150 replies to this topic

[Solar Sausage]

Captain_Bobski, on 09 December 2009 - 09:26 , said:

Ah, but do you agree that the Sun is responsible for the current warming trend?



CB

Yes. But not all of it??

[Captain_Bobski]

Pete Tattum, on 09 December 2009 - 09:28 , said:

Yes. But not all of it??

Good answer!

CB

[Devonian]

Captain_Bobski, on 09 December 2009 - 09:18 , said:

I suppose you could describe the LI simplistically by saying that the Sun is the main driver of climate: it is responsible for the current warming trend. That trend is modulated by ENSO, albedo, vulcanism and everything else.



CB

Does it not depend upon which of the various estimates of past solar output you use?

Btw, i hope I don't have to stress the Sun warms us from near absolute zero to ~18C bla bla?

[Boar Wrinklestorm]

Devonian, on 09 December 2009 - 10:12 , said:

Does it not depend upon which of the various estimates of past solar output you use?

Sunspot count is used as a proxy for solar activity because it doesn't require any sophistication and the reliable record starts in 1749.

[Devonian]

VillagePlank, on 09 December 2009 - 10:31 , said:

Sunspot count is used as a proxy for solar activity because it doesn't require any sophistication and the reliable record starts in 1749.

Ok, I think I knew that and had forgotten

So, if this solar cycle is low, or perhaps very low, numbered and the atmosphere keeps warming you're going to have to put a number on the 'lag'? But, you have and it's a big number?

So, what predictions does the LI, can the LI, make?

[Captain_Bobski]

Devonian, on 09 December 2009 - 10:12 , said:

Does it not depend upon which of the various estimates of past solar output you use?

Btw, i hope I don't have to stress the Sun warms us from near absolute zero to ~18C bla bla?

Well, we have pretty reliable records of sunspot count going back to the mid-17th Century. Using sunspots seems to give a good correlation. One of the problems with other measures of solar activity (TSI, magnetic flux and so on) is that you need to rely almost exclusively on proxies. So, using sunspots avoids the proxy issue (at least if you go back to around 1650 - going back further than that does rely on proxies...more on this later).

I am well aware that the Sun warms us from a bit above absolute zero to around -18C. Bear in mind that the -18C figure is an average, though. The Sun actually has the potential to heat us to well over +100C - or one half of us at least - a swing of over 400C. Given this fact I find it quite easy to imagine that the Sun can be responsible for a 1C warming trend over 100 years, especially during the last 60 years of extremely high activity.



CB

[Captain_Bobski]

Devonian, on 09 December 2009 - 11:07 , said:

Ok, I think I knew that and had forgotten

So, if this solar cycle is low, or perhaps very low, numbered and the atmosphere keeps warming you're going to have to put a number on the 'lag'? But, you have and it's a big number?

So, what predictions does the LI, can the LI, make?

It's a bit hard to put a definitive number on the lag because, due to the Stefan-Boltzmann law, the lag is constantly changing. It is determined by the input into the system and the amount of energy within the system. Since these two factors are constantly changing, the lag is constantly changing too.

The main prediciton that the LI has so far made is that temperatures should start to go down within the next 3-5 years (if I remember correctly). This is assuming that the next solar cycle is a moderate to low one, and that ENSO does nothing spectacularly unusual, and that there isn't a major volcanic eruption, and that albedo doesn't change significantly.



CB

[Roger J Smith]

Just wanted to say, I'm having a look at all this for the first time, and I noted the pdf of the paper on the last page of the main thread.

I will summarize my own approach to long-term climate change, and note that at this early stage, I am not sure if a "leaky integrator" concept resides in it or not, or if it perhaps should and could thereby be improved.

My approach has been to assume that the atmosphere is a partially fixed and partially mobile response to external drivers. The first assumption is one that everyone in this field makes, that the general set-up is controlled by the Sun's nearly constant output of heat. From that broad assumption that is not in question in any concept, different people then assess different importances to second-order variations in solar output, and search for any other external forcing factors that they consider relevant.

