Jump to content
Sign in to follow this  
Roger J Smith

Roger Smith's developing LRF model

Recommended Posts

Roger Smith's developing LRF model

a discussion thread that will gradually give readers an overview

_________________________________________________________________

Introduction

Several years ago I published some of my theory on Net-weather and some readers are generally familiar with the concepts. Because there have been some faint indications of progress in some of the derived long-range forecasts, there is an expressed interest in how these forecasts are made. At the same time, I would like to share some of these concepts so that other people can use them. I don't really see a commercial application at this point in my life (my age being 63) and given also that I reside a long way from the U.K. in western Canada. Perhaps someone would show an interest in developing the research under an arrangement but my guess is that these concepts are going to be resisted by the more established (and funding) portions of the atmospheric sciences community until some point further down the road when the forecasts become so reliable that we reach a "resistance is futile" situation. I don't claim that we are at that point yet. This is because the theory has not yet identified all possible variables and is not fully operational. The only reason I even discuss it or mention it in connection with my forecasts (which I could just produce as my educated guesses, no reasons given) is that despite an obvious lack of full development, these forecasts have been about as good on the whole as any other methodology can realistically claim (the field is unfortunately full of inflated claims which tends to obscure the real situation, which in my humble opinion is somewhat similar to the six blind men and the elephant paradigm, if you're familiar with that).

So, what I propose to do in this thread in the course of the new year 2013, is to expose quite gradually the various components of the research model that I have developed over a long period of time, but which has really expanded considerably since I was able to incorporate daily CET data to match what was already available for the North American component (170 years of data for Toronto, Canada) and to build on the detail possible from what I was using earlier, monthly CET statistics.

What I intend to show, after a brief overview of how the thread will be formatted, will be the various "index values" created from these daily data sets, and some discussion in each case of how the North American and the U.K. data sets interact as illustrations of any theoretical content of the research model. So to simplify, you will see two sets of graphs showing signals derived from these data sets, with the U.K. data set providing the basis for normalizing the time scale on the North American data set. To illustrate what that means, I need to refer to the specifics of each data set. The U.K. set runs from 1772 to 2011 or 240 years. I have 2012 data in the files but have generated "index values" for the 240-year period and subsets. The Toronto data runs from 1841 to present.

As many readers would already know, one component of my research is lunar-atmospheric interactions. I should say from the outset that these are now considered a secondary part of the theory, the solar system magnetic field variations are considered primary. But the lunar components are easier to visualize in timing, so to illustrate how the two data sets are going to be normalized for comparison, the first concept of importance is that the start year defines the time frame. In 1772, the first new moon was on the 5th of January, whereas in 1841 it was on the 22nd of January. The synodic period of the Moon is 29.53 days (greater precision is used in setting up the index values) but the declination cycle of the Moon is a shorter 27.32 days. The first :"southern max" in 1772 was on the 4th and the first one in 1841 on the 19th. So for example, the declination "index values" derived from the daily data are only directly comparable if we take the Toronto data from day 16 of that data set signal and compare to the UK data. Otherwise the two data sets are timed almost opposite. So by normalization of data, the procedure will be to take the UK data as defining the signals (the 1772 timing is good from my point of view, having the data run from before southern max to the next similar date is just about what you would want) and then the Toronto data are adjusted from the raw signals to the same time frame as the UK data.

As we will start with that example, I will post both the raw Toronto data signal and the adjusted one so you can see what happens.

But before we get into looking at any signals or data, I just wanted to establish some ground rules for the graphs that will be used in this discussion. The main point to be stressed is that they will all be on the same scale, so you can visually compare the signals. That scale will be from +1.0 C deg at the top of the graph to -1.0 at the bottom with 0.0 of course running through the middle of the graph. Those anomalies are relative to the entire data set (in the UK case, 1772 to 2011 daily means were calculated and the actuals were converted to anomalies from those). I am going to make these data sets available in excel files later this year, and I may charge a small user fee because there is a lot of work that went into creating them, however, I will probably proceed on the basis of a voluntary donation. By the way, the daily normals for the CET are already published on Net-weather, see my thread link in the signature line below and follow the link. You can also see some daily extremes and interesting normals for recent years and the Dalton minimum period.

