Jump to content
Snow?
Local
Radar
Cold?
IGNORED

Leaky Integrator Discussion


Admiral_Bobski

Recommended Posts

Posted
  • Location: Mytholmroyd, West Yorks.......
  • Weather Preferences: Hot & Sunny, Cold & Snowy
  • Location: Mytholmroyd, West Yorks.......

The output of such a set of assertions is a 0.91 correlation. For comparative reasons the CO2 hypothesis sits at 0.71.

How can we account for that?

I'd have to fall back on our lack of understanding of the system and it's workings V.P.!

I'd hate to try and clobber the figures so they fit the picture we have but would prefer to find out why our sums do not add up.

Maybe CO2 didn't start from a position of 'zero' as we tend to assume (climate was 'cooling' and accelerating in that cooling and not roughly stable as we had assumed)?

Maybe it started out at, say, -2 (and was heavily handicapped on top of that).

It's a bit R.I.R.O. and until the programmer has all the correct data to input then the final figure will be out.

Link to comment
Share on other sites

Posted
  • Location: Near Newton Abbot or east Dartmoor, Devon
  • Location: Near Newton Abbot or east Dartmoor, Devon

I think that that's a little unfair. At most, the LI would suggest that the GhG effect is more or less constant. Not that it doesn't exist, not that a complete new era of science is upon us. Indeed, the whole CO2 hypothesis is based upon measurement that the average temperature of the moon is 45C. A faulty assumption for sure, but I still won't discount it.

Not sure it's based on the average temp of the moon - though that does help make the case. Isn't it based on the physical properties of some gasses?

I've tried. Nearly everyone isn't interested in alternate hypothesis'

Ok fair enough, and I'd expect that to be the case.

Ok, how about if you're right, you need (to bolster your case) to explain why the anthro GH effect isn't as found by science to be.

Link to comment
Share on other sites

Posted
  • Location: Rochester, Kent
  • Location: Rochester, Kent

I'd have to fall back on our lack of understanding of the system and it's workings V.P.!

I'd hate to try and clobber the figures so they fit the picture we have but would prefer to find out why our sums do not add up.

Maybe CO2 didn't start from a position of 'zero' as we tend to assume (climate was 'cooling' and accelerating in that cooling and not roughly stable as we had assumed)?

Maybe it started out at, say, -2 (and was heavily handicapped on top of that).

It's a bit R.I.R.O. and until the programmer has all the correct data to input then the final figure will be out.

My immediate (and less than well thought-out answer) is that anyone who tries a deterministic approach to climate modelling is doomed to failure. Anecdotally, we can see that in action - we know that the inter-decadel variance over the last ten years simply wasn't expected. I am sure models will be tweaked and output will correspondingly be improved.

There are assumptions in the LI. Once I finish the certainty work - I have to write, and document, the random number generator since I can't afford MatLab, and MS products simply aren't strong enough, I will be able to assign some sort of cetainty to the 0.91 correlation figure. Until then, the set of assertions, above, is all I have; and any claims to the contrary are erroneous.

Besides that, the question I would be asking - is what if the certainty work turns out to be strong? This is a really simple model (as models go). I guess most will cross that bridge if and when it comes to that.

Not sure it's based on the average temp of the moon - though that does help make the case. Isn't it based on the physical properties of some gasses?

Yes, but the paper makes the logarithmic conclusions based on a mean temperature of the moon compared to the Earth over an extended period of time. It is then put together, fitted using least linear squares and a hypothesis reached.

Ok, how about if you're right, you need (to bolster your case) to explain why the anthro GH effect isn't as found by science to be.

I don't know that it isn't. It could well be that sun activity is a causal effect of CO2 and CO2 is the driver that causes sunspot activity to be so well correlated. In which case, like we all agree, we need to stop polluting because there are unteneble risks associated with such behaviour

Edited by VillagePlank
Link to comment
Share on other sites

Posted
  • Location: Mytholmroyd, West Yorks.......
  • Weather Preferences: Hot & Sunny, Cold & Snowy
  • Location: Mytholmroyd, West Yorks.......

My immediate (and less than well thought-out answer) is that anyone who tries a deterministic approach to climate modelling is doomed to failure. Anecdotally, we can see that in action - we know that the inter-decadal variance over the last ten years simply wasn't expected. I am sure models will be tweaked and output will correspondingly be improved.

There are assumptions in the LI. Once I finish the certainty work - I have to write, and document, the random number generator since I can't afford MatLab, and MS products simply aren't strong enough, I will be able to assign some sort of certainty to the 0.91 correlation figure. Until then, the set of assertions, above, is all I have; and any claims to the contrary are erroneous.

Besides that, the question I would be asking - is what if the certainty work turns out to be true? This is a really simple model (as models go). I guess most will cross that bridge if and when it comes to that.