I have adopted the somewhat outlier position that solar activity is a second "effect" of some primary cause that produces both atmospheric variations and solar activity variations.

In other words, sunspots don't influence weather patterns, they correlate with weather patterns for some overarching reason. That overarching reason (from my research) is the interplay of solar system magnetic field sectors. The way in which these interact leads to both weather and solar variations.

Now that is one rather broad statement and another one would be the caveat that if solar heat production is reduced significantly by lack of sunspot activity, as indicated by research, then a longer solar cycle such as the oscillations on a 200-year time scale may become more of a factor in a climate model.

What I'm getting at is this -- the large 11-year cycle variations of an active sun period are difficult to correlate with climate response. Just consider the position in solar cycles of these four winters, 1947, 1963, 1987 and if it proves cold, 2009-10, and you'll see that right away, cold UK winters are randomly distributed against solar cycle. In North America where we have more cold winters to work with, the distribution is quasi-random, but as I have 168 years of temperature records in a data base, and a solar activity data base, I was able to correlate mean annual temperature with solar activity; the correlation was something like -0.2 ... barely significant (a positive correlation is expected, the hypothesis that temperature depends on solar activity predicts that warmer weather will occur near solar maximum, actually, the data show a slightly opposite tendency).

I have no doubt that the correlation for a longer time scale is more like +0.6 or better, so I would certainly be factoring long-term solar activity into any finalized climate model. As all of my practical work involves LRFs for the upcoming months or seasons, that has not been a very necessary step in my work, but as Fred and I constantly keep hinting, we are expecting "the big chill" factor to emerge soon if this solar quiet continues much longer.

Let me back up slightly here to mention that I have already posted on NW some evidence of a 20-year (actually 19.86 year) cycle of solar variation linked closely to interactions between Jupiter and Saturn. It takes that long for Jupiter to pass Saturn and therefore, for similar orientations of J-field and S-field sectors to interact. This seems to correlate also with the Hale cycle of alternating polarity of sunspot peaks. Since the Maunder minimum period (1650-1710) the Sun has never gone totally quiet but has missed a few regular peaks and gone into a slower, weaker modality. It did this from 1798 to 1827 and to a lesser extent from 1873 to 1915. Now, Jupiter passes Saturn in this modern sunspot era in years from about 1722 to 2000 (note the even second last digit) and the two planets are opposite each other from about 1732 to 2010. A quick inspection of sunspot maxima will show that quite often, the peaks occur just before these alignments, and display a secondary peak just after them (think of 1968, 1972 for example). My analysis shows that overall, the two kinds of peaks (aligned, opposite) are equal in magnitude. Even in the less reliable medieval data, the same pattern emerges in strong activity periods.

Sometimes these peaks disappear and some second-order forcing from Jupiter alone, which exists as a second-order variable in the long-term data, takes over the modulation on a 12-year pulse. That seemed to be the case in the Dalton minimum, the pulse became 1804, 1816, 1828 before the stronger pulse returned into dominance. Then again, in the next minimum (does it have a name?) we see 1883, 1895, 1907 as a good fit. An incomplete part of this research has to be admitted here, as Jupiter was in a different part of its orbit for these two series.

The implication here is that occasionally, some aspect of the J-S interaction weakens or fails altogether. The culprit is more likely to be Saturn, a planet twice as far from the Sun as Jupiter, possessing a strong but not as strong magnetic field. The usual sign that the strong pulse is about to weaken is an unusually long cycle (think 1787 with its very long fadeout, and 2001 was similar, 1870 was less marked in this regard). Then the first of however many weak cycles will take 12-13 years to arrive and will be offset from the normal position relative to the J-S interaction.

For example, after 1787, the next peak should have been around 1797-99 to maintain regular service. There was a slower and weaker return to action in the period 1801-04. The weakness in this cycle then increased with a total absence of activity around 1807-11 and the totally offset 1816 mini-peak.