Now, by having the same graph ranges between +1.0 and -1,0 all the signals can be compared directly. Sometimes there will be graphs showing small signals on an expanded scale. These will be clearly identified below the graphs and I will colour code the graphs so that "normal signal strength" graphs will have red and blue bars in the graphs whereas augmented signal graphs will use orange and green instead. So if you see a graph that has an orange-green colour scheme, be alerted to the fact that the signal is defined by a smaller range on the graph than plus or minus one C degree, and that range will be identified both on the graph and in the text below it.

There will also be some graphs showing historical trends of signals. These will generally fall into the same signal strength range (-1 to +1 C) with the components generally moving upward with time. However, some graphical presentations will require that the more recent data be placed on graphs with a range from -0,.5 to +1.5 ... this will be comparable in strength of signal and the placement of the zero line will be the variable changing but the colour code will maintain the plus or minus aspect of the signal, relative to the longer time scale. So these particular "recent warming" signals will be comparable to the larger sets visually even though the graph range will be shifted upward.

There are a few cases where signals would expand a range beyond +1 to -1 C anomalies. I will then present the signals on larger (vertical) graphs than most, but these cases are rare for the entire period of record signals. I should qualify that by saying that signals derived from long periods such as the Jupiter synodic year (398.9 days) stay within the +1 to -1 region on graphs only if you take running 3-5 day means, if you take daily means for these longer cycles for which there are fewer cases in the period of record, you can easily start to get into the +2 to -2 range on particular days. I find the 11-day running mean a good way to narrow the signal range and bring more coherence to the visual presentation, so the convention to be followed is that any signal based on longer than the lunar-cycle 27-30 day time frame will be 7-day running mean for cycles of 75 to 200 days and 11-day running means for longer than 200 days.

The research model may still be using the raw data rather than the running means. This is not intended to be a total exposition of the details of how forecasts are made, but a general illustration, that will say in total (and over a rather long slow exposition period in 2013) here's what goes into the research model, all of these various index values for specific components, and here is my latest thinking on what the signals show us in theoretical terms. That part will be open to discussion and I am certainly hoping to advance my own understanding by putting this out for discussion -- this is not intended to be a one-way lecture but a real discussion. I know, if only they had those for the AGW theory, right? Well I'm open to it.

And that leads to the final point of this introduction, what about global warming, climate change etc? How does all that interact with the research model? I think you'll find the time intervals interesting, the general answer seems to be that both data sets have warmed about 0.7 to 1.2 C since the 19th century and that a larger warming has been observed in the interval 1990 to 2006 than either before or after that. The interesting part is, can this research theory explain any of that warming, in other words, do any of the index values create any of the warming, or is it all imposed from the postulated non-natural variability factor of human activity? My findings on that are mixed. Some of the longer cycles appear to be contributing to a natural peak of warmth on the time scale being discussed. But I suspect that the AGW effect is about half to two-thirds of the recent temperature upward drift. What is really intriguing is that some of the natural signals show different forms of increase, which may help us to answer those difficult questions about how the atmosphere is processing the warming signal. My general finding there is that some of the signals may be co-operating with the AGW background signal more directly than others. Some show a tendency to be more resilient -- in particular, anything retrograde and cooling in its contribution tends to act as though natural warming was not taking place. This may account for an observed increase in variability in recent years. We'll get into that as the year of data discussion and presentation unfolds.

