Codes and data to analyse trends of saproxylic beetle and butterfly species richness between 1930 and 2021 in Switzerland

This repository contains codes and additional data which were used in the analyses for the following publication:

Neff F, Bollmann K, Chittaro Y, Gossner MM, Herzog F, Korner-Nievergelt F, Litsios G, Martínez-Núñez C, Moretti M, Rey E, Sanchez A, Knop E (2026). Ninety-year trends reveal sharpest insect declines in the mid-twentieth century. Nature Ecology & Evolution. https://doi.org/10.1038/s41559-026-03074-6.

In this study, based on species records data, mean annual occupancy was estimated for 595 saproxylic beetle species and 216 buttefly species (Papilionoidea, incl. Zygaenidae moths) in six biozones in Switzerland. The data cover the years 1930–2021 (occupancy estimates were determined per two-year interval; 1930/31, 1932/33, etc.). Trends in saproxylic beetle and butterfly species richness were reconstructed for the biozones and the whole of Switzerland. Richness trends were related to a set of environmental variables (land use, climate) and to species traits (size, specialisation, temperature niche).

The following R code files are included in the folder R_Code:

• R_Main_analyses.R: Contains all the main analyses necessary to reproduce figures and numbers in the manuscript. Based on processed data (occupancy estimates, drivers, traits).

• R_Drivers.R: Contains codes used to calculate the driver variables from the raw data

• R_Traits.R: Contains codes used to calculate the trait variables from the raw data

• R_Occupancy_detection_models_sapro.R: Contains the codes used to fit the occupancy-detection models for the saproxylic beetle species. The models were fitted on an HPC cluster.

• R_Occupancy_detection_models_butter.R: Contains the codes used to fit the occupancy-detection models for the butterfly species. The models were fitted on an HPC cluster.

• f_occ_det.R: Helper function used to fit occupancy-detection models in Stan (called by the previous files).

The following Stan code file is included in the folder Stan_Code:

• Stan_occ_det_cmdstan.stan: Stan model code for the occupancy-detection models (called by the respective R code files). The basic structure of the code is based on code published by Hiroki Itô.

The sources of external data are indicated in the code files. Occupancy estimates are available from Zenodo (https://doi.org/10.5281/zenodo.17255265). The following additional data files necessary to reproduce the analyses are available from the folder Data:

• splist_sapro.txt: Text file containing the names of the 595 saproxylic beetle species analysed

• splist_butter.txt: Text file containing the names of the 216 butterfly species analysed

• Traits/Traits_saproxylic_beetles.csv: Categorical trait data for all saproxylic beetle species. Each row is a species. Contains the following columns:

  • Name_std: species name

  • size_cat: body size category (categorical: small, medium, large)

  • stenotopy: habitat specialisation (categorical: eurytopic, oligotopic, stenotopic)

  • food_spec: food specialisation (categorical: polyphagous, oligophagous, monophagous)

  • Tniche: temperature niche (categorical: cold, intermediate, warm)

• Traits/Traits_butterflies.csv: Categorical trait data for all butterfly species. Each row is a species. Contains the following columns:

  • Name_std: species name

  • size_cat: body size category (categorical: small, medium, large)

  • stenotopy: habitat specialisation (categorical: eurytopic, oligotopic, stenotopic)

  • food_spec: food specialisation (categorical: polyphagous, oligophagous, monophagous)

  • Tniche: temperature niche (categorical: cold, intermediate, warm)

  • voltinism: number of generations per year (categorical: 0.5, 1, 2, 3)

  • hibernation: overwintering stage (categorical: egg, egg/larva, larva, larva/pupa, pupa, adult)

• Drivers/Drivers.csv: Processed data on environmental variables. Each row contains data for one biozone and interval. Contains the following columns:

  • zone: biozone

  • year_mean: Mean year of the interval

  • length_interval: Length of the intervals, either 8 (eight years) or 12 (twelve years)

  • forest_area_change: estimated change in forest area (real number)

  • forest_area_abs: estimated absolute forest area (proportion) (real number)

  • wood_harvest_int_change: estimated change in wood harvest intensity (real number)

  • after_storm: factor denoting whether the interval follows a storm event (boolean)

  • human_pop_change: estimated change in human population density (real number)

  • mechanisation_abs: estimated absolute mechanisation degree (real number)

  • mechanisation_period: factor denoting whether the intervals falls within the major mechanisation period (boolean)

  • grassland_area_change: estimated change in grassland area (real number)

  • grassland_area_abs: estimated absolute grassland area (proportion) (real number)

  • temperature_change: estimated mean annual temperature change (real number)

  • temperature_abs: estimate absolute mean annual temperature anomaly (real number)

  Most variables are based on data in the file 'Drivers/Census_data_raw.csv' (see sources and licences therein). The variables 'temperature_change' and 'temperature_abs' were derived from gridded, reconstructed climate data provided by MeteoSwiss (https://www.meteoswiss.admin.ch; TrecabsM; licensed under CC BY 4.0).

