File name: ReadMe-Krokar vegetation resurvey This (ReadMe) file was created on 2026-08-17 by Janez Kermavnar ------------------- GENERAL INFORMATION ------------------- Name of the dataset: Dataset on site, plot and sampling information and vegetation data in the Krokar primeval forest reserve Authors Name and surname: Janez Kermavnar ORCID: https://orcid.org/0000-0001-8052-4653 Institution/Affiliation: Slovenian Forestry Institute, Vecna pot 2, 1000 Ljubljana, Slovenia ROR organization identifier: https://ror.org/0232eqz57 Email: janez.kermavnar@gozdis.si Name and surname: Lado Kutnar ORCID: https://orcid.org/0000-0001-9785-1263 Institution/Affiliation: Slovenian Forestry Institute, Vecna pot 2, 1000 Ljubljana, Slovenia ROR organization identifier: https://ror.org/0232eqz57 Email: lado.kutnar@gozdis.si Date of data collection: from 2024-06-06 to 2024-07-12 (resurvey), period 1975-1982 (baseline survey) Type of data: c-data set / series Type of research data: j-observational data Geographical location of data collection: Kočevsko region, Krokar forest reserve: 45.5414° N, 14.7692° E Information on the funders/programmes/projects that made the data collection possible: Slovenian Research and Innovation Agency (ARIS): postdoctoral project Long-term changes of forest vegetation caused by global and local environmental change drivers (No. Z4-4543), Research Programme “Forest biology, ecology and technology” (No. P4–0107) License: CC BY 4.0 DOI: 10.20315/data1012 ----------------------------- SHARING/ACCESSING INFORMATION ----------------------------- Data licences/restrictions: The data will be made available once the related paper has been published. Links to publications that cite or use the data: Not yet available. Links to other publicly available data sites: / Links to ancillary databases: / Was the data obtained from another source? YES: Hočevar et al. 1985 and Hočevar et al. 1995 ---------------------- VIEWING DATA AND FILES ---------------------- File list: a) site_information.csv: A list of 16 variables describing general information about the studied site, including some plot information. b) plot_metadata.csv: General information of the plots, including WGS84 latitude and longitude (taken with GPS during resurvey), altitude, years of sampling and plot size. c) sampling_protocols_definitions.csv: Definitions of field names for vegetation_data_releves.csv, including definitions on surveyed vegetation layers, used cover classes according to Piskernik method and nomenclature sources. d) vegetation_data_releves.csv: Species cover values per layer. The matrix contains a total of 198 rows (species) and 100 columns (releves), each plot being listed twice (B = baseline survey, R1 = resurvey). Numbering of plots is consistent with maps provided in Hočevar et al. 1995. Abundance scale is explained in sampling_protocols_definitions.csv. e) ecological_indicator_values.csv: Ecological indicator values for each species recorded in the herb layer and used in the analysis (some species were removed beforehand, see the main text for details). Indicator values for light, temperature, continentality, soil moisture (humidity), soil acidity (reaction), soil nutrients (nitrogen) were compiled from Pignatti et al. (2005), because they revised Ellenberg's values for the flora of Italy and the northeastern region of Italy has a similar flora to that of Slovenia, including also majority of Illyrian (Dinaric) species. "x" means indifferent. File ratio, if relevant: / Additional related data collected that were not included in this dataset: NO Are there multiple versions of the dataset? NO -------------------------- METHODOLOGICAL INFORMATION -------------------------- Description of the methods used to collect/obtain the data: Floristic surveys were carried out in the late 1970s and early 1980s as part of a comprehensive national inventory of primeval forest reserves in Slovenia (Hočevar et al. 1985, 1995) using Piskernik phytosociological method. Across the entire area of the reserve, 7 m × 7 m semi-permanent plots with reliable approximate