Jessica,

Not sure of your question or if it for me. Do you seek monthly temperature averages for Jan and Feb of 1993 and 1981 for Melbourne central?

These are easily found at
http://www.australianweathernews.com/recent_AWN...

Cheers Geoff.

Why did'nt you add 1993 and 1981.

Geoff #7

BTW, I think it quite silly to ruminate about Global Warming having anything to do with these fires, not that many intelligent people have. One swallow doth not a summer make.

Trouble is that there are people who see only the magnitude of the disaster and fail to notice that it is a long tail event. We had same large bushfires here at the same time in Wollemi but threatened nobody and were just allowed to burn.

David, I see a complex of events melding in to the meaning of Exceptional Circumstances. I have alluded to fires needing good years of growth for fuel, but also dry years to increase burning/fire temperature. A fire the year after a big fire in the same spot will tend to be smaller. Lightning can cause fires and increased lightning might correlate with increased rain (I don't know), which alters the chance of both putting out a fire and starting one. The runoff from fire ravaged land is less likely to soak in and so the following growth season is affected, except that some seeds need fire to germinate and some fires can lead to better growth. The water that runs off more rapidly to sea instead of soaking in, is less available to form clouds by evaporation, hence changing water vapour feedback on radiation. Fires can also change soil fertilization and surface albedo and both affect regrowth, possibly rainfall, possibly temperature. ...... The whole lot is interrelated, under the umbrella of Exceptional Circumstances prediction, which I think is not fruitful because the effects are not well understood. Evidence that all relevant factors are not all understood quantitatively is shown by the DECR that is uncertain in some degrees that might not matter, sometimes in degrees that will. Yes, I have read it and I found it unconvincing.

BTW, I think it quite silly to ruminate about Global Warming having anything to do with these fires, not that many intelligent people have. One swallow doth not a summer make.

Geoff makes an interesting point, and I am actually quite surprised as I though there was a run of days at +40C. The assumption of the the DECR work is that the increased exceptional circumstances result from an increase in the mean of distribution of random events, tending to skew hot days up say. This might work on droughts say, that go on over a period of time, but Geoff illustrates that the bushfires occurred on a day so exceptional, they are a classic long-tailed event. In this case the increase in probability of the even as a result of movement of the mean would be miniscule.

I think it would be the case that the kind of model for exceptional events used by the DECR would be pointless. Rather a model of a logistic distribution using a shape parameter, rather than mean, would make more sense, if shape actually did change for any rason.

Nick,

The report has been called the "Drought Exceptional Circumstances Report". It takes a set of exceptional climate/weather circumstances to produce bushfires like these. Drought is one factor equated by some to bushfires. Others say a period of good growth is a precursor, to build up fuel. I have not read which of these somewhat opposing factors is the more important in prediction.

The whole thrust of my observations was that forecasting climate-related exceptional circumstances is a bit pointless. If you have data that refute this assertion, I'm a willing listener.

Show me a climate projection of a decade or more that has been shown correct. Just one.

Geoff,
It's silly to try to link this with the report of Hennessy and co. They made no attempt to predict specific wildfire events, which as you say is an impossible task based on climate data.

Not sure where to insert this.

Temperatures in Melbourne at BOM Central Weather Station are somewhat sensitive to discuss because of the loss of more than 200 lives in bushfires on 7th Feb with a maximum of 46.4 deg C (115.5 deg F). Some of the fires were below 50 km from the station and below 20 km from some populous suburbs.

Here is a comparison of temperatures at Melbourne Central Weather Station for each February since year 2000.



FEB BY YEAR, Tmax, Tmin, Tmean Deg C

2000, 30.1, 17.7, 23.9
2001, 29.0, 18.3, 23.7
2002, 25.4, 15.2, 20.3
2003, 26.1, 16.9, 21.5
2004, 26.0, 15.5, 20.8
2005, 24.4, 14.6, 19.5
2006, 25.3, 15.8, 20.6
2007, 29.7, 16.5, 23.1
2008, 24.9, 15.9, 20.4
2009, 28.1, 16.5, 22.3

This month of horrendous fires highlights the difficulty in forecasting catastrophic events. In the table, February 2009 does not look abnormal in any predictive way. A forward model made in year 2000 would have zero possibility of predicting fires in year 2007, in February 2007, to the day, or at all.

It takes a particular coming together of several short-term events to make such a catastrophe. It is beyond both the resolution of weather events and the mathematics of forecasting to make a model of any value. It follows that one should not make models predicting exceptional circumstances. However, in 2008 we had the joint BOM/CSIRO Drought Exceptional Circumstances Report with K Hennessey as principal author.

