Leveraging machine learning and citizen science data to describe flowering phenology across South Africa
Bibliographic record
Abstract
Societal Impact Statement Recent shifts in flowering times are an index of, and a response to, human driven climate change. However, most information on these flowering changes is heavily skewed to the northern hemisphere. This imbalance limits our understanding of how climate change is affecting ecosystems, including the mismatches of flowering times between species, increased risks from frost or drought, and shifts in crop growing seasons. Here we show how this knowledge gap can be addressed by combining modern machine learning tools with large‐scale data collected by citizen scientists. We present a new map showing flowering patterns across South Africa, revealing complex regional differences and wide variation between plant species. Summary Phenology—the timing of recurring life history events—is strongly linked to climate. Shifts in phenology have important implications for trophic interactions, ecosystem functioning, and community ecology. However, data on plant phenology can be time consuming to collect and current records are biased across space and taxonomy. Here, we evaluate the performance of convolutional neural networks (CNN), a machine learning tool, for classifying flowering phenology on a very large and taxonomically diverse dataset of citizen science images sourced from the iNaturalist database. Our analysis focuses on plants listed in the National Botanical Gardens (NBGs) of South Africa, a country famed for its floristic diversity (approximately 21,000 species), but poorly represented in phenological databases. We trained our CNN on a dataset of 10,000 images, representing several thousand species across >200 families, and applied the trained model to >1.8 million iNaturalist records. We were able to correctly classify images with >90% accuracy. Using metadata associated with each image, we then reconstructed the timing of peak flower production and length of the flowering season across South Africa and within each National Botanical Garden. Our analysis illustrates how machine learning tools can leverage the vast wealth of citizen science biodiversity data to describe large‐scale phenological dynamics. We suggest such approaches may be particularly valuable where data on plant phenology is currently lacking.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.005 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".