The popular science documentary program Horizon on the BBC recently asked this.

The episode was presented by comedian Ben Miller and to immediately digress from the actual point of the post, Ben Miller starred in a comedy sketch that appeared on many climate skeptic blogs on the apparent assumption that it was an anti-man made global warming sketch and somehow supported the denier position. I can only assume that such people not only lack critical reasoning but satire is also outside their area of understanding.

Miller may be a comedian but was a physicist studying at Oxford and fully accepts AGW. The sketch was more about the ‘nanny state’ in the UK so perhaps Americans can be forgiven somewhat for not always understanding this and British humour, but there is no excuse for the many references to it from the UK. One can only assume that such people will clutch anything they think lends any support to their beliefs when the science, so clearly, does not.

Anyway…

The premise of the Horizon documentary was that Miller got into a conversation with a skeptic and was put in a position where he had to explain why a one degree rise in temperature was significant when temperature changes many times this occur every day. He found he was unable to give a satisfactory answer.

This resulted in him going on a quest to find out exactly what temperature was, how to accurately measure it and find an answer to the skeptics question. His quest involved going back to his research lab at Oxford, building a thermometer and fitting a Met Office temperature gauge at his own home.

Near the end of the program a scientist shows him how significant a single degree in warming is be by using a simple bell cure / normal distribution graph. I thought that it was such a simple way to explain things that I would repeat it here and expand on it.

Figure 1. Normal distribution graph

Figure 1 shows a standard normal distribution graph similar to the one drawn in the program. This type of graph is found all through the natural world when things are compared. It can represent height, with the number of people of average height in the middle and the shorter and taller people to each side; it can represent the number blooms of a type of flower at any given time, the time that insects hatch, IQ etc. And of course it can represent climate related data like temperature and rain fall.

Figure 2. Perception of 'Normal'

Figure 2 is coloured to show what we might perceive as 'normal' in this distribution. The values at the bottom are not as important as the shape and area which represents totals in each group. For example if the graph was of annual temperatures experienced, the green area would be considered within the normal range of temperatures with cold winters to the left (-1) and warm summers to the right (1). The green area represents 68% of the total - 34% to each side of the centre line.

The yellow areas would still be considered normal but would represent the less common cold snowy winters to the left or the heat wave summers to the right. The ‘not since I was a lad’ stuff and probably the winters we have seen in the northern hemisphere of late. Just to reiterate the yellow areas would not represent abnormality, just less common events considered within natural variability. If the graph represented not temperature but a countries rain fall then these areas would be the intermittent droughts and floods also all within natural variability, with the more extreme the closer to each end. The yellow areas represent 28% of the total – each patch being 14%. The total represented by both green and yellow covers 95% of all data, so will represent 95% of all normally recorded temperatures or rain fall depending on the data used to construct the graph.

Only the red areas, the final 5%, represent extreme events. These would be cold or heat, droughts or floods not seen in living memory. These and anything even further to the left or right are the true extreme records in the data with just 2.5% represented on each side. They can be thought of as events that occur only once or twice within a whole data set.

Figure 3. A climate shift will move the average of the recorded data

What happens if the climate shifts as it is predicted to? If figure 3 represents temperature then an increase of temperature of one degree would move the normal distribution from the blue curve along the line toward the position of the red curve. Note that this is for visual demonstration purposes and it is very unlikely that a single degree would move the distribution so far. However we can imagine that with the increase in global temperatures already recorded from the pre-industrial we are already two thirds on our way to that one degree. It can be immediately seen from the graph that there is a new average. That extremes not previously recorded have been added to the data set.

Figure 4. New average compared to old

If we use the same colouring criteria for ‘normal’ on this shifted graph we get figure 4. It should be noted that most of the new graph is still mainly green or yellow – which would be considered within the normal range. Most people living in that shifted world will see little difference most of the time. But it is clear from the red that one type of extreme is much more common. Things that are once in a century event have become more like once a generation. But it is also important to note that there is still red on the opposite side of the graph. So even though one extreme event is much more common, droughts or temperature extremes, the opposite is still a possibility. This is why cold weather does not disprove global warming because the number of cold related events does not cancel out the more numerous warm related events.

The overall point of this visual representation is not the values used but to show just how important one degree is – it may go unnoticed most of the time by most people but it has a very disproportionate effect when compared to the ‘normal’ that the environment is used to experiencing. The increase in extreme events is costly not just in monetary terms but also personal loss and environmental damage. Data is now starting to support that such a shift is occurring