• Drivers/Census_data_raw.csv: Raw data from different censuses, which were used to calculate some of the environmental variables. At the level of Swiss cantons. Contains the following columns:

  • year: year of the data

  • canton: name of the canton

  • value: value of the censused variable (real number)

  • parameter: name of the censused variable. One of the following:

    • "forest: area" (forest area in ha)

    • "forest: wood harvest" (wood harvest in m3)

    • "grassland area" (grassland area in ha)

    • "human population" (human population)

    • "tractors" (number of tractors)

  • forest_owner: for forest data, denotes whether data is on publicly or privately owned forests

  • source: source of the data

• Other/GBIF/d_taxonomy_GBIF_sapro.csv: File used to match GBIF taxon keys to taxonomic names used in the present analyses. Used during temperature niche calculation of saproxylic beetles.

• Other/biogeographic_zones.csv: File containing information on the biozone at the level of 5x5 km squares. Also includes condensed data on the fine-scale biogeographic region (raw data available from https://data.geo.admin.ch, see 'ch.bafu.biogeographische_regionen') and the average elevation (raw data from the Federal Office of Topography swisstopo, avilable from https://data.geo.admin.ch, see 'ch.swisstopo.swissalti3d'). Contains the following columns:

  • five_km2_ID: ID of the 5x5 km squares. Based on Swiss coordinate system LV95 (EPSG: 2056)

  • zone: Name of the biozone of the respective square

  • biogeo12: Name of the fine-scale biogeographic region (12 levels; in German). Based on data from the Federal Office for Environment (https://data.geo.admin.ch, see 'ch.bafu.biogeographische_regionen')

  • height: Average height of the 5x5 km square in meters above sea level. Based on data from the Federal Office of Topography swisstopo (https://data.geo.admin.ch, see 'ch.swisstopo.swissalti3d')

The folder Data/RAW_occdet contains data necessary to reproduce the occupancy-detection models. These data are not necessary to reproduce the main analyses, as the aggregated occupancy estimates are also available. The folder contains the following files:

• d_expertprojs_butter.csv: File containing the project IDs of projects that recorded butterflies, which were defined as expert projects (see manuscript for definition)

• d_expertprojs_sapro.csv: File containing the project IDs of projects that recorded saproxylic beetles, which were defined as expert projects (see manuscript for definition)

• d_expertprojs_butter.csv: File containing the observer IDs of people that recorded butterflies, which were defined as experts (see manuscript for definition)

• d_expertprojs_sapro.csv: File containing the observer IDs of people that recorded saproxylic beetles, which were defined as experts (see manuscript for definition)

• d_records_butter.csv: Butterfly species records, which were extracted from the info fauna database and used in occupancy-detection models. The data are anonymised and only contain records included in the modelling. The file contains the following columns:

  • Name_std: species name

  • five_km2_ID: ID of the 5x5 km square, in which the species was recorded. Corresponds to longitude and latitude in the Swiss coordinate system (LV95; EPSG: 2056)

  • two_A: two-year interval of the record. Always denotes the first year of the interval.

  • MUS: Denotes whether the recorded specimen was digitized from a museum collection ("Yes")

  • visit_ID: anonymised ID of the visit (see manuscript for definition)

  • LEG: anonymised ID of the observer

  • PROJET: anonymised ID of the project. If the record was not part of a project, the field is empty

• d_records_sapro.csv: (Saproxylic) beetle species records, which were extracted from the info fauna database and used in occupancy-detection models. The data are anonymised and only contain records included in the modelling. The file contains the following columns:

  • Name_std: species name

  • Tax_group: Larger taxonomic group to which the species was attributed

  • five_km2_ID: ID of the 5x5 km square, in which the species was recorded. Corresponds to longitude and latitude in the Swiss coordinate system (LV95; EPSG: 2056)

  • two_A: two-year interval of the record. Always denotes the first year of the interval.

  • MET_SMRY: Sampling method used. One of "Flight", "Light", "Ground", "Emergence", "Other traps". If no trap was used (or trap type was not specified), the field is empty

  • MUS: Denotes whether the recorded specimen was digitized from a museum collection ("Yes")

  • visit_ID: anonymised ID of the visit (see manuscript for definition)

  • LEG: anonymised ID of the observer

  • PROJET: anonymised ID of the project. If the record was not part of a project, the field is empty

License

The research data in the folder 'Data' is licensed under CC BY 4.0. For the codes in the folders 'R_Code' and 'Stan_Code', the supplied LICENSE applies.