THE CLIMATE ADJUSTMENT PROCESS - Wind Resource Assessment: A Practical Guide to Developing a Wind Project - page 164

the nearest upwind turbine is closer than about 20 rotor diameters to the mast, then
the data may not be usable. It is also wise to mistrust published data summaries and
preprocessed data files. Whenever possible, the raw data from the towers should be
obtained so the analyst can carry out his or her own QC.
12.3.2 Surface Weather Stations
The mainstay of MCP remains surface weather stations. In the United States, ASOS
stations are generally preferred for MCP because they are well documented, and, with
the exception of the replacement of cup anemometers by ultrasonic IFW anemometers,
their instrumentation, maintenance, and data-recording protocols have remained con-
sistent since the ASOS installation date. There are approximately 900 ASOS stations
in the United States. Although their geographic distribution is uneven, they provide
reasonably good coverage in most parts of the country. Similar networks exist in most
other countries, although data quality and availability vary.
Data are reported hourly at most stations. Each speed value is not a true average
for the hour but generally represents a 2-min average recorded some minutes before
the top of the hour. Thus, even if a station is very near a wind project site, a good
correlation of hourly wind speeds should not be expected, and daily means should be
employed instead.
Although many surface stations provide a consistent picture of wind speed trends
in their respective regions, numerous issues can create inhomogeneities. Three com-
mon problems are encroaching urbanization, land clearing (e.g., deforestation), and
tree growth. Surface stations located in built-up areas are especially suspect. The ana-
lyst should also be alert for stations that may be poorly or inconsistently maintained.
These are often apparent because of widely varying rates of data recovery or fre-
quent or long gaps in the data record. Regardless of the situation, the homogeneity
of each station's data record must be evaluated case by case, usually by comparing
trends from different stations in the same region. Plotting the ratio of monthly or
annual mean wind speeds for different pairs of reference stations can be a useful tool
for spotting suspicious trends. Where two apparently reliable stations exhibit differ-
ent trends, one upward and the other downward, and no other stations are available
for comparison, it is usually the safer course to reject the downward-trending sta-
tion because most problems with surface stations are likely to cause a decrease in
speed.
12.3.3 Rawinsonde Stations
Data from instrumented weather balloons (Fig. 12-7), known as rawinsondes , can
sometimes be useful for MCP. One advantage of rawinsonde observations is that they
are generally taken well above the land surface (at both fixed and variable heights
defined by atmospheric pressure) and so are largely insulated from changes in land
cover. Another advantage is that they can be taken at or near the height of ridgetop
wind project sites and thus may provide a better correlation with the target tower than
relatively sheltered surface weather stations. The lowest mandatory monitoring levels