I've mentioned then that the graphs will be directly comparable on the vertical (temperature) scale. So far I have not generated any pressure or rainfall index values for Europe, I do have some rather extensive ones for North America to compare with the temperature index values. That goes into the theoretical side of the discussion, a comparison of temperature and these other elements gives us a better idea what the signals mean in terms of sensible weather. Meanwhile, I should mention that the time scale or horizontal axis of the graphs will be comparable mainly in terms of process over different time scales. The signal period will define the time intervals likely to be used, as already mentioned. A lunar signal will always be in the time range 27-30 days and so the horizontal scale refers to that number of days. A solar system magnetic field signal is often in the range of about 120 to 400 or even 800 days or longer. Many are in the range of 367 to 400 days. That tends to make an 11-d running mean about the same scale as the daily scale for lunar signals. For the J-year of 398.9 days, I am going to show the data in 40 groups of 10-d averages. This will scale the Jupiter signal to a similar intensity potential to the lunar signals, and you'll see by direct comparison that this key planetary component of the Solar System Magnetic Field is about ten times larger than any individual lunar signal (but there are several of those, so more like three times larger than the entire lunar contibution to the research model).

Another way of normalizing the data is to define the Jupiter data to be the standard setting for other external signals, which is accomplished by locating the opposition dates for those other variables at the same point in the cycle (in percentage terms) to Jupiter. In the specific case of the two data sets, here's what this entails. For Jupiter, the 1772 opposition date was 20 August. For the other data set, the 1841 date was 5 June. So what this means is that the J-year signal for the CET data will run from 1 Jan 1772 to about 2 Feb 1773 (with about 220 of those "J-years" in the total data set) and the Jupiter opposition date (where we pass Jupiter, in other words) will be about 58% of the way into the data set. Other data sets can be visually compared by placing their opposition dates at a similar point, the raw data might place the opposition anywhere in the range of the graph, but then that can be adjusted so that all other similar index values will have the same orientation to the geometry of the solar system. Just keep that in the back of your mind for now, specific examples will follow as we get into the data. Note then that the Toronto data generating a J-year signal will on the gross scale have to be adjusted so that the earlier 1841 opposition date is shifted back by 77 days which is about 19% of a cycle, to make the two sets directly comparable. This point will be explained in more detail when we get to that part of the data, but J-year signals are even more interesting on a segmented basis. In general terms, what I mean by segments in this whole discussion is related to orbital position, a J-year segment would refer (for example) to all the cases where Jupiter is located in a similar range of its orbital positions. The other meaning that might be suspected for segments would be time segments like for example 1772 to 1851 (the first third of the CET data) but I use the term interval for that in this entire discussion, so let's be clear then on these definitions:

segment is related to orbital position range of a postulated signal producer and

interval of data will refer to quarters, thirds, halves, tenths or whatever portion of a long-period data set.

The next section I plan to post is an overview of theory, just so when we get to the data graphs, we have some basic understanding of a theoretical framework for the research. Some will be impatient to see the graphs and results, but bear with me, I want to do this in a coherent way that will minimize confusion about what is being discussed as we go through those data sets.

My plan is to post something new each day with a few breaks here and there, and I will be posting mainly at this time of day (late evening here, morning there) so you might expect to find something new each day about this time of day (1030 GMT or so). So look in again tomorrow and there should be a post on the basic theory involved in the research. That will no doubt be somewhat obscure without illustrations, so the rest of the month or year, however long we need, will be dedicated to bringing forward all the necessary illustrations from the actual data. Let me know as we proceed if this is either too fast or too slow for your liking, if too slow, I can try to post more than once a day, if too fast, we'll see what works best. I don't want to move on to new content until everyone actively following is satisfied they understand what has been said or posted.

-- Roger Smith

Edited by Roger J Smith
  • Like 9

Share this post


Link to post
Share on other sites

Wow... that post was a book. Fortunately I am a historian... :-) however I'll need to reread it several times I suspect to absorb the idea.

First question - what "tools" do we need to get started here. Computer - got. Excel - got. The various signals/data you refer to - are there sites for these, or are we talking books? I'm jumping the gun I expect, but want to get my teeth into this.

