https://doi.org/10.5061/dryad.3ffbg79v2

Description of the data and file structure

The provided dataset contains the original input data and scripts used to study interactive effects of macro- and microclimate on alpha diversity across taxonomic groups.

The present datasets contain the relevant scripts, input data, and results.

Description of the data and file structure:

We conducted our analyses in R 4.5.0 (R Core Team 2025). We only used open source packages available on CRAN and GitHub. All analysis was conducted within one R project, which comes with a set workspace and relative folder paths, ensuring input data can be accessed with the provided code on all devices. We thus recommend to open the scripts directly via the R project by first opening the Rproj-file (01_analysis_revised.Rproj) and to follow the following folder structure:

MAIN_FOLDER (e.g. "analysis") containing the following folders and files (files are printed in italic):

• 01_input (uploaded as file "01_input.zip")
  • biodiv
  • study_plots

• 02_Rda (uploaded as file "02_Rda.zip")

• 02_scripts (uploaded as file "02_scripts.zip")

• 03_output ( (uploaded as file "03_output.zip")
  • alpha_div_models
  • biodiv
  • climate
  • envdata

• 01_analysis.Rproj

• renv.lock

We created the project with a renv environment (Version 1.1.4; Ushey & Wickham 2025) to record all used package versions. renv.lock is the lockfile, which contains metadata about all used packages, so that the same versions can be re-installed on a new device.

After opening the R project file on a new device, users will need to run the script 0_initiate_renv.R to install and initiate the renv package. This will generate a project-specific library within the same folder as the R project file. Then run renv::restore() to install all packages with version control into the project library. For help using renv see: https://rstudio.github.io/renv/articles/renv.html.

---

The folder 01_input contains raw data that our results are based on, including biodiversity, climate, and other environmental data. Note that raw data for macro- and microclimate were provided to reviewers, but are not publicly available here. All input data was processed in the provided scripts only.

Descriptions of files in subfolder biodiv (only columns relevant in analysis are described):

01_biodiv_bacteria.csv and 01_biodiv_fungi.csv: raw metabarcoding output data for soil bacteria and soil fungi, respectively, recorded from soil samples. Empty cells arose when an OTU was unknown in any of the reference databases (NCBI, RDP, Silva) in columns regarding taxon identifications (not relevant for analyses as only number of OTUs was used).

• OTU: Operational Taxonomic Unit (used as proxy for a species)
• Run2022...: columns with number of reads per OTU (one column per sample)

01_biodiv_bacteria_IDs.csv: codes used by the laboratory metabarcoding soil microbial data to assign each soil sample to a study plot.

• AIM.sampleID: sample IDs used by metabarcoding laboratory

• Sample.name.labeling: ID of the study plot the sample was collected from

01_biodiv_bacteria_samplingdates.csv: sampling dates of soil samples per plot.

• plot: ID of the study plot the sample was collected from
• collection_date: date of collection in format dd.mm.yyyy

03_biodiv_plants.csv: raw biodiversity data of plants recorded within a square area of 16 m². Note: column "expert" documenting the taxonomist in charge of each plot was removed for data protection.

• plot: ID of the study plot
• date: recording date in format yyyy-mm-dd
• layer: vegetation layer (categorical, one of cryptogame, herb, shrub, tree1, tree2)
• spec: identified plant species
• cf: column indicating whether species could be determined with certainty ("nein" indicates "yes, with certainty")
• cov16: cover of the plant species within a quadratic area of 16 m² (in %)
• cov200: cover of the plant species within a quadratic area of 200 m² (in %) - removed since this only applied to forest plots
• comment: any additional information noted by the taxonomists (empty when no comment was made)

04_biodiv_iso_myriapoda.csv: raw biodiversity data of woodlice, centipedes, and millipedes collected with pitfall traps. Note: Column "expert" documenting the taxonomist in charge of each sample was removed for data protection.

• plot: ID of the study plot

• trap: ID of the pitfall trap (one of BF1, BF2, or BF3)

• sampling: number of the sampling (continuous integer summarizing the number of samplings per plot). In a few cases, the traps were not found on the scheduled collection dates so that realized collection dates written on the collection bottles were not in line with the field protocol for the corresponding sampling period (no clear distinction between two adjacent samplings). In such cases, sampling is "NA". However, the column "sampling" was only used to check the number of different samples as initial data check, but did not form part of analyses).

