Posts Tagged ‘Weather forecasts’

Weather predictions- final check

January 3, 2014

This is the final post in my series of checking a hunch that temperature change indicates a weather change 160 days later.

Back on November 3, I predicted:

“December
2 to 10 unstable; 12-13-14; 16-17-18-19; 22-23-24; 26 to 31 unstable.”

I did not change this in December.

This is how  I went:

Dec2013 predictions check

I’ve marked with green bars the predicted dates of unsettled weather as above.  Red bars show the actual times.  They match.

And finally, here’s my graph showing predicted weather events for April to June.  Again, green bars indicate dates when weather events may be expected.Apr-June 2014

I will leave this topic for now, not because the method doesn’t work (it does!), but to concentrate on other interests.

Weather predictions: December

December 1, 2013

At the start of November, I said:

“November
5 to 10 unstable; 13 to 21 unstable with several events; 26-27-28-29-30 unstable.”

All correct, 1 miss.  Instability with some very wild storms marked much of November especially in the South-East of the state.

Now I suppose anyone could have predicted storms for November.  But remember, back in August I had said:

“November

5-6-7, 9-10, 13-14-15, 17-18-19-20, 27-28-29.”

Here’s a chart showing August predictions in light green and early November predictions in dark green.octdec13resultsnov

5 right, I miss.  I should have stuck with my original predictions!

So the method is holding.

Predictions for December to 31 March remain the same as I predicted last month.  As well, I expect weather events around these dates in April and May (+/- 1 day):

2,4,7,11,15,20,23,25, May 1, 8,11.

April should have unstable weather, and I would not be surprised if we get significant rain.

 

Weather Predictions November – March

November 3, 2013

At the start of October, I made the following predictions:

October
Sep 30- 1-2, 8-9-10, 14-15-16, 20-21-22-23, 25?, 28-29-30-31-Nov 1 (possibly 2 events).

Changes were detected on:- 2-3, 8-9, 13-14, 17-18, 22-23- 24, 28- 29-30-31 (the”possibly 2 events” was true indeed with a series of storm fronts that persisted.)Oct 13 results & pred

Dark red is Tmin, pink is Tmax, black is inverted pressure, blue is rain, green is the 160 day lagged Tmin, yellow is 40 day lagged pressure.  I may try 40 day lagged temperature in future.   November and December predictions are shown as red ovals.  So there were 5 correct predictions, 1 wrong, and 1 miss (although the wrong event would have been a “correct” if it had been a day earlier).  Not bad for 160 days out, I console myself.  I also suspect there has been a slight acceleration of weather systems, so perhaps I should be looking at events around 158 days ahead instead of 160.  No matter for now but I will keep an eye on it.

Predictions for November and December (slightly adjusted from last month):

November
5 to 10 unstable; 13 to 21 unstable with several events; 26-27-28-29-30 unstable.
December
2 to 10 unstable; 12-13-14; 16-17-18-19; 22-23-24; 26 to 31 unstable.

 January to March:

Here is the graph I use for predicting weather changes, which shows 2nd derivative, or acceleration/deceleration, of minimum temperature.

jan-mar 2014 predictions

Think of a cool change moving through from the west.  A large temperature differential, or a fast moving change, both have the same effect.  When the green line goes below zero, the temperature has decelerated, and a change (or unsettled weather) occurs in the time period indicated.  Several rapid oscillations appear to relate to unsettled weather.  My only problem is I am averaging over a very wide region, and the time lag may be changing.

There is always something to learn.

October- December Predictions

October 2, 2013

At the start of last month, I said:

“A vigorous change came through the Capricornia district today, and further events are expected on:  4, 7, 9-10, 13, 16-17-18, 23-24-25, and 30- October 2.”

Results in September have not been as successful.  Changes occurred on September 2, 10, 13-14, 17, 25, 27, and storms brought rain to some places in the last two days.

6 right, 3 wrong, and 1 miss (which in hindsight I should have seen.)

75% right, 25% wrong, and a miss- not good.