Share this post


Link to post
Share on other sites

Thanks for starting this thread, Roger. It's much appreciated.

However sceptical of your methods I may be (similar to what WE was saying in the other thread) your results really have been quite impressive, over the years. You may well be 'onto something, mate. But, whatever it is, it's light-years' outwith my understanding!

Keep up the good work!Posted Image

Share this post


Link to post
Share on other sites

Thanks, just a general comment in response to Catacol's question above, readers won't need too much "at hand" to follow this, I plan to break it down into small pieces and not to move on if there is discussion showing uncertainty about what has been posted, so at some later stage if you're interested in the method and the research there could be files available and perhaps an e-book of some kind. This thread is basically a prototype for that e-book I suppose.

Now before we get into the actual data I should outline the general concepts of this theory, but bearing in mind, the graphs and discussions of signals will serve as some kind of illustration (proof might be too strong a word to use) of these concepts -- anyone could write up any theoretical paradigm about this sort of relationship but there would be little point in studying it, if the concepts were not based on demonstrable relationships in data sets. As an example, perhaps you might hear that it always rains at full moon or something like that, not a concept that I would endorse, but fairly easy to demonstrate or refute using weather data.

I am going through this basic overview of the theory just to set the ground work for the later discussions, any questions about "what is the proof?" should wait until we get into that detailed discussion of all the signals over the next few weeks. Also, I realize that even the more curious readers have only a limited amount of time that they can give to this discussion and so we need to move forward in small steps that won't overwhelm the flow of your daily activities, or mine (I currently have about ten forecasting duties around the internet and various other things to occupy my time, but I am taking a pause from new research in this effort to catch up through forcing myself to produce graphs -- I can of course look at these any time I want but many of them are not yet saved to files, however I do have notes on them).

So here's that basic overview, any sort of concept unique to this theory is in bold print so you can easily find it if you stumble across it again down the road in this thread.

I started into this alternative research in 1980 due to a chance encounter with a guy at Accu-weather, not the famous Joe B although he was in the room, but one of the more senior people there -- he just casually mentioned one day that he thought every time there was a full moon, a big storm hit the east coast (of the U.S.A.). I had never thought once about that subject despite having been active in forecasting for about five years. My work up to that point had been the very short range "nuts and bolts" of meteorology, issuing short term forecasts for clients interested in snowfall, rainfall amounts, wind speeds, air quality conditions, etc. I had some interest in global climate and historical climatology and that mainly in the U.K. tradition having read whatever I could find of Manley, Lamb etc. But I wasn't at that point into forecasting long-range. JH and a few other readers will realize that in 1980, 3-5 days was considered a very long way out and forecast models for that time range were not as reliable as we might consider them today. Longer term outlooks were available but considered to be a shot in the dark.

However, that remark got me interested in possible interactions between the Moon and our atmosphere. I found almost nothing in any literature about the subject, except for a few negative results a long time before I was born. Somebody had done a study on weather records for Berlin and found little correlation. As it turned out, my later study convinced me that this was probably about the worst place anyone could have chosen to look for lunar effects, as Berlin turns out to be about mid-way between two timing lines that had the researcher chosen either London or Moscow he might have found something using the same methods. But there were a few people who had found what they believed to be minor associations between declination (of the Moon) and weather patterns. So I went in that direction, because my early look at weather data with an astronomical almanac available, showed me that the associations between lunar phase and weather might not be very promising.

