Computer simulations of the weather workings of the entire planet will be able to make forecasts to within a few kilometers accuracy, helping predict the effects of deadly weather systems.

But the world may have to wait 20 to 40 years' for such accurate information on weather events like El Nino as computer capacity grows, a senior British scientist said Thursday.

If we step forward 20 to 40 years into the future of climate science, it is conceivable we can have climate models down to a scale of a few kilometers' resolution, Alan Thorpe, director general of the UK-based European Center for Medium-Range Weather Forecasts (ECMWF), told reporters.

That would add a huge amount of information to this variability question.

A climate model is a computer-based version of the Earth's climate system, based on physics and complex equations. Such models can be used for weather forecasting, understanding the climate and projecting climate change.

A model with a very fine resolution could produce more accurate results but this depends on computer capacity.

Thorpe said some climate models are now nearing a resolution of 100 km, compared to around 300 km 10 to 15 years ago.

We are running global weather picture models at a 16 km resolution already so we have the science and the models to reduce the problem of high resolution but we need the computer power to do it, Thorpe said.

It would cost up to 200 million pounds to buy a top-end super computer, he added, which is around 7 percent of the UK's yearly science budget of 3 billion pounds.

The impact of climate change needs to be seen as sufficiently important to society to devote this level of resource to it, Thorpe said.

Some experts warn that some of the most devastating impacts of climate change could be felt before and during the period 2030 to 2050.

Some climate models have been criticized for not being accurate enough or not predicting extreme events far enough into the future.

Thorpe said ECMWF scientists are doing a lot of research into so-called tipping points, when there is a rapid change in the climate which is irreversible or which would take a long time to reverse.

Inevitably, those are the aspects of the system we have to worry about most because they are not linear behavior. How many of those there are is still an open question, he added.

If we devoted the whole of the science budget to these questions we could make more rapid progress but we are doing a lot of research on these areas.

Some tipping points are seen happening in the coming decades, such as the loss of summer Arctic sea ice or the loss of the Amazon rainforest.

(Editing by William Hardy)