• start: start date of the collection period covered by the pitfall trap in format yyyy-mm-dd
• end: end date of the collection period covered by the pitfall trap in format yyyy-mm-dd

• species: collected species
• abundance: number of specimens per species. "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species"). For plots with no collected specimen throughout all samplings, species number was set to zero in analysis (i.e. true zeros rather than missing data). The trap on plot F150 was damaged during the sampling period used to identify woodlice, centipedes, and millipedes - here, NAs indicate data deficiency and were thus not set to zero.
• taxon: taxon of the species, woodlice (Isopoda) were identified here. "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species")
• class: class of the species, centipedes (Chilopoda) and millipedes (Diplopoda) were identified here. "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species")
• rank: column indicating the level of identification (one of "species"/"genus"/"tribe"/"family"). "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species")
• comment: any additional information noted by the taxonomists or field workers (empty when no comment was made)

05_biodiv_heteroptera.csv, 06_biodiv_orthoptera.csv, 08_biodiv_coleoptera.csv, 09_biodiv_formicidae.csv: raw biodiversity data of true bugs (heteroptera), grasshoppers (orthoptera), beetles (coleoptera), and ants (formicidae) collected with pitfall traps. Notes: 1. Column "expert" documenting the taxonomist in charge of each sample was removed for data protection. 2. In some cases, traps were destroyed and/or samples were disregarded due to a mistake during sorting that did not allow us to identfy which sample belonged to which plot afterwards. In these cases, columns were set to "NA" to indicte data deficiency. 3. In the case of orthoptera, one specimen was determined but the taxonomist could not attribute it to a sample anymore, therefore all columns regarding plot information were data deficient (set to "NA").

• plot: ID of the study plot
• trap: ID of the pitfall trap (one of BF1, BF2, or BF3)
• sampling: number of the sampling (continuous integer summarizing the number of samplings per plot). In a few cases, the traps were not found on the scheduled collection dates so that realized collection dates written on the collection bottles were not in line with the field protocol for the corresponding sampling period (no clear distinction between two adjacent samplings). In such cases, sampling is "NA". However, the column "sampling" was only used to check the number of different samples as initial data check, but did not form part of analyses).
• start: start date of the collection period covered by the pitfall trap in format yyyy-mm-dd
• end: end date of the collection period covered by the pitfall trap in format yyyy-mm-dd
• species: collected species
• abundance: number of specimens per species. "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species"). For plots with no collected specimen throughout all samplings, species number was set to zero in analysis (i.e. true zeros rather than missing data).
• comment: any additional information noted by the taxonomists or field workers (empty/"NA" when no comment was made)

07_biodiv_moths.csv: raw biodiversity data of moths collected with light traps. Note: column "expert" documenting the taxonomist in charge of each sample was removed for data protection.

• plot: ID of the study plot

• date: collection date in format yyyy-mm-dd

• species: collected species

• abundance: number of specimens per species
• comment: any additional information noted by the taxonomists or field workers (empty when no comment was made)

10_biodiv_arthropods_MT.csv: raw metabarcoding output data for flying, crawling and jumping arthropods collected in Malaise traps. Cells containing "NA" arose when an BIN was unknown in any of the reference databases (BOLD, GenBank) in columns regarding taxon identifications (not relevant for analyses as only number of BINs was used).

• BIN_Nr: Barcode Index Number (used as proxy for a species)
• [columns 22 to 1091]: columns with number of reads per BIN (one column per sample)

10_biodiv_arthropods_MT_codes.csv: codes used by the laboratory metabarcoding insect data to assign each Malaise sample to a study plot and sampling event.

