Researchers from UNSW and the Botanic Garden of Sydney are using AI to comprehend and mitigate the impact of climate change on plant life. They trained an AI system to extract information from millions of plant specimens in herbaria worldwide.
Herbarium collections act as time capsules for plant specimens, but manual examination is no longer feasible due to the large influx of specimens. AI transforms these collections, effectively documenting how changing environmental conditions influence plant life on Earth.
Machine learning algorithms identify trends that elude human researchers, providing insights into plant evolution, adaptation, and predictions for future environmental changes.
Herbarium collections are being scanned to create high-resolution digital copies, enabling faster processing and logging of specimen characteristics.
A collaboration between the Botanic Gardens and UNSW resulted in the development of models and algorithms to detect and measure leaf size in plant images from two plant subfamilies.
The machine learning algorithm achieved an acceptable level of accuracy in examining the relationship between leaf size and climate, disproving a commonly observed pattern across species.
Contrary to the notion that leaf size increases in warmer climates, the study found that factors other than climate play a significant role in determining leaf size within a single species.
While the pattern holds true across different species, it does not apply within a single species due to gene flow, which weakens plant adaptation on a local scale.
This discovery highlights the valuable applications of AI in botany, unveiling previously hidden insights.