THE CLIMATE ADJUSTMENT PROCESS - Wind Resource Assessment: A Practical Guide to Developing a Wind Project - page 171

standard, it is hard to beat a bulk linear regression. It is easy to apply, and its estimates
of the long-term mean wind speed are about as accurate as any linear method. It is
not suitable for predicting the speed distribution; for that, scaling the observed wind
speeds to the predicted mean is recommended so long as there is at least 9 months of
valid on-site data. More complicated methods are best left to analysts with time and
experience, who can fully understand the potential benefits and pitfalls of applying
them to the available data sample.

12.5 QUESTIONS FOR DISCUSSION

1. Suppose you observe a significant declining trend in a reference station's data
record over a 15-year period, and you are confident of the consistency and
quality of the measurements. What are some possible causes of this trend? Given
each cause you identify, what are the implications of using this station for MCP?
2. Discuss the potential risks of using a site data record of less than 1 year in
MCP.
3. You have a 5-year reference mast at your project site and two masts with 1 year
of data at each. When correlating the short-term masts to the reference mast, one
yields r 2

8, while the other yields r 2

5. Discuss the potential causes of
this difference. If the problem cannot be remedied, what are the uncertainties in
the projected 5-year mean for each mast? Assume an interannual variability

= 0 . = 0 .

σ

of 4%.
4. What are the relative advantages and disadvantages of surface and upper-air
data for MCP? Are there certain project conditions that would cause you to
favor one over the other? Are there certain cases where modeled data may be
preferable to any source of measured data?
5. Besides distance, what are some other factors that could influence the strength
of the wind speed relationship between a target site and a long-term reference
station?
6. The majority of surface observation stations measure the wind speed at 10 m
height, while wind energy assessments are primarily conducted at much greater
heights. Discuss the potential effects of this on an MCP analysis and how those
effects can be mitigated.
7. You have 1 year of data from your site, for which the annualized mean wind
speed is 6.51 m/s. Your MCP analysis (using linear regression) yields the long-
term (LT) mean speeds shown in the table at the top of the following page for
10 different prospective reference stations. Which stations would you include
in your analysis, and why? Can you think of two methods of arriving at your
final long-term mean speed estimate?
8. You are considering a multiple linear regression with two reference stations.
Station A's record starts in 1995 and station B's in 1980. Do you think it is
reasonable to go back to 1980? Why or why not? Describe the process you
might use to decide which years to use as the start of both station A's and