Paludification reduces black spruce growth rate but does not alter ecophysiological mechanisms in Canadian boreal forested peatlands


 Background Paludification is widespread in the boreal biome, inducing tree growth decline in forested peatlands following the development of thick organic layers over the mineral soil. However, the ecophysiological processes involved remain poorly documented and little is known about the interactions between tree growth mechanisms and site conditions in these ecosystems. We investigated changes in stem growth and main ecophysiological processes in a black spruce forested peatland in eastern Canada by combining peat-based and tree-ring stable isotope analyses. These were conducted at three sampling sites located along a paludification gradient with different peat thicknesses.Results Organic layer thickening induces black spruce growth decline without altering tree ecophysiological mechanisms. A 40% increase in water use efficiency, or the ratio of carbon assimilated to water losses, was observed at the three sites from 1920 to the 1980s, but did not translate into enhanced tree growth. A clear shift in the 1980s revealed a decline in black spruce sensitivity to climate and rising atmospheric CO2 concentration, regardless of the organic layer thickness. Water table reconstructions revealed an important drawdown in the last few decades at the three sites, but we found no evidence of an influence of water table variations on stem growth.Conclusions This study shows that paludification induces black spruce growth decline without altering tree metabolism in boreal forested peatlands. This underlines that changes in water use efficiency are decoupled from changes in carbon allocation, which are constrained by site, or even tree-specific strategies to access water and nutrients from belowground. Our findings indicate that dynamic changes in edaphic conditions need to be considered in process models. Otherwise, failing to account for the degree of paludification can lead to misleading forest productivity predictions and result in considerable overestimations of aboveground carbon stocks from trees in the boreal regions.

a high precision altimeter (ZIPLEVEL PRO-2000) and an Oakfield probe. Water table depths were 119 measured at the same intervals a few hours after holes were dug to make sure that the water table 120 level had stabilized. Twenty black spruce trees were also sampled at each site within a 10 m radius 121 of the collected peat core. Only dominant and codominant trees with straight stems and no visible 122 scars were selected. Peat thickness was measured at the bottom of each sampled tree to validate the 123 concordance with the mean peat thickness of the site. The diameter at breast height (DBH) and the 124 height of selected trees were measured and cross-sections were collected at standard height (1.3 m). 125 The root system of one black spruce per site was excavated to verify the depth of the rooting zone, 126 and to identify the growth substrate (i.e., mineral or organic matter). Moreover, tree aboveground 127

Isotopic analysis of tree rings 197
Black spruce ecophysiological response to rising ambient CO 2 concentration and climate variability 198 was evaluated from carbon (δ 13 C) and oxygen (δ 18 O) isotopic ratio analyses. These were performed 199 on five trees per site and from two wood strips per tree (i.e., a total of 30 samples). Sample where δ 13 C air is the carbon isotope ratio of the atmosphere and δ 13 C tree is the isotopic value of the 210 tree ring.
where A is the rate of CO 2 assimilation, g s is the stomatal conductance, and the constant 1.6 226 represents the ratio of water vapor and CO 2 diffusivity in air. Equation 3 shows that the difference 227 between c a and c i is related to the ratio of assimilation (A) to stomatal conductance (g s ).  (Table 1). Tree-ring analyses revealed even-aged stands covering the period 1839-238 2018 CE at each site (see sample depth in Fig. 4 for tree age variability). Radiocarbon dating of the 239 most recent charcoal layer indicates that the last fire event occurred between 0 and 290 cal yr BP 240 (median age: 175-179 cal yr BP; Table S2.1). These results suggest that trees were from the first 241 cohort that grew after the last local fire, which most likely occurred around 200-250 years ago 242 (~1800 CE). The depth of the uppermost charcoal layer in the peat profile indicates that black 243 spruce established in a residual organic layer of 15, 45, and 65 cm at sites CAS0, CAS50, and 244 CAS100 respectively. The root system excavation of the three selected trees suggests that roots 245 reached the mineral soil at CAS0 and CAS50, but were restricted to the organic layer at CAS100. 246 247   (Table 1). Mean DBHs of 10.4, 9.4, and 5.6 cm were calculated 270 for CAS0, CAS50, and CAS100 respectively. BAIs also indicate a decrease in stem growth with 271 increasing peat thickness (Fig. 4, S2.4). Trees from CAS0 added a greater wood surface with age, 272 especially since 1940, resulting in an increasing BAI trend (mean BAI=70 mm 2 ). At CAS50, tree 273 radial growth was more limited (mean BAI=51 mm 2 ). In contrast, trees from CAS100 maintained 274 relatively constant BAI values, resulting in decreased wood production (mean BAI=40 mm 2 ). for sites CAS0, CAS50, and CAS100 respectively. Results from CAS0, CAS50, and CAS100 are shown in black, red, and blue respectively. 304 305 Trends in δ 13 C, δ 18 O, and iWUE 306 The δ 13 C-derived ecophysiological parameters do not differ between the three sites over the 1919-307 2000 period (Fig. 6a). Over time, black spruce trees used two different strategies in response to 308 rising c a . A substantial increase in iWUE was first observed until the 1980s (c a ≈ 340 ppm), along 309 with relatively stable intercellular CO 2 concentration (c i ). During this period, iWUE increased by 310 Our study demonstrated that the degree of paludification considerably altered growth conditions 328 and site fertility, but did not influence intrinsic water use efficiency of black spruce trees. Indeed, 329 sites with the thickest organic matter accumulation were characterized by dominant trees that grew 330 slower, presented smaller heights and diameters (DBH), and had a lower tree density comparatively 331 to the least paludified site (Table 1). Surprisingly, however, δ 13 C-derived parameters are almost 332 identical in all sites (Fig. 6), both in terms of average iWUE levels or temporal variations, 333 suggesting that the ratio of photosynthesis to stomatal conductance is unaltered by the degree of 334 paludification. We therefore refute our research hypothesis, and cannot ascertain a clear and direct 335 effect of increased peat accumulation on black spruce water use efficiency and growth mechanisms.

