https://doi.org/10.5061/dryad.rn8pk0pmn

Description of the data and file structure

The collection of the soil seed bank was carried out in June 2021, in 12 plots of 40 m x 50 m.

The surface layer of soil (first 8 cm) was collected within each sampling unit, using a wooden frame (25 cm x 25 cm x 8 cm = 0.005 m³). Three simple samples were collected per plot and then homogenized to form one composite sample per plot, which was then divided to cover the four different treatments.

To assess the impact of climate change on the soil seed bank, two distinct scenarios were used: the current scenario (Cur) and the future scenario (RCP8.5). The first was represented by the current climate of the Pacotuba National Forest. Climate data for this scenario were established using the climatic normals from the nearest weather station to the remnant provided by the National Institute of Meteorology (INMET), located in the district of Rive, municipality of Alegre, Espírito Santo, Brazil.

The second scenario (future) was based on the projection of climate change from the Representative Concentration Pathways 8.5 (RCP8.5) for the period 2081 to 2100, described in the fifth report (AR5) of the Intergovernmental Panel on Climate Change (IPCC).

Two greenhouses were used (one scenario per greenhouse), each containing 24 trays of soil seed bank distributed across four treatments: Cur (current scenario without water restriction), Cur_WR (current scenario with water restriction), RCP8.5 (future scenario without water restriction), and RCP8.5_WR (future scenario with water restriction). Within each scenario, a completely randomized design (CRD) was adopted, with the levels of water availability and different climatic scenarios as treatments.

For all individuals, four functional traits were measured: leaf area (m²), stalk dry mass (g), leaf dry mass (g), and root dry mass (g). The leaf area for each scenario established in this study was determined using a LI-3100 leaf area meter (Li-Cor, USA).

To determine the stalk, leaf, and root dry mass production for each treatment, plant material from each individual was sectioned and separately placed in paper bags for a forced-air oven at 65°C until reaching constant mass. Roots were washed to remove excess soil. After drying, the dry mass of each individual was weighed on a digital analytical balance with a precision of 0.0001 g. These analyses were conducted at the end of the fifth month, which was the total period of emergence for all plants in the trays, including those that were transplanted.

Data analyses

The structural parameters (abundance and frequency), used for the structural description of the community, were calculated according to Mueller-Dombois & Ellenberg (1974) for the seed bank of each scenario studied using R software version 3.2.2 (R Core Team, 2023). To estimate the diversity and evenness of species in each scenario, Shannon, Simpson, and Pielou's evenness indices (Magurran, 2013) were calculated using the "vegan" package (Oksanen et al., 2019).

Species richness in the four sampled environments was evaluated in relation to the number of individuals and sample units using individual rarefaction and extrapolation curves with the "iNEXT" package (Hsieh et al., 2016). Richness was constructed using the first Hill number (species richness, q=0) (Chao et al., 2014). Extrapolations were made from abundance data, considering between two and three times the total sample size per environment type (Colwell et al., 2012). Rarefaction was estimated as the mean of 100 replicated bootstrap executions to estimate 95% confidence intervals. Whenever the 95% confidence intervals did not overlap, the number of species differed significantly at p < 0.05 (Colwell et al., 2012).

A Permutational Multivariate analysis of variance (PERMANOVA; Anderson, 2001) was run to test for differences in the taxonomic composition of the seed bank between the scenarios (factor with four levels). The significance was obtained using the function ‘adonis2’ from the “vegan” package (Oksanen et al., 2019). To test for the homogeneity of variances between the scenarios we performed Permutational Analysis of multivariate Dispersions (PERMDISP; Anderson et al., 2006). We used the function ‘betadisper’ also from the “vegan” package (Oksanen et al., 2019). A Principal Coordinates Analysis (PCoA) was used to visualize the results in a biplot, based on a distance matrix of Bray-Curtis.

The data were tested to check if they met the assumptions of residual normality and variance homogeneity. Since the data did not meet the assumptions of residual normality (Shapiro-Wilk test) and variance homogeneity (Levene's homoscedasticity test), the variable was subjected to the non-parametric Kruskal-Wallis multiple comparisons test (p<0.05) to detect significant differences between the treatments. This analysis, along with the preliminary tests of normality and homoscedasticity, was conducted using the "stats" package. All analyses described in this section were carried out in R software version 4.3.0 (R Core Team, 2023).

Data description and file structure: R scripts:

01 - functional_traits.R: R script that reads functional trait data and compares results between treatments.

02 - nmds.R: R script that determines species richness and abundance of treatments using indexes.

03 - phytosociology.R: R script that determines area, frequency and density values ​​for each treatment.

04 - rarefaction.curves.R: R script for developing extrapolation and rarefaction curves of species richness and diversity, based on the number of individuals and sampling units of treatments.

The data used are described in each R script.