My first apparent discovery in this search for possible natural signals from external sources was that declination was more important than phase. It was just the accident that the Moon is always full at a northern declination maximum in mid-winter, that had led to the original observation that interested me. As we move into the spring, that northern maximum falls earlier than full moon by 2-3 days a month until it catches up to new moon in June. My research into Toronto's extensive weather records (1841 to present) convinced me that northern maximum (which I shortened to N Max) was a powerful signal of some kind that would predictably set off strong cyclogenesis near where I was in the Great Lakes region, and as I mapped out various cases I could see them clustering on what I then called a timing line which seemed to run not north-south but more like NNW-SSE then more to the SE as it passed into the Atlantic around South Carolina. So the position of this timing line was apparently from the central arctic down the west side of Hudson Bay, through western Lake Superior, then down towards Chicago and across the Ohio valley into the Atlantic. Over a few years I gradually developed a system for timing other weather events and the timing seemed to be whenever the Moon passed one of these points -- the two declination max and min (N Max and S Max), larger stars along the ecliptic plane (Regulus, Spica and a pair that are opposite each other in the sky, Antares and Aldebaran). That led me to realize that the source of the energy at the declination max and min was probably not the Moon's declination at all, but the fact that at those points the Moon is crossing the galactic equator. In other words, while there might not be necessarily one star there to supply gravitational energy, there was a source that might be comparable in strength from many more distant stars (the N Max situation also includes the possibility that nearby stars that are a bit off the ecliptic path such as stars in Orion and Sirius might be contributing). By studying a few more distant cases, I came to the conclusion that the Moon needed to move within about 10 degrees of gravitational sources to set up what seemed to be an interference pattern, if the star was further off the path than that, the effects seemed to be weak or zero.

Also I found that low pressure waves were forming at times equivalent to the Moon's alignments with Jupiter, Saturn and Mars, and not just at times when the Moon moved past them, but at the alternate time when the Moon was opposite them. Taking Jupiter as an example, I identified these events as JC and JO from the astronomical terms conjunction (Jupiter near Moon in sky) and opposition (in this case it means on the opposite side, but this may confuse some because when we talk about the earth moving past an outer planet, that is also called an opposition). So the summary is that I developed a series of "astronomical events" involving the Moon and various gravitational energy sources, and this included various pairs such as the one just described. And these would always fall at the same point in the declination cycle for fixed stars, and at slowly changing points for planets which are moving forward in the same direction as the Moon only a lot slower as we see the motion from earth.

A few years into the research (I was working full time at non-weather-related jobs throughout most of the 32 years since the start of this research) I had identified a system of nine timing lines that radiated outward from the magnetic poles. Since the one in eastern North America was of primary interest to me then, I called it timing line one and that placed timing line three across western Iceland, bisecting Ireland NW-SE then off into France and the Mediterranean. On each of these nine timing lines, the weather events or astronomical events seemed to occur on the same general timetable. However, as I studied the details, I began to find that the signals were scattered up and down the timing lines in ways that seemed unrelated to lunar orbital variables. In other words, I could say there was always a strong low on the timing lines at northern max, but predicting its latitude on that timing line was not very easy to do from just lunar variables.

Something else had to be working perhaps independently to produce the storm tracks, and I knew enough about basic meteorology to understand that this meant, in effect, something had to be producing ridges and troughs in the upper atmosphere, and shifting the jet stream out of just one "normal" position. At first, somewhat fixated on the Moon, I searched for answers in declination cycles of 18.6 years which change the range of the Moon's north-south wanderings. We are currently approaching a minimum in that cycle and in 2015 the Moon will always stay within 20 degrees of the equator. By the way, did you know that our Moon is unlike all the other solar system satellites in that it moves along the ecliptic plane (the apparent path of the Sun through the sky) and not around the planetary equator. I have concluded that our weather might be considerably less active if the Moon went around the equator like Jupiter's moons and Saturn's moons etc. But as it goes around the ecliptic plane, it also rides above and below that by 5 degrees every "sidereal month" and this cycle is just a bit shorter than the declination cycle, hence the northern latitude max happens about .12 days earlier each month; that leads to the 18.6 year cycle. At this point in the long cycle, the latitude max is a few days after southern max, when it moves back to southern max we will be at declination minimum. Then in 2024 we will be back to the declination maximum when the winter full moon is at about 29 degrees declination.

You can follow all these lunar cycles in any astronomy program, but the basic idea is that the Moon has three other significant cycles, sidereal (27.32 days), anomalistic (perigee to perigee, about 27.55 days) and synodic (full moon to full moon, or new moon to new moon, about 29.53 days).