1920-1980s: iWUE increases (active response) 350
From the 1920s and until the 1980s, a ~40% iWUE increase was observed at each site, regardless 351 of the accumulated organic layer thickness. This significant increase, which occurred over a short 352 period of time, is among the highest recorded; most studies report iWUE increases of 20-30% over

Stem growth is decoupled from iWUE variations 394
Our results indicate that lower radial growth rates are found with increasing peat accumulation 395 (Fig. 4). This effect was also reported in the black spruce feather moss domain of the James Bay established is thereby critical in determining tree nutrient uptake and aboveground biomass and Johnson 1990). Trees from CAS0 and CAS50 both established in a less than 50 cm-thick 407 organic horizon, and roots could therefore easily reach the mineral soil, resulting in higher radial 408 growth rates at these sites. On the other hand, tree roots were unable reach the mineral soil layer at 409 CAS100 as black spruce trees established in a 67 cm-deep organic layer. ratio, and consequently, iWUE values. Actually, we suspect that this proportional adjustment in the 435 A/g s ratio might be an important process driving interactions between iWUE and growth rates in a 436 paludified context. As a supporting evidence for this, we found that black spruce tree ring cellulose 437 from the least paludified site (CAS0) was significantly more depleted in 18 O compared to that of 438 other sites (Fig. 6b). Unsurprisingly, CAS0 is also the site where radial growth rates are the highest. 439 Increased evapotranspiration rates are probably required to sustain enhanced carbon assimilation 440 and growth rates, forcing g s to level up and proportionally adjust to increases in A (matching the 441 ratio of other sites). Consequently, higher evapotranspiration rates cause black spruce to pump more Difference in peat thickness along the paludification gradient did not have a noticable influence on 447 tree response to climate. Before the 1980s, tree growth from all sites showed negative correlations 448 with previous summer temperatures (Fig. 5). These climate-growth relationships have previously induce stem growth decline, but this tree response to paludification is not reflected in black spruce 496 ecophysiological mechanisms. The increasing iWUE trends observed could suggest increasing 497 carbon assimilation and radial growth rates at the three sites. However, radial growth of black 498 spruce trees clearly declined with organic layer thickness, resulting in different tree aboveground 499 biomass between the study sites. This underlines that changes in iWUE are not necessarily related 500 to changes in carbon use efficiency because of site conditions (Manzoni et al. 2018). Consequently, 501 dynamic changes in edaphic conditions need to be considered in process models (Guiot et al. 2014). 502 Otherwise, based on tree ecophysiological parameters alone, comparable growth between the three 503 study sites would have been assumed. Our results thus suggest that failing to account for degree of 504 paludification in interpreting tree growth mechanisms can lead to misleading forest productivity