Bushfires in Australia
Authorities confirmed that at least 108 people died in fires described by Prime Minister Kevin Rudd as "hell in all its fury", gave warning that the number of dead was expected to climb higher as they continue the grim search for at least 100 people who remain unaccounted for in the ruins of houses and the shells of charred cars.
Authorities said more destruction was possible, with 31 fires still burning in the state, although a cooler front was aiding firefighters. It was estimated it could take weeks to bring some of the fires under control.
Mr Rudd said the bush fires, some of which police believe were deliberately lit, constituted "mass murder", with media reporting the death toll could reach 170 as authorities searched hundreds of burnt-out homes.
The infernos, which started on Saturday in a period of record temperatures, destroyed at least 750 homes as they burned through 350,000 hectares of towns and farmland in the state of Victoria in a day and night already dubbed "Black Saturday". Entire towns have been wiped off the map, with aerial images of the region showing an eerie moonscape of blackened earth and collapsed homes where fields and communities used to stand.
Most of the damage was done by two massive fires - one that virtually wiped out towns northeast of Melbourne including Kinglake and Marysville with a 60-mile front - and a second inferno that raced across Gippsland. Burnt-out cars were strewn across the road outside Kinglake, testimony to residents who made doomed attempts to escape.
GCP hypothesis test: I have not been able to learn when the fires started or the time of the tragic loss of life, so we will use a 24 hour period that definitely includes most of the early part of the tragedy. 11:00:00 Saturday February 7, 2009 in Australia/Melbourne converts to 00:00:00 Saturday February 7, 2009 in UTC. We will make the formal prediction the 24 hour UTC day, which includes 11:00 am Saturday to 11:00 Sunday, Melbourne time. The resulting Chisquare is 85961.001 on 86400 df, for p = 0.855 and corresponding Z = -1.056. I have marked an unusual peak in the middle of this data sequence beginning at midnight with a very strong positive departure that persists for half an hour then turns to a persisting negative trend for the rest of the 24 hour period.
It is important to keep in mind that we have only a tiny statistical effect, so that it is always hard to distinguish signal from noise. This means that every "success" might be largely driven by chance, and every "null" might include a real signal overwhelmed by noise. In the long run, a real effect can be identified only by patiently accumulating replications of similar analyses.