Data set for "Drought response of the boreal forest carbon sink is driven by understory-tree composition"

This data set is a compilation of 1) environmental conditions, 2) biometric- and chamber-based annual CO₂ fluxes, 3) vegetation phenological greenness, and 4) forest-floor environmental conditions, all measured over the Krycklan Catchment Study (KCS, https://www.slu.se/Krycklan), a multi-scale long-term monitored boreal catchment spanning 68 km² in northern Sweden.

The environmental measurements cover the period 1991–2020. Specifically, meteorological conditions measured close to the central part of the KCS at the Svartberget reference climate station (64°14′N, 19°46′E, 225 m.a.s.l.) included air temperature at 1.7 m above ground (Ta, °C), global radiation at 1.7 m above ground (Rg, MJ m^-2), and precipitation (P, mm). Drought conditions were characterized by the Standardized Precipitation Evapotranspiration Index (SPEI) computed at 3-month time scale. SPEI was retrieved from the 0.5° gridded dataset supplied in the Global SPEI Database (SPEIbase v2.8, https://spei.csic.es/database.html). The data set comprises monthly values obtained during the long-term reference period 1991–2020 (LT_91–20), the baseline period 2016–2017 (BL_16–17), and the drought year 2018 (D₁₈). The standardized anomaly (ɀ-score) was used to identify extreme environmental measurements during both the BL_16–17 and D₁₈ periods relative to the LT_91–20 period.

Annual CO₂ flux estimates were collected in 50 forest stands located across the KCS during the period 2016–2018 using biometric- and chamber-based methods. However, to prevent confounding effects, one forest stand that was subjected to thinning operations in spring 2018 was excluded from the analysis. The selected forest stands encompassed different landscape attributes such as 1) soil type (i.e., sediment and till), 2) dominant tree species (i.e., pine and spruce), and 3) stand age classes (i.e., initiation, young, middle-aged, mature, and old-growth stands). The annual CO₂ fluxes included the net ecosystem production (NEP) and its component fluxes, i.e., net primary production (NPP), total heterotrophic respiration (RH), net primary production of trees (NPP_t) and its above- and belowground components (ANPP_t and BNPP_t, respectively), and net primary production of understory (NPP_u) and its above- and belowground components (ANPP_u and BNPP_u, respectively). The impact of drought on annual CO₂ fluxes was evaluated by calculating both the absolute and relative anomalies (∆X and δX, respectively) of D₁₈ relative to BL_16–17. To identify the temporal shift of the dominant contributor to ∆NEP, a moving-window correlation was conducted between the absolute anomaly of NEP (∆NEP) and the absolute anomalies of understory and tree NPP (∆NPP_u and ∆NPP_t, respectively), using a 7-forest-stand window with 1-forest-stand step.

The study assessed the phenological greenness of the understory and trees in a ⁓110 years-old mixed-species forest stand in the central part of the KCS from 2016 to 2018. The greenness index (gcc) was derived from hourly images collected through digital repeat photography at the Integrated Carbon Observation System (ICOS) Svartberget ecosystem station (SE-Svb, 64°15′N, 19°46′E, 270 m.a.s.l., https://www.icos-sweden.se/svartberget). Web cameras were used to capture images below- and above-tree canopy to define the gcc index for understory (gcc_u) and trees (gcc_t), respectively. The gcc_u and gcc_t values were then normalized (0–1) to describe the seasonal minimum and maximum of vegetation biomass development. A locally estimated scatterplot smoothing (loess) curve fit was then used through the normalized data points to improve visualization. The impact of drought on mean estimates of gcc_u and gcc_t during the growing season was evaluated by calculating the absolute and relative anomalies (∆X and δX, respectively) of D₁₈ relative to BL_16–17.

Environmental conditions at the forest-floor interface were measured in each of the 50 forest stands located across the KCS during the period 2016–2018. As before, one forest stand that was subjected to thinning operations in spring 2018 was excluded from the analysis to prevent confounding effects. The measured conditions included the below-canopy air temperature (Ta_bc, °C), soil temperature at 10 cm depth (Ts, °C), and soil volumetric water content at 5 cm depth (SWC, %). The data set includes mean monthly and mean May-August values estimated during the BL_16–17 and D₁₈ periods, for which the absolute and relative anomalies (∆X and δX, respectively) were calculated.

This data set consists of four Microsoft Excel workbooks:

1_dataset_environmental_conditions.xlxs

2_dataset_biometric_&_chamber-based_CO2_fluxes.xlxs

3_dataset_vegetation_phenological_greenness.xlxs

4_dataset_forest-floor_environmental_conditions.xlxs

Further details can be found in Martínez-García et al. “Drought response of the boreal forest carbon sink is driven by understory-tree composition” (Nature Geoscience, https://doi.org/10.1038/s41561-024-01374-9).

Contact information:

Ph.D. Eduardo Martínez García^1,2 (eduardo.martinez@slu.se, eduardo.martinezgarcia@luke.fi, edu.martinez.garcia@gmail.com)

Professor Matthias Peichl¹ (matthias.peichl@slu.se)

¹ Department of Forest Ecology and Management, Swedish University of Agricultural Sciences (SLU), Skogsmarksgränd 17, SE-901 83, Umeå, Sweden

² Natural Resources Institute Finland (Luke), Latokartanonkaari 9, FI-00790, Helsinki, Finland