Sandy Harrison with the take home messages to remember and understand when using  models

Sandy Harrison with the ‘take home messages’ to remember and understand when using climate models

Paul Valdes (University of Bristol, Bristol)
Sandy Harrison (Macquarie University, Sydney)


Why are palaeoclimate models important?

Palaeoclimate modelling can help us to improve our interpretations of past (palaeoenvironmental) data – such as fossil records, ice core evidence and marine records. This allows us to address important questions such as:

  • Are the changes that we see at the present day, and in the record of past environmental change, consistent with current theories of earth system processes?
  • Are the models realistic in their sensitivity and response to forcing mechanisms (such as orbital changes)?

Through developing models of past environmental change, using palaeoenvironmental datasets, we have a very valuable opportunity to test our understanding of earth system processes. This allows us to model, and therefore better understand, the ways in which environmental changes might play out across different parts of the earth system – at the present, and in the future.


It is important to remember that all models are simplifications of what is a very complex Earth system, comprising multiple components (such as: lithosphere, biosphere, hydrosphere, and cryosphere). Each model differs in: the level of detail (resolution); the components included; and the approximations and assumptions made. It is all about how useful a model is for a particular purpose.


What have been the major developments in modelling?

The first major climate model was developed in 1976, and demonstrated earth surface temperature using CLIMAP data. Initially, models focused only on atmospheric datasets, with a view to develop our understanding of the mechanisms of change. Since the 1970s, major advances have been made in the field of earth system modelling. This has been driven by both technological advances (giving us greater computing power), but also through vast increases in the availability in empirical data (ie: measurements taken in the field), which are used to ‘feed’ the model. Models have become much more comprehensive (to incorporate not only atmospheric data, but also oceanic and terrestrial parameters, for example), and now have a much higher resolution. The major developments in modelling approaches have improved our understanding of:

  • Spatial variations in earth system dynamics
  • Temporal variations in earth system dynamics
  • Sensitivity in the earth system

The key questions now is: ‘can we quantify these variations and sensitivities?’


What are the major challenges?
  • While models can simulate climate conditions (in the past, at the present, and in the future), can they effectively simulate climate change? This is a key area of our understanding of models that needs to be addressed over the coming years.
  • It will be important to establish the suitability of each model for its particular purpose, through considering its ability to simulate climate at appropriate temporal and spatial scales. For example, we cannot produce a long-term global model of air temperature, when it has been developed using only one year’s worth of air temperature measurements from a single location.
  • All models have both positive aspects and limitations which need to be taken into account when we develop our interpretations from the models.
  • Through using palaeoclimate models together with palaeoclimatic data, we can begin to better understand the areas in which our models perform well, so that we can develop more robust projections of future climate change scenarios.