Paludification reduces black spruce growth rate but does not alter tree water use efficiency in Canadian boreal forested peatlands

Background Black spruce (Picea mariana (Mill.) BSP)-forested peatlands are widespread ecosystems in boreal North America in which peat accumulation, known as the paludification process, has been shown to induce forest growth decline. The continuously evolving environmental conditions (e.g., water table rise, increasing peat thickness) in paludified forests may require tree growth mechanism adjustments over time. In this study, we investigate tree ecophysiological mechanisms along a paludification gradient in a boreal forested peatland of eastern Canada by combining peat-based and tree-ring analyses. Carbon and oxygen stable isotopes in tree rings are used to document changes in carbon assimilation rates, stomatal conductance, and water use efficiency. In addition, paleohydrological analyses are performed to evaluate the dynamical ecophysiological adjustments of black spruce trees to site-specific water table variations. Results Increasing peat accumulation considerably impacts forest growth, but no significant differences in tree water use efficiency (iWUE) are found between the study sites. Tree-ring isotopic analysis indicates no iWUE decrease over the last 100 years, but rather an important increase at each site up to the 1980s, before iWUE stabilized. Surprisingly, inferred basal area increments do not reflect such trends. Therefore, iWUE variations do not reflect tree ecophysiological adjustments required by changes in growing conditions. Local water table variations induce no changes in ecophysiological mechanisms, but a synchronous shift in iWUE is observed at all sites in the mid-1980s. Conclusions Our study shows that paludification induces black spruce growth decline without altering tree water use efficiency in boreal forested peatlands. These findings highlight that failing to account for paludification-related carbon use and allocation could result in the overestimation of aboveground biomass production in paludified sites. Further research on carbon allocation strategies is of utmost importance to understand the carbon sink capacity of these widespread ecosystems in the context of climate change, and to make appropriate forest management decisions in the boreal biome. Supplementary Information The online version contains supplementary material available at 10.1186/s40663-021-00307-x.


Tree rings and climate analysis
Ring-width series were standardized using a negative exponential curve to remove cambial age trends (Fritts 1976). Standardization was performed on all individual series before constructing a mean standardized chronology for each site. Daily climate data (mean temperature and total precipitation) from 1950 to 2013 were retrieved from the interpolated gridded climate dataset of McKenney et al. (2011). Pearson correlation coefficients were calculated between standardized ring-width series and monthly climate data from March to September of both the current year and the year preceding ring formation. Because of time series autocorrelation, effective numbers of degrees of freedom were calculated to generate adjusted p-values (Hu et al. 2017).

Isotopic analysis of tree rings
Black spruce ecophysiological mechanisms were investigated through δ 13 C and δ 18 O analyses.
These analyses were performed on cross-sections from five trees per site. For each selected crosssection, two wood strips of 3 × 10 mm were cut and finely sanded on all sides. A five-year resolution covering a 100-year period (1919-2018) was considered. After having carefully crossdated the wood strips, tree rings were cut with a razor blade under a binocular microscope. Rings of the same years were pooled together in equal amounts, resulting in 20 subsamples of five years per site, which were then grinded using a mixer mill (Retsch MM400) to ensure homogeneity (Borella et al. 1998). Alpha-cellulose was then extracted as suggested for black spruce samples (Bégin et al. 2015) following the protocol used by Naulier et al. (2014).
All isotopic analyses were carried out at the Light stable isotope geochemistry laboratory of the GEOTOP Research Center (Université du Québec à Montréal, Canada). Stable isotope ratios of carbon and oxygen were analyzed with an isotope ratio mass spectrometer (Isoprime 100 for δ 13 C and Isoprime VisIon for δ 18 O) coupled to an elemental analyser (Elementar Vario MicroCube for δ 13 C and Elementar Vario PyroCube for δ 18 O) in continuous flow mode. Results were normalized with three internal standards on NBS19-LSVEC and VSMOW-SLAP scales for δ 13 C and δ 18 O respectively. Results are reported in ‰ (± 0.1‰ for δ 13 C and ± 0.3‰ for δ 18 O) relative to VPDB for carbon isotopic ratios and to VSMOW for oxygen isotopic ratios. Because of the high combustion of 13 C-depleted fossil fuels since the industrial period (~1850 CE), atmospheric 13 CO2 concentration is significantly decreasing, which causes a declining trend in tree ring δ 13 C values. This Suess effect was therefore corrected as proposed by McCarroll and Loader (2004).

Testate amoeba analysis
Testate amoeba shells were extracted from 1 cm 3 peat subsamples following the standard protocol of Booth et al. (2010). Subsamples were gently boiled in distilled water and washed through 300 and 15 µm sieves. The material remaining in the 15 μm mesh was stained and mounted on glass slides before being analysed under an optical microscope (400 × magnification). Testate amoebae were identified following the taxonomy of Mitchell (2002), Siemensma (2018), and Charman et al. (2000), with the modifications of Booth and Sullivan (2007).
The dataset of the transfer function used to reconstruct WTD variations (Lamarre et al. 2013) was improved by adding 40 surface samples collected along the study transect, and 40 others sampled in another forested peatland in the study area. Peat surface samples of approximately 10 cm 3 were cut with a serrated knife, following the method described in Lamarre et al. (2013). Samples were collected in lawns and hummocks, as these are the only microforms found in our sites.

Macrofossil analysis
Subsamples of 4 cm 3 were gently boiled in a 5% KOH solution before being washed through a 125 µm mesh sieve, following the protocol of Mauquoy et al. (2010). Macrofossils were analysed in a gridded Petri dish under a stereomicroscope (10-40 × magnification), using Lévesque et al. (1988) and Mauquoy and van Geel (2007) for plant identification. The relative abundances of the main peat components (e.g., Sphagnum, ligneous, Cyperaceae) were estimated visually and expressed as volume percentages, and vascular plant remains (e.g., seeds, needles, leaves) were counted. The degree of plant material decomposition was also determined visually using an index ranging from 1 (poorly decomposed) to 5 (highly decomposed). Macroscopic charcoal particles (>0.5 mm) were analysed at 1 cm intervals along the three peat cores to identify past local fire events (Beaulne et al. 2021).       Results from CAS0, CAS50, and CAS100 are shown in black, red, and blue respectively.