Can we trust climate predictions?

Article by Hazell Ransome.

Scientists are constantly working to check [1] and improve [2] how accurately their models make predictions.

Let’s discuss two ways they can test climate models.

Firstly, scientists can use historic data to check if their model can simulate past climates [3]. If their models can do this then they are more likely to correctly simulate future climates [4].

However, there are challenges by using data to test models. For example, if the data that isn’t collected directly (scientists weren’t using thermometers to measure air temperatures 500 years ago!) there is some uncertainty [5].

Furthermore, sometimes there is not enough data about certain aspects of the climate system. Therefore predictions and simulations of those aspects, such as rain, deep ocean temperatures and Antarctic sea ice [6], are less reliable[3,6].

Therefore it is often better to see whether a model is good by comparing it to other, different climate models, rather than past data [7].

Comparing model predictions allows scientists to see if their model has roughly the correct answer [8,9], and by using the average value of multiple models scientists reduce the uncertainty in predictions [10].


Join our Newsletter!


Climate Science Ltd
Company Nr: 12370672
Registered in England & Wales
Mail: [email protected]


15 Hope Close
United Kingdom

Climate Science is registered as a non-profit company limited by guarantee in England and Wales.

Copyright © 2019-2020 Climate Science Ltd. All rights reserved.

Climate Science uses Cookies to ensure you get the best experience on our website.