Search for: All records

ABSTRACT Herbarium specimens represent critical historical records of plant phenology, yet automating annotation of reproductive structures remains challenging given the diversity of floral morphologies, specimen age and quality, and image quality. Here, we present a machine learning pipeline that uses an ensemble modeling approach to detect flowers on herbarium specimens and deliver these data to the phenology research community. After testing multiple strategies for generating training data, we found in-house expert-curated annotations were essential for producing reliable results. Expert validation found relatively strong accuracy for detecting present floral structures, but still had moderately high false negative rates. Applying the ensemble to our filtered final image dataset of 22 million records resulted in 11.1 million records labeled with flowers present. However, only 2.9 million of these contained complete metadata necessary for downstream phenology research, highlighting the need for full label digitization efforts. Still, this dataset represents a large compilation of historical herbarium-derived phenology records available as a resource for the phenology community. We end by demonstrating how integrating these machine-labeled records into Phenobase, a publicly-available phenology database, expands taxonomic and temporal coverage for large-scale phenological analyses, and discuss remaining challenges and next steps. 

The National Ecological Observatory Network (NEON) is a long-term monitoring program at the continental scale designed to understand and forecast ecological responses to environmental change at local to broad scales. However, despite robust and nearly continuous collections, NEON mosquito data have been underused in downstream analyses. Here, we provide species-level estimated abundances for nighttime collected female mosquitoes derived from the mosquitoes sampled from CO2 traps (DP1.10043.001) (RELEASE-2024; NEON, 2024). By including zero counts, our derived data complement existing data sets and provide an analysis-ready time series useful for investigating mosquito phenology, abundances, and diversity at the species or community level. We also outline a set of considerations specific to filtering NEON mosquito data by sex and for day or nighttime collections, highlighting factors that could introduce uncertainty to abundance estimates. Along with the data set, we provide an R Markdown file that includes annotated code and documents our data filtering and QC/QA steps, as well as data files used to filter the mosquito data based on QC/QA criteria. All files are freely available for download through the Environmental Data Initiative data portal. Our reproducible and fully documented workflow can be easily adapted for specific needs or other NEON surveillance data. Our work aims to enhance the accessibility and use of NEON’s rich, long-term monitoring data. 

Abstract Plant phenology plays a fundamental role in shaping ecosystems, and global change‐induced shifts in phenology have cascading impacts on species interactions and ecosystem structure and function. Detailed, high‐quality observations of when plants undergo seasonal transitions such as leaf‐out, flowering and fruiting are critical for tracking causes and consequences of phenology shifts, but these data are often sparse and biased globally. These data gaps limit broader generalizations and forecasting improvements in the face of continuing disturbance. One solution to closing such gaps is to document phenology on field images taken by public participants. iNaturalist, in particular, provides global‐scale research‐grade data and is expanding rapidly.Here we utilize over 53 million field images of plants and millions of human annotations from iNaturalist—data spanning all angiosperms and drawn from across the globe—to train a computer vision model (PhenoVision) to detect the presence of fruits and flowers. PhenoVision utilizes a vision transformer architecture pretrained with a masked autoencoder to improve classification success, and it achieves high accuracy on held‐out test images for flower (98.5%) and fruit presence (95%), as well as a high level of agreement with an expert annotator (98.6% for flowers and 90.4% for fruits).Key to producing research‐ready phenology data is post‐calibration tuning and validation focused on reducing noise inherent in field photographs, and maximizing the true positive rate. We also develop a standardized set of quality metrics and metadata so that results can be used effectively by the community. Finally, we showcase how this effort vastly increases phenology data coverage, including regions of the globe where data have been limited before.Our end products are tuned models, new data resources and an application streamlining discovery and use of those data for the broader research and management community. We close by discussing next steps, including automating phenology annotations, adding new phenology targets, for example leaf phenology, and further integration with other resources to form a global central database integrating all in situ plant phenology resources. 

Abstract Climatic change is dramatically altering phenology but generalities regarding tempo and mode of response remain limited. Here we present a general model framework incorporating spring temperature, velocity of spring warming, and species’ thermal requirements for predicting phenological response to warming. A key prediction of this framework is that species active earlier in the season and located in warmer regions where spring temperature velocity is lowest show strongest sensitivity to climatic change and greatest advancement in response to warming. We test this prediction using plant phenology datasets collected in the 1850s and 2010s. Our results strikingly confirm model predictions, showing that while temperature sensitivity is higher in regions with low temperature velocity, the greatest realized change in phenological onset is northern areas where warming rates have been fastest. Our framework offers enhanced utility for predicting phenological sensitivity and responsiveness in temperate regions and across multiple plant species and potentially other groups. 

This Level 2 data package contains species level estimated abundances, including zero counts, and estimated mean number of female mosquitoes per trap derived from the NEON Mosquitoes sampled from CO2 traps (DP1.10043.001), RELEASE-2024 Level 0 data (https://doi.org/10.48443/3cyq-6v47). The data set includes mosquito records of traps collecting mosquito samples at night, for up to 24 trap hours, across a total of 20 terrestrial core and 27 terrestrial gradient sites from 2014 to 2022. To ensure high confidence in abundance estimates, records were only included when at least 90% of collected individuals were identified to sex, and 90% of female specimens were identified to species. Information across multiple QC/QA fields within the NEON mosquito data was evaluated to identify and exclude records where confidence in estimated abundances may have been compromised. Species level zero counts were added for all species collected at least once within the sampling year and trap location. Additionally, species level zero counts were included for trap events where only male mosquitoes had been collected or where QC/QA remarks indicated traps were inactive due to cold temperatures. The data set provides an analysis ready time series of estimated abundances across NEON sites and plots. An R Markdown file that contains descriptions of the QC/QA and data filtering steps along with annotated code, as well as data tables used to filter active and inactive trap events based on QC/QA fields, are published with the data package. Any questions about this data package should be directed to Amely Bauer listed under contacts. 

It was shown in [10] that that there exist strongly dense free subgroups in any semisimple algebraic group over a large enough field. These are nonabelian free subgroups all of whose subgroups are either cyclic or Zariski-dense. Here we show that the same is true for as long as the transcendence degree of the field is at least 1 in characteristic 0 and transcendence degree at least 2 in positive characteristic. 

Abstract PremisePlant 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. MethodsWe 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. ResultsPost‐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. DiscussionThese data, now accessible through the Phenobase web application, establish a foundation for near real‐time monitoring of leaf phenology, supporting global‐scale synthesis analyses.