Anyway, the point of this is that solar activity only seems to be a big climate factor when regular strong activity fades and irregular weak activity (the 25% second modality) arrives. However, some strong series are stronger than others, and this may have a sort of cumulative, slow effect that would certainly suggest a lag. I think we have seen two kinds of lags from the 20th century warm period suggested by solar activity theory. One would be the continuation of warm temperatures to about mid-2007, and another would be the response of arctic ice depletion which had its "high water mark" in late 2007. If we are indeed now sliding into a colder period due to lower solar activity, the lag time may be about 3-7 years as the atmosphere loses remnants of the warmer regime.

This post is getting a bit long, I have some more to add which I will continue to do (thankfully today is a quiet weather day after a big storm marathon past two days). ...

[Rob]

can someone please explain what a leaky integrator is?

[Captain_Bobski]

Hi Roger

Nice to see you on this thread - I was curious to see what you would think of the LI! If you haven't done so already then I would highly recommend reading through the full pinned thread at the top of the main climate board. In that thread VP has put a great deal of effort (for which I am ever grateful) into going through the basics of the LI a step at a time. In the early pages you will see a series of graphs which use official sunspot data. The leaky integrator function produces a graph with a surprising correlation to observed temperature trends over the past 100+ years.

We went on to put in some volcanic data, ENSO data and albedo data (the latter using sea ice extent as a proxy for albedo, since sea ice will be the biggest determining factor in changing albedo), and what we got out the other end was a graph with a 0.91 correlation with observed 20th Century temperature - a result which I still find rather astonishing (especially since VP points out that, on the same basis, the CO2 hypothesis has only a 0.71 correlation)! At present, VP is working on the certainty aspect of that correlation.

If you have any questions, comments or suggestions then they would be very welcome

CB

PS - Rob: everything is explained in the pinned thread on the main climate board. It's basically a mathematical articulation of a concept also known as the "leaky bucket". Imagine a bucket with a hole in it, into which you are pouring water. The leaky integrator describes what happens to the volume of water within the bucket as the amounts of water in and out change. As I said above, it's all explained really well here: http://forum.netweat...aky-integrator/

Edited by Captain_Bobski, 10 December 2009 - 22:15 .

[Roger J Smith]

Excuse me for rambling, but I assume the points will be made, somewhat anecdotally.

Okay, I have pretty much exhausted my overview on solar activity as a possible factor in my research model. The bottom line is that atmospheric responses to solar system magnetic field positions will have their own integrity and whatever happens to solar activity (during regular strong cycle periods) will be independent of that and there won't be a cause and effect relationship on that time scale. The longer time scale, I buy into ... and who wouldn't, seeing the colder weather in the CET from its starting date to about 1710?

The Dalton min, I suppose, has a generally consistent look to the Maunder on about half the scale, and the un-named later min is certainly no counter-example but the response there is obscured by the volcanic dust veil factor that was major from late 1883 to perhaps 1888.

So, I should return to the larger question, how to model external variables into what would otherwise be a steady-state climate based on relatively constant solar heat input (with a seasonal modulation of course). That variation requires first of all a designed "meteorological grid" which slightly redefines the terrestrial grid to account for the geomagnetic field. The north pole of this met grid has always been deflected by the actual position of the NMP, and this has been converging in recent years. On a very broad scale, the geomagnetic assumption "explains" global warming -- the NMP was at 68 N in 1840 and is now at 84 N. Thus the distorted position of the north meteorological pole must have varied in sync, perhaps from 79 N to 87 N using a half-way balance. Given some assumption of wandering polar vortex locations, this would scatter southward-displaced polar vortex positions around the Atlantic and North American sectors with a trend northward over time, a gross indicator of warming. The devil is in the details -- where to position the field sectors, how these change over time in latitude, and how that change affects climate response.

We just saw in North America how a "warming climate" can feed back unexpectedly cold weather. In October, the jet stream took an unusually northward path across the continent. Pacific air masses took a route across southern Canada and brought very warm anomalies to some regions near Hudson Bay. But this warm air went so far north that the northeast US was forced into a temporary period of relative cold as air masses drained out of this warm pool, retrograded and supported an unusual Gulf stream eddy. This was probably part of some large-scale warming pattern, but it gave a week of record cold temperatures related mostly to upwelling of cold water offshore. These are difficult non-linear factors to model from any perspective, the logic of them is somewhat like saying if you hose down your frozen driveway, most of it will warm up, but some random chunk of ice will swirl across a given location and cool that down.

