pep725

pep725: an R package for Pan-European Phenological Data Analysis

pep725 is an R package for analyzing phenological data from the PEP725 Pan-European Phenology Database and beyond. It provides tools for data exploration, quality assessment, climatological baseline calculation, trend analysis and visualization of phenology-climate relationships.

Citation

If you use this R package, please cite: Templ and Templ (2026). pep725: An R package for Pan-European Phenological Data Analysis. The Journal of Open Source Software, under review. Until it gets published, please cite: Templ et al. (2018). Pan European Phenological database (PEP725): a single point of access for European data. Int. J. Biometeorology, 62(6): 1109-1113. doi:10.1007/s00484-018-1512-8

Features

Data Access:

• Import PEP725 data files with automatic cleaning and formatting
• Download pre-generated synthetic datasets for learning and testing
• Generate synthetic data from your own datasets for teaching and sharing

Core Analysis:

• pheno_normals() - Calculate climatological baselines (WMO-style reference periods)
• pheno_anomaly() - Detect deviations from normal phenology
• pep_quality() - Assess data quality and assign reliability grades
• pheno_gradient() - Quantify elevation and latitude gradients
• pheno_synchrony() - Measure spatial coherence of phenological events

Quality & Validation:

• pep_completeness() - Assess species/phase coverage across stations
• pep_check_phases() - Validate phenological phase sequences
• pep_flag_outliers() - Detect outliers (30-day rule, MAD, IQR, z-score)

Advanced Analysis:

• pheno_combine() - Combined time series from multi-station data (robust/mixed/OLS)
• pheno_trend_turning() - Sequential Mann-Kendall trend turning point detection
• pheno_pls() - PLS regression for temperature-sensitive phenological periods

Visualization:

• pheno_plot_timeseries() - DOY trend plots over time
• plot_phenology_trends() - Robust regression trend analysis
• pheno_leaflet() - Interactive maps for station exploration and selection
• pheno_map() - Static maps of station networks

Installation

# Install from GitHub
# install.packages("devtools")
devtools::install_github("matthias-da/pep725")

Quick Start

library(pep725)

# Download synthetic data (cached locally after first download)
pep <- pep_download()

# Explore the data
print(pep)
summary(pep)
pep_coverage(pep)

# Filter to a species
wheat <- pep[species == "Triticum aestivum"]

# Calculate phenological normals
normals <- pheno_normals(
  wheat,
  period = 1990:2015,
  by = c("country", "phase_id"),
  min_years = 10
)
print(normals)

# Detect anomalies
anomalies <- pheno_anomaly(
  wheat,
  baseline_period = 1990:2010,
  by = c("country", "phase_id")
)
summary(anomalies)

# Assess data quality
quality <- pep_quality(wheat, by = c("s_id", "phase_id"))
summary(quality)

Data Options

Option	Best For	Requires Internet
pep_download()	Learning, tutorials, reproducible examples	First use only
data(pep_seed)	Quick offline tests, minimal examples	No
pep_simulate()	Creating shareable datasets from your own data	No
pep_import()	Research with actual PEP725 observations	For download
as.pep()	Working with your own phenological data	No

Why synthetic data? The original PEP725 database requires registration and has usage restrictions. Synthetic data preserves statistical properties while being freely shareable for tutorials and reproducible workflows.

Working with your own phenological data

You can use all analysis functions on any phenological dataset — it does not have to come from PEP725. Convert your data to a pep object with as.pep():

library(data.table)

my_data <- data.table(
  s_id     = c(1, 1, 2, 2),
  lon      = c(8.5, 8.5, 9.1, 9.1),
  lat      = c(47.4, 47.4, 46.9, 46.9),
  alt      = c(400, 400, 550, 550),
  genus    = "Prunus",
  species  = "Prunus avium",
  phase_id = 60,
  year     = c(2022, 2023, 2022, 2023),
  day      = c(105, 98, 112, 103)
)

my_pep <- as.pep(my_data)

# Now use any pep725 analysis function
normals <- pheno_normals(my_pep, period = 2022:2023, min_years = 2)

Required columns: s_id, lon, lat, genus, species, phase_id, year, day. Optional: alt, country, subspecies.

Spatial Analysis

# Elevation gradient analysis
gradient <- pheno_gradient(
  pep,
  variable = "alt",
  species = "Triticum aestivum",
  phase_id = 60,
  method = "robust"
)
print(gradient)
plot(gradient)

# Spatial synchrony
synchrony <- pheno_synchrony(
  pep,
  species = "Triticum aestivum",
  phase_id = 60,
  by = c("country", "year"),
  min_stations = 3
)
summary(synchrony)

Interactive Mapping

# Launch interactive map for station selection
selected <- pheno_leaflet(pep, label_col = "species")

# Use selected stations for analysis
pep_subset <- pep[s_id %in% selected$s_id]

Built-in Datasets

Dataset	Description
pep_seed	Small seed dataset for quick tests (1,319 rows)

Vignettes

The package includes detailed vignettes:

• Getting Started with pep725 - Data loading, exploration, and basic usage
• Phenological Analysis - Normals, anomalies, quality assessment, and visualization
• Spatial Phenological Patterns - Gradients, synchrony, and mapping
• Data Quality Assessment - Quality grading, outlier detection, and completeness

vignette("getting-started", package = "pep725")
vignette("phenological-analysis", package = "pep725")
vignette("spatial-patterns", package = "pep725")
vignette("data-quality", package = "pep725")

Requirements

• R >= 4.1
• Google Maps API key required for pheno_map() static maps:

  ggmap::register_google(key = "your_api_key")

License

GPL-3

Authors

• Matthias Templ (maintainer)
• Barbara Templ

Acknowledgements

Dr. Barbara Templ was supported by the Berner Nachwuchsförderung (BNF) program of the University of Bern for developing the pep725 R package.

Links

• GitHub Repository
• Report Issues
• PEP725 Database