location (sensu Kapfer et al. 2017) were arranged on approximate 100 m × 100 m grid (more than 70 plots in total). The exact locations of the historical plots were not available, but they were associated with detailed vegetation maps. In each plot, all vascular plants were recorded in the herb, shrub and tree layers. The vegetation layers were defined in accordance with the Piskernik method (Hočevar et al. 1985, 1995), as follows: (1) tree layer (trees and shrubs exceeding 5 m in height), (2) shrub layer (trees, shrubs and woody vines between 0.5 and 5 m in height), (3) herb layer (all herbaceous plants, irrespective of height, and trees, shrubs and woody vines below 0.5 m in height). The tree (overstory) layer was recorded within a larger circular plot with a radius of 20 m, extending from the center of each vegetation plot (Kermavnar and Kutnar 2025). The tree layer was further split into two distinct height classes: the upper tree or canopy layer, comprising trees with heights exceeding ca. 20 m, and the lower tree or subcanopy layer, which included trees with heights ranging from 5 to ca. 20 m. The abundance of each plant species was estimated using the following scale: e = 1 specimen; r = 2–5 specimens; + = 6–10 specimens; x = >11 specimens and <10 % cover; 1 = 11–20 % cover; 2 = 21–40 %; 3 = 41–60 %; 4 = 61–80 %; and 5 = 81–100 %. The species nomenclature follows National Flora (Martinčič et al. 2007). Fifty plots were resurveyed in June and July 2024. We first did georeferencing in QGIS software to obtain the coordinates of the plot centers based on a detailed map (Hočevar et al. 1995). During our field work, several variables (i.e. altitude, slope aspect and steepness, local topographic position, cover of rocks on the surface) were essential for orientation and for achieving the most accurate plot relocation possible. Plot locations were also double checked to ensure that the distance and bearing between plots matched the original grid. Garmin GPS was used to obtain plot coordinates reported in plot_metadata.csv. Data processing methods: In the plot × species matrix, cover estimates for each species were converted from the scale used in the field surveys to respective mid-point cover values (in accordance with the Piskernik method), as follows: e = 0.05 %, r = 0.1 %, + = 1 %, x = 5 %, 1 = 15 %, 2 = 30 %, 3 = 50 %, 4 = 70 %, 5 = 90 %. Due to differences in sampling dates, we excluded seven early spring ephemerals (geophytes) that typically reach their peak development before broadleaf trees form a dense canopy. The total cover of each vegetation layer (upper tree, lower tree, shrub, herb) per plot was calculated by summing the percentage cover of all species present. Given that species may overlap within the plot, this sum can surpass 100 %. The main part of the analysis focused on the changes in the herb layer. Alpha diversity was calculated as the plot-level number of plant species (species richness). Additionally, the Shannon diversity index, Pielou's evenness (equitability) and Simpson dominance were calculated for each plot. The diversity indices were calculated using the vegan package in R (Oksanen et al. 2022). The total number of recorded plant species across all plots (i.e. species pool) was considered as gamma diversity. For the herb layer, two separate analyses were conducted: (i) for the whole dataset and (ii) only for the herbaceous component by excluding phanerophytes. Based on plant species recorded in the herb layer, we calculated community-weighted means for ecological indicator values (EIVs), which allow inference into the underlying drivers of understory dynamics during the survey period (Diekmann 2003; Hédl et al. 2017). These were calculated according to Pignatti (2005), who revised Ellenberg's values for the flora of Italy (the northeastern region of Italy has a similar flora to that of Slovenia) for light, temperature, continentality, soil moisture, soil acidity/pH and soil nutrients. EIVs were computed in two ways: (i) as community-weighted means based on herb-layer species and (ii) as