Ken Ring's Long Range Forecast
2006 April 08 to 30: Analysis

Disclaimer

My Introductory Comments

Forecasts
8th, 9th, 10th, 11th, 12th, 13th, 14th,
15th, 16th, 17th, 18th, 19th, 20th, 21st,
22nd, 23rd, 24th, 25th, 26th, 27th, 28th,
29th, 30th.

Daily Results Summary Table

Concluding Remarks

Return to Main Forecasts Only Page.

Return to Synoptic Charts and Satpics Page.

My Introductory Comments

I have changed the format, so the first page contains Ken's forecasts only - those who are interested may check out his website <www.predictweather.com> for more information on his methods.

I have included a separate page with 6 hourly BoM MSLP synoptic charts and satellite images along with each days forecast so anyone can assess for themselves how well he did.

This page is for my analysis of Ken's forecasts.

(Added the above, removed some redundant paragraphs 29th May)

As this is a long range forecast the rules I use will be a little more relaxed than those that would apply to say a BoM forecast for the next day.

My interest at this stage is in seeing if Ken's lunar cycle based forecasts are able to describe the weather in the general sense of overall patterns, rather than being a tool for precise descriptions of conditions at specific places and times.

In fact, my assessment will be somewhat lenient, in that I will be allowing a fair amount of leeway in what I consider to be a successful prediction - for example, I will consider Ken's "H" to apply to a High pressure cell as a whole or even ridging rather than just the central location of the High pressure cell, as from reading the forecasts I think that is how Ken is using it some of the time, however such use may carry a penalty in my assessment.

Each forecast will be assessed by being broken down into discrete predictions, and assessed under two categories 'place' and 'time', with a 'hit' scoring "1" in both, and a 'miss' scoring "0" in both.

Ken allows +-1 day leeway in his forecasts, so I will be bearing that in mind, so a situation successfully forecast within the leeway will still count as a successful prediction, however a 'time' penalty of "0.5" will apply.

If a prediction is correct in a general sense, but the location is a bit out, a 'place' penalty of "0.5" will apply.

I will also add implied predictions in [square brackets] where it is obvious from the forecasts that there is continuity across intervening days, and these will also be included in the assessment.

I hope to encourage Ken to be a little more precise in his descriptions and use of terminology in future forecasts, so that a "High" may be qualified at times by terms such as "strong High" or "weak High", with prominent extending ridges being identified by location, a "Low" at times qualified with "weak" or "deep, with troughs or fronts being indentified by location, etc..

In short, what I would like to see is a more complete description of the chart with all main features identified and located clearly - if Ken is prepared to do this, I'm quite prepared to continue assessing his monthly forecasts, and to develop the means of assessing them more accurately.

If Ken can learn to better describe his forecasts in a little more detail they will be more easily subject to verification so giving him a greater chance of gaining further acceptance of his work should it indeed verify, and it may also improve the usefulness of his forecasts for his clients - note that I make no claims as to the accuracy of Ken's forecasts or how useful his clients find them to be.

Update 15th April:

Ken Ring has expanded his forecasts for the rest of the month starting on the 16th and posted them on the auspacwx list: 15 April 2006 12:49:11 AM. This was in response to my request that he do so, and I appreciate the fact that he has. I have incorporated his expanded forecasts below.

James Holbeach from the auspacwx list has suggested I introduce some location definitions, which seems like a good idea to me - thanks James.

For the purposes of this forecast, where Ken has not been forwewarned of where these areas are, they will be broad by definition - only those over water areas mentioned in Ken's forecasts are considered at this stage:

Coral Sea: E of Australia, N of 28S, W of a curve running from the NE tip of NZ through the axis of the islands of Vanuatu and the axis of the Solomon Islands.
Tasman Sea: E of Australia, S of 28S, N of a line from extreme S of Tasmania to extreme S of Sth Is of New Zealand, W of New Zealand and line from NE tip of NZ to Vanuatu.
The Bight: S of continental Australia, W of Tasmania.

For future forecasts I will prepare a map for Ken giving specific regions including breaking larger areas into specific parts to make both the forecasts and the analysis less ambiguous.

James Holbeach has also suggested that I apply some controls as to what consitutes a Low - he suggested less than 1003 hPa, however that appears too arbitrary to me as it removes a low from it's wider context, and would negate the 1008-1009 hPa Coral Sea Low early in the period that was significant enough for the BoM to put on their charts and mention in their TC outlooks for several days.

