INTEGRATING TOOLS FOR SURFACE WATER QUALITY MANAGEMENT - Integrating GIS, Remote Sensing, and Mathematical Modelling for Surface Water Quality Management - page 73

The main structure of all these water quality DSS's depends on the structured databases and
computational frameworks. These computational frameworks comprise water quality modeling tools
of different properties and approaches, which depend on many factors such as the addressed water
quality modeling problems, scale of application, parameters of concern and management objectives.
Data is the key issue in developing a reliable DSS that could be used in solving water quality problems
and managing water resource. A well structured and designed DSS can still give misinterpretation and
wrong information for decision makers if the data used is not with accepted quality. Despite big
progress made in recent years with respect to collection and storage of data, including usage of remote
sensing technology, availability of reliable, credible and consistent data has been and will remain a
problem for next years. Collection and storing of data requires not only technical and technological
infrastructure, but also high investment in measurement networks and in processes of data validation
and verification. Therefore without significant efforts and investments, which have to be carried out by
governments and international agencies, it is hard to believe that any significant process will be
achieved in developing reliable DSS for aiding the decision making process of water quality.
3.5.7. Uncertainties in the applied DSS tools
Modeling and decision support tools (e.g., integrated assessment models, optimization algorithms, and
multicriteria decision analysis tools) are being used increasingly for comparative analysis and
uncertainty assessment of environmental management alternatives. If such tools are to provide
effective decision support, the uncertainties associated with all aspects of the decision-making process
need to be explicitly considered. However, as models become more complex to better represent
integrated environmental, social and economic systems, achieving this goal becomes more difficult.
The new high expectations for the aquatic environment, incorporated into the current wave of
regulations, is prompting additional complexity with regard to modeling spatial variability, micro-
pollutants and ecological indicators (Somlyody et al. 1998; Thomann 1998). Facilitated by improved
computational resources, there is a trend for spatial discretisation to be increased, multi-media models
to be developed (e.g. Havnø et al. 1995), and for traditional water quality determinands to be broken
down into constituent species (Chapra 1999). As a consequence, the typical number of modelled
components has risen exponentially over the past years, and this growth is expected to continue
(Thomann 1998). Despite the increasing expectations placed upon water quality models, contemporary
deterministic models, when audited, frequently fail to predict the most local and basic biological
indicators with a reasonable degree of precision (e.g. Jorgensen et al. 1986). Even when models are
claimed to be 'reliable' following audits, a very significant margin of error is allowed (e.g. Hartigan et
al. 1983).
Uncertainty analysis for model simulation is of growing importance in the field of water quality
management. The importance of this concern is provided by the recent awareness over health risks
from improper disposal of toxic wastes as well as continuing emphasis of risk assessment (Radwan et
al., 2004). Mechanistic modeling of physical systems is often complicated by the presence of
uncertainties, which could be classified into different categories according to the study limitations.
Uncertainties in water quality modeling can be classified as Natural, Model, and Data uncertainties.
Model input, model structure, model parameters are examples of uncertainties that could be associated
with water quality modeling. As presented by Radwan (2002), different approaches for representation
of uncertainty are applied; classical set theory, probability theory, fuzzy set theory and rough set
theory. The most widely used uncertainty representation approach in transport - transformation
modeling is the probabilistic modeling approach.
McIntyre et al., (2003), stated that while additional model complexity might be expected to improve
the precision of model results, this has proven to be unfounded in a variety of studies (e.g. Gardner et
al. 1980; Van der Perk 1997; also see Young et al. 1996). Furthermore, future driving forces such as