whole-community unweighted means based on species occurring in all vegetation layers, with woody species present in more than one layer counted only once. The shift in species composition was explored using non-metric multidimensional scaling (NMDS) with two dimensions. The species cover data were square-root transformed prior to computing Bray-Curtis dissimilarities to down‑weight dominant species. A permutational multivariate analysis of variance (PERMANOVA; Anderson 2001) was applied to test for differences in species composition of the herb layer between the baseline survey and recent resurvey using the adonis2 function in the vegan package with 999 permutations. The assessment of changes in beta diversity was conducted through the test for homogeneity of multivariate dispersions (PERMDISP; Anderson et al. 2006) using betadisper function in the vegan package. Furthermore, a model-based approach (Baeten et al. 2014) was employed to identify species with significant contribution to potential taxonomic homogenization or differentiation. This method quantifies changes in community heterogeneity over time using presence-absence data, providing a general indication of community convergence or divergence. PERMANOVA results were interpreted alongside PERMDISP to distinguish centroid shifts from dispersion changes. Both of these multivariate tests were run separately based on abundance data (Bray-Curtis dissimilarity) and presence-absence data (Jaccard dissimilarity), first for all species in the herb layer and then using herbaceous species only. The magnitude of compositional shift for each plot was quantified by calculating the Euclidean distance based on NMDS scores. In addition, we calculated within-plot temporal dissimilarity index (Bray-Curtis) as a measure of compositional change over time. Indicator species analysis (Dufrêne and Legendre 1997) was performed to assess the change in species occurrences, distinguishing between “winners” (significant association with the recent resurvey) and “losers” (associated mostly with the baseline survey). This analysis was conducted separately for the herb and shrub layers by using the indicspecies package (De Cáceres et al. 2024). Software information: R software version 4.5.3 (R Core Team, 2026). Experimental conditions: natural conditions ------------------------------------- DATA-SPECIFIC INFORMATION: [site_information, plot_metadata, sampling_protocols_definitions, vegetation_data_releves, ecological_indicator_values] ------------------------------------- List of variables: site_information: Site_variable, Value plot_metadata: Plot_origID, Latitude, Longitude, Altitude, Year_baseline_survey, Year_resurvey_R1, Plot_size sampling_protocols_definitions: Field_name, Explanation, Veg_layer, Definition, Class_symbol, Class_midpoint, Taxonomy, Value vegetation_data_releves: Species_name, Veg_layer, Plot_origID, Krokar_1, Krokar_1, Krokar_2, Krokar_2, Krokar_3, Krokar_3, Krokar_4, Krokar_4, Krokar_5, Krokar_5, Krokar_6, Krokar_6, Krokar_7, Krokar_7, Krokar_8, Krokar_8, Krokar_9, Krokar_9, Krokar_10, Krokar_10, Krokar_12, Krokar_12, Krokar_13, Krokar_13, Krokar_14, Krokar_14, Krokar_15, Krokar_15, Krokar_16, Krokar_16, Krokar_17, Krokar_17, Krokar_18, Krokar_18, Krokar_19, Krokar_19, Krokar_20, Krokar_20, Krokar_21, Krokar_21, Krokar_22, Krokar_22, Krokar_23, Krokar_23, Krokar_24, Krokar_24, Krokar_27, Krokar_27, Krokar_28, Krokar_28, Krokar_29, Krokar_29, Krokar_30, Krokar_30, Krokar_31, Krokar_31, Krokar_32, Krokar_32, Krokar_33, Krokar_33, Krokar_34, Krokar_34, Krokar_35, Krokar_35, Krokar_36, Krokar_36, Krokar_37, Krokar_37, Krokar_38, Krokar_38, Krokar_39, Krokar_39, Krokar_40, Krokar_40, Krokar_41, Krokar_41, Krokar_42, Krokar_42, Krokar_43, Krokar_43, Krokar_44, Krokar_44, Krokar_45, Krokar_45, Krokar_46, Krokar_46, Krokar_51, Krokar_51, Krokar_52, Krokar_52, Krokar_53, Krokar_53, Krokar_54, Krokar_54, Krokar_56, Krokar_56, Krokar_58, Krokar_58, Krokar_59, Krokar_59 ecological_indicator_values: Species_name, Light, Temperature, Continentality, Moisture/Humidity, Acidity, Nutrients