Hi everyone, the OPI theory seems interesting. But what I don't understand, is what kind of weather pattern occuring in October creates a negative OPI number? High pressure there and low pressure here etc. Thanks.

it's to do with how organised the polar vortex looks to become as the month progresses. a disorganised vortex as we enter November will promote a better chance of a negative AO in the months that follow. there are other factors at play which will promote feedback mechanisms which also promote a -AO. 

Hi everyone, the OPI theory seems interesting. But what I don't understand, is what kind of weather pattern occuring in October creates a negative OPI number? High pressure there and low pressure here etc. Thanks.

it's to do with how organised the polar vortex looks to become as the month progresses. a disorganised vortex as we enter November will promote a better chance of a negative AO in the months that follow. there are other factors at play which will promote feedback mechanisms which also promote a -AO. 

I think low pressure south of Alaska and over southern Greenland and western Russia are signs that the Polar Vortex is having trouble becoming organized.

But this is a subject about which we're all learning as we go along, I think.

 

---

Cheers for the explainations folks.

---

Based on the GFS12z run it's now down below -3 again. I'm surprised because at a glance the 6z and 12z runs both look pretty flat, though TBH I've not looked at them in depth. If Riccardo is around I'd be interested to know what the key differences are between today's 6z and 12z GFS runs in terms of the OPI.

i think it's to do with the angle of the vortex Brian. I wish I could get a better handle on this but I believe that the lower the angle, the more likely it is that a split vortex will ensue rather than an organised one.

At -0.19 based on the GFS6z. Interesting to see what it changes to once the 12z data is being used.

---

What is the mechanism that's proposed to explain the correlation? r=0.91 is very high, higher than almost any other correlation in the world of climate forecasting even in the tropics (I recall El Nino and regional rainfall correlations are around r=0.7/0.8 at best). Is it simply persistence, i.e. once the atmosphere is organised in a particular way it then remains so for a few months? Or is there something specific about October?

For an index that is atmosphere based, where there has traditionally been thought to be far less persistence than in the ocean, it seems too good to be true. But certainly intriguing.

With such a strong statistical link present in actual values, I would expect dynamical models to reproduce the correlation too and therefore their ensemble forecasts would be much more accurate than they are.

Finally, many winters are a mix with the AO swinging from strongly positive to strongly negative from month to month. Given the 0.91 correlation this would imply the ONI would have to be roughly neutral in any year where the AO swings around intra-seasonally, but that seems counter-intuitive if we consider what dynamical forcing might explain the relationship.

Given how the index is described I'd have thought it could quite easily be independently replicated by sourcing the raw data.

---

I don't really know anything about the OPI, but all these negative numbers sounds exciting!

---

anyone know what the proven track record is here OPI + vs OPO - , so how many mild winters have seen with a + & likewise for a negative or is this a untested / unproven theory model at present time?

 

either way if the minus continues then I am supporting the theory and only time will tell?

 

interesting none the less, just 16 more days to go.

 

thanks

The OPI numbers are based on a GFS run but which is the important figure, is it the actual figure at the end of the month or the trend throughout the month. Anyone know?

---

by the end of tomorrow, we will have more solid data than forecast data re the month as a whole. once we get through the upcoming weekend, the OPI graph should begin to home in on its likely result.  heights are definitely on the drop through the remainder of the month (which would be expected). how far can this take the OPI towards neutral ?

What is the mechanism that's proposed to explain the correlation? r=0.91 is very high, higher than almost any other correlation in the world of climate forecasting even in the tropics (I recall El Nino and regional rainfall correlations are around r=0.7/0.8 at best). Is it simply persistence, i.e. once the atmosphere is organised in a particular way it then remains so for a few months? Or is there something specific about October?

For an index that is atmosphere based, where there has traditionally been thought to be far less persistence than in the ocean, it seems too good to be true. But certainly intriguing.

With such a strong statistical link present in actual values, I would expect dynamical models to reproduce the correlation too and therefore their ensemble forecasts would be much more accurate than they are.

Finally, many winters are a mix with the AO swinging from strongly positive to strongly negative from month to month. Given the 0.91 correlation this would imply the ONI would have to be roughly neutral in any year where the AO swings around intra-seasonally, but that seems counter-intuitive if we consider what dynamical forcing might explain the relationship.

Given how the index is described I'd have thought it could quite easily be independently replicated by sourcing the raw data.

I don't think we're talking about persistence of the atmospheric pattern here. Quite the opposite, in fact. What we're looking for are conditions this month which are associated with an amplified jet in the troposphere. Over the coming weeks and months this should lead to warmth rising northwards (to promote northern blocking) and, at some point, upwards into the stratosphere (your Sudden Stratospheric Warming).

There's no way of knowing when these events will present themselves - a coupled system of oceans and atmosphere is too complex for that - hence your observation that the AO swings from positive to negative during the winter from one month to the next. But there's a fair case that this analysis of October patterns can tell you whether or not a negative AO will manifest itself at some point(s) between December and March.

Disclaimer: O level physics grade B 1977 may not be enough for me to know what I'm on about.

---

-2.86 on the 6z - a stonker!