The value of this report has been severely questioned on this site by Dr Stockwell. I suggest that this table is another nail in the coffin of the modellers. The reason is that predictive models have little value when they fail to predict.

In the short term, there were warnings some 4 days in advance of the fires. Then later in the month and into March, there were more warnings which turned out to be non-events in comparison. So kudos for the win a few days before 7th Feb, but no such praise for longer terms and near misses.

Additionally, here are the daily Tmax and Tmin for each day in February.

1, 33.8, 20.3
2, 28.5, 20.8
3, 30.2, 20.5
4, 30.2, 16.8
5, 29.2, 19.8
6, 33.1, 17.9
7, 46.4, 18.7
8, 22.5, 18.6
9, 21.4, 16.1
10, 21.2, 14.1
11, 20.2, 14.3
12, 23.7, 14.4
13, 28, 12.8
14, 28.4, 14
15, 29.6, 14.8
16, 30.1, 16
17, 31.7, 15.8
18, 27.8, 17
19, 30.5, 16.8
20, 24.6, 18
21, 23.8, 17.4
22, 25.5, 16.9
23, 35, 16.2
24, 21.3, 14.3
25, 21.6, 14.1
26, 31.2, 14.2
27, 34, 14.3
28, 23.2, 16.1


Quite an ordinary February, except for one day, with its night being about the same as the week before. So how do you predict that one day, which was really from about 3 pm to 9 pm, when the main damage was done?

Not sure where to insert this.

Temperatures in Melbourne at BOM Central Weather Station are somewhat sensitive to discuss because of the loss of more than 200 lives in bushfires on 7th Feb with a maximum of 46.4 deg C (115.5 deg F). Some of the fires were below 50 km from the station and below 20 km from some populous suburbs.

Here is a comparison of temperatures at Melbourne Central Weather Station for each February since year 2000.



FEB BY YEAR, Tmax, Tmin, Tmean Deg C

2000, 30.1, 17.7, 23.9
2001, 29.0, 18.3, 23.7
2002, 25.4, 15.2, 20.3
2003, 26.1, 16.9, 21.5
2004, 26.0, 15.5, 20.8
2005, 24.4, 14.6, 19.5
2006, 25.3, 15.8, 20.6
2007, 29.7, 16.5, 23.1
2008, 24.9, 15.9, 20.4
2009, 28.1, 16.5, 22.3

This month of horrendous fires highlights the difficulty in forecasting catastrophic events. In the table, February 2009 does not look abnormal in any predictive way. A forward model made in year 2000 would have zero possibility of predicting fires in year 2007, in February 2007, to the day, or at all.

It takes a particular coming together of several short-term events to make such a catastrophe. It is beyond both the resolution of weather events and the mathematics of forecasting to make a model of any value. It follows that one should not make models predicting exceptional circumstances. However, in 2008 we had the joint BOM/CSIRO Drought Exceptional Circumstances Report with K Hennessey as principal author.

The value of this report has been severely questioned on this site by Dr Stockwell. I suggest that this table is another nail in the coffin of the modellers. The reason is that predictive models have little value when they fail to predict.

Additionally, here are the daily Tmax and Tmin for each day in February.

1 33.8 20.3
2 28.5 20.8
3 30.2 20.5
4 30.2 16.8
5 29.2 19.8
6 33.1 17.9
7 46.4 18.7
8 22.5 18.6
9 21.4 16.1
10 21.2 14.1
11 20.2 14.3
12 23.7 14.4
13 28 12.8
14 28.4 14
15 29.6 14.8
16 30.1 16
17 31.7 15.8
18 27.8 17
19 30.5 16.8
20 24.6 18
21 23.8 17.4
22 25.5 16.9
23 35 16.2
24 21.3 14.3
25 21.6 14.1
26 31.2 14.2
27 34 14.3
28 23.2 16.1


Quite an ordinary February, except for one day, with its night being about the same as the week before. So how do you predict that one day, which was really from about 3 pm to 9 pm, when the main damage was done?

It would be unusual if February temps did not go up, but it will not be much, maybe .03-.04 for satellite. However, beginning in March it will go in reverse.

The recent SSW (sudden stratospheric warming) tells the story, and as La Nina rears its ugly head, any hope for GW to ramp up this year will be lost.

Met O can be expected to move the goal posts yet again, and of course will blame it on "unexpected" ENSO changes with no mention of PDO.