The quiet hamlets of Vermont are transformed at this time of year. Verdant hills begin to pop with new hues: scarlet and crimson, gold and orange. Comparisons to a blazing fire are hard to avoid.

Tourists follow these colours, staying in quaint inns and snapping photographs of red barns in green fields, framed by bursts of autumn leaves. Known as “leaf peepers” in the US, they flock to New England every autumn where they wind their ways along normally quiet country roads. Elsewhere they seek out golden larches in the Alps, wild cherry in Britain, and maple and ginkgo in Japan.

We still don't have a good handle on all the factors that affect colour

As temperatures rise across the globe, these traditions are under threat. Already trees in many areas are changing colour several days later than they did two decades ago. One recent study predicts delays of seven to 20 days by the year 2100 in the US.

But it’s not just that those who appreciate autumnal colours will have to wait a little longer.

Higher temperatures and new patterns of precipitation are putting pressure on some tree species, either directly or via altered conditions for disease-bearing fungi, bacteria and insects. Landscapes are already changing as species behind some of the most dramatic of autumn’s colour shows move north and to higher latitudes. Scientists say that the more frequent droughts and greater cloud cover that some areas will see thanks to climate change is likely to reduce the production of leaf pigments that provide some of the season’s most vivid colours.

While researchers are beginning to piece together the complex relationship between climate and these autumnal changes, they admit there is still much to learn.

"We still don't have a good handle on all the factors that affect colour," says Howard Neufeld, a plant physiologist who maintains a popular autumn colour report at Appalachian State University in North Carolina. "Like, if it rains a lot in May, does that affect the fall?"

Leaves that turn yellow and orange in the autumn do so through a process of subtraction. The green pigment chlorophyll, which converts sunlight to energy while trees are growing, drains from the leaves at the end of summer as metabolism slows to a near-halt. This leaves behind yellow and orange pigments called xanthophylls and carotenoids.

Red pigments are more of a puzzle. The fiery colours of the sugar maple and other trees that turn red are due to anthocyanins, pigments that are produced after the growing season ends.

Scientists are unsure precisely why trees make them. Some say they're a signal to insects, warning potential parasites that the tree is armed with chemical defences. Another, not necessarily conflicting, theory is that anthocyanins protect the leaves from light damage. In the autumn, trees withdraw remaining nutrients from the leaves into the twigs for storage until the next growing season. Visible light can damage this transport system. Anthocyanins, which are capable of absorbing light in the yellow-green range, form part of some trees’ defences against such damage, Neufeld says.

Regardless of the uncertainties remaining around the roles of these brilliant red pigments, they may well be more at greater risk from climate change than yellows and oranges. Anthocyanins are manufactured from sugars. In cool temperatures, trees keep these sugars in the leaf, but when it's warmer, the sugars shift to other parts of the plant.

Reds can also be dampened by other human phenomena. The use of large amounts of fertilisers dumps excess nitrogen into soils, potentially causing trees to store less sugar in their leaves, and, again, dulling autumn reds.

"Anything that lowers the sugar levels will reduce the red colours," says Neufeld.

Increased drought frequency could also dampen autumn displays. This can already be seen in years when trees drop their leaves prematurely, resulting in poor autumn colour seasons.

Climate change models predict rainfall with increase in some areas and decrease in others. Yet the colour shows might also be spoilt in areas that become wetter because photosynthesis is necessary to make those red anthocyanins, and more rain means more clouds and less photosynthesis.

The study of the basic triggers of autumnal leaf changes, never mind the impact of climate change on these, is still in its infancy in part because scientists have traditionally focused their attentions on the seasonal changes in March and April than those later in the year.

People have this habit of linking together events like a warm weekend in September with the great colours they saw in the fall and not being very scientific about it.

"For some reason that I don't really understand, people have always thought that when leaves emerge in spring, and when flowers come out in spring, is a much more important thing to understand than what causes leaves to turn colour and fall off," says Andrew Richardson, a biologist who studies the impact of climate change on ecosystems at Harvard University, in Cambridge, Massachusetts.

Richardson notes that for every study of the timing of autumn colours, there are probably ten studies about the timing of spring buds, and that dubious folk knowledge has traditionally filled the gap.

“People have this habit of linking together events like a warm weekend in September with the great colours they saw in the fall and not being very scientific about it."

Increasingly, however, scientists have more data to work with, due to camera-based observations and the US National Phenology Network (NPN), a citizen science project that collects observations of seasonal changes. Some researchers have also turned to natural variation to predict climate's effects on autumn, comparing trees on high, cool slopes to the same species lower down. Autumn colours are trickier to study than spring buds, says Richardson, because one good storm can take down huge quantities of leaves and skew the data overnight, but nonetheless some clear trends are beginning to emerge.