• Code: sample IDs used by metabarcoding laboratory
• Plot: ID of the study plot the sample was collected from
• Sampling_No: number of the sampling event (continuous integer summarizing the number of samplings per plot)
• Date_Start: start date of the collection period covered by the Malaise trap during the sampling event in format dd.mm.yyyy
• Date_End: end date of the collection period covered by the Malaise trap during the sampling event in format dd.mm.yyyy
• No_samples: number of samples (collection bottles) that were pooled for metabarcoding

11_biodiv_birds.csv: raw biodiversity data of birds recorded with sound recorders. Note: column "identifier" documenting the taxonomist in charge of each recording was removed for data protection.

• plot: ID of the study plot

• sampling: number of the sampling event (continuous integer summarizing the number of samplings per plot)

• date: recording date in format yyyy-mm-dd

• time: recording time referring to the first two minutes of each hour (000 = first full hour, 060 = second full hour, 120 = third full hour, 180 = fourth full hour, 240 = fifth full hour, 300 = sixths full hour)
• spec_sci: recorded vocalizing species

• replaced: column indicating whether record snip from first two minutes of the hour was replaced by another snip
• comment: comment, e.g. documenting why a snip was replaced ("NA" when n comment was made)

12_biodiv_bats.csv: raw biodiversity data of bats recorded with sound recorders. Notes: 1. column "expert" documenting the taxonomist in charge of each recording was removed for data protection. 2. In a few cases, single recorded calls could not be determined due to a technical error (indicated by "Spec." in column "spec_abbr" and "NA" in the other columns regarding species identity). These records were disregarded in the analyses as NAs were removed.

• plot: ID of the study plot

• date: recording date in format yyyy-mm-dd

• time: time of recorded call in format hh:mm:ss. "NA" in cases where no specimen was contained in the sample (indicated with "no_species" in column "species"). For plots with no collected specimen throughout all samplings, species number was set to zero in analysis (i.e. true zeros rather than missing data).

• spec_sci: recorded species or species group
• spec_abbr: abbreviated species name

• selected: column indicating whether the recording night was selected for analysis in case of multiple available nights (based on randomized selection), binary, only data with "TRUE" were used
• comment: any additional information noted by the taxonomists or field workers (empty when no comment was made)

biodiv_pitfall_aug_dates.csv and biodiv_pitfall_dates.csv: sampling dates of pitfall traps per sampling event and plot that were relevant for woodlice, millipedes, and centipedes (August only) and the remaining ground-arthropod taxa, respectively.

• plot: ID of the study plot
• trap: ID of the pitfall trap (one of BF1, BF2, or BF3)
• sampling: number of the sampling (continuous integer summarizing the number of samplings per plot)
• start: start date of the collection period covered by the pitfall trap in format yyyy-mm-dd
• end: end date of the collection period covered by the pitfall trap in format yyyy-mm-dd. Value was set to "NA" in cases where samples were disregarded or traps were destroyed (noted in column "comment")
• comment: any additional information noted by field workers (empty when no comment was made)

Descriptions of files in subfolder study_plots (only columns relevant in analysis are described):

study_plots_base_data.csv: basic information about the study plots (i.e. information that was gathered during plot selection). Note that plot coordinates (i.e. columns E and N) were removed from the published dataset.

• plot: ID of the study plot (longer version)
• plot_short: ID of the study plot (short version)
• E: X coordinate of the station (UTM32; EPSG 25832)
• N: Y coordinate of the station (UTM32; EPSG 25832)
• elevation: elevation of the study plot (m asl)
• category: habitat category, one of anthropogenic (= anthropogenic open habitats), open (= natural open habitats) or forest (= forested habitats)
• type: type of dominant vegetation, one of pasture (= anthropogenic open habitats), rock/herb/shrub (= rock-, grass-, or shrub-dominated natural open habitats), gap/establishment/optimal/plenter/terminal (= one of the five development stages of forest habitats)

• elev_belt: elevational belt of the plot, one of submontane, montane, subalpine, or alpine
• tour: grouping parameter clustering plots in spatial proximity (fine clustering)
• tour_adjusted: grouping parameter clustering plots in spatial proximity (rough clustering)

study_plots_LAI.csv: Leaf Area Index (LAI) of each plot at different radii around the plot center

• plot_short: ID of the study plot

• buffer_radius: radius around the plot center used to estimate LAI
• LAI_value: LAI value. "NA" for plots A001 and A002, which were located outside of the national park and thus, no LiDAR data was available and LAI was data deficient. However, these two plots were excluded from the entire analysis.

