sdtaylor / phenology_gradients

0 stars 0 forks source link

Opportunistically collected photographs can be used to estimate large-scale phenological trends

This is the code and primary data for the above study.

The analysis develops and tests a method, the Weibull Grid, to estimate flowering onset across large-scales using only observations of flowering presence. It uses a Weibull estimator (from the phest package) integrated into a spatially explicit ensemble. The method is tested with a simulation study and then applied to actually phenology data derived from iNaturalist.

Code for the Weibull Grid model and making flowering simulations is in the flowergrids R package.

Data sources:

iNaturalist
Phenocam
USA National Phenology Network*
iDigBio*
*These data sources were only used to show historic trends in data collection, not in the main analysis.

Repository files

data/ : All original data used in the analysis.

results/ : Results from the simulation analysis

inat_data_prep/ : Scripts and config files for downloading and scoring the iNaturalist photographs for the two species.

prior_study_phenology : Notes and data on prior measurements of onset for the two species analysed.

install_packages.R : Install the packages used in this analysis

config.R : Setup for the simulation analysis parameters.

run_estimators.R : Runs the simulation analysis.

analysis_simulation_errors.R : Produce the figures and tables related to the simulations.

conceptual_figures.R : Create Figure 1 from the introduction.

phenocam_site_explore.R : Mapping out the different phenocam sites for selection.

process_phenocam_data.R : Downloading and process the phenocam data used in the analysis

inaturalist_onset_models.R : Fit the iNaturalist models for the 2 species and create the figure comparing them to phenocam onset

supplemental_figure_simulation_example.R : Create the supplemental figures showing the example simulations.

other_dataset_comparison.R : Create Figure 5 showing the sample size of different data sources over time.