We describe an approach to environmental monitoring that has been deve
loped to deal with future pulp mills in Australia. We propose decision
criteria that balance the chance of missing impacts and the chance of
falsely accusing a proponent of environmental damage. Rather than foc
using on either Type I or Type II statistical errors, we fix the ratio
of the two error rates according to perceived costs of making each er
ror. As monitoring is scaled up or down, risks of both errors rise and
fall proportionately, in contrast to more traditional approaches, in
which one error rate is fixed. We describe the steps necessary to impl
ement a monitoring program using these criteria Our emphasis is on gui
delines that allow the flexibility to deal with monitoring a range of
point source discharges in coastal environments that vary widely. (C)
1997 IAWQ.