In a study published in the journal Global Ecology & Biogeography in July, US researchers used data from NPN and from 3,000 acres of trees managed by Harvard University for research purposes to build a model designed to predict autumn colour changes. Whereas the observations from NPN go back only to 2009 but are scattered across the country, those from the Harvard Forest go back to 1990 but cover only one site.

"We could make the model by using the long-term data and then we can verify the model by using the short-term, spatially wide data," says lead author Se-Jong Jeong, of NASA's Jet Propulsion Laboratory in California.

Jeong and his colleague David Megvigy of Princeton University, US found they could largely predict colour change with two variables: daylight length and temperature. Tree species mattered, as well. For example, trees with high shade tolerance responded faster to cold temperatures by changing colours more quickly than trees with low shade tolerance.

Autumnal leaf colour changes have moved back by an average of about five days in the last 18 years, Jeong says. By 2100, the delay will almost certainly be more pronounced. Under an optimistic scenario of only 1.5 to 2 degrees C warming by then, the delay is likely to be an extra week or so, according to the researchers’ model. Under a "business as usual" scenario without a significant cut in carbon emissions, temperatures are predicted to reach 5.6 degrees C above pre-industrial levels, autumnal leaf colour changes will be delayed nearly three full weeks compared with today.

If you change the trees, then you change the colour scheme

Of course delayed colour change on its own is not necessarily disastrous. Last year, research by Richardson and colleagues also observed delayed autumn colours in Harvard Forest, with warm September temperatures driving the biggest delays. However, their predictions, published in the journal PLoS ONE, suggested the delay might actually improve colour change for leaf peepers. Because autumn will stay warmer longer, they calculated, trees will stay colourful longer before shedding their leaves. And more leaves will be in colour at the same time.

While studies of North America's autumn colours are scarce, they are even thinner on the ground elsewhere. Delays in autumn leaf colour changes have been observed in Europe too, however extensions of growing seasons are being driven more by earlier springs than later autumns, says Yann Vitasse, of the University of Basel in Switzerland.

In Japan and China, where climate-driven changes are also more noticeable in spring, leaves are colouring later too, says Hideyuki Doi, of the Institute for Sustainable Science and Development at Hiroshima University. Autumn air temperature is the most significant factor influencing colouration, says Doi, who has studied leaf colour change in Asian deciduous trees such as ginkgos, which turn yellow-gold and maples, which turn red. Trees at lower latitude respond more strongly to increasing temperatures than those higher up, his research has found.

Autumn leaves aren't the only seasonal colours affected by climate change. The Japanese apricot or Chinese plum (Prunes mume) creates a shower of pink blossoms when it flowers between January and March. These trees are blooming an average of a week earlier than they did in 1953, Doi reported in a 2007 paper in the journal Climate Research.

However, beyond the models, plenty of uncertainties remain about the precise influences of leaf colour depth and brightness, and therefore what autumns of the future will look like.

A further complication which the models fail to take into account is that trees aren't static. They move on a slower timescale than humans, of course, but they can migrate up slopes or to new areas to find their temperature comfort zones.

US Forest Service research has found that 70% of tree species in the northeastern US are already showing signs of a northward march. Among the likely climate refugees is the sugar maple, the species that lends New England much of its fiery, tourist-attracting red. If the prediction prove to be accurate, tourists looking for an autumnal colour show might in future be better off heading for Canada.

"If you change the trees, then you change the colour scheme," says Neufeld.

In the Rocky Mountains of the western US, the mountains get most of their autumn colour from aspen stands, trees which turn a shimmering, breath-taking gold. As of 2006, 138,000 acres of aspen forest had been affected by "sudden aspen decline," with more than half of mature trees dying in some areas, according to the Colorado Forest Restoration Institute. Prolonged dry spells seem to be leaving aspen vulnerable to disease and insects, and climate change is expected to increase the frequency of droughts in the American West.

Tree species don't only march north to escape climate change, they also climb upwards. This can be seen in Vermont where the border between colour-changing hardwood forest and high-elevation spruce forest is already moving up mountain slopes.

"There's a lot less land area as the mountains come to a peak, so it's harder to get up on the mountain roads, which could make tourism kind of difficult," says Neufeld.

Migrating species, tree die-offs, and delays to leaf colour changes will trigger a host of complex related changes. The trout lily, a group of spring perennials that grow in the forests of North America, Europe and Asia, for example, could miss its early-spring window of sunlight as trees leaf out sooner.

The questions swirling around the future of autumn leaves highlight the interconnectedness of the forest ecosystem. A leaf falls late; the growing season shifts and lengthens; an unobtrusive flower fails to blossom. The complexity of these systems means scientists cannot be sure precisely how these changes will play out.

It seems increasingly likely however that the forests of New England and others currently showing off their seasonal colours won’t be quite so resplendent in the years to come.