The University of Manchester · FACULTY OF ENVIRONMENTAL SCIENCE

EART60061 Chap.6 The Modelling Process: Abstraction to Validation

- one subject, every graph, every model, every mark
7 Chapters4-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 6 of 10 · EART60061

The Modelling Process: Abstraction to Validation

The Modelling Process supplies the vocabulary used for every case study in the prediction weeks. A model simplifies reality by leaving things out, needed in environmental science because systems are large, variable and rarely open to experiment, so fitted lines fail through insufficient data or poor understanding.

The process runs from the real world, through abstractions and assumptions, to a conceptual model; formulation turns it into a mathematical model; solution is analytical or numerical, and numerical solutions need verification; output is then calibrated and validated against observations, and failure sends the modeller back to the assumptions.

The chapter also separates deterministic from probabilistic models and contrasts what an engineer and a researcher expect of a model. It closes with the habit the lecturer stresses most: recognising where abstractions and assumptions were made in models written by other people, because those choices shape every output you will use in later units and in employment.

In the exercises that follow, each of these terms is attached to a concrete action, so the vocabulary becomes a checklist rather than a list of definitions.

In this chapter

What this chapter covers

  • 01

    Why environmental prediction needs models

  • 02

    Abstractions and assumptions

  • 03

    Conceptual and mathematical models

  • 04

    Analytical and numerical solutions

  • 05

    Verification, calibration and validation

  • 06

    Deterministic and probabilistic models

  • 07

    Engineering and research expectations of a model

Worked example · free

Label the stages of a stream temperature model

Q [6 marks]. A student models summer stream temperature. Label each action: treating the reach as one mixed volume; assuming heat gain proportional to sunshine hours; writing the heat balance as an equation; stepping it forward hourly in Excel; repeating with half-hour steps; tuning a shading factor to match logger data. Marks used here come from AskSia for self-assessment, not from the official assessment.
The generalised modelling process runs clockwise from the real world through a conceptual model and a mathematical model to an output; the dashed return arrow iThe generalised modelling process runs clockwise from the real world through a conceptual model and a mathematical model to an output; the dashed return arrow is calibration and validation against observations, and verification applies only when the solution is numerical.Real worldConceptual modelMathematical modelOutputabstractions and assumptionsformulationsolutionverification if numericalcalibrationvalidation
The generalised modelling process runs clockwise from the real world through a conceptual model and a mathematical model to an output; the dashed return arrow is calibration and validation against observations, and verification applies only when the solution is numerical.
  • 1Treating the reach as one mixed volume is an abstraction.
  • 1Assuming heat gain is proportional to sunshine hours is an assumption.
  • 1Writing the heat balance as an equation is formulation of a mathematical model.
  • 1Stepping the equation forward hourly is a numerical solution.
  • 1Repeating with half-hour steps to see whether the answer changes is verification.
  • 1Tuning the shading factor to match logger data is calibration; testing on a new period would be validation.
The actions follow the generalised process: abstraction and assumption, formulation, numerical solution, verification and calibration, with validation as the next step on independent data.
Sia tip — Ask whether a step changes the equations, checks the arithmetic of the solution, or compares output with the real world; each points to a different term.
Glossary

Key terms

Abstraction
A deliberate simplification of the real world, such as dropping a dimension or merging zones, made so a model can be formulated.
Calibration
Adjusting uncertain parameter values within plausible limits to improve the match between model output and observations.
Probabilistic model
A model run many times with parameters drawn from ranges, whose result is the spread of possible outcomes.
Formulation
Turning a conceptual model into mathematical equations that can be solved and tested against data.
Deterministic model
A model with fixed parameters that seeks a single correct answer, even if it is run many times.
FAQ

The Modelling Process: Abstraction to Validation FAQ

What is the difference between verification and validation?

Verification checks how accurately a numerical solution approximates the exact solution of the equations, for example by halving the time step. Validation checks whether the model output matches observations of the real world, which tests whether the model's assumptions hold.

Why not make the conceptual model as detailed as possible?

A very complicated conceptual model may be impossible to formulate or solve, and a model that cannot be solved cannot be tested. The advice is to start simple, find through validation whether it is too simple, and add complexity only when needed.

When is a probabilistic model worth using?

When many future scenarios are possible or parameters such as diffusion coefficients are known only as ranges. Running the model many times with different values gives a range of outcomes and their likelihood, which is more useful than a single answer from one causal chain.

Why can a model that fails still be useful?

For an engineer a model that cannot predict may have no value, but for a researcher its failure tests the assumptions. It can reveal which processes dominate, whether constants are accurate and which parameters the system is sensitive to, and it shows where the conceptual model needs more complexity.

Where does verification fit in the modelling cycle?

Verification belongs to the solution step and applies only when the solution is numerical, because cutting time or distance into steps makes the answer approximate. Analytical solutions are exact, so they need no verification, although their output still needs calibration and validation against observations.

Study strategy

Exam move

Draw the modelling cycle from memory, then annotate it with one action from each case study: the rating curve, rainforest regrowth, the soil carbon compartments and the lake spill. For each, write the main abstraction, the solution type and the check performed. Practise distinguishing calibration from verification with short scenarios until the choice is immediate.

Test yourself with a map: list what it abstracts, such as scale and a dimension, and then do the same for the soil carbon model.

A+Everything unlocked
Unlocks this Bible + all 41 of your subjects at The University of Manchester, plus 1,000+ more Bibles across the AskSia library.
Sia - your EART60061 tutor, unlimited, worked the way the exam marks it
The full 4-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works