Data from: Effects of climate and forest development on habitat specialization and biodiversity in Central European mountain forests
Data files
Dec 04, 2024 version files 82.06 MB
-
bacteria_2021.csv
12.33 MB
-
bats_2021_zero_nights.csv
1.58 KB
-
bats_2021.csv
248.90 KB
-
birds_2021.csv
527.44 KB
-
chilopoda_2021.csv
13.26 KB
-
coleoptera_2021.csv
433.61 KB
-
diplopoda_2021.csv
29.33 KB
-
dist_forest_plots.csv
181.05 KB
-
forest_plots.csv
5.82 KB
-
formicidae_2021.csv
19.04 KB
-
fungi_2021.csv
6.31 MB
-
insects_2021.csv
56.49 MB
-
isopoda_2021.csv
20.60 KB
-
large_mammals_2021_dates.csv
187.16 KB
-
large_mammals_2021.csv
34.53 KB
-
lidar_canopy.csv
2.70 KB
-
lighttrap_2021_dates.csv
3.47 KB
-
microbe_2021_IDs.csv
9.46 KB
-
microclim.Rda
4.51 MB
-
moths_2021.csv
207.03 KB
-
pitfall_2021_dates.csv
41.08 KB
-
plants_2021_dates.csv
2.56 KB
-
plants_2021.csv
388.95 KB
-
README.md
4.21 KB
-
small_mammals_2021.csv
20.76 KB
-
soil_2021_dates.csv
2.56 KB
-
species_lookup_lifeform.csv
37.85 KB
Abstract
Mountain forests are biodiversity hotspots with competing hypotheses proposed to explain elevational trends in species richness. The altitudinal-niche-breadth-hypothesis suggests decreasing specialization with elevation, which could lead to decreasing species richness and weaker differences in species richness and beta diversity among habitat types with increasing elevation. Testing these predictions for bacteria, fungi, plants, arthropods, and vertebrates, we found decreasing habitat specialization (represented by forest developmental stages) with elevation in mountain forests of the Northern Alps – supporting the altitudinal-niche-breadth-hypothesis. Species richness decreased with elevation only for arthropods, whereas changes in beta diversity varied among taxa. Along the forest developmental gradient, species richness mainly followed a U-shaped pattern which remained stable along elevation. This highlights the importance of early and late developmental stages for biodiversity and indicates that climate change may alter community composition not only through distributional shifts along elevation but also across forest developmental stages.
README: Effects of climate and forest development on habitat specialization and biodiversity in Central European mountain forests
The repository contains 26 data and 5 software files (4 R scripts and 1 renv.lock file). The analysis and thus the R scripts are divided into the 3 steps: data preparation and metric calculation with an additional script for simulating communities for null models, model fitting and checking, and model predictions and preparation of figures and tables. Data files contain biodiversity data of 14 species groups (during data preparation pooled into 6 taxa), data characterizing the study plots (elevation, forest development stage), LiDAR data used to quantify canopy cover, temperature data describing microclimatic conditions, and additional files needed for data processing (plant trait data) or supplementary material.
The data was collected at 150 plots along gradients of forest development (5 stages: gap, establishment, optimum, plenter, terminal) and elevation (ranging between 605 m and 1725 m asl) in the German Alps (Berchtesgaden National Park) and was used to test the interactive effect of climate and forest development on biodiversity in mountain forests.
Description of the data and file structure
[bacteria/fungi/plants/isopoda/diplopoda/chilopoda/formicidae/coleoptera/moths/insects/birds/bats/small_mammals/large_mammals]_2021.csv: Each file contains at least the plot ID and species names found at the respective plot. Bacteria, fungi, and insect data was determined through metabarcoding and are structured differently. For bacteria and fungi, we use OTUs (Operational Taxonomic Units) as surrogates for species. Besides the column containing the OTU classification, the files contain additional taxonomic information (e.g. phyllum, order, family, genus, etc.) and multiple columns (name starting with "run") each representing an individual sample and containing information about the number of reads of each OTU found in the respective sample. For insects, we use BINs (Barcode Index Numbers) instead of OTUs as surrogates for species. The file also contains additional taxonomic information, two columns containing the OTU and BIN classification, respectively, and multiple columns (names contain the plot IDs) representing an individual sample.
microbe_2021_IDs.csv: Links the bacteria and fungi samples with the plot IDs.
bats_2021_zeronights.csv: This file contains plot ID, date, and alpha (= species richness), which describes absence nights for bats, as the main file bats_2021.csv only includes presence nights.
forest_plots.csv: This file contains plot IDs, forest development stage, elevation, elevation zone and sampling area.
species_lookup_lifeform.csv: This file contains the lifeform data of plants (column PlantGrowthForm) gathered from TRY database by Braziunas et al. (2024).
dist_forest_plots.csv: This file contains spatial distances between all pairs of plots used as covariable in the beta diversity models.
With all files described until here, one can reproduce all main analyses presented in the main part of the paper. All files described in the following are needed to reproduce the supplemental material.
lidar_canopy.csv: This file contains the plot ID and a column showing the percentage of lidar returns above 5 m (named pzabove5), which we used to quantify canopy cover.
microclim.rda: This file contains temperature measures from Tomst TMS-4 loggers. The file contains the plot IDs, 3 columns (T1-T3) containing temperature measures at different heights (-6, 2 and 15 cm), date and time stamps of each measurement, as well as start and end date of insect trap exposition at each plot.
[pitfall/lighttrap/soil/plants/large_mammals]_2021_dates.csv: These files contain sampling dates and plot IDs to generate a figure showing all sampling dates.
[canopy/spec_rich/microclim]_spatial_autocor.png: These figures show the results of the tests for spatial autocorrelation not provided with the code.
Code/Software
All code is written in R (version 4.3.2) and we provide a renv.lock file that enables everybody to load each R package in the exact same version we used for the analysis.
Methods
We collected data from 14 taxonomic groups across 150 plots at Berchtesgaden National Park. For the selection of the study plots, we differentiated five forest development stages (gap, establishment, optimum, plenter, terminal) replicated ten times each across three elevation zones (submontane, montane, subalpine), ranging between 605 and 1725 m asl. Field sampling was conducted in 2021 and 2022. Bacterial and fungal communities in the soil were identified through metabarcoding of four soil samples from each plot, consisting of mineral and organic soil samples taken at approximately 3 m distance from the plot centre in each of the four cardinal directions. Percent cover of plants by species was recorded on a 200 m2 quadratic area, distinguishing the herb (<1 m height) and shrub layer (>1-5 m height). Arthropods from pitfall and light traps were identified by taxonomists, while those from Malaise traps were analysed by metabarcoding. Due to the different underlying species concepts, we analysed arthropods identified by taxonomists (“arthropodsTAX”) and those identified by metabarcoding separately and refer to the latter as “insectsBIN”, since insects made up 96% of all arthropod BINs in Malaise traps. We recorded birds through passive acoustic recorders in the morning hours around sunrise, and experts identified species based on these recordings. Bats were recorded using ultrasonic recorders, identified to species level through an automated software (batIdent, ecoObs, Nuremberg, Germany), and subsequently evaluated by an expert. Small mammals (i.e. mice, voles, dormice and shrews) were caught in pitfall traps and large mammals were recorded through wildlife cameras. We used LiDAR data from 2021 to quantify canopy cover at plot level (r = 12.62 m) as the proportion of returns above 5 m. To quantify microclimate, we used Tomst TMS‑4 loggers to record temperature every 15 minutes at -6, 2 and 15 cm at the plot centre.