Yes, but the paper makes the logarithmic conclusions based on a mean temperature of the moon compared to the Earth over an extended period of time. It is then put together, fitted using least linear squares and a hypothesis reached.

And I'd have to agree with most of that. The fact we seem to be permanently 'running to catch up' with the reality we observe must show us something? The fact that many of these surprises appear to compound and not mitigate our problems has me of the opinion that we do not understand the complexity of the forcing at play and not that we have mis-understood the forcing at play. It would be far to coincidental for a multitude of natural forcings to be conspiring to bring us to this point whilst we are midst recognising that climate is shifting out of kilter.

We should never distance ourselves from the importance of all the planets natural checks and balances, nor the role they have in forcing climate, but we should also not 'shelve' the major forcing we have placed into the system. Should we not be seeking to blend our understandings of both and not be in this 'either/or' game?

Link to comment
Share on other sites

Posted
  • Location: Rochester, Kent
  • Location: Rochester, Kent

And I'd have to agree with most of that. The fact we seem to be permanently 'running to catch up' with the reality we observe must show us something? The fact that many of these surprises appear to compound and not mitigate our problems has me of the opinion that we do not understand the complexity of the forcing at play and not that we have mis-understood the forcing at play. It would be far to coincidental for a multitude of natural forcings to be conspiring to bring us to this point whilst we are midst recognising that climate is shifting out of kilter.

We should never distance ourselves from the importance of all the planets natural checks and balances, nor the role they have in forcing climate, but we should also not 'shelve' the major forcing we have placed into the system. Should we not be seeking to blend our understandings of both and not be in this 'either/or' game?

Consider the LI as a model of the sum of all forcings, not a replacement for any.

It is easy to presume, for instance, and as I said a moment ago, that sun activity forces CO2 levels up, and therefore warms the planet with some sort of lag. There is no evidence that this is the case, but it may well be true. It is also a possibility that ANY of the other two of the three drivers: ENSO, sea-ice, are related to CO2. Volcanism, clearly, isn't. It could well be that the GhG effect is effectively a constant. I don't know, and the LI doesn't say anything, directly, about it either. It simply correlates sunspots with global temperatures in a very strong way.

There is no re-writing of science here - perhaps, simply, trying to understand the sum of all things. But even that's too grandiose, I think ...

Edited by VillagePlank
Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

What nobody has actually addressed yet is the fact that the LI comes up with this 0.91 correlation. Doesn't anybody but me and VP find that interesting? Amazing, even?

No figures were massaged, no data cherry-picked and no clever after-the-fact alterations made.

All of the data used are from legitimate sources (Hadley, NOAA, etc.) and publicly-available. The only real assumption made was that solar activity (proxied by sunspot count) was the biggest contributor, and the three other factors (ENSO, vulcanicity and albedo) were chosen because they were clearly relevant to the phenomenon.

Does anyone have anything to say about the correlation factor?

CB

Link to comment
Share on other sites

Posted
  • Location: Rochester, Kent
  • Location: Rochester, Kent

What nobody has actually addressed yet is the fact that the LI comes up with this 0.91 correlation. Doesn't anybody but me and VP find that interesting? Amazing, even?

No figures were massaged, no data cherry-picked and no clever after-the-fact alterations made.

All of the data used are from legitimate sources (Hadley, NOAA, etc.) and publicly-available. The only real assumption made was that solar activity (proxied by sunspot count) was the biggest contributor, and the three other factors (ENSO, vulcanicity and albedo) were chosen because they were clearly relevant to the phenomenon.

Does anyone have anything to say about the correlation factor?

CB

Missed an important point there, CB: it is also reproducible. Hence the reason for the run-through nearly a year ago. Still need to do the certainty work, but preliminaries are promising.

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

Missed an important point there, CB: it is also reproducible. Hence the reason for the run-through nearly a year ago. Still need to do the certainty work, but preliminaries are promising.

Good point, thereby making it all the more interesting (and, dare I say it, amazing).

I'm not sure what point GW is making about Arctic sea-ice being "a different beastie" from what it once was. I don't see how that has any bearing on the LI - we have used arctic sea ice as a proxy for albedo and so, more or less, ice extent is all that is relevant to the LI. The fact that the ice may be thinner or whatever is of no enormous concern, especially as its changes (into "a different beastie") have taken place only within the last 5-10 years, and the LI goes back well over a hundred years.

I am also a little unsure about GW's remark, "we should also not 'shelve' the major forcing [CO2] we have placed into the system." Calling CO2 the "major forcing" is presuming that CO2 is in fact the (or a) major forcing. The point of the LI is to see if we can take the single biggest input of energy (the Sun) and produce a result comparable with observations without any presumptions!

CB

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

What nobody has actually addressed yet is the fact that the LI comes up with this 0.91 correlation. Doesn't anybody but me and VP find that interesting? Amazing, even?

