Page 200 - Soil and water contamination, 2nd edition

Systems and models 187

into the Guadiamar river, which polluted the river water and approximately 5000 ha of
marshlands including parts of the Doñana national park (see Grimalt et al., 1999). Apart
from the immediate effects, these spill also had long-term effects on the river ecosystem , since
heavy metals are retained in the natural environment as they are readily adsorbed by fine
sediments and deposited on the river bed and floodplains . Despite the extensive clean-up
of the flooded areas of the Doñana national park in the months following the Aznalcóllar
accident, much of the flooded area still shows elevated concentrations of arsenic and other
heavy metals far above tolerable limits (Galán et al., 2002). Long-term environmental models
may help us to understand where the contaminants are stored in the river sediments and how
they are dispersed further downstream in the river system.
It is clear that the pollution of river water and sediments, whether due to accidental
releases as described above or to long-term emissions, have implications far beyond
the increased concentrations of chemicals in water and sediment, since the pollutants
accumulate in the food chain. If we can quantify the spatial and temporal distribution of the
environmental concentrations of pollutants in water and sediment and their transfer rates
into the food web via the various pathways, we may also be able to assess the exposure of
humans and other organisms to pollutants. As you will recall, that was the second purpose
for developing mathematical models.
The third purpose – of predicting future conditions under various scenarios of
environmental change and management strategies – is particularly relevant for supporting
decisions in the policymaking process. Because there are too many uncertainties in the factors
that determine future pollutant emissions and dispersion (for example, future weather and
climate conditions, socio-economic development, and occurrence of hazardous releases) it is
not realistic to forecast future distributions of pollutant concentrations in the environment.
So, instead of forecasting the future, it is usual to use scenarios; these can be defined as a
hypothetical, generally intelligible descriptions of sequences of possible future events. The
scenarios can then serve as input for mathematical models and, accordingly, help answer
‘what if’ questions. Examples of such questions are: What are the effects on nutrient runoff
from large river basins if temperature and precipitation patterns change due to global
warming or if landuse changes due to socio-economic development? What are the effects
on regional groundwater quality if urbanisation intensifies? If the what-if question relates to
possible management actions, such as reducing groundwater abstraction rates or expanding
wastewater treatment facilities, models may also help to evaluate the impact of such actions
on water quality.
The fourth purpose – of reducing soil and water quality monitoring costs – is related to
the high costs of sampling and laboratory analysis. If we are able to predict environmental
concentrations of contaminating substances at unsampled locations or points in time
with sufficient accuracy, we do not need to sample at these locations or points in time.
These samples therefore become redundant in monitoring strategies and may be omitted
without loss of information. In this way, environmental models may aid in the economic
optimisation of monitoring networks of soil, groundwater, and surface water quality.
The fifth purpose, to gain an improved understanding of the controlling mechanisms,
is especially important in a research and development setting. Traditional models of
contaminant transport address biochemical oxygen demand dynamics in river water, and
the development of a pollution plume for a single chemical in groundwater. Recently, much
effort has been put into detailing the biogeochemical speciation and reactions occurring
during transport and into integrating them into water quality models (e.g. Hunter et al.,
1998). On the other hand, chemical transport modelling is often very time-consuming if
the model includes the interactions amongst several environmental compartments (soil,
groundwater, surface water, sediment, and biota) and the area modelled is large. Therefore,
methods are being sought for effectively identifying the driving forces and for upscaling

10/1/2013 6:44:42 PM
Soil and Water.indd 199
Soil and Water.indd 199 10/1/2013 6:44:42 PM