimate (POE) and the diameter at breast height (DBH) are available or missing, respectively. We conclude with a general recommendation to calculate BAI based on both POE and DBH information using the Bakker method. Further, analyses with BAI should always consider the size or age of each tree, especially when trends or growth rates are compared for trees of different age classes. For climate-growth analyses, the detrending of BAI is always necessary. By addressing these key issues, we aim to improve the reliability and consistency of BAI-based studies in dendrochronology and climate-growth research.