Decreasing stem growth in common European tree species despite earlier growth onset

Recent findings suggest that global warming is altering the timing of trees’ phenological activities including earlier emergence from winter dormancy. While early-season warming can boost carbon uptake, tree growth does not seem to benefit. The underlying mechanisms and the altered intra- and inter-annual growth dynamics, as well as their interaction with environmental factors, remain poorly understood. We analysed daily-resolved dendrometer data from 228 trees across 48 Swiss forest sites over 2012–2022 to examine stem radial growth timing, intra- and inter-annual dynamics, and environmental controls for five tree species. We examined how weekly tree growth is related to envrionmental variables including day length, temperature, preciipitation, vapor pressure deficit, soil water potential and their interactions. We found a significant negative growth trend for Picea abies, Abies alba, and Fagus sylvatica across a wide climatic gradient. The reduction in growth was associated with the decrease in number of days with growth. The positive effect of higher temperatures in spring was canceled out by a negative effect towards the end of the growth period. Overall, such negative effect of increased temperature at annual scale was strongest in Pinus sylvestris and persisted over 2012-2022. To provide a comprehensive understanding of the data analysis process, we have uploaded an R script that details all analyses and datasets used. The script includes thorough documentation of the analyses corresponding to each figure, table, and supplementary material. The dataset contains information on tree growth, site characteristics, and the environmental conditions in which the trees are growing. It is available for reuse in other scientific studies, with no legal or ethical restrictions.

Methods
The data is a part of an extensive tree growth and drought monitoring network (www.treenet.info) in Switzerland. In this network, radial stem growth is derived from measurements of stem radius changes of trees with high-precision point dendrometers on mainly mature dominant trees. Tree radial growth was measured at about 1.3 m above the ground. The data was recorded and transmitted with Decentlab data transfer nodes (Decentlab GmbH, Duebendorf, Switzerland) with a logging resolution <1 µm at every 5-10 mins. The measurements were processed to a 10-min time aligned data set using the R-package ‘treenetproc’, including outlier removal, jump correction and linear interpolation of short gaps. Radial growth data, extracted with the zero-growth assumption were aggregated on a daily, weekly, monthly, and yearly sum basis. As growth strongly varies within individual trees, across years, species, sites and climatic regions, annual growth was standardized by converting it to relative annual growth. This was done by calculating the difference between each tree’s annual growth to its tree-specific long-term average, calculated over the available data from 2012 to 2022.

Air temperature (°C) data was obtained from weather stations www.meteoswiss.admin.ch, mean distance: 8 km, max: 15 km) or were measured at the site and aggregated to daily values. Air temperature and relative humidity (%) at the sites were measured at 2 m height within the forest stands. Precipitation was extracted from MeteoSwiss model generated CombiPrecip combining data from weather stations and precipitation radar. The vapour pressure deficit (VPD, in kPa) was calculated using temperature and relative humidity. Soil water potential (SWP) was measured with dielectric MPS-2/MPS-6 sensors (Decagon Devices, Pullman, US) at 10-20 cm soil depth at each site and the values were corrected for soil temperature fluctuations. All analyses, including statistical model runs and figures, were produced using the R statistical software.