I am going to simply say any Low that appears on a BoM chart or is mentioned in any form of BoM forecast or bulletin within a 24 hour period is valid, at least for this analysis. I will give further consideration to this issue for future forecasts, as tighter controls are needed.

Update 28th May:

Jane ONeill questioned with my definition of the Coral Sea which includes the Solomon Sea in WeatherZone Forums - you can see her comments here and here.

Jane does have a valid point in that it is not strictly correct, however after giving this much consideration, whist I'm inclined to chop up the Coral Sea into more strictly correct areas as I said here in the WeatherZone Forums, it is unclear whether Ken Ring gave this much attention before making his forecasts so I will err on the side of leniency and stay with my earlier definition for this analysis.

As all the synoptic charts and satpic images are included in the Synoptic Charts and Satpics Page, everyone is free to undertake their own analysis with their own bounderies.

| Top of Page |

From: ken@weatherman.co.nz
Subject: [austpacwx] Longrange for rest of April
Date: 8 April 2006 7:28:21 AM
Includes Update: 15 April 2006 12:49:11 AM

8th: A L forms in Coral Sea as fronts cross The Bight

The weak Low in the Coral Sea was slowly developing all day in the satpics moving SSW to S from nearer the 'tail' of PNG, and had developed sufficiently later in the day for the BoM to it include in the 10pm chart, so a good prediction both in place and time.

A High moving east dominates the eastern bight while only one front is crossing into the area S of the bight, however another will do so during tomorrow by which time the first front will be decaying while slipping SE, so partial place and timing errors.

| Top of Page |

9th: A H approaches Carnavon and a H crosses SA to VIC [Low remains in Coral Sea]

The western High is still a long way west and is extending a ridge towards WA, albeit a fair way S of Carnarvon, and extends a ridge across most of S WA by 10pm, so a partial error in place.

The eastern High has indeed crossed SE SA into Vic, so a good prediction in both place and time.

[The weak Low remained in the Coral Sea, so a good prediction both in place and time.]

| Top of Page |

10th: A H drifts down and then E towards Perth as the further E another H crosses TAS. [Low remains in Coral Sea]

The western High did indeed drift a little S then late in the period moved significantly E towards Perth, so a good prediction in both time and place.

The eastern High crossed over S NSW and then moved up the coast with a ridge extending S that crossed Tas, so a partial error in place but good in time.

[The weak Low remained in the Coral Sea, so a good prediction both in place and time.]

| Top of Page |

11th: Another L develops in the Coral Sea as a H crosses the Bight to cojoin the eastern H now in the Tasman. [First Low remains in Coral Sea]

From the BoM's Qld. Tropical Cyclone Outlook issued at 2:35pm on Tuesday the 11th of April 2006: "A monsoonal trough occurs across the central Coral sea with weak embedded low centres along trough axis." - also the satpics show a weak low developing off Lucinda which moves away from the coast - this verifies Ken's forecast for the original Low plus a new Low in the Coral Sea, even though the BoM has shown only the trough on some charts, so a double prediction both in place and time.

The High off WA seems to have stalled out in the Indian Ocean, however as the day progresses it builds a ridge of higher pressure right across southern coastal Australia that 'conjoins' to the eastern High - I am going to be lenient on this as the 'conjoins' description does indicate a ridge more than a cell - if that magic word was not there it would have been a bust, however a place penalty will apply.
Update: on tomorrows charts a High does indeed move across the Bight! I'm separating it from the 'conjoins' ridge, and a time penalty will apply for the Bight High.

The eastern High has moved into the Tasman, so a good prediction both in place and time.

| Top of Page |

12th: Ls in the Coral Sea begin bunching together, an H sits over TAS and another H approaches WA behind a front that crosses the Bight with a L near Melbourne.

The Coral Sea Lows are no longer evident in the charts or mentioned elsewhere, however you can see the remnant lows bunching closer as the trough weakens over towards New Caledonia in the satpics - they would be more correctly described as disturbances by later in the day - I'm reluctant to assess this as anything but a bust, but am doing so with place and time penalties in keeping with my policy of leniency.
Update: I have tightened the assessment criterea so that a system must appear in at least one chart or be mentioned in forecasts or other bulletins to qualify, so this now becomes a bust - now changed below.

There is no High sitting over Tas, however a weak High develops over S NSW and Vic, so a place error applies.

IO High is approaching WA, so good in both place and time.

A strong front is crossing the Bight, so good in place and time.

There is no Low near Melbourne, however a weak Low devlops on the SE NSW coast and drifts up the coast during the day - a place penalty applies.
Update: As I have tightened things up a bit this is now a bust - corrected below.

| Top of Page |

13th: The H approaching WA expands, the L in the Coral consolidates and the H moves east off TAS.

