1)
Probability that U
or more upward trending sites occur in a random selection of size U+D, assuming that direction of trend at
each site is independent, and that upward trending and downward trending sites
are equally likely.
2)
Probability that U+D
or more of the sites would show trends at the given significance level,
assuming that the measurements at each site are independent and approximately
normally distributed.
3)
Probability that D
or more downward trending sites occur in a random selection of size U+D, assuming that direction of trend at
each site is independent, and that upward trending and downward trending sites
are equally likely.
[Gilbert, 1987]. Richard O. Gilbert, Statistical Methods for Environmental Pollution Monitoring. New York: Van Nostrand Reinhold.