But on the statistical index level, the field theory gives good results. Both Fred and I take this approach from somewhat different perspectives, but the experience of the past two years is that we are converging on similar solutions, and these solutions are non-random. I wouldn't want to present them as finished solutions like astronomical tide tables, but we feel that we are narrowing uncertainty. This is where I am fascinated by this leaky integrator concept, because an ongoing question is always, will the analogues need to be adjusted in some predictable way? One thing that I caught on to after several years was that the met grid had shifted northwest (not north) and that analogues would therefore tend to run a bit earlier if retrograde activity was involved (patterns shifting northwest will arrive earlier as well as at a higher latitude). You may recall that in Feb 2007 there was a good example of retrogression shooting far to the north of analogues, and this seemed to confuse even the global models which no doubt are programmed to recognize similar analogues.

I'm sort of feeling my around here in a free-association way, since I'm groping for contact with this LI concept. If you've read David Dilley's work, you'll know that he is quite adamant that you can model all atmospheric variation from one source, lunar tidal and gravitational effects. I have that as a factor in my model, but I see it as a component, not the absolute driver. The research certainly indicates a process at work, but it also shows the same significance or even more for field sectors. So in my research model, I would be giving lunar effects about 30% of the variability, and field sector positioning about 40% ... that leaves 30% which is what I am trying to understand, because I am convinced that the 70% assigned to the two main drivers is valid. This 30% is not influenced by volcanic dust, for example -- that is more like a temporary cooling of all systems, not a forcing factor for weather patterns as such. I would make the same argument in the AGW context, sure we would expect a warming of the atmosphere in general but not a very big one, and not one that would structurally alter the atmosphere (at least not very much, obviously there would be feedback questions to consider).

These 30% non-external (because when you've eliminated magnetic, solar, lunar and planetary, what else is there?) factors are tougher to model in the predictive sense, because first you need a reason to predict them. For example, you could name things like the El Nino, and assume that it had an effect on climate systems that was independent of any accidental correlation already assumed in your external drivers (and there is some, I've identified these and reported on them). If you don't know when an El Nino will occur in the period say 2020-30, then how can you model it into a LRF for that period? There are obviously many other internal system processes that need to be considered. The philosophical question is, are these processes or descriptions. You know what I mean, when the average non-research met is asked why last month was so mild, he will usually say "because the jet stream was coming from the south" but of course that tells me nothing useful for my research into making LRFs, that is just a different way of describing the warmth. It is not cause and effect in the sense of explaining process, just in the sense of explaining effect.

To sum up this rather sketchy response to LI, I would say this ... my approach is basically to assume a base-line for given locations and add warming to that base-line from whatever field, lunar or obvious feedback processes were identified in research, or what we've taken to calling signals (although we speak of cold signals, which is essentially the presence of blocking or the absence of warm signals). This is how I make a long-range forecast in general, I assume that it will be at least as warm as the coldest period of recent record, and then look for all reasons to add warmth until there are none left. This is essentially an energy-level model. I suppose in this model, a leaky integrator approach would be to consider the existence of an open window rather than a leaky barrel.

I hope people reading all this will agree that the question on the table here is not whether my research model is valid or invalid, that's for a different thread. The question on the table is, can such an approach benefit from considering the LI theorem? In other words, is the model conceptually incomplete without reference to LI, or, has it somehow incorporated the LI concept without naming it as such? That's what I need to understand. If the model is just plain wrong, that's not going to affect this discussion at all. If the model is valid and breaking new ground, then this LI concept becomes all the more significant because a developing model needs improvement factors to become more useful (and thereby credible).

I'll return perhaps tomorrow to see what the thinking is about any relevance of one paradigm to the other paradigm.

[Boar Wrinklestorm]

Quite a lot there, and it may take me a few days to absorb it all.