No figures were massaged, no data cherry-picked and no clever after-the-fact alterations made.

All of the data used are from legitimate sources (Hadley, NOAA, etc.) and publicly-available. The only real assumption made was that solar activity (proxied by sunspot count) was the biggest contributor, and the three other factors (ENSO, vulcanicity and albedo) were chosen because they were clearly relevant to the phenomenon.

Does anyone have anything to say about the correlation factor?

CB

Anybody...?pardon.gif

CB

Link to comment
Share on other sites

Posted
  • Location: Mytholmroyd, West Yorks.......
  • Weather Preferences: Hot & Sunny, Cold & Snowy
  • Location: Mytholmroyd, West Yorks.......

Me sir,me sir,me sir!!!

Yes I do find both the efforts involved (money where your mouth is kinda stuff) and the fact that the 0.91 figure is the result (without 'cherry picking to enable it to be so).

Sadly until I fully understand (to the point where I am fully capable of both visualising the thing and can comfortably manipulate it) I can still draw the parallel with those Victorian caged singing birds which reproduce nature but in a wholly different way to nature itself (though sharing some components).

Please be patient with me though, it's turned Chrimbo in this house and I've turned Grinch...........

Link to comment
Share on other sites

Posted
  • Location: Beccles, Suffolk.
  • Weather Preferences: Thunder, snow, heat, sunshine...
  • Location: Beccles, Suffolk.

Yes. I, for one, find the correlation pretty compelling... :blush:

If I might ask a (silly?) question: would it be possible to get a first estimate of GHG forcing, by somehow combining the GHG function with that for the cumulative function for the natural effects? Could such a combination of functions be 'manipulated' using valid statistical methodology to get a correlation even better than 0.91???

More to the point: would such an exercise be of any use whatsoever?? :pardon::whistling:

Link to comment
Share on other sites

Posted
  • Location: Rochester, Kent
  • Location: Rochester, Kent

Yes. I, for one, find the correlation pretty compelling... :blush:

If I might ask a (silly?) question: would it be possible to get a first estimate of GHG forcing, by somehow combining the GHG function with that for the cumulative function for the natural effects? Could such a combination of functions be 'manipulated' using valid statistical methodology to get a correlation even better than 0.91???

More to the point: would such an exercise be of any use whatsoever?? :pardon::whistling:

That problem is akin to "can we take out ENSO forcing from the HadCRU dataset" ... I'm sure it's possible, but I think it would be treading new/untested/arguable statistical ground to do so.

To my mind, the best (everyone is right) way forward is to presume that one of the factors Sunspots/ENSO/sea-ice/Volcanism is somehow modified by GhG's. At least two of them are strong contenders: ENSO, and Sea-Ice, which means that the LI might well be summing the GhG effect anyway.

In that respect, it is possible to show how much each forcing adds to the temperature on an annual basis (already done, but something for publication, not for here)

Regardless, sun activity, and it's associated hysteresis, is, by far the biggest driver behind the LI hypothesis given that other factors are simply used to remove or add extra effect of the sun on an annual basis (so they effectively modify sunspot count)

Edited by VillagePlank
Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

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

Link to comment
Share on other sites

Posted
  • Location: Beccles, Suffolk.
  • Weather Preferences: Thunder, snow, heat, sunshine...
  • Location: Beccles, Suffolk.

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

And, that is how I see things, too... :)

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

And, that is how I see things, too... biggrin.gif

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

:)

CB

Link to comment
Share on other sites

Posted
  • Location: Beccles, Suffolk.
  • Weather Preferences: Thunder, snow, heat, sunshine...
  • Location: Beccles, Suffolk.

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

:)

CB

Yes. But not all of it?? :)

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

Yes. But not all of it?? good.gif

Good answer! :)

CB

Link to comment
Share on other sites

Posted
  • Location: Near Newton Abbot or east Dartmoor, Devon
  • Location: Near Newton Abbot or east Dartmoor, Devon

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?

Link to comment
Share on other sites

Posted
  • Location: Rochester, Kent
  • Location: Rochester, Kent

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.

Link to comment
Share on other sites

Posted
  • Location: Near Newton Abbot or east Dartmoor, Devon
  • Location: Near Newton Abbot or east Dartmoor, Devon

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 :yahoo:

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?

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

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.

:yahoo:

CB

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

Ok, I think I knew that and had forgotten whistling.gif

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.

:unsure:

CB

Link to comment
Share on other sites

Posted
  • Location: Rossland BC Canada
  • Location: Rossland BC Canada

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). ... :cray:

Link to comment
Share on other sites

Posted
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey
  • Location: A small planet somewhere in the vicinity of Guildford, Surrey

Hi Roger smile.gif

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 smile.gif

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/

smile.gif

Edited by Captain_Bobski
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...