I have just got a reply from BOM (see below) which I wanted to post asap so others much more knowledgeable than me could digest and consider.
Dear Neil,
The Bureau's use of apparent temperature is described at
http://www.bom.gov.au/info/thermal_stress/index.shtml. From the table displayed there your calculations would appear to be broadly correct. The paper cited (Steadman 1994) is available at
http://www.bom.gov.au/amoj/docs/1994/steadman.pdf.
Having looked at the Steadman paper, I would make the following observations:
- the apparent temperature curves have been fitted over a temperature range of -5 to +45 C (representing the typical Australian range of conditions) - with any curve-fitting exercise one might expect the least reliable fit to be at the ends.
- the apparent temperature is described as being based on someone moving at walking pace (i.e. experiencing an effective wind speed of a few km/h even if the winds are calm). As someone who has a certain amount of experience as a long-distance runner in extreme cold conditions I can certainly testify as to how much difference it can make going into the wind as opposed to with it.
From what I've seen of the Steadman paper I would not recommend the use of that formula for temperatures significantly below 0 C as it is not calibrated through most of that range. The Canadian index (
http://www.ec.gc.ca/meteo-weather/defau ... 5FBF816A-1) is more appropriate for that situation.
Blair
National Climate Centre
Australian Bureau of Meteorology
Login 
Register
Enhancement requests
Search
The BOM formula is the one I'm using. I can now see why the apparent temp is so low when there's no wind - makes sense now. The mention of sub-zero temperatures is interesting. I might switch to the Canadian formula for these conditions. Although we only get sub-zero temperatures for a few weeks a year here (in the warmer SW), the apparent temperature is rather more important when it's very cold. Something else to go on the software wish list!