study_plots_lidar_metrics.csv: basic LiDAR metrics calculated for different radii around the plot center

• plot: ID of the study plot
• radius: radius around the plot center used to summarize LiDAR metrics
• pzabove1: percentage of returns > 1 m above the ground, used to calculate returns between 1 and 2 meters above the ground as pz_1_to_2m = pzabove1-pzabove2
• pzabove2: percentage of returns > 2 m above the ground
• pground: percentage of returns from ground level

study_plots_ndvi.csv: NDVI of each plot at different radii around the plot center

• plot: ID of the study plot
• RADIUS: radius around the plot center used to summarize NDVI metrics
• MEAN: mean of NDVI within the given buffer radius around the plot center
• SD: SD of NDVI within the given buffer radius around the plot center

study_plots_soil_chemistry.csv: soil chemistry per plot based on soil samples from the humus and mineral layer. Note that various variables could not be measured for one sample of mineral soil on plot N006 and the plot was thus data deficient regarding soil chemistry in the following analyses.

• Plot_ID: short version of plot IDs (without leading "BGD_BIO_")
• Gewicht: weight of sample in g
• Volumen: volume of sample in cm³. Unknown for gravel samples (thus, "NA" for "not available")
• Flaechen: sampled area in cm². Unknown for gravel samples (thus, "NA" for "not available")
• TRD: density (Trockenraumdichte) of the soil substrate in g cm ^-3. Could not be determined for gravel samples due to missing volume and area (thus, "NA" for "not available")
• pH>6,3: column indicating whether measured pH was > 6.3 (1 indicates yes) --> if so, calciumcarbonate was measured and anorganic carbon had to be substracted from total carbon content to estimate organic carbon
• pH_H2O: pH value measured in distilled water (measured using a glass electrode)
• N: nitrogen content in %, measured with an elemental analyzer (Vario EL cube, Elementar, Germany). Could not be measured for one sample of mineral soil on plot N006, but reasons were not documented by the laboratory, therefore data deficient and set to "NA".
• Cges: total carbon content in %, measured with an elemental analyzer (Vario EL cube, Elementar, Germany). Could not be measured for one sample of mineral soil on plot N006, but reasons were not documented by the laboratory, therefore data deficient and set to "NA".
• Canorg: anorganic carbon content in % determined based on calciumcarbonate content. Could not be measured for one sample of mineral soil on plot N006, but reasons were not documented by the laboratory, therefore data deficient and set to "NA".
• Corg: organic carbon content in %, calculated by substracting anorganic carbon from total carbon content. Could not be measured for one sample of mineral soil on plot N006, but reasons were not documented by the laboratory, therefore data deficient and set to "NA".

study_plots_topo_metrics.csv: topography of each study plot at different radii around the plot centre.

• radius: radius around the plot center used to summarize topography metrics
• plot: ID of the study plot
• mean_easterness: mean easterness within the buffer radius around the plot center
• mean_northerness: mean northerness within the buffer radius around the plot center
• mean_slope: mean slope within the buffer radius around the plot center
• mean_tpi: mean TPI (topographic position index) within the buffer radius around the plot center

study_plots_vegetation_cover_16m2_substituted_A005.csv: information on the cover of different substrate types and vegetation layers within a quadratic area of 16 m² and 200 m² around the plot center (the latter only for forests). Mapped in the field by experts. We replaced NAs for A005 with values from "A005_VN_00" (3x3m) since vegetation cover data were missing for the extended area of 16 m² at this plot. At the other pastures, the vegetation cover was quite similar between the 9 m² and 16 m² area. Note: column "taxonomist" documenting the taxonomist in charge of each plot was removed for data protection.

• plot_short: ID of the study plot
• area_m2: reference area (200 m² or 16 m², we only used data from 16 m² area)
• date: survey date in fromat dd.mm.yyyy
• tree1_cover: total cover of plants within the upper tree layer (> 15 m)
• tree2_cover: total cover of plants within the lower tree layer (> 5-15 m)
• shrub_cover: total cover of plants within the shrub layer (1-5 m)
• herb_cover: total cover of plants within the herb layer (< 1 m)

---

The folder 02_scripts contains all scripts used to perform data analysis, from data preparation to linear models and final illustration of results. The names of the scripts follows the order of the individual calculation steps and they should thus be run in consecutive order.