After exhausting all possible explanations, by about the mid to late 1980s, I started to notice that some patterns were discernible in temperatures over the cycle of Jupiter's synodic year of 13.1 months or 398.8 days. That's the time between Jupiter oppositions, on average. Jupiter takes 12 years to orbit the Sun and so it has moved about one-twelfth of an earth year forward each time we catch up to it. So this realization led me to the concept of field sectors in the solar system magnetic field, and that seemed to be supported by developing research about radial sectors in the solar wind. It made sense to me that if anything could modulate the solar wind (make it other than uniformly radiating out from the Sun) it would be the two largest planets in the solar system. I also noticed that the alignment of Jupiter and Saturn every 9.93 years (every 19.86 years Jupiter passes Saturn, and halfway in between they are opposite each other in the solar system) was related to the sunspot cycle, at least statistically, and so I began to research what if any predictive value there might be to a study of solar system magnetic field sectors. The research led me to theorize that each planet (most of my early research was only done with Jupiter and Saturn in mind) generated or at least was blessed with a four-sector system in which two sectors were almost linear and the other two were curved ahead of the straight line joining the planets with the Sun. This four-field system showed up as four distinct warmings over the 398.8-day J-year period in the Toronto data.

By this point, I had committed to the concept that timing line one must represent a strong point in the magnetic-oriented weather system and so not only was it accidentally named timing line one from my location near it, but it had the significance of being a location at which primary effects would be seen -- in other words, whatever I postulated might be the connection between atmosphere and solar system field sector, the effect would fall onto timing line one, and similar effects elsewhere must occur as the result of either prograde or retrograde drift of these effects. Today, I have a more complex understanding of this, namely, that some motion of effects takes place but also that some timing lines, probably 1, 3, 6 and 8, were acting as primaries and that external energy would concentrate near them in a four-wave pattern. But this becomes so complicated that it is easier to say simply that the atmosphere is responding to external signals in a complex pattern and that we can decode the pattern by studying signals at well-chosen locations in a grid -- and that led to the realization that the human life span had become my biggest single obstacle to totally unravelling this very complicated mystery of a global or at least hemispheric atmospheric response to the external signals.

To the extent that I could easily summarize the concepts of motion of prograde and retrograde field sectors, I would say what's in the next paragraph, but I should say first that the whole point of that dynamic form of the theory can only be understood from the data, then it seems a lot easier to follow. From first theory, most of this seems very obscure.

Retrograde field sectors are presumed (from observation and data analysis) to be linked to field sectors moving faster than earth and therefore linked to Venus and Mercury. These interactions, by the way, are not presumed to be gravitational, but electro-magnetic. It's not just the limited size of the planetary disk that interferes with the solar wind but presumably a larger magnetic interference pattern. For these retrograde parts of the larger system, latitude is key, and the geometry of the orbits of Venus and Mercury assure that winter retrograde events will be at a high latitude and generally gaining in latitude as they move west. This is interesting in very long-range climate terms because there's a long cycle by which all of this reverses over cycles of about 26,000 years so that one could speculate that in ice age climates, there may be times when winter retrogression is stronger at lower latitudes than it is nowadays.

Anyway, I have come to associate the strat warm phenomenon, strong ridge-building in high latitudes and other winter synoptic favourites, with those times when the retrograde index values are high and especially when they reach a critically high latitude, in layman's terms, the field effects in the atmosphere may tend to shoot over the north pole and suddenly disappear from the system for periods of several days and I believe this may produce the circulation reversing processes that other workers are studying under totally different paradigms. Of course I am very open to collaboration about that in either direction as I learn more about their research. But this tends to give my methodology the advantage of having some concrete timing available, and at a longer lead time. The same general concept can also be invoked to explain why the blocking can sometimes break down rapidly as the signals return from their polar overshoot and resume westward motion, usually by this time well to the west of timing line one. So for the retrograde index values, the primacy of timing line one seems very significant, and my data profiles tend to confirm that the process at timing line three is earlier than at timing line one, indicating westward motion.

