Soil carbon sequestration after afforestation of former cropland: oak and Norway spruce chronosequences repeatedly sampled after 13 and 22 years

Afforestation of agricultural land is one of the main nature-based solutions to reduce emissions of CO₂ to the atmosphere while possibly increasing soil carbon (SOC) stocks. However, information on high-resolution temporal dynamics in SOC are scarce. SOC sequestration after afforestation of former cropland has commonly been studied by the chronosequence approach. The advantage of such space-for-time substitution for estimating slow SOC change processes must be balanced against the spatial variation introduced. However, no previous studies extended the chronosequence approach with multiple repeated inventories for comparison and validation of observed SOC dynamics.

We conducted a long-term combined chronosequence/resampling study in a former cropland area afforested with oak (Quercus robur) and Norway spruce (Picea abies) over the past 50 years. Soil sampling was carried out in 1998, 2011 and 2022 in the same oak and spruce afforestation chronosequences to reveal inferred and true temporal trends in forest floor and mineral soil SOC to 25 cm depth. Sampling in adjacent cropland and a 200-year-old forest served as references for the overall SOC sequestration rates. The C sequestration in woody biomass was quantified to estimate the contribution of SOC stocks to ecosystem C sequestration. The objective was to study the decadal patterns in post-agricultural SOC change in afforested oak and Norway spruce by i) comparing chronosequence trends in forest floor and top mineral soil C stocks within and across the three sampling campaigns, ii) quantifying current rates of SOC stock change at stand level based on multiple sampling campaigns.

Forest floor C stocks followed a non-linear trend and levelled off after about 30 years towards 8.6 ± 1.2 Mg C ha^-1 for spruce and 3.4 ± 0.9 Mg C ha^-1 for oak. The chronosequence trajectory was largely validated by resampling, and decadal rates of forest floor C sequestration approached 0 after about 40 years. Chronosequence trends in topsoil SOC were similar for oak and spruce and increased across the three inventories by 0.29 ± 0.05 Mg C ha^-1 yr^-1 to a C stock equivalent to 75% of that in the 200-year-old forest after about 50 years. However, there was no detectable topsoil SOC change along the three chronosequences based on individual inventories. Repeated sampling revealed further temporal and species-specific dynamics. SOC sequestration rates in the periods 1998-2011 and 2011-2022 were positive in most stands, and particularly increased with stand age in the spruce stands older than 20 years. Norway spruce also sequestered most C in biomass.

We conclude that contrasting temporal change in forest floor and mineral soil C sequestration rates indicates a shift in C source-sink strength over 50 years. Three decades of forest floor C sequestration is shifted to increasing mineral soil C sequestration, and sequestration rates in both soil compartments were greater in Norway spruce than in oak. The chronosequence approach across all three inventories provided the best estimate of mineral soil C trajectories since afforestation, but repeated sampling revealed significant stand- and species-specific dynamics in soil C change even within a homogeneous former cropland area.