Identifying and Evaluating Impacts - Mining and the Environment: From Ore to Metal - page 445

Modelling input data can be incomplete or inaccurate. Another concern about using quan-
titative methods relates to the accuracy of the input data. There is the saying in modelling,
'garbage in, garbage out'. Natural systems have a high degree of natural variability. To
represent nature's response to an imposed impact adequately, it is often necessary to collect
a wide range of data over long time periods to represent natural data averages and their
variation. In most cases, this is impractical and data collection is limited to one year at the
best to capture seasonal variation. Errors are introduced. For example, the year in which
the baseline was measured may have been an unusual year. This is particularly true for
faunal populations which may experience large l uctuations for reasons that are not always
apparent.
Model limitations do exist. For aspects such as noise or emissions to air, the science behind
pollution propagation is well-understood and quantitative impact predictions can be made
with coni dence. However, this holds true for few aspects in nature. Most ecosystems
are complex, and to quantify how they, and elements within them, will react to project
impacts is difi cult and uncertain. Where uncertainty exists, the causes and implications of
the uncertainty need to be clearly stated.
Standards can conl ict. One of the advantages of quantitative approaches is that they
allow comparison with standards, if they exist, providing a transparent evaluation of sig-
nii cance. In the i rst instant, predicted impacts are compared to the relevant standards of
the host region. Sometimes local standards are not established. In other cases, local stand-
ards exist but are not applicable to the specii c type of industry, or discharge. Occasionally,
local standards may be wrong. In all three cases reference to established international
standards becomes helpful in discussions with regulatory authorities.
Model presentation can create the illusion of certainty and objectivity. As is true for all tech-
nical reports, presenting outcomes of the impact assessment study should be concise and
carefully organized. Tables and charts should be prepared when needed to enhance the
presentation and highlight information. Many readers of environmental documents are
visually oriented while others will rely more heavily on the narrative text. Written text
should accompany each table or chart to assist the reader in understanding the table or
graphic. The original source for data for the compilation of charts and tables should be
clearly identii ed. However, the danger is real that word processing and visual tools com-
bined with creative imagination suggest an accuracy that often does not rel ect reality. Care
is necessary in presenting quantitative results. Limitations, assumptions, and uncertainties
should be clearly stated, which, contrary to common belief, should actually strengthen the
credibility of the impact assessment.

'garbage in, garbage out'.

9.12 DEALING WITH UNCERTAINTIES AND RISKS

A number of project aspects introduce a degree of uncertainty related to impact predic-
tion. Environmental Risk Assessment (ERA) is the scientii c method of confronting and
expressing uncertainty in predicting the future. Risk is the chance of some degree of dam-
age in some unit of time, or the probability or frequency of occurrence of an event with
a certain range of adverse consequences. These probabilistic expressions, as opposed to a
single (mean or expected) value, are what distinguish ERA from mere impact assessment.
The irreducible uncertainties leave uncertain the calculated absolute value of a risk, often
by as much as an order of magnitude or more. Risk is therefore expressed and communi-
cated as a mean plus its standard deviation, in addition to any upper bound or worst case
calculation that may be used for purposes of conservative policy. The following summa-
rizes the concept of an environmental risk assessment study.

Risk is therefore expressed and
communicated as a mean plus its
standard deviation, in addition to
any upper bound or worst case
calculation that may be used for
purposes of conservative policy.