18
ERDC/CRREL TR-02-10
In our analysis of the wind data, we have taken a conservative approach, in
that wind data that are suspect are marked as missing and excluded from all cal-
culations. For example, any period greater than 24 h when wind speeds do not
exceed the minimum measurable wind speed threshold (0.4 m s1) were marked
as missing. While this may seem overly restrictive, close inspection of the data
showed that episodes of minimum wind speed occurrence were bimodal in dura-
tion, either lasting less than 10 h or else persisting for several days. Typically
these longer periods were abruptly terminated when wind speed and/or tempera-
ture increased enough to free the instruments. Wind direction values were treated
in a similar manner, though there were many fewer periods in which the wind
direction met the discard criterion of less than 10 variance in a 15-h period. In a
few cases winds were very light at all sites simultaneously for several hours, and
these data were not excluded. The net result of this quality filtering is most likely
a bias for higher mean winds than in reality, since data excluded during periods
of nonfunctional wind instrumentation are typically characterized by light winds.
A less-frequent problem, partial icing of the anemometer resulting in lower wind
speeds than occurred, is usually not detectable from the data alone and hence
cannot be filtered.
In addition to being more difficult to measure than temperature, a climatol-
ogy of the vector quantity wind is also much more challenging to present and
discuss in a coherent fashion. Winds exhibit influences on several scales from
synoptic pressure gradients to mesoscale influences (e.g., sea breezes and terrain-
induced drainage flows) to very localized conditions (e.g., vegetation state and
snow depth) and in turn can influence other environmental parameters in several
ways (e.g., desiccation of vegetation and saltation, sublimation, and deposition of
snow). Therefore, we present wind data in several ways, with an emphasis on
how it may affect the redistribution of snow, the predominant surface cover for
most of the year.
Table 4 contains monthly mean and record maximum wind speeds for each
month. (The record wind speeds are, of course, the extrema for only the five-year
period considered here.)
Wind speed is summarized in a manner analogous to our approach to tem-
perature (Fig. 6), with 11-day running averages of daily mean wind speed and the
maximum wind speed for each calendar day displayed. The dispersion (also
known as the coefficient of variation, i.e., the standard deviation from the mean
for each day, normalized by the mean) is also shown in Figure 6.
To summarize wind direction, data were binned by octant considering three
threshold speeds: 1.0, 4.5, and 10 m s1 (Fig. 7) and presented as wind roses. The
1.0-m s1 threshold is indicative of all winds experienced at the site. (Winds less
Previous Page
