Tracking vegetation phenology across diverse biomes using Version 3.0 of the PhenoCam Dataset

Authors: YOUNG, ADAM - Neon, Inc; MILLIMAN, THOMAS - University Of New Hampshire; HUFKENS, KOEN - Institute Of Medical Immunology; BALLOU, KEITH - Northern Arizona University; COFFEY, CHRISTOPHER - Northern Arizona University; BEGAY, KAI - Northern Arizona University; FELL, MICHAEL - Northern Arizona University; MOSTAFA, JAVADIAN - Northern Arizona University; POST, ALISON - University Of Colorado; SCHÄDEL, CHRISTINA - Woodwell Climate Research Center; VLADICK, ZAKARY - Northern Arizona University; ZIMMERMAN, OSCAR - Northern Arizona University; Browning, Dawn; FLORIAN, CHRISTOPHER - National Ecological Observatory Network (NEON); MOON, MINKYU - Kangwon National University; SANCLEMENTS, MICHAEL - National Ecological Observatory Network (NEON); SEYEDNASROLLAH, BIJAN - Northern Arizona University; FRIEDL, MARK - Boston University; RICHARDSON, ANDREW - Northern Arizona University

Submitted to: Earth System Science Data
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/26/2025
Publication Date: 11/26/2025
Citation: Young, A.M., Milliman, T., Hufkens, K., Ballou, K., Coffey, C., Begay, K., Fell, M., Mostafa, J., Post, A.K., Schädel, C., Vladick, Z., Zimmerman, O., Browning, D.M., Florian, C.R., Moon, M., SanClements, M.D., Seyednasrollah, B., Friedl, M.A., Richardson, A.D. 2025. Tracking vegetation phenology across diverse biomes using Version 3.0 of the PhenoCam Dataset. Earth System Science Data. 17(11):6531-6556.

Interpretive Summary: USDA scientist in Las Cruces, NM and partners have expanded PhenoCam Dataset, a national dataset that provides daily, ground-based imagery of vegetation across forests, rangelands, farms, ranches, and other natural landscapes. This release includes over 700 camera sites and more than 4,800 site-years of data (through 2023), offering detailed records of seasonal plant growth that help document how crops and rangeland vegetation respond to regional weather conditions. The long-term dataset can help farmers, ranchers, and land-grant educators better understand year-to-year variability in plant productivity and inform outreach, planning, and conservation efforts. Its simplified, user-friendly format also increases access for rural educators, students, and local land managers looking for reliable science-based tools to support agricultural stewardship.  

Technical Abstract: Vegetation phenology plays a significant role in driving seasonal patterns in land-atmosphere interactions and ecosystem productivity, and is a key factor to consider when modeling or investigating ecological and land-surface dynamics. To integrate phenology in ecological research ultimately requires the application of carefully curated and quality controlled phenological datasets that span multiple years and include a wide range of different ecosystems and plant functional types. By using digital cameras to record images of plant canopies every 30'min, pixel-level information from the visible red-green-blue color channels can be quantified to evaluate canopy greenness (defined as the green chromatic coordinate, GCC), and how it varies in space and time. These phenological cameras (i.e., “PhenoCams”) offer a pragmatic and effective way to measure and provide phenology data for both research and education. Here, in this dataset descriptor, we present the PhenoCam dataset version 3 (V3.0), providing significant updates relative to prior releases. PhenoCam V3.0 includes 738 unique sites and a total of 4805.5 site years, a 170'% increase relative to PhenoCam V2.0 (1783 site years), with notable expansion of network coverage for evergreen broadleaf forests, understory vegetation, grasslands, wetlands, and agricultural systems. Furthermore, in this updated release, we now include a PhenoCam-based estimate of the normalized difference vegetation index (cameraNDVI), calculated from back-to-back visible and visible+near-infrared images acquired from approximately 75'% of cameras in the network, which utilize a sliding infrared cut filter. Both GCC and cameraNDVI showed similar, but somewhat unique, patterns in canopy greenness and VIS vs. NIR reflectance, across various ecosystems, indicating their consistent ability to record phenological variability. However, we did find that at most sites, GCC time series had less variability and fewer outliers, representing a smoother signal of canopy greenness and phenology. Overall, PhenoCam greenness as measured by both GCC and cameraNDVI provides expanded opportunities for studying phenology and tracking ecological changes, with potential applications to the evaluation of satellite data products, earth system and ecosystem modeling, and understanding phenologically mediated ecosystem processes. The PhenoCam V3.0 data release is publicly available for download from the Oak Ridge National Lab Distributed Active Archive Center: the source imagery used to derive phenology information is available at https://doi.org/10.3334/ORNLDAAC/2364 (Ballou et al., 2025), and the summarized phenology data are available at https://doi.org/10.3334/ORNLDAAC/2389 (Zimmerman et al., 2025).