Descriptions of files:

01_preparation_biodiv_data.R: Script used to prepare the raw biodiversity data.

01_preparation_environment_data.R: Script used to curate the environmental data.

01_visualization_environment_data.R: Script used to visualize the environmental data (mainly for S3).

02_biodiv_indices_alpha.R: Script used to calculate biodiversity indices including standardized Hill numbers.

03_exploration.R: Script used to explore climate data, correlations between environmental drivers and raw trends in biodiversity data before fitting the models.

04_model_alpha_GAMs_q0.R: Script used to model alpha diversity of order q = 0.

04_model_alpha_GAMs_q1.R: Script used to model alpha diversity of order q = 1.

04_model_alpha_GAMs_q2.R: Script used to model alpha diversity of order q = 2.

05_figures_main.R: Script used to generate figures used in the main body of the manuscript.

05_explainedVariance_predictors.R: Script used to generate additional figures illustrating the explained deviance of the single predictors per taxonomic group.

05_tables.R: Script used to generate tables of model outputs.

---

The folder 02_Rda contains Rda-files that were stored as interim results after time-intensive preparation steps or from scripts that cannot be run to enable follow-up analyses.

Descriptions of files (only columns relevant in analysis are described):

analysis_alpha_gam_[habitat]_[orderQ]_coefficients.Rda: model coefficients of the different GAMs applied to all habitats (one of "overall", "forest", or "open") and Hill numbers (orderQ, one of 0, 1, or 2).

• group: abbreviated taxonomic group
• family: error distribution family (one of Gaussian or Negbinomial) - one value per model
• Hill_order_q: order of q, one of 0, 1, or 2 - one value per model
• predictor: predictor
• edf: effective degrees of freedom of the predictor. Not applicable ("NA") for predictors "tour" or "tour_fine" as these were parametric.
• Ref.df: reference degrees of freedom. Not applicable ("NA") for predictors "tour" or "tour_fine" as these were parametric.
• Chi.sq: Chi-squared value of predictor (not applicable ("NA") for models with Gaussian error family and for predictors "tour" or "tour_fine")
• F: F-value of predictor (not applicable for models with negative binomial error family and "NA" for these models)
• p_value: p-value of predictor
• dev_expl_predictor: deviance explained by the predictor
• intercept: model intercept - one value per model
• AIC_model: AIC of the model - one value per model
• residual_df: residual degrees of freedom - one value per model
• N_model: sample size the model was based on - one value per model
• dev_expl_model: total deviance explained by the model - one value per model
• R2adj_model: adjusted Rsquared - one value per model

analysis_alpha_gam_[orderQ]_[habitat].Rda: list of GAMs per taxonomic group, specific for each habitat (one of "overall", "forest", or "open") and Hill number (orderQ, one of 0, 1, or 2).

biodiv_all_long.Rda: list of curated biodiversity data with one data frame per taxonomic group, long format

• plot_short: ID of the study plot
• date_start: start date of sampling event/data collection in format yyyy-mm-dd
• date_end: end date of sampling event/data collection in format yyyy-mm-dd (can be the same as start date for taxonomic groups that were collected during one day)
• OTU / BIN / species: species or alternative species identifier (i.e. OTU for bacteria and fungi and BIN for flyong arthropods)
• reads / cover / abundance / n_snips / minute_count: abundance measure (reads for bacteria, fungi, and flying insects; cover for plants; abundance for ground-dwelling arthropods and moths; n_snips for birds and minute_count for bats)
• method: collection/recording method
• duration_days: duration of the data collection in days
• group1: Taxonomic group
• group2: Larger taxonomic group

biodiv_all_mat.Rda: species (columns) x study plot (rows) matrices with one data frame per taxonomic group containing species abundances per plot. Data deficient plots (in cases where samples were lost or traps were destroyed) contain "NA" in all cells, plots with true zero values (i.e. no species observed despite successful sampling) contain zeros.

div_diversity_data_indices_harmonized.Rda: standardized alpha diversity data with coverage-based standardized Hill numbers for all taxonomic groups except understory plants. For understory plants, the table contains observed species richness (q = 0), exponential of Shannon diversity (q = 1) and inverse of Simpson concentration index (q = 2).

