4.1 Identifying change
In order to compare two sets of probabilities with each other and decide whether they are different, we need to measure or predict differences between distributions in every area of the science and engineering of climate change. So we need to:
- measure past climate change
- understand the causes of past climate change
- predict future climate change with different levels of greenhouse gases or geoengineering
- test whether climate predictions are similar to the real world.
For all of the above, we must decide if two distributions are ‘different enough’ to call this a change.
If the chance of your house being flooded each year increases from 1% to 2%, does this mean you are guaranteed to be flooded every 50 years? And should you look for a new home?
No, and it depends!
This is difficult because the change might be very subtle or slow, so we must make decisions to help us decide, such as:
- which statistical technique should we use to compare them?
- how large a difference ‘counts’ as a climate change?
- how many years of data do we need to feel confident in our conclusion?
These questions do not always have easy answers – the decisions may be somewhat arbitrary or subjective – so they are often contested by climate sceptics.