My first response is that sunspots is simply used as a proxy for something else. The LI does introduce a new driver, hysteresis, in that the absolute quantitative effect of all drivers is easily overridden by energy already in the system and that such reasoning can account, with a strong correlation, for the modern warming.

In terms of the primary external driver, sunspot count, I can see two options:

(i) Something affects the sun and then the sunspot count affects climate
(ii) Something else affects sunspots and climate at the same time

(ii) is my favourite option - ie that sunspot count correlates well with the climate because of some effect that affects the climate and sunspots at the same time. A sort of mirror effect, if you will.

Particularly, I have tried integrating the LI using total solar insolation and it simply doesn't work. This is, of course, in line with all current solar/climate research, but, this also strongly implies, then, that there is something else of which sunspot count is indicating that correlates well with climate. I do not know what that is.

In terms of where I am ... I am conducting due-diligence (randomised testing of the data and method (which is not fully disclosed) to determine statistical certainties of the correlation) I will report in due course. I hope to finish the source-code at some point this weekend, and get going on it. If the certainty is above or beyond the 1.96sd then I will write up and attempt to publish as a curiosity (so avoiding AGW politics entirely)

It is perhaps notable that whilst there are hints as to where hysteresis might be hiding, or why sunspot count seems to have a strong correlation, the LI makes no attempt to assign a physical explanation. I must leave that for others, more talented that I.

Edited by VillagePlank, 11 December 2009 - 08:48 .

[Boar Wrinklestorm]

Rob, on 10 December 2009 - 21:51 , said:

can someone please explain what a leaky integrator is?

The essential argument is this:

(i) External drivers raise the temperature of the Earth
(ii) The Earth is not as efficient as losing temperature as it is in gaining it by a very slight margin

Sounds familiar, doesn't it? Anyway ...

(iii) Such heat accumulates and the rate of heat gain or loss is directly proportional to the heat in the system - this is the leaky integrator bit.

The approach has been, where possible, to conduct this via first principles - even to the point of showing certain properties of dynamical systems to be true under most circumstances.

Edited by VillagePlank, 11 December 2009 - 10:22 .

[jethro]

VP have you considered the Earth's magnetic field? There are quite strong links between field strength/dipole movement and climate, field strength and Solar wind etc. It might be worth a look; here's an old thread I started ages ago, some of the links may be dead but there's quite a lot of info in there.

http://forum.netweat...magnetic-field/

[Captain_Bobski]

I've been giving a lot of thought to this issue of Mechanism, so I thought I'd let you in on where I've got to so far.

We've been talking about hysteresivity being an essential part of the Stefan-Boltmann law. The SB law obviously is all about black body radiation. Since SB incorporates the Planck constant, it must hold true down to the atomic scale (since the Planck constant is a quantum-mechanical constant). This means that the LI concept must also hold true down to the atomic scale.

So, every single component of the Earth's physical structure must be involved in hysteresis, down to invidual atoms.

On the scale of an atom, the amount of energy involved is obviously tiny, but when you take every single atom into account it can add up to a large amount. So here's my thinking on how the process works...

Atoms can only take on energy in discrete packets - called quanta - and the more energetic the atom is the more packets of energy are required to raise the atom to its next energy state. Left to its own devices, the atom will happily give up its extra energy - it will only retain that energy state if the energy input remains high enough to prevent it from releasing the energy (or, to be more accurate, it happily drops an energy state only to be immediately raised back up to that energy state by the next wave of quanta). Because the gaps between energy states increase the more energetic the atom gets, it becomes harder and harder to raise the energy state and easier and easier to drop an energy state.

This means that as an atom becomes more energetic its emissivity necessarily increases.

What SB tells us is that the absorptivity of a body increases at a slightly faster rate than its emissivity does. So far I can only tie this into the atomic model if there are lots of atoms around which are all undergoing similar energy bombardment. Here's how it works:

As an atom becomes more energetic it relinquishes that energy more readily. For the purposes of this explanation, we'll call our atom Fred.

Fred is at energy state 1 when he is suddenly bombarded by quanta. There are enough quanta to raise his energy state (ES) to 2. Fred is quite lazy and likes being at ES1, so he tries to give the quanta back so that he can return there, but every time he gives the quanta back more energy comes in and he ends up back at ES2.

