Environmental Engineering Reference
In-Depth Information
Finally, when such systems are proprietary systems of either research groups or
projects, it may not be possible for third parties to re-use the system for further
development.
A basic requirement of any biophysical model is that it must be able to simulate
the processes which influence significantly the behaviour of a system, and particularly
those aspects that relate to the purpose of the model. An obvious example is that
the model must not be restricted to potential production if its intended use is to
study water-limited agriculture This example, however, is at a “high” level, meaning
that simulating water-limited production may require:
•
The simulation of a different number of approaches; and
•
Even different approaches for the simulation of the same process, when environ-
mental conditions change
As an example of the first point, studying the impact of mulching requires soil
evaporation models which react to soil cover beyond that given by canopy cover,
and the fate of the mulching material must also be simulated. An example of the
second point can be simulating the water budget of conditions typified by peak
evapotranspiration of 5 mm day
−1
on a deep soil compared with conditions where
the peak evapotranspiration is 12 mm day
−1
on a shallow, cracking soil. The former
case can be simulated with simpler, yet still adequate, approaches compared to the
latter. Moreover, some approaches may demand inputs which may not be easily
available, thus compromising its operational use. Also, as peer reviewed publications
may propose alternative options for modelling processes with the same assumptions;
tests need to be carried out to assess performance and reliability throughout the
range of operational conditions. Finally, effective simulation of a biophysical
system, no matter what level of simplification is chosen to simulate its behaviour,
requires expertise in different domains. This is a demanding task that requires a
multi-team effort for system analysis and model development. All these reasons,
argue for a flexible and modular simulation system, and provide, in effect, a
specification for the simulation system described in this chapter.
The advent of component-based software engineering has enabled the development
of scalable, robust, large-scale applications in a variety of domains, including
agro-ecological modelling. In systems analysis, it is common to deal with the
complexity of an entire system by considering it to consist of linked sub-systems.
This leads naturally to thinking of models as being made of sub-models. Such a
conceptual model can be implemented as a computer model composed of connected
component models. This type of implementation has at least two major advantages.
First, new models can be constructed by connecting existing component models
of known and guaranteed quality to new component models. This has the potential
to increase the speed of development. Secondly, the predictive capabilities of
two different component models can be compared, as opposed to only comparing
whole simulation systems. Further, common and frequently used functionalities,
such as numerical integration, visualization and statistical ex-post analysis, can
be implemented as generic tools which are developed once and shared by all the
model developers.
