Scorching summers such as the one in 2003 look set to become more common in England and Wales, a study suggests.
And devastating rains such as in Britain's worst winter in 2013-14 may be less likely in the decades ahead.
Work by the Met Office has calculated the odds of particular weather scenarios striking in future years.
The computer simulations-based study, in the journal Nature Climate Change, finds that milder winters and drier summers will also become more likely.
The work draws on a major analysis, known as UKCP09, released back in 2009 which offered projections of the future British climate divided into 30-year periods.
This new research instead provides a more detailed focus by giving projections for winters and summers in each individual year from now until the end of the century.
The aim is to take more account of the fact that Britain's weather is notoriously variable - fluctuating for natural reasons year to year regardless of human-induced climate change.
The authors of the research looked only at data from England and Wales; the analysis did not take into account Scotland or Northern Ireland.
A parallel goal is to make clear that a trend to warmer temperatures does not mean that extremes of cold or rainfall are made impossible - instead, weather that seems to buck the prevailing remains on the cards, if less likely as the century progresses.
The 2009 study had suggested that the country faced a future of milder, wetter winters and hotter, drier summers - and the Met Office faced fierce criticism when shortly afterwards Britain was suddenly plunged into the bitterly cold winter of 2009-10.
Met Office scientists acknowledge that there was confusion in the public mind about the "apparent contradiction" of hearing a 30-year projection for milder winters only to endure the reality of ice and snow.
The authors of the new study hope that coming up with odds for different scenarios for weather in individual seasons will more useful.
The paper says that the new approach has two advantages: "First, it allows fair comparisons with recent weather events, for instance showing that recent cold winters are within projected ranges.
"Second, it allows the projections to be expressed in terms of the extreme hot, cold, wet or dry seasons that impact society, providing a better idea of adaptation needs".
Some key conclusions from the study include:
- By 2100, the chances of a summer being hotter than the one in 2003 are 89% - that's odds of roughly 9-out-of-10
- There is still a 35-40% chance of getting a wetter-than-average summer until 2035 but that risk falls to 20% by 2100
- The chances of a winter with the same kind of rainfall as in 2013-14 fall to just under 10% by the end of the century
- And the odds of a very cold winter similar to 2009-10 fall to less than 1% over the same period
A co-author of the report, David Sexton, said that basing the projections on 30-year averages, as in the UKCP09 study, risked giving the impression to people that those weather conditions would apply to every single year.
"When I talk to people, they remember the hot summer of 2003 or the wet winter of 2013-14 and they know they were extreme seasons - people can make tangible links to those impacts, they mean something to them personally, and the 30-year averages don't make sense to people in the same way."
Follow David on Twitter.