The bombardment of quanta increases a bit, but not by enough to raise him to ES3 (Fred breathes a sigh of relief). This means that Fred can throw all that extra energy away and remain at ES2 (which he doesn't especially like, but he has to deal with). Unfortunately for Fred, his neighbours (Tom, Dick and Harry) are also all at ES2 and they think the same thing.

So Tom, Dick and Harry all throw away the extra energy they have received from this increased bombardment. Fred finds that when the extra energy from the bombardment hits him, he also gets hit with the extra energy from Tom, Dick and Harry. Sadly, the extra energy from the bombardment plus the extra energy from Tom, Dick and Harry is enough to push him up to ES3.

Now somewhat annoyed, Fred starts to throw away bigger amounts of energy, but every time he does he finds that the total energy coming at him is keeping him at ES3. However, because he's throwing away more energy, it is not long before Tom finds himself at ES3 as well. Soon, Dick and Harry are in the same situation.

With the energy bombardment remaining even, Fred, Tom, Dick and Harry remain at ES3. Crucially, though, the energy bombardment alone is not enough to get them to ES3 - only the bombardment plus the "extra" from the other atoms is enough to keep them there.

Note also that it takes some time for the knock-on effect to take place - Fred reaches ES3 first, followed some time later by Tom, then later still by Dick and Harry.

Does that make sense to anyone, or are there any objections?



CB

[Boar Wrinklestorm]

Captain_Bobski, on 11 December 2009 - 11:06 , said:

So, every single component of the Earth's physical structure must be involved in hysteresis, down to invidual atoms.

Blimey CB - more reading to do!

Yes, and that is the consequence of the explicit first part of the PDF. If a system is dynamical (and the vast majority of natural systems are) then it must, necessarily, exhibit hysteresis to one degree or another.

This part is essential to exploiting the LI as an abstraction of the SB law. I am working on a proof that it holds for all values except one special case (which is in the PDF)

Edited by VillagePlank, 11 December 2009 - 11:17 .

[Captain_Bobski]

VillagePlank, on 11 December 2009 - 11:13 , said:

Blimey CB - more reading to do!

Yes, and that is the consequence of the explicit first part of the PDF. If a system is dynamical (and the vast majority of natural systems are) then it must, necessarily, exhibit hysteresis to one degree or another.

This part is essential to exploiting the LI as an abstraction of the SB law. I am working on a proof that it holds for all values except one special case (which is in the PDF)

You might want to avoid using words like "exploiting", VP - people might think we're using some kind of "trick" to "hide" something!!



CB

EDIT - I'm glad I've given you some more food for thought, but I'm sorry about it at the same time!

Edited by Captain_Bobski, 11 December 2009 - 11:19 .

[Boar Wrinklestorm]

Captain_Bobski, on 11 December 2009 - 11:18 , said:

I'm glad I've given you some more food for thought ...

If the energy exchange at such a low-level leaves something behind in some form or another, then there will be an excess of mass/energy, and we can stop calling such an excess 'dark' ...

(Just kidding ...)

Edited by VillagePlank, 11 December 2009 - 11:28 .

[Solar Sausage]

Thanks CB...Good stuff!!!

Is that the mechanism of hysteresis...if so it makes sense (to me) both in terms of QMS and SMS...And, it's bloody good revision!!

Just a quickie?? When you chose the names Tom, Dick and Harry - were you alluding to Quantum Tunneling??

Time for a Great Escape, methinks...

Anyhoo. Good work, guys!

[Captain_Bobski]

Pete Tattum, on 11 December 2009 - 11:43 , said:

Thanks CB...Good stuff!!!

Is that the mechanism of hysteresis...if so it makes sense (to me) both in terms of QMS and SMS...And, it's bloody good revision!!

Just a quickie?? When you chose the names Tom, Dick and Harry - were you alluding to Quantum Tunneling??

Time for a Great Escape, methinks...

Anyhoo. Good work, guys!

I don't know whether I should ROFL, PMSL, groan or kill myself...



CB