Turning a new leaf: PhenoVision provides leaf phenology data at the global scale
7 affiliations
- 1Florida Museum of Natural History University of Florida Gainesville Florida USA
- 2USA National Phenology Network, School of Natural Resources and the Environment University of Arizona Tucson Arizona USA
- 3iNaturalist San Rafael California USA
- 4Berkeley Natural History Museums University of California Berkeley California USA
- 5Department of Botany University of Wisconsin–Madison Madison Wisconsin USA
- 6Department of Biological Sciences University of Alberta Edmonton Alberta Canada
- 7Alberta Machine Intelligence Institute Edmonton Alberta Canada
Abstract
Premise: Plant phenology dictates many aspects of community function and ecosystem dynamics. Yet, global phenology data are still limited, especially in areas lacking monitoring programs. Here we present a new data resource, PhenoVision–Leaf, which extends a computer vision pipeline utilizing iNaturalist digital image vouchers to produce global‐scale leaf phenophase data for deciduous woody genera.
Methods: We first discuss our implementation of a new human annotation framework for leaf phenology on iNaturalist, aligning with phenophase definitions used by the larger phenology community. We then showcase the use of 165,988 crowdsourced annotated records to train a Vision Transformer model with a two‐stage regime to maximize precision across single‐ and multi‐image records. This approach extends PhenoVision from scoring individual images to aggregating at the iNaturalist record level, better aligning with human annotation processes.
Results: Post‐hoc validation showed high performance for detecting present green and colored leaves (>98% precision) and reasonable precision for breaking leaf buds (>87% precision). Applying PhenoVision–Leaf to over six million iNaturalist records yielded 5.6 million record‐level phenology observations across 6500 species and 57 families, filling geographic and taxonomic gaps.
Discussion: These data, now accessible through the Phenobase web application, establish a foundation for near real‐time monitoring of leaf phenology, supporting global‐scale synthesis analyses.
Related
- Genotype-by-sequencing of global hyacinth bean (Lablab purpureus) collections provides genomic resources for crop conservation and improvement
- iNaturalist records reveal shifts in phenology of caterpillar (Lepidoptera) activity in response to warmer spring temperatures
- Advancement of spring arrival and delayed autumn departure in Ruby-throated Hummingbird (Archilochus colubris) migration phenology over more than a century
- Effect of Short-Term, Above-Optimal Temperature on the Phenology of Partially or Fully Vernalized Seedlings and Its Consequences on Grain Yield in Wheat (Triticum aestivum L.)
- Latitudinal Variation in the Phenology of the Emergent Species Phragmites australis and the Submerged Species Potamogeton crispus in Shallow Lakes of the North China Plain
BibTeX
@article{grady2026turning,
title = {{Turning a new leaf: PhenoVision provides leaf phenology data at the global scale}},
author = {Erin L. Grady and Ellen G. Denny and Carrie E. Seltzer and John Deck and Daijiang Li and Russell Dinnage and Robert P. Guralnick},
journal = {Applications in Plant Sciences},
year = {2026},
month = jun,
doi = {10.1002/aps3.70067},
url = {https://doi.org/10.1002/aps3.70067},
note = {CC BY-NC 4.0},
}
RIS
TY - JOUR
TI - Turning a new leaf: PhenoVision provides leaf phenology data at the global scale
AU - Erin L. Grady
AU - Ellen G. Denny
AU - Carrie E. Seltzer
AU - John Deck
AU - Daijiang Li
AU - Russell Dinnage
AU - Robert P. Guralnick
PY - 2026
DA - 2026/06/24
JO - Applications in Plant Sciences
DO - 10.1002/aps3.70067
UR - https://doi.org/10.1002/aps3.70067
AB - Premise: Plant phenology dictates many aspects of community function and ecosystem dynamics. Yet, global phenology data are still limited, especially in areas lacking monitoring programs. Here we present a new data resource, PhenoVision–Leaf, which extends a computer vision pipeline utilizing iNaturalist digital image vouchers to produce global‐scale leaf phenophase data for deciduous woody genera. Methods: We first discuss our implementation of a new human annotation framework for leaf phenology on iNaturalist, aligning with phenophase definitions used by the larger phenology community. We then showcase the use of 165,988 crowdsourced annotated records to train a Vision Transformer model with a two‐stage regime to maximize precision across single‐ and multi‐image records. This approach extends PhenoVision from scoring individual images to aggregating at the iNaturalist record level, better aligning with human annotation processes. Results: Post‐hoc validation showed high performance for detecting present green and colored leaves (>98% precision) and reasonable precision for breaking leaf buds (>87% precision). Applying PhenoVision–Leaf to over six million iNaturalist records yielded 5.6 million record‐level phenology observations across 6500 species and 57 families, filling geographic and taxonomic gaps. Discussion: These data, now accessible through the Phenobase web application, establish a foundation for near real‐time monitoring of leaf phenology, supporting global‐scale synthesis analyses.
SN - 2168-0450
ER -