The High approaching WA does grow in pressure and extent, so good in both place and time.

A Low is consolidating over the Solomon islands whilst Coral Sea convection is increasing - because this Low is over the Solomons and is a new one I'm giving place and time penalties to reduce it's effect in the total.

The weak High over Vic / NSW actually moves north, so to lower it's influence in the results place and time penalties apply.

| Top of Page |

14th: The H is over Nullarbor, the L remains in the Coral and strong W winds cross TAS.

A High is over the Bight S of the Nullarbor, so a place penalty applies.

A Low over the Solomons moves SW into the Solomon Sea (considered part of Coral Sea here).

A front with NW winds predeeding it and W to SW following crosses TAS, so I consider this good enough.

| Top of Page |

15th: The L remains in the Coral, an H covers Melbourne.

The Low remains in the southern Solomon Sea (considered as Coral Sea here).

The High moves slowly E over the eastern Bight, but does not cover Melbourne until tomorrow, so place and time penalties apply.

| Top of Page |

16th: The H over Melbourne moves to cover Sydney as another H approaches Perth behind a L near Carnarvon. [Low remains in Coral Sea] Low begins moving SW. Isolated storm Cape York. A heat low forms over N Kimberley extending a trough down the coast until 17th. Cold front crosses SA.

High moves over Victoria and to the NSW central coast - that's close enough for me, so good prediction in place and time.

Another High approaches Perth, so good prediction in place and time.

Low near Carnarvon is borne out by the trough and shape of synoptic lines on charts, and is in the last chart of the day, so good prediction in place and time.

The Low in the southern Solomon Sea (considered as Coral Sea here) moves SW into Coral Sea (proper), so a good prediction in place and time.

There is possible storm activity over Cape York in the satpics, however I'm not going to include this prediction as I have no way to confirm or disprove it.

There is an implied trough down the WA NW coast from the N Kimberley by the shape of the synoptic lines, however it is indefinite enough that place and time penalties apply.

Cold front crosses SA, but not till the 17th, so place and time penalties apply.

| Top of Page |

17th: The H begins to affect Perth, the L moves NE from Carnarvon, whilst the L in the Coral remains as does the H over E NSW. A front moves through TAS. Possible fog in Port Curtis.

High develops in ridge near Perth, so good prediction in place and time.

Low moved NE from Carnarvon yesterday and SW today, so place and time penalties apply.

Low in Coral Sea becomes TC Monica, so good prediction in place and time.

H over E NSW moves slowly into Tasman, so good prediction in time but place penalty applies.

A front moves over SA towards Tas but does not get there till tomorrow, so place and time penalties apply.

Possible fog in Port Curtis - no idea, so not included.

| Top of Page |

18th: Another H approaches Perth, the L moves from TAS towards Sydney, whilst the L in the Coral remains but starts moving SW. Further Ls develop near Bourke and Cape Byron which move S resulting in a L off the coast. At the same time a H near TAS. Isolated storm in SE QSLD. A frontal system develops causing the H over WA to retrogress. Cold front crosses SA. Rain in TAS today and tomorrow.

A High is approaching Perth, so good prediction in place and time.

There is no Low moving from near Tassie towards Sydney, so a bust forecast.

Low in Coral Sea is covered by TC Monica approaching N Qld, so good prediction in place and time.

No Low near Bourke, so bust forecast.

No Low moving S near Cape Byron, so bust forecast.

No High near Tassie, so bust forecast.

No storms in SE Qld, so bust forecast.

Front S of WA but H remains stationary, so place and time penalties apply.

Fronts cross SA yesterday and tomorrow but not today, so place and time penalties apply.

Front brings rain to Tas, so good prediction in place and time.

| Top of Page |

19th: Most of the W half of Australia fine, a L passes through TAS and SE of NSW. [Low remains in Coral Sea] Storm North Coast and Moreton Coast. The front in WA weakens and moves to the SE.

Most of the W half of Australia is fine, so good prediction in place and time.

No Low moving from near Tas to SE of NSW, so bust forecast.

TC Monica moves onto C York, so good prediction in time but place penalty applies.

No storms NE NSW or Moreton Coast, so bust forecast.

Yesterdays WA front has long since moved SE, so place and time penalties apply.

| Top of Page |

20th: A L approaches WA from the W, a H crosses SW into VIC. [Low remains in Coral Sea]

L approaches WA, so good prediction in place and time.

No High near Vic, so bust forecast.

