Environmental modelling

[edit] Introduction

Environmental modelling involves the application of multidisciplinary knowledge to explain, explore and predict the Earth’s response to environmental change, both natural and human-induced. See also Building energy models.

Environmental models can be used to study many things, such as:

• Climate.
• Coastal changes.
• Hydro-ecological systems.
• Ocean circulation.
• Surface and groundwater.
• Terrestrial carbon.
• The behaviour of enclosed spaces.
• The behaviour of spaces around buildings.

According to the Environment Protection Agency (EPA) (2009a), a model is defined as:

Models can be used to improve understanding of natural systems and their reactions to changing conditions. They can also help inform decisions and policy.

Models are becoming increasingly sophisticated as computational power increases and our knowledge of processes and behaviours improves, but they will never completely replicate the full complexity of environmental system, and must be based on simplifications of, and assumptions about, environmental processes. Despite these limitations, models can be invaluable tools in helping diagnose what has taken place, to examine the causes of behaviour and to forecast outcomes and future events.

Before beginning modelling, it is important to identify the limitations and boundaries of available models, or models that can be created, how they can be applied and to which systems and situations.

When developing a model the following questions should be considered:

• What are the processes that the model is attempting to reproduce or include?
• What is the time scale for these processes?
• What is the spatial scale of these processes?
• How reliable will the results be?
• How will the results be used?
• Do the benefits of modelling outweigh the cost?
• Is an alternative means of assessment available?

[edit] Types of model

There are a number of different kinds of model, including:

• Empirical: Relying on observed relationships in experimental data.
• Mechanistic: Including the underlying mechanisms and processes between the variables.
• Deterministic: Changes in model outputs are due to changes in model components, meaning that repeated tests under constant conditions will produce consistent results.
• Probabilistic: Utilising the entire range of input data to develop a probability distribution of model output rather than a single point value.
• Dynamic: Predict the way a system may change over time or space.
• Static: Predict the way a system may change as the value of an independent variable changes.

[edit] Model life-cycle

The model life-cycle may include a number of stages.

Identification:

• Determine the correct decision-related questions and establish the modelling objectives.
• Define the purpose of the modelling activity.
• Specify the context of the model application.

Development:

• Develop the conceptual model that reflects the underlying science of the included processes.
• Develop the mathematical representation of that science.

Evaluation:

• Peer review.
• Formal testing to ensure the correct encoding of model expressions.
• Comparison with empirical data to test model outputs.

Application:

• Run the model and analyse the outputs.

[edit] Data quality

It is important that data upon which environmental modelling is based is of a high quality. Data which is of poor quality will not yield model results of a higher quality.

Some of the indicators of data quality include:

• Precision.
• Bias.
• Representativeness.
• Comparability.
• Completeness.
• Sensitivity.

[edit] Model availability

Modelling complex systems has in the past been carried out by experts who have a sound understanding of the processes involved, and a good grasp of the sort of input data that is required and the outputs that are likely to be generated. In effect, the model would, to a certain extent, simply confirm what they already expected. If the model produced unexpected results, they would re-assess the inputs, the model and the outputs to understand why.

As computers have become increasingly powerful, more data has become available, and software developers have begun to give models more user-friendly front ends, and visually attractive outputs, modelling has become accessible to non-experts. Whilst this can be positive, in that it allows greater use of sophisticated analytical tools, it can also be very dangerous, as the inexperienced modeller is more likely to accept model outputs as 'facts' rather than simply a contributing part of a wider analytical process that should be regarded with healthy scepticism.

[edit] Related articles on Designing Buildings

• Building energy models.
• Carbon plan.
• Civil engineers must report climate-change risk.
• Climate change science.
• Computational fluid dynamics.
• Construction environmental management plan.
• Conventions for calculating linear thermal transmittance and temperature factors.
• Dynamic thermal modelling of closed loop geothermal heat pump systems.
• Energy targets for buildings.
• Environment.
• Environmental impact.
• Environmental impact assessment EIA.
• Models in the construction industry
• The design of temporary structures and wind adjacent to tall buildings.
• The thermal behaviour of spaces enclosed by fabric membranes (Thesis).
• Thermal behaviour of architectural fabric structures.