Finding Simplicity in Complexity in Biogeochemical Modelling - Environmental Modelling: Finding Simplicity in Complexity - page 277

It is never true that 'the model is wrong!' - if something
is wrong then it applies to your mental model, your
own understanding!

Ecosystem

Indicator

Many
species

Selection
model

Simplification is necessary in order to sort variables and
gain a basic understanding of the system functionalities.

•

Limit

Interpretation
model

Bioassays
observations

Data collected without purpose is not research; it is
'redundant information with potential for use' until a
problem is found that defines a use for a specified part
of it.

•

Model System

models

Mechanisms
Experiments
Observations

Basic principles and driving forces need to be identified
in order to determine effectively the required level of
detail in the model building.

Input
data

Deposition •

Interpretation
models
Interpretation
models

Sampling
Observations

CL
EX Deposition

models

Data

Make sure that the calibration of the model is robust, in
order to avoid telling the model what to give you (the
input becomes the output).

Collectors

•

Figure 17.10 The 'critical-load model' model is not only the
specific computer code chosen, but in reality all the infor-
mation reinterpretations that actually take place before the
computer code can come into play. The computer codes like
MAGIC or PROFILE are just one of the components of the
'critical load model'.

A transparent and testable model is required for an
effective communication of the model to others. Sim-
pler models are easier to communicate than complex
ones.

•

In terms of making research a learning process, the
authors encourage researchers actively to build their own
model using user-friendly software such as CONSIDEO,
STELLA, POWERSIM, MATLAB and MAPLE.

often forget about mentioning as a model. This can be
repeated for all the other values we use to create our input
data files. Computer codes like MAGIC or PROFILE are
just some of the components of the 'critical load model'.
Uncertainty arises at many places in this scheme, and
the computer codes are not necessarily where most of
the uncertainty is generated. The experience from the
European critical-load mapping programme is that use
and modification of these 'non-models' consumed 75%
of work time, and that running MAGIC or SAFE only
occupied 15% of the time.

References

Alcamo, J., Shaw, R. and Hordijk, L. (1990) The RAINS Model of
Acidification, Science and Strategies in Europe , Kluwer Academic
Publishers, Dordrecht.
Berge, E. and Jakobsen, H.A. (1998) A regional scale multi-layer
model for the calculation of long-term transport and deposition
of air pollution in Europe. Tellus , 50B , 205-23.
Belyazid, S., Bailey S., Sverdrup H. (2010) Past and future effects of
atmospheric deposition on the forest ecosystem at the Hubbard
Brook experimental forest: simulations with the dynamic model
ForSAFE, in Modelling of Pollutants in Complex Environmental
Systems , (ed. G. Hanrahan), International Labmate Limited, St
Albans, volume 2, pp. 357-77.
Botkin, D.B., Janak, J.F. and Wallis, J.R. (1972) Some ecological
consequences of a computer model of forest growth. Journal of
Ecology , 60 , 849-72.
Brakke, D.F., Henriksen, A. and Norton, S.A. (1990) A vari-
able F-factor to explain changes in base cation concentrations
as a function of strong acid deposition. Verhandlungen des
Internationalen Verein Limnologie , 24 , 146-9.
Chen, C.W. (1993) The Response of Plants to Interacting Stresses:
PGSM

17.9 Conclusion

The question asked or the issue investigated defines the
model to be used for any biogeochemical problem or
issue. The chosen model defines the data needed to create
the learning process required to produce the answer. No
modelling starts by assembling 'all' data, and adding more
does not give more clarity but less. The best model is the
model that answers the question asked with the necessary
amount of accuracy with the smallest cost or effort. This
will vary depending on the question asked. Thus:

Version 1.3 Model Documentation , Electric Power

All models should start with problem definition and
explanation of system boundaries.

•

Research Institute, Palo Alto, CA.
Cosby, B.J., Wright, R.F., Hornberger, G.M. and Galloway, J.N.
(1985) Modelling the effects of acidic deposition, assessment
of a lumped-parameter model of soil water and stream water
chemistry. Water Resources Research , 21 , 51-63.

Understanding of a system implies a mental model.
There are no models without a mental model, and the
quality of the mental model is what defines the quality.

•