No Coral Sea Low (TC Monica moves into Gulf), so bust forecast.

| Top of Page |

21st: The L affects Perth, a H is over Melborne and most of NSW, the L (now a TC) in the Coral affects the Gold Coast. Another L develops off the south QSLD coast with more gales and heavy rain. This slowly weakens over the next few days so that by 24th only a trough off the coast remains.. Strong gale force winds Curtis Coast. Low press off the NW coast brings scatt shrs and thunderstms Learmonth to Esperance.

Low affects Perth, so good prediction in place and time.

No High near Vic or NSW, so bust forecast.

No Coral Sea Low, so bust forecast.

No SE Qld Low, so bust forecast.

No gales Curtis Coast, so bust forecast.

Another Low off WA and rain band Learmonth to Esperance, so good prediction in place and time.

| Top of Page |

22nd: The L crosses the interior of WA, the L in the Coral moves south, another H approaches WA from the W. Cold front crosses SA.

Low crosses SW WA, so good enough prediction in place and time.

No Coral Sea Low, so bust forecast.

H approaches WA, so good prediction in place and time.

No cold fronts near SA, so bust forecast.

| Top of Page |

23rd: The H moves onto WA, the L from the Coral crosses Fiji. Strong winds on QSLD S coast cease. Rain in TAS.

WA High stays well and truly offshore, so bust forecast.

No Coral Sea Low, so bust forecast.

There were no strong winds on QSLD S coast to cease, so bust forecast.

Satpic looks cold and showery in Tas, so good prediction in place and time.

| Top of Page |

24th: The H crosses the Bight while the L over Fiji moves to the W. Fog in Darling Downs. Cold front over TAS

H moving slowly from S of Adelaide towards Tas, so place and time penalties apply.

No chart for Fiji, so not assessed.

Don't know about fog in Darling Downs, so not assessed.

No cold front anywhere near Tassie for several days, so bust forecast.

| Top of Page |

25th The H reaches the middle of the Bight, a front moves through Sydney.

Both predictions are clear busts.

| Top of Page |

26th: The H remains stationary in the Bight. Cold front enters SW VIC, reaching Sydney by 27th. Cold front crosses SA and TAS.

First two predictions are clear busts, front moves into SA so place and time penalties apply.

| Top of Page |

27th: The whole southern half of Australia is mostly fine. [The H remains stationary in the Bight.] From now until May 3rd the H in the Bight moves slowly east. Strong SE winds along Moreton South. High press over NSW. From now till May 3rd an anticyclone crosses Bight and Tasman.

All predictions are clear busts.

| Top of Page |

28th: The whole southern half of Australia is mostly fine. [The H remains stationary in the Bight.]

Both predictions are clear busts.

| Top of Page |

29th: The whole southern half of Australia is mostly fine. [The H remains stationary in the Bight.]

Both predictions are clear busts.

| Top of Page |

30th: A L is over QSLD, a front moves over TAS, the H remains in the Bight. Isolated storms in SW QSLD. Thunderstorms may cause flooding in SA.

Trough NW to SE over Qld, so place and time penalties apply.

Front crossed S Tas yesterday, so place and time penalties apply..

High in Bight, so good prediction in place and time.

Storms SE Qld, not SW Qld, so place and time penalties apply.

Showers E SA but nothing that looks like flooding from storms, so place and time penalties apply.

Ken Ring
www.predictweather.com

| Top of Page |

| Top of Page |

Concluding Remarks

Overall, it was an interesting exercise!

The 47% + 50% = 49% final result, whilst not definitive for some quite important reasons, certainly shows that at least in this case that Ken has underperformed compared to claims of about 80 to 85 percent accuracy on his website.

Closer examination reveals that from the 8th to the 17th inclusive Ken was doing quite well (76% + 83% = 80%), however it seems that from the 18th onwards 'the wheels fell off' - one could speculate that the analog year(s) he was using for his forecasts changed for some reason that he missed, however that is for Ken to determine.

Some important points need to be made about this assessment:

1. There were no 'control' forecasts to enable determining skill. Perhaps suitable controls could be made by generating parallel forecasts using synoptic charts from the same dates in selected previous years, as the 'solar' or seasonal component will be similar, however the 'lunar' component will be different.

2. The forecasts are often vague and the assessment is subjective - forecasts giving Lat and Lon coordinates of all High and Low pressure cells to the nearest whole degree would enable far more objective assessments to be made, including determining how accurate the forecasts are for various degrees of error.

I leave any further conclusions to be drawn up to the reader.

| Top of Page |