Just a Little Psychic...

That's what weather forecasters used to believe, that they were just a little psychic. But now they depend more and more on computer predictions. There are intrinsic limits to numerical weather forecasting, first detailed by Edward Lorenz in the ground-breaking paper "Deterministic non-periodic flow," [Journal of the Atmospheric Sciences, vol. 20, pages 130–141 (1963)]. Lorenz discovered that tiny changes in the beginning parameters of weather systems lead to large changes a little later, the so-called "Butterfly Effect". More poetically, the flap of a butterfly's wing in Australia might plausibly lead to a hurricane in New Brunswick.

I recently discovered a new weather product and have been pretty pleased with the results. First, what's wrong with what we have?

As you know, aviation terminal forecasts (TAFs) discuss wind, visibility, ceiling, and precipitation for 24 or 30 hours, depending on the location. They are only good within a short radius of the airport. There is no long-range forecast involved, and there is no mention of temperature or dewpoint. I've always been surprised by the decision to ignore temperature: it would be nice to know if there is going to be a density altitude problem or a need to have the airplane preheated.

Commercial providers like The Weather Channel, Weather Underground, and Accuweather provide long range forecasts and temperatures, but make little or no mention of ceiling or visibility.

What I would like would be a longer-range forecast including long-range ceiling, visibility, wind, and temperature. And I think I've found one: the National Ocean and Atmospheric Administration (NOAA) put out several such products called LAMP. LAMP stands for Localized Aviation MOS Program (trust the government to nest an acronym inside an acronym: MOS stands for Model Output Statistics).

My favorite is the NAM-MOS forecast, available here. Another acronym? NAM is the current North American Mesoscale model; you can read about other models in Wikipedia.

Here is a typical NAM-MOS forecast.



There's a lot of information there, and it's not the easiest thing to read. (Maybe this would be a good iPhone app?) The top says that it's the forecast for Boise on 28 January 2009. The columns on the left are, in order

HR: hour

N/X: maX and miN temperature

TMP: temp

DPT: dewpoint

CLD: cloud cover

WDR: wind direcion (nearest 10 degrees)

WSP: wind speed

P06, P12: precipitation probabilities

Q06, Q12: quantitative precipitation probabilities

T06, T12: thunderstorm probabilities

SNW: snowfall (categorical)

CIG: ceiling (categorical)

VIS: visibility (categorical)

OBV: obstruction to visibility

I highlighted the ones I focus on in red. It's great to be able to forecast that, for example, density altitude might be very high in three days, and therefore decide not to invite a fourth friend along.

The categorical outlooks are listed at the NOAS MOS website, but the short of it is that 5 or more means VFR. (Category 5 ceiling means between 2100 and 3000 feet AGL, category 6 is between 3100 and 6500, etc.)

So now I know that the forecast for Saturday morning in Boise is for a temperature of 26 (no preheat required), light winds out of the southeast, and high (ceiling category 7) overcast clouds with no obstruction to visibility. Twin Fall is forecast to have similar weather, while Pocatello is forecast to have marginal VFR conditions in fog.

You have to take this with a grain of salt, because it is a forecast, and there are plenty of butterflies in Patagonia. But it's much better than what I've had before. Give it a try.