 |
For Environment Canada, Crystal Ball
is a Crucial Tool for Chemical Risk Assessment
|
APPLICATION: Environmental risk assessment of chemicals
SUMMARY: Environment Canada applied Crystal Ball in an investigation
of the effects of ammonia in aquatic environments. Crystal Ball
was used to generate distributions for the raw hydrological and
effluent data and to combine these distributions to generate a set
of hypothetical river conditions for a plume model.
RESULTS: With the aid of Crystal Ball, Environment Canada
has determined that ammonia released in large quantities under some
river conditions is toxic to aquatic life. A risk management process
will be developed to discuss reduction strategies with Canadian
municipalities.
Miles Constable, a Senior Toxic Substances Officer for Environment
Canada in Edmonton, coordinates teams of professionals conducting
ecological risk assessments of chemicals on Environment Canada's
Priority Substances List. The list is developed jointly by Environment
Canada and Health Canada to address chemicals suspected of being
either inherently hazardous or released in such large quantities
as to pose a hazard to the environment. This assessment process
determines the entry, exposure, effects and risks of chemicals to
Canada's environment.
If a chemical released into the environment is determined to be
toxic, Environment Canada negotiates with the industrial or municipal
sector releasing the chemical to reduce quantities released. This
can have a major economic impact on that sector, as reduction strategies
can vary from simple technological improvements in facilities to
"virtual elimination" of the chemical from the Canadian environment,
in essence a ban on the release of the chemical.
With such serious consequences, Environment Canada must conduct
a balanced assessment of the hazards that a chemical poses to various
Canadian ecosystems. This balanced assessment includes an open and
transparent accounting of the uncertainties that exist regarding
the effects of the chemical to biota. In 1997, Environment Canada
established a methodology or "tool box" to help conduct probabilistic
risk assessments. When Miles Constable was tasked with putting together
a team to complete a risk assessment of ammonia in aquatic environments,
his first decision was to hire an environmental statistics consultant
to help develop the assessment. The consultant recommended Crystal
Ball as a powerful statistical tool that could propagate sources
of variability and incertitude through environmental exposure models.
Ammonia is a common component of sewage effluents that are typically
released into waterways. Only three case studies were available
for use: two had very good field monitoring data and one had a good
ammonia profile. Miles Constable decided that it would be necessary
to model the sewage dispersion profile for the third case study
as a typical Canadian river because of the lack of in-stream monitoring
data. Crystal Ball formed a key component of this modeling effort.
The modeling case study had a good profile of ammonia concentrations
through a sewage effluent plume for a single sampling. Miles Constable,
his group, and the consultant developed a set of spreadsheets containing
hydrological data (depth, width, flow rate, temperature and pH),
and sewage effluent data (effluent flow rates, ammonia concentrations,
and temperature) from government monitoring records and the sewage
treatment plant. They wanted to estimate the ammonia concentrations
in the plume under a large number of hypothetical flow conditions
that could be expected in the river.
Crystal Ball was used to determine the statistical distributions
inherent in the hydrological and sewage effluent data. Once the
distributions were established for the most important variables,
Crystal Ball was used to generate 500 data sets of hypothetical
flow and effluent conditions based on the distributions. They then
entered the 500 data sets into a plume-modeling program (CORMIX
3) to estimate ammonia concentrations downstream of the outfall
for each of the river flow conditions.
Crystal Ball proved very useful for generating distributions for
the raw hydrological and effluent data and combining these distributions
to generate a set of hypothetical river conditions. The group developed
a series of ammonia distribution predictions for each month of a
hypothetical average year and were able to use this information
in predicting the toxic risk to aquatic biota exposed to the sewage
effluent plume. "While it was tedious work to put together the data
set," said Constable, "it was also intriguing to see if the approach
would work and if we had enough data to properly generate an average
monthly flow and effluent regime for the modeled location."
The Environment Canada team then used Crystal Ball to develop larger
data sets for their final risk analyses and to generate the final
predictions of risk. First, they used Crystal Ball to generate Cumulative
Density Functions (CDF) of the probability of being in a specific
range of concentration. They also generated a relationship between
toxicity and ammonia concentration for a number of different species.
Next, they combined these two CDFs to provide a relationship of
probability of a toxic impact versus the percentage of species affected.
This was the crux of the project, to determine the risk of a toxic
impact to an aquatic community from the release of sewage effluents.
For this final analysis they chose to use sets of 300 ranges for
the CDFs. They then ran Crystal Ball with the number of trials set
to 10,000 to develop a large enough data set to ensure a good distribution
of data across the ranges. Crystal Ball was used to predict the
probability of an ammonia concentration being within a given range
and the percentage of species that would be affected.
The result was a set of probabilities of a toxic impact at different
distances or locations from the sewage outfalls. This enabled Environment
Canada to show that ammonia, under some conditions, posed an unacceptable
risk to aquatic biota, and was thus toxic. Following the official
release of the assessment report, a risk management process will
initiate discussions with Canadian municipalities on risk reduction
strategies.
Miles Constable cites the intuitive user interface, the distribution
selection gallery, the speed of simulations, and the linkage to
Microsoft Excel as his favorite Crystal Ball features. "With modern
PCs, we could run multiple tests in a day, " he said. "We would
not have been able to conduct our probabilistic risk assessment
without the use of Crystal Ball. It was really the crucial tool
for the job. For those not current with statistics or with Crystal
Ball I would strongly recommend a training course on the program.
It may surprise you with its capabilities."
Visit our Environmental
industry page!
|
