Eight Million Plant Specimens Show That Flowering Times Are Drifting
herbarium specimens
Pressed and dried plants kept with collection records.
pollinators
Animals that carry pollen between flowers.
digitization
Turning physical specimens and labels into online images and data.
What happened
A new 2026 report from the Royal Botanic Gardens, Kew, a UK institution that studies plants and fungi, examines how digital tools are changing biodiversity science. More than 400 scientists from 40 countries contributed. One global study used artificial intelligence, or AI, to analyze eight million digitized plant specimens. The records cover roughly the past century. Flowering times shifted earlier or later by an average of 2.5 days per decade. This does not mean every plant now blooms earlier. The direction and size of change vary by place and species. Kew’s 2026 report
How old specimens became useful
A herbarium specimen is a plant pressed, dried, labeled, and stored. Its label can record where and when people collected it. Digitization turns the specimen and label into an online image. AI can scan many images, find flowers, and compare patterns across years. Kew has completed digitization of 7.4 million plant and fungal specimens in its own collections. Researchers can now compare historical records without traveling to London.
Why it matters
Flowering is tied to pollinators and animals that depend on plants. If flowers and pollinators shift at different speeds, they can fall out of sync. That may reduce pollination, interrupt food chains, and weaken ecosystems. In India’s Western Ghats, synchronized flowering in one forest tree fell from 79% in the 1950s to 47% in the 1990s. This suggests timing changes can affect relationships between species, not just calendars.
What the study found
The report describes this as the first comprehensive global study of flowering-time change. Shifts were strongest in tropical areas. Tropical flowering did not depend on temperature alone. Rainfall patterns mattered too. In the Canadian Arctic, different plant species changed in different directions. The flowering season also became shorter. Together, these findings show that climate change is altering nature’s timing in complex ways.
What remains uncertain
The world holds an estimated 406 million herbarium specimens. Fewer than 16% have been imaged and made available online. Digitized collections are also concentrated in some regions. That creates blind spots. The 2.5-day average is therefore a global clue, not a perfect picture of every place. Scientists still need to learn how temperature, rain, and other conditions affect each species. They also need more evidence linking a timing shift to long-term ecosystem damage. AI can speed the search, but human experts must check its results.
What to watch next
Researchers want to digitize more collections, especially in data-poor regions. They can then combine specimen records with observations of pollinators, rainfall, and local seasons. That could help conservation teams find vulnerable relationships earlier. The main story is not that AI solved climate change. It is that old collections can reveal changes that one specimen alone could never show.
AI Finds Climate Clues in Eight Million Plant Specimens
📰 Full story: Eight Million Plant Specimens Show That Flowering Times Are Drifting
Old plant collections are showing how nature’s calendar is changing.
AI
A computer system that finds patterns in many examples.
herbarium specimens
Pressed and dried plants kept with their labels.
pollinators
Animals that carry pollen between flowers.
💡 The gist
- AI studied 8 million plant specimens.
- Flowering times shifted 2.5 days per decade.
- Different places changed in different ways.
How did scientists do this?
The Royal Botanic Gardens, Kew, published a 2026 report. It is a UK institution that studies plants and fungi. More than 400 scientists helped. They studied herbarium specimens. These are pressed and dried plants. Their labels show collection places and dates. Scientists made digital images of the specimens. AI, a computer system that finds patterns, helped read millions of images. It looked for flowers and collection dates. This let researchers compare plants from different years and places. Kew also put 7.4 million plant and fungal specimens online. More detail is in Kew’s report.
Why does flowering time matter?
Flowers and pollinators depend on each other. Pollinators move pollen between flowers. This helps plants make seeds. If flowers change timing, pollinators may arrive too early or too late. Food chains can also be disturbed. In tropical areas, rainfall mattered as well as temperature. That makes changes harder to predict.
What did they find?
Over the last century, flowering shifted earlier or later. The average shift was 2.5 days per decade. Some places became earlier. Others became later. In India’s Western Ghats, synchronized flowering in one tree fell from 79% in the 1950s to 47% in the 1990s. In the Canadian Arctic, different plants changed differently. The flowering season also became shorter.
What is still unclear?
The world has about 406 million herbarium specimens. Fewer than 16% are online. Many regions have fewer digital records. So the global average is not a perfect picture. Scientists still need local data. They must check AI’s results too.
What comes next?
Researchers want to digitize more collections. They especially want records from places with little data. They can compare plants, insects, rainfall, and seasons. This may help conservation teams protect nature earlier. AI found a large clue. Human scientists still decide what it means.
A Computer Finds Changes in Flower Seasons
📰 Full story: Eight Million Plant Specimens Show That Flowering Times Are Drifting
A computer found small changes in when flowers open.
AI
A computer that looks for patterns.
Royal Botanic Gardens, Kew
A place in Britain that studies plants.
pollen
Tiny powder that flowers use to make seeds.
Looking at old plant pictures
The Royal Botanic Gardens, Kew, studies plants in Britain. It saved many old plant samples. People pressed plants flat and dried them. Then they made pictures of the samples. AI, a computer that looks for patterns, studied eight million pictures. The pictures showed when flowers opened. Every ten years, flowering shifted by 2.5 days on average. Some places changed earlier. Some places changed later. Rain and warmth can change flower timing. Tiny insects carry pollen between flowers. They need flowers to be open. If timing changes, insects may miss flowers. Scientists need more pictures from many places. AI helps count clues. People check the clues.