• plot_short: ID of the study plot
• group: taxonomic group
• Hill0: standardized Hill diversity at q = 0 for all taxa but understory plants, species richness for understory plants
• Hill1: standardized Hill diversity at q = 1 for all taxa but understory plants, exponential of Shannon diversity for understory plants
• Hill2: standardized Hill diversity at q = 2 for all taxa but understory plants, inverse of Simpson concentration index for understory plants

div_indices_hill_covstand_default.Rda: list of standardized alpha diversity data with coverage-based standardized Hill numbers for all taxonomic groups except understory plants with default settings, i.e. standardized to the smallest sample coverage obtained when sample size was doubled.

• Assemblage: ID of the study plot
• m: sample size to which the sample was standardized
• Method: applied standardization method, one of "Rarefaction" or "Extrapolation"
• Order.q: order of q
• SC: obtained sample coverage after standardization
• qD: estimated Hill diversity
• qD.LCL: lower confidence interval of estimated Hill diversity
• qD.UCL: upper confidence interval of estimated Hill diversity

div_indices_hill_covstand_manual.Rda: list of standardized alpha diversity data with coverage-based standardized Hill numbers for all taxonomic groups except understory plants with manually determined target sample coverage based on distribution of observed sample coverages across study plots - same columns as div_indices_hill_covstand_default.Rda.

div_indices_raw_prepooled.Rda: observed diversity indices after pooling all ground-arthropod groups.

• plot_short: ID of the study plot
• group: taxonomic group
• spec_richness: species richness
• shannon: Shannon diversity
• exp_shannon: exponential of Shannon diversity
• invsimpson: inverse of Simpson concentration index

div_indices_hill_prepooled.Rda: list of outputs from running iNEXT() for each taxonomic group after pooling all ground-arthropods. See function documentation for details.

div_modeldata_alpha_div.Rda: curated dataset containing all data required to run our GAMs

• plot_short: ID of the study plot
• group: taxonomic group (abbreviated)
• tax_group: taxonomic group (spelled out)
• Hill0: standardized Hill diversity at q = 0 for all taxa except understory plants, species richness for understory plants
• Hill1: standardized Hill diversity at q = 1 for all taxa except understory plants, exponential of Shannon diversity for understory plants
• Hill2: standardized Hill diversity at q = 2 for all taxa except understory plants, inverse of Simpson concentration index for understory plants
• elevation: elevation of the study plot (in m asl)
• category: habitat category of the study plot (one of "anthropogenic", "open" or "forest")
• type: type of dominant vegetation, one of pasture (= anthropogenic open habitats), rock/herb/shrub (= rock-, grass-, or shrub-dominated natural open habitats), gap/establishment/optimal/plenter/terminal (= one of the five development stages of forest habitats)
• elev_belt: elevational belt of the plot, one of submontane, montane, subalpine, or alpine
• tour: grouping parameter clustering plots in spatial proximity (fine clustering)
• T_macro_mean: mean macroclimatic summer temperature (°C)
• slope_air: slope of microclimatic temperature 15 cm above the ground versus macroclimatic temperature, metric for microclimatic buffering
• vegetation_PCA1: Axis 1 of the PCA representing vegetation characteristics (including the herb layer for all groups but understory plants, excluding the herb layer for understory plants)
• vegetation_PCA2: Axis 2 of the PCA representing vegetation characteristics (including the herb layer for all groups but understory plants, excluding the herb layer for understory plants)
• soil_PCA1: Axis 1 of PCA representing the soil conditions. Data deficient ("NA") for plot N006 (see above in description regarding soil chemistry data).
• soil_PCA2: Axis 2 of PCA representing the soil conditions. Data deficient ("NA") for plot N006 (see above in description regarding soil chemistry data).

div_modeldata_alpha_div_list.Rda: curated dataset containing all data required to run our GAMs (in long format) within a list, including spatial grouping to adjust for autocorrelation

