Weather reporting on television

Thanks for your nice note.

(For those of you from out of town, Geoff Fox is a charmingly
energetic TV weather forecaster in Connecticut, who also writes
an interesting weblog http://www.geofffox.com/ )

Here are a few thoughts about TV weather reporting from
someone who doesn’t have a television:

Other than astronomy, weather produces incredibily interesting
images and data. Spoken words are often less interesting and
less intense, except for the 60 words (30 seconds) saying what
has happened and what will happen tomorrow, which could be a
graphic anyway.

Thus intense visuals, nearly independent of the spoken voice,
could be beautiful and informative. It must be possible to
construct graphics that depict today and tomorrow in context of
this year, or what it was like this time last year. The graphics
would stand nearly alone, as the voice describes what will
happen tomorrow.

And then images: perhaps an intense (there’s that word again)
pastiche of short clips, similar to the end of Annie Hall, could be
constructed (with some explanatory typography) of the day’s
weather events in the state, or nation, or world. Television is
keyed to low bandwidth of the voice (it is that low bandwidth of
the voice that leads to the chronic impatience of television).
Oddly, television is often thin visually and much like radio, except
the viewer sees someone talking. The idea is that television
should take advantage of its inherent visuality. So perhaps the
personality of the weather person could be heard rather than
seen, as beautiful visuals show up on the screen.

Thus weather reporting should have the visual richness and intensity of commercials, which do exploit television’s inherent visuality and usually avoid talking heads.

A few years ago I was stranded in a hotel in Washington DC
during a snowstorm and tried to find out what was happening by
looking at the Weather Channel. It was hard to imagine how the
content per minute could be thinner and so unvisual for such an
interesting topic. I longed for the richness of the internet, which
does not march along to slow linearity of the voice, and which
could provide what I needed to know along with wonderful
visuals.

Finally, two underlying ideas might help the audience
understand the inherent difficulties in forecasting the weather:

(1) If you watch a storm on radar moving across the state, it is
clear that dramatic weather events are often patchy; that is, a lot
of rain one place but nothing 10 miles away. So the forecast will
be right in one place but not in an adjacent area.

(2) A pretty good forecast on average is that tomorrow will be like
today; or that 5 days from now will be like that day a year ago.
That is, naive persistence-forecasting is not all that bad.

I can’t speak for weather forecasting methods or presentation formats in other countries but I can say that the varied outputs of the BBC system for UK viewers are usually first class.

When we stop to consider:

1. the astounding complexity of the natural processes being measured, described and interpreted;

2. the range of frequently opposing needs and different levels of understanding of weather viewers; and

3. the sheer time pressures on weather professionals (between and during live transmissions)

then I think we should welcome cautiously whatever newly emerging methods are being tried and tested if these might have mutual benefits for presenters and viewers alike. As I’ve said in another context (the evaluation of Powerpoint), the effectiveness of different methods of presentation could – and should – be evaluated using the ‘controlled trial’ format. Thus, Group A (exposed to 3d graphics) are compared with Group B (‘standard’ graphics) and results of their test scores (for understanding/recall of the message) are compared.

Notwithstanding the very legitimate concerns of scaling, imaging and ‘flying’ through models, I wonder if we might reserve judgement until we see what the end product looks like in action?

I’m old enough to remember the broadcasts of the 1960s with ‘stick on’ magnetic symbols that must have been seen as avant garde for the times. It’s a great tribute to the resourcefulness of the BBC that today we have, on the one hand, satellite and rainfall radar plots capable of broadcast within 30 minutes or so of capture and yet, on the other hand, the now institutionalised and strangely soporific ‘shipping forecast’ for radio listeners only.

I work in the field of Public Health where timeliness of data usually means ‘a couple of years ago’ and where many methods of information presentation – in any medium – are often grossly deficient.

In practice, I have learned more about optimising presentation techniques for combining complex information with an acceptable format by observing television weather forecasters than any other suitable ‘model’ I can think of.

With careful observation and a fair wind, we can still learn a lot from the weatherman/woman.