For the rest of the year, I am predicting weather events (storms, heat followed by cool changes, dry fronts moving through) at the following times:

October
Sep 30- 1-2, 8-9-10, 14-15-16, 20-21-22-23, 25?, 28-29-30-31-Nov 1 (possibly 2 events).
November
4; 6-7; 9-10; 13-14-15; 18-19-20-21; 27-28-29.
December
5-6-7; 8-9-10; 12-13-14; 17-18-19; 22-23-24; 26-27-28.

For January-March 2014:  If I get within a day or two of these I’ll be happy.

January:

5, 9?, 14, 20-26 unsettled.

February:

6, 10, 16?, 21-24, 27?.

March:

4-9 (unsettled).

Concerning how hot September was:

Yes, it was hot.  Many daily and monthly records were broken.

For the 10 sites I monitor, all but Rockhampton, Maryborough, and Amberley broke daily and monthly records.  Rockhampton equalled the monthly record but had no chance of breaking the 37.9 (corrected to 37.4) daily record mentioned in the 1953 Commonwealth Year Book, p.44.  (And Brisbane was not even close to 38.3 set in 1943.)  Inland was very hot.

The 365 day mean of maxima has been above average all year.

Weather Outlook 14 May 2013- Update 16 May!

May 14, 2013

(Update:- Oops!  I’ve found a glitch in my calculations which I have corrected.  This has the effect of changing and adding some predicted events- see below.)

Firstly, my apologies for not attending to this blog more regularly.  Retirement certainly doesn’t mean having more time.

This is what I predicted back in April:

Meanwhile, refined predictions for April to early May (>90% certain) and long range predictions to early September (about these dates, plus several others), for weather changes (cold change, pressure drop, or rain):-

April:

3-4-5-6-7-8-9; 10-11-12; 13-14-15 (A succession of small changes)

17-18-19-20; 22-23-24

25-2627-28

May:

April 30-1-2-3-4-5

7-8-9-10-11-12-13-14 (unsettled for the first 2 weeks).

April-may check

(Pink is Tmax, dark red is Tmin, blue is rain, black is inverted mean sea level pressure.)

11 Correct, 1 Wrong.  91.6%. (I could be persuaded that there were only 10 major changes, which makes it 83%).  Moreover, my prediction of a succession of small changes in early April and unsettled conditions for the first two weeks of May was definitely correct.

So far this year, my short term (up to 40 days) predictions fared as follows:

Predictions of changes

Correct

No change

Unpredicted changes

33

32

1

2

That’s 91.4 %.

160 day predictions (from 27 November and 2 January)

“April: 7, 9, 11, 13, 15, 19, 27.

May: 4, 9, 13, 19, 29”

up to today, were for 25 events so far this year, of which 23 occurred as predicted (later refinements removed one of the predicted events).   92% correct for the year, 80% for April and May so far.  That is also 67.6% of all events.

I note the BOM forecast is for fine weather this weekend, whereas from 2 January I have been predicting a change for 19 May.

My predictions for sub-tropical Queensland for the rest of May and up to mid-October (temperature change, pressure drop, or rain) are as follows:

Unsettled with several successive events for the next several weeks:

May

17-18-19-20, 21-22-23, 24-25-26-27-28-29-30-31-1-2 June

June 5-6-7-8-9-10-11, 13-14-15-16-17, 20-21-22-23, 27-28-29-30

July

2-3-4-5-6-7, 9-1011-12, 15-16-17-18-19-20-21, 25-2627-28-29-30

August

1-2-3-4-5-6-78-9, 11-1213-14, 15-16-17-18-19-20-21, 23-24-25, 27-28-29-3031

September

1-2-3-4-5-678-9, 12-1314-15, 18-19-20, 22-23-24-25-26-27-28

October

September 30-1-2, 8-910-11, 14-15-16.

Do you know of anyone else predicting a weather event (probably storms) for 15 October 2013?

I am still working on an objective test for my predictions, which I am close to finalising.  One thing I have found is that the frequency of weather changes does vary from about 22 to 38 in a 90 day period, which means there are often lengthy periods of changes occurring at least once every three days, which would make accurate forecasting much more difficult.  In a future post I will also show how my 160 day predictions are even more accurate for individual locations than when averaged over a whole region.