• plot_short: ID of the study plot
• group: taxonomic group (abbreviated)
• tax_group: taxonomic group (spelled out)
• lifestyle_group: thermoregulatory strategy (spelled out, not used in alayses)
• elevation: elevation of the study plot (in m asl)
• category: habitat category of the study plot (one of "anthropogenic", "open" or "forest")
• habitat: habitat type of the study plot as ordered factor (one of "Open habitat" or "Forest")
• type: type of dominant vegetation, one of pasture (= anthropogenic open habitats), rock/herb/shrub (= rock-, grass-, or shrub-dominated natural open habitats), gap/establishment/optimal/plenter/terminal (= one of the five development stages of forest habitats)
• elev_belt: elevational belt of the plot, one of submontane, montane, subalpine, or alpine
• tour_fine: grouping parameter clustering plots in spatial proximity (fine clustering)
• T_macro_mean: mean macroclimatic summer temperature (°C)
• slope_air: slope of microclimatic temperature 15 cm above the ground versus macroclimatic temperature, metric for microclimatic buffering
• vegetation_PCA1: Axis 1 of the PCA representing vegetation characteristics (including the herb layer for all groups but understory plants, excluding the herb layer for understory plants)
• vegetation_PCA2: Axis 2 of the PCA representing vegetation characteristics (including the herb layer for all groups but understory plants, excluding the herb layer for understory plants)
• soil_PCA1: Axis 1 of PCA representing the soil conditions. Data deficient ("NA") for plot N006 (see above in description regarding soil chemistry data).
• soil_PCA2: Axis 2 of PCA representing the soil conditions. Data deficient ("NA") for plot N006 (see above in description regarding soil chemistry data).
• tour: grouping parameter clustering plots in spatial proximity (fine clustering)
• name: Order of q for Hill diversity (one of "Hill0", "Hill1", or "Hill2")
• qD: standardized Hill diversity at the given order of q for all taxa except understory plants; species richness for understory plants
• Order.q: Order of q for Hill diversity as numeric variable
• order: Order of q for Hill diversity as character used for labels in figures (one of "q = 0", "q = 1", or "q = 2")
• E: X coordinate of the plot (UTM32; EPSG 25832)
• N: Y coordinate of the plot (UTM32; EPSG 25832)
• tour: grouping parameter clustering plots in spatial proximity (normal clustering)
• tour_rough: grouping parameter clustering plots in spatial proximity (rough clustering, not used)

---

The folder 03_output contains generated tables and figures. Note that figures published in the main body and Supporting information of the article related to this data repository (DOI 10.1002/ecog.07984) are not included in this data repository, but can be reproduced using the provided R scripts. NOTE: original figures were saved as png and svg for documentation, the svg file was then manually edited (i.e. no script) to add clustering of thermoregulatory strategies and exported as high- and low-resolution png files. Points or predicted trends were not edited.

--

Descriptions of files in subfolder alpha_div_models:

subfolder cross_validation contains csv files with results from 10-fold cross-validation for each model per habitat and order of q.

subfolder model_diagnostics contains full model diagnostics per taxonomic group for the overall model across habitats as well as habitat-specific models. Includes histograms, observed vs. predicted density, DHARMa diagnostics plots, spatial autocorrelation plots, as well as concurvity matrices.

subfolder model_summaries contains model outputs of the GAMs per taxonomic group for the overall model across habitats as well as habitat-specific models. Same columns as those described for analysis_alpha_gam_[habitat]_[orderQ]_coefficients.Rda files.

--

Subfolder biodiv contains plots of raw diversity trends along the studied gradients as well as summary statistics.

--

Subfolder climate contains a folder with raw (uncleaned) microclimate time-series data, which was used for a manual quality check (subfolder microclimate_check) and different summary statistics, as well as the estimated microclimate metrics used in the analyses (see scripts for processing steps).

--

Subfolder envdata contains PCA results of soil and vegeation variables as well as correlations among environmental predictors used in the GAMs.

Code/software

We used the open-source software R 4.3.2 (R Core Team 2023) for all data preparation and analyses.