-3.26 after 12z and these forecasts are coming towards months end now of course.

---

What is the mechanism that's proposed to explain the correlation? r=0.91 is very high, higher than almost any other correlation in the world of climate forecasting even in the tropics (I recall El Nino and regional rainfall correlations are around r=0.7/0.8 at best). Is it simply persistence, i.e. once the atmosphere is organised in a particular way it then remains so for a few months? Or is there something specific about October?

For an index that is atmosphere based, where there has traditionally been thought to be far less persistence than in the ocean, it seems too good to be true. But certainly intriguing.

With such a strong statistical link present in actual values, I would expect dynamical models to reproduce the correlation too and therefore their ensemble forecasts would be much more accurate than they are.

Finally, many winters are a mix with the AO swinging from strongly positive to strongly negative from month to month. Given the 0.91 correlation this would imply the ONI would have to be roughly neutral in any year where the AO swings around intra-seasonally, but that seems counter-intuitive if we consider what dynamical forcing might explain the relationship.

Given how the index is described I'd have thought it could quite easily be independently replicated by sourcing the raw data.

Tim I agree that a correlation coefficient of r=0.91 is incredibly(if not suspiciously high).The AO is linked to the NAO though i,m bot sure how strong the link is. The NAO  DJF CET correlation is around 0..6 ie it explains c 40% of the winter temperature variability.So even if the 0.9 is correct some statistics about  the further link between AO and DJF CET are required before we get too excited.

OPI = -3.36 on the 00z this morning.

Let's look at a most extreme-case scenario, where the figure for the next 10 days of the GFS is -4.0. (In reality it must be higher).

Let "n" be the average OPI for each of the 16 days of October for which we have actual, banked data.

((-4.0 x 10) + 16n)) / 26 = -3.36

So, -40 + 16n = 26 x -3.36

So, 16n = -87.36 + 40

So, n = -47.36 / 16

So, n = -2.96

Shocked smiley.

Which means that with over half the month gone, the OPI has already shown an unchangeable, actual, non-GFS figure of -2.96. And that's with the 00z GFS being insanely blocked, which it isn't. In reality the figure will be lower than than -3.0.

Even more shocked smiley.

Bear in mind that with a negative QBO (as we have this year), any figure below -1.5 suggests a negative Arctic Oscillation for all three winter months.

Very exciting.

(Disclaimer: A level maths grade C 1978; it's a very long time ago and I would be grateful if someone younger could check that I've formulated the algebra correctly.)

@Roger: At The Other Place Murr has suggested that a positive QBO is the reason why negative OPIs haven't corresponded with cold UK winters in certain years.

I know that you have found that the QBO in general hasn't shown as much correspondence with cold or mild UK winters as some people imagine, but perhaps in combination with this kind of atmospheric set-up, there is a more robust connection

---

OPI = -3.36 on the 00z this morning.

Let's look at a most extreme-case scenario, where the figure for the next 10 days of the GFS is -4.0. (In reality it must be higher).

Let "n" be the average OPI for each of the 16 days of October for which we have actual, banked data.

((-4.0 x 10) + 16n)) / 26 = -3.36

So, -40 + 16n = 26 x -3.36

So, 16n = -87.36 + 40

So, n = -47.36 / 16

So, n = -2.96

Shocked smiley.

Which means that with over half the month gone, the OPI has already shown an unchangeable, actual, non-GFS figure of -2.96. And that's with the 00z GFS being insanely blocked, which it isn't. In reality the figure will be less than -3.0.

Even more shocked smiley.

Bear in mind that with a negative QBO (as we have this year), any figure below -1.5 suggests a negative Arctic Oscillation for all three winter months.

Very exciting.

(Disclaimer: A level maths grade C 1978; it's a very long time ago and I would be grateful if someone younger could check that I've formulated the algebra correctly.)

Sorry it's all beyond me, LOL!

---

Gavin P: I worked as a teacher once many years ago. If I had a quid for every time I've heard those words ...

Just focus on that OPI figure of <-2.96 from the first 16 days of October. That is going to take a hell of a lot of shifting upwards in the remaining 15 days for it to become positive.

---

I wonder what the index was last year?  Mind you a negative AO doesn't necessarily equate to cold over the UK but it increases the chances of it.

Pretty interesting though, I do believe there is an element of persistance which begins to show in late Autumn (i.e blocking over Greenland, SSTs) so it will be interesting to see what happens.

---

I wonder what the index was last year?  Mind you a negative AO doesn't necessarily equate to cold over the UK but it increases the chances of it.

Pretty interesting though, I do believe there is an element of persistance which begins to show in late Autumn (i.e blocking over Greenland, SSTs) so it will be interesting to see what happens.

Riccardo published a value of +1.6 for October 2013 on 7th November 2013 - see link below

http://www.americanwx.com/bb/index.php/topic/41639-seasonal-winter-20132014-forecast-based-on-opi-index/

---

Still in negative territory based on yesterday's GFS12z run.

---

Given we're under the final ten day's of October and every day this month has been negative, surely the only question now is HOW negative will be the OPI be this October?

---