Between weather and climate: the predictability gap

Illustration of the region of the Pacific Ocean in which the El Nino climate event is focussed. The event affects global climate and measurably affects the rotation of the Earth.

Illustration of the region of the Pacific Ocean in which the El Nino climate event is focussed. The event affects global climate and measurably affects the rotation of the Earth. However we cannot predict this vast event even 6 months in advance.  Picture is from Ars Technica

[UPDATE: Monday 12th May 2014: Richard Gilham wrote to highlight the Met Office’s work in trying to fill the ‘predictability gap’ – which they refer to as ‘seasonal to decadal forecasting‘. ]

The ease with which the future may (or may not) be predicted is a subject of continuing personal fascination.

When it comes to weather and climate I recently noticed there is a curious blindness in our forecasts in between the ranges of weather forecasts and climate forecasts.

  • A weather forecast involves saying (for instance) whether it will rain at a particular time and in a particular place. Advanced measurements and advanced computer models allow weather forecasts over the UK to be accurate over a range from a few days (when things are complicated) to a week or two (when things are simple).
  • However even though climate forecasts are more complex than weather forecasts and reach out much farther into the future, they are in some ways easier than weather forecasts because there is no need to predict exactly when or exactly where it will rain. So, for example, we are confident that in 10 years time, the average daily temperature for July in the United Kingdom will be close to 15 ºC.

[Aside: Do take a look at this excellent Met Office page which summarises ‘normal’ UK climate and has data for the last 100 years of so!]

This ‘predictability gap’  is at the heart of the questions about the ‘hiatus’ in global mean temperature. Of course the reason for the gap can be easily stated:

  • Weather forecasts begin with a very detailed picture of the existing weather and use  the laws of physics to run this detailed picture forward in time. Eventually, the lack of detailed knowledge of the initial state makes the forecast inaccurate.
  • Climate forecasts are based on models of the entire Earth including many layers in the oceans and the land and of course many layers in the atmosphere. Also included are the changing distance from the Sun and known solar cycles. We expect Climate models to predict the average of quantities such as the average July temperature – not the temperature in any particular year.

So there is a gap which I sometimes find shocking given the advanced state of development of both weather forecasts and climate models. And this can be  confusing when some specific weather events take on special significance. This year I am keeping my eye on two weather/climate events:

  • The extent of Arctic Sea Ice at its minimum value in September
  • The possibility of an El Nino event in the Southern Hemisphere summer.

Arctic Sea Ice Extent after the summer melt

Chart showing the average amount of sea ice (black line) versus the month of the year. The grey region shows the typical variability and the  green line shows this years data: what will be the value in September 2014? It's impossible to say.

Chart showing the average amount of sea ice (black line) versus the month of the year. The grey region shows the typical variability and the green line shows this years data: what will be the value in September 2014? It’s impossible to say. Click the image for larger version or click here for a live version of the graph

We can confidently predict that the extent of Arctic Sea Ice will reach its minimum extent in late September. We predict this not based upon climate models or weather forecasts but on statistical analysis of previous behaviour.

However, we cannot predict whether it will reach a new record minimum, even though we are sure that the trend will mean that in less than 50 years, the summer minimum of Arctic Sea Ice extent will be zero.

El Nino

The El Nino event  is the largest ‘regular’ climate event on the Earth, strongly affecting rainfall and temperatures across the Pacific Ocean and having detectable affects across the entire planet.

But even though we now monitor the globe hour-by-hour, and day-by-day, we really don’t know even six months in advance if this year will be an ‘El Nino’ year.

Climate models do predict an El Nino-like behaviour – a prediction which emerges from the models rather than being programmed into them. But they predict that it will occur typically every two to seven years – not which year will happen

Currently NOAA estimate the chance of an El Nino Event at the end of this year is just greeter than 50% but really, we just don’t know.

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: