Comments on “Inconsistencies in the ‘New’ Windchill Chart at Low Wind Speeds”

Abstract

Shitzer and de Dear (2006) used the regression equation that accompanies our 2001 windchill model (Bluestein and Osczevski 2002) to calculate windchill equivalent temperatures (WCET) at low wind speeds. They were surprised to find that WCET was not equal to the air temperature when the air was “still,” according to their definition of calm. They also found an abrupt transition in WCET above a minimum wind speed. To circumvent these inconsistencies, they proposed a correction; however, these inconsistencies do not exist. The problem is not in our model of windchill, but in their understanding of it. Their definition of still air as “any wind speed that is equal to, or below, the assumed threshold of 1.34 m s ” (Shitzer and de Dear 2006, p. 788) is the only real inconsistency. That is not how we defined it in the 2001 model of windchill, where “still air” or “calm” means no wind at all. There is no minimum applicable wind speed for calculating WCET; it can be found for any wind, right down to dead calm (Fig. 1). Because they used a wind speed of 1.34 m s 1 (4.8 km h ) in our regression equation, the WCET they thought they were calculating for someone walking in still air was actually for a person walking into a light wind. It is therefore neither inconsistent nor surprising that they calculated a value that was colder than the air temperature. The other “inconsistency” they found, the abrupt transition in WCET values, also results from their misinterpretation of what we meant by still air. WCET actually varies smoothly with wind speed (Fig. 1). Even if Shitzer and de Dear (2006) had used the regression equation properly, it would not have worked well at very low wind speeds. The regression equation was only intended for operational use. As Fig. 1 shows, it is a good approximation to the model output at wind speeds above 3 km h 1 (0.8 m s ). Lower wind speeds are not easily measured and are never reported anyway. We regret that the regression equations were published without an explicit warning of this limitation in our early paper. We did point it out in a later, more detailed paper (Osczevski and Bluestein 2005). The full model must be used to calculate WCET in this range. At wind speeds close to zero, WCET may be slightly warmer than the air temperature. This is due to the greater radiant heat transfer in the reference still-air condition, which assumes that there is no water vapor in the air. Any factor that increases the severity of the reference condition, such as greater radiant heat transfer, increases WCET. WCET is equal to air temperature at zero wind speed if the same relative humidity, that is, 50%, is assumed for the still-air condition as for other combinations of wind and temperature. Especially when there is no wind, radiant heat loss is significant: about one-half as large as the convective heat loss caused by walking in cold, still air. The supposed problems reported by Shitzer and de Dear (2006) do not exist and do not demonstrate any need to correct the windchill calculation method or the existing charts. We could make the model more elegant by increasing the relative humidity in the still-air condition to 50%, making WCET exactly equal to air temperature at zero wind speed. However, this adjustment would have no little or no operational impact. Corresponding author address: Randall Osczevski, DRDC Toronto, P.O. Box 2000, 1133 Sheppard Ave. W., Toronto, ON, Canada M3M 3B9. E-mail: randall.osczevski@drdc-rddc.gc.ca OCTOBER 2008 N O T E S A N D C O R R E S P O N D E N C E 2737