Many other components of the model are prograde and the signals are comparable in different timing sectors. I am still trying to unravel whether this is due to a uniform dispersion into the atmosphere or the coincidence of two data sets being one-quarter of the total period apart in longitude (hence they are in sync but in different field sectors).

The general theory to be remembered as we move forward is that large-scale atmospheric features are taken to be reflections of the particular structure of the solar system magnetic field, especially the sectors close to the earth at any given time. In broad general terms, a ridge on timing line one indicates that the earth is moving through an enhanced region (in space) of solar wind, and a trough (larger than the shallow mean-500-mb trough) on that timing line indicates a weaker outflow than average. These variations may only be on the order of half a per cent or so, we're not talking about the Sun being highly variable in its output. There may also be unresolved (in this research) variations in the type of solar wind responsible for these effects, and perhaps this is all to do with geomagnetic variables.

My philosophy at this point is that perhaps I can explain some of the variations, and perhaps some are still mysterious, but given the complexities and time scales of human life etc, the most promising path forward is creating a working system of forecasting based on index values, and not an esoteric understanding of that which may require centuries to understand in any detail, which may also require space monitoring and the research work of trained astro-physicists (while I have become very familiar with solar system astronomy, I don't have any great educational background in theoretical physics or magnetic field analysis, what I know about that tends to be the general level encountered among weather people).

I have done some work on the postulated gravitational interactions involving the Moon's interference patterns, and the best fit for all the observed effects suggests an interaction that operates not over distance or distance squared as with gravitational energy or force, but over a scale of sixth root of mass over distance, a function which greatly reduces contrasts between sources in terms of their different masses and distances. This may have something to do with gravitational waves. But the more practical problem for my research is to relate the lunar events to this other body of work on field sectors and atmospheric responses.

That is also better served by practical, empirical research that assumes an interaction and searches for the details in case studies of analogous past years. In other words, I am dedicating my research time to getting a working system together and leaving for some unlucky future researcher(s) the complicated questions of what makes this run. I have my general suspicions but there is so much detailed work to be done on creating working index value data sets that I have tended since about 2006 to move entirely in that direction and away from the theoretical reasoning. This has been promising in these apparently intermediate stages and I think we're seeing improvements in these forecasts each year but of course I'm aware of errors that crop up, my expectation is a very slow improvement in terms of percentage of days on the right side of normal in a forecast period, moving up hopefully through the 60s and perhaps into the 70s. When this technique gets into the 80s on that sort of validation, it will be edging into the reliable frame. I have had some really encouraging hits with this method in recent years, notably, the severe heat waves in the U.S. last summer which I predicted and which I believed that I understood from the theoretical perspective, hence there was even a fairly good second-order prediction of how the pattern would evolve over time. It was, by the way, the result of prograde and retrograde warmings overlapping over Kansas leading to extreme heat and drought.

Okay, I realize that this overview is a bit disjointed perhaps and I can promise to answer a few questions but would also say, if you get the general idea here, that the forecasts are based on the combination of many different (and to some extent independent) index values, then the real interest should be in examining these index values and how strong or weak the signals are. What are all these external signals forcing our atmosphere into patterns that produce anomalous weather patterns? Is it true that if the earth was alone in the solar system, the weather would perhaps be almost featureless? Probably not entirely so, and this leads to my final observation in this overview. My research is based on the assumption that external signals are an important variable but I have never been able to convince myself (let alone anyone else) that these external signals are the whole story and that if you can find all of them (already I have a list of over a hundred) you can model all variations in the atmosphere. There very well may be large percentages of variation that are only explained by teleconnections of processes within the earth-ocean-atmosphere system and that are independent of this research. There may be questions of solar variation that are not handled by these index values and (worse still) that could remain quasi-random for predictive purposes. On the other hand, and this would be lucky for me or more likely for whoever is doing this research in 2040 or 2080, perhaps this system will expand out to cover almost all variation. We know for certain that major volcanic events would interfere with this prediction system, and so that's one factor that cannot be handled by this theory. We suspect that changes in ocean circulation may be partly related and therefore partly predictable -- I will try to illustrate that towards the end of the entire discussion -- but perhaps some elements of ocean-atmosphere coupling remain entirely outside the scope of this research. The only way to find out will be to continue on with the identification of valid index values and the study of validation statistics on the forecasting. I honestly don't know where it might lead or end up. This may be already the theoretical limit, or we may see better results in the future.

