Seeing this headline “Perth faces dry winter” – made me check BoM 3 month Outlook rainfall predictions again. They are as hopeless as ever – for month after month from November 2009 – the BoM has failed to predict the dominant wet over most of Australia.
Starting with the Outlook for summer (Dec-Feb) – the links below let you compare
the BoM prediction with real world rainfall. I am simply looking for the BoM to correctly predict the broad shape of the rain decile patterns – but I do not see that they are close to doing that over the past six months. If a private business was wrong this often – they would soon go broke.
Dec-Feb prediction
Dec-Feb actual rain deciles – the Sth Aust wet never happened – central Aust was not average it was WET. The predicted N Qld dry never happened, it was in fact wetter than average. The very dry WA coast was not predicted. A few points for the tiny dry area on Cape York. You get the picture – I will leave it for readers to work through the links for themselves. Check you own locality – I would appreciate hearing how you rate the BoM there. For my region ACT, the BoM predicted below average rain every Outlook but the actual rain was greater in each of the 4 cases. Of course the BoM puts out temperature Outlooks too, maybe somebody can check how they have performed for the 4 Outlooks since November 2009.
Jan-Mar prediction
Jan-Mar actual rain deciles
Feb-Apr prediction
Feb-Apr actual rain deciles
Mar-May prediction
Mar-May actual rain deciles
And the media still report these BoM press releases as if they mean something ?
IMHO the BoM should be disbanded and a weather forecast section salvaged out of it. Failures like I have demonstrated over years now are signs of a dysfunctional and unscientific org past its use by date.
you will enjoy this Why scientists get it wrong and NZ climate crisis gets worse
www.quadrant.org.au/blogs/doomed-planet
Warwick, I have looked at rainfall records near where I live going back to 1893. I have a manual gauge and an electronic one which I look at daily
The area is the Sunshine Coast. Here is the result
Period Prediction% Average (>100yrs -mm) Actual (mm)
Dec09-Feb10 45 698 577
Jan10-Mar10 30 810 939
Feb10-Apr10 50 739 992
Mar10-May10 50 610 739
BOM got it wrong for the Jan-Mar period. This is the normal wet period. The rainfall often varies considerably in within a short distance. In March several areas on the Sunshine coast had floods. My daughter 10kms away had over 350mm one day when we had 250mm (2March). The record rainfall at Buderim occurred in 1898 when 1100mm fell on the three days 8 to 10Jan.
Peter – thanks for your input. You said the last two BoM predictions were for 50 – I disagree with that – they were both for less than 50.
Here are my takes on the BoM results for the Sunshine Coast region – using their own published maps it looks to me like four failures. Tell me if you think I am being fair or not.
Dec-Feb prediction – mostly less than 45% with small patches just over 45%
Dec-Feb actual rain deciles – average rain – a clear fail
Jan-Mar prediction – rain less than 35% of average.
Jan-Mar actual rain deciles – over average rain – so a pretty big failure
Feb-Apr prediction – looking at the positions of the BoM contours, their prediction was for ~46-47% of average rain in the Sunshine Coast.
Feb-Apr actual rain deciles – rain solidly above average – a clear failure again
Mar-May prediction – is for clearly a little less than 50% chance of average rain – maybe 48 or 49% chance.
Mar-May actual rain deciles – clearly above average – another clear failure
Of course you’re not being fair, since you’re intentionally misreading BoM’s charts.
No, their prediction was that there was a 46-47% chance of exceeding the median rainfall, which is an entirely different thing – in 46 years out of 100 with the same prevailing climate pattern, there will be greater than median rainfall. In the other 54 years there will be less than median rainfall.
I think you need to ponder why BoM is getting it so wrong, considering the data sets they’re using to make their predictions.