Many important decisions are made using a “calibrated environmental model”; this may be a groundwater model,
surface water model, or a model which integrates both systems. Expensive contracts are signed for the production, review and use of such models. Implied in many of these contracts, and in the proposals that win
these contracts, is that good models are complex models, that better models are even more complex models, and that a model’s ability to replicate historical system behaviour ensures its ability to predict
future system behaviour. Is this indeed the right recipe for model-based decision-making? If not, then how should models be best developed and deployed to inform decisions and resolve conflicts?
This workshop has two purposes. First it explains and demonstrates the capabilities of the PEST suite of software in model
calibration and in calibration-constrained uncertainty analysis. In a basic and easily-understood way, it introduces the theory behind PEST and demonstrates application of this theory, with numerous examples
from the real world.
The discussion is then broadened to explore issues that calibration and uncertainty analysis expose. In particular, it
investigates how an environmental model can provide receptacles for the two types of information on which decisions pertaining to environmental management are necessarily based, namely expert knowledge and the
historical behaviour of that system. The issues of appropriate complexity and the need for “good calibration” are discussed. Particular attention is focussed on the fact that even the most complex
model is a grossly imperfect replicator of real-world environmental system behaviour. Conclusions are drawn which question some long-established beliefs of what constitutes “a good model”.
Suggestions are made of what should be asked of an environmental model when used to support the making of important management decisions.
Who Should Attend?
This course is intended for both modellers and managers of groundwater and/or surface water systems. Modellers will gain
insight into the theory and capabilities of the PEST software suite as demonstrated through examples specific to California. Managers will understand more about what can be expected of models, and learn to
discriminate fact from fantasy when presented with claims of a model’s ability to predict the future.
Registration Information
If you register for the 2015 Annual Meeting, the PEST
Workshop is included in the registration fee. If you only want to attend the PEST Workshop, the registration fee is $250. The PEST registration fee includes CWEMF membership for 2015. You can register for the
Annual Meeting HERE or the stand-alone PEST Workshop at the PayPal
link below:
|