Scientists update earthquake probabilities for Canterbury - 4/11/2011

Updated figures issued today are based on a combination of two different statistical forecasting models. Until recently, the earthquake probabilities have been dominated by what is known as a short-term clustering model. However, in October the medium-term clustering model started to exert a bigger influence on the earthquake probabilities.

"Because the medium-term model estimates slightly higher probabilities than the short-term model used up until now, this has caused a slight increase in the combined figure," said Dr Matt Gerstenberger, a senior natural hazard scientist at GNS Science.

Updated aftershock probabilities released today, which are slightly increased on September's figures, are the result of the change in mathematical modelling rather than an increased risk of earthquakes.

"Nothing has changed inside the earth to increase the risk of an earthquake - it's just that there is a change in the way the probabilities are calculated," said Dr Kelvin Berryman, Manager of the Natural Hazards Research Platform.

An important aspect of this modelling is that these figures are for the entire aftershock zone, and not just for Christchurch city. The zone extends from Hororata in the west to large parts of Banks Peninsula, and from Kaiapoi in the north to Lincoln in the south.

The combined earthquake forecast model for the region will continue to be used from now on.

The updated figures based on the combined model show for the 12 months to October 15 next year there is:

• an 88% probability of a magnitude 5.0 to 5.4 aftershock - up from 78% in September
• a 46% probability of a magnitude 5.5 to 5.9 quake - up from 37% in September
• a 15% probability of a magnitude 6.0 to 6.4 quake - up from 10% in September
• a 4% probability of a magnitude 6.5 to 6.9 quake - unchanged from September
• a 1%, or one in 100 chance, of a magnitude 7.0 or higher - down from 2% in September.

Dr Berryman said the recent magnitude 5.5 earthquake had placed a small amount of upward pressure on the probabilities, but it was just one of many inputs.

Questions

Q1: How is the combined model calculated?
In simple terms, it is the average of the short-term and medium-term models.

Q2: Have figures from the combined model been used anywhere?
Yes, they were provided to the Canterbury Earthquakes Royal Commission and they have been used in the review of Building Codes for Canterbury.

Q3: Why have the probabilities increased when only a month ago you said they were decreasing?
This is mainly due to the increasing effect of the medium-term model in the combined estimates. The recent magnitude 5.5 earthquake has also put slight upward pressure on the probabilities.

Q4: Why didn't you use the combined model from the start?
The medium-term clustering model has very little effect on the probabilities in the first year of an aftershock sequence. This model is now making a significant contribution to the combined estimates.

Q5: Has this modelling technique been peer reviewed?
This modelling technique has been published in peer-reviewed science journals. It was also included in the GNS Science submission to the Canterbury Earthquakes Royal Commission. The GNS Science submission was internationally peer reviewed.

Q6: Is the use of the combined forecasting model a sign that the aftershock sequence has entered a new phase?
No. The aftershocks are behaving as expected. They are declining in frequency quite quickly, but now the possibility of triggered earthquakes at greater distances is becoming a prominent component of the total hazard, so that is being reflected in the more complex forecast model.