Once you get used to what I am showing in the days to come, you might have insights that are valid as to how to improve the model's overall accuracy, and at some point those results could become good enough to convince many skeptical observers that this is actually the best way forward in the challenging field of long-range forecasting. I hope so, but at this point in my life, I mostly want to get this material into a format where, should I happen to lose my life or at least my ability to do detailed research, at least somebody else could pick up and move forward. So that's the main motivation for the thread.

The next installment (which may be in two days rather than one) will begin to show graphs of the lunar signals. It may take a week or two to move through those before we get into the solar system magnetic field type signals that I now believe are the bigger and more significant part of the theory and the forecasting (but the interplay between them and the lunar signals is the way forward if you want to speculate about details in patterns). So I expect that if I can keep at this on a regular basis, we might have the more significant parts of the technique covered by mid to late February.

  • Like 3

Share this post


Link to post
Share on other sites

Are you able to go backwards in the record, and test your theories against years such as the winters of 62/63 or 88/89 here in the UK... or look at specific events such as recent SSWs? Assuming it is possible to do this, what kind of success % in broad terms do you see across the historical record?

I need to read up a lot more as to what the solar wind actually is... but I assume the 11 year (and other) solar cycle is a part of your model? If so how do you quantify predicitions of cycle taking into account the variation we can see already in the current SC24 where prediction was way off and the cycle seems pretty quiet... if not very quiet?

Finally from this post it is clear to me I need to get a better 3D appreciation of the solar system. Do you have a suggested text or site that would start me off here, bearing in mind I know virtually nothing about it (although by chance my Dad gave me a rather impressive looking telescope for Xmas!!)?

Thanks for your time on this.

Share this post


Link to post
Share on other sites

I will get into posting graphs over the weekend, have some just about ready to go ... t.b.h. feeling a bit less energetic than usual all day, there's a nasty bug going around here and I think it is making a run on my anti-virus program (vitamins etc) ... and so I won't burden you with any new reading. Except to say about the questions above, yes I have back-tested the research model, I am not up to this past century yet, working my way rather steadily through the 1800s after finding reasonably good results in the decades checked so far. My method for doing this will change when I get to about 1890, to test with only the data already recorded, as well as the operational current model. It isn't worth trying to use only expired data near the beginning of the series, you pretty well have to use future data for those years if you're going to test them at all. So of course that makes the test semi-dependent on what is being tested, but as the year in question moves forward it takes up a smaller percentage of the data. The real test in back-casting will be those years you mentioned and other extreme cases. I have looked at them informally (meaning just the larger index values) and things looked promising. My methodology did give a warning on the cold spell in late Nov and Dec 2010 as well. However, it suggested reloads of that severe cold that failed to materialize later in January (except weakly).

The 11-year solar cycle is really a composite of two regimes, a faster 10-year cycle that dominates active centuries (the period from 1917 to 1989 was 10.3 years) and a slower irregular cycle that appears in weaker periods like the Dalton (note the long intervals 1801, 1816, 1830) and the occasional low or absent activity of a singularity like the Maunder Minimum. This appears to have been the case in the more distant past as well, from indirect evidence such as auroral records. The year to year or month to month variations are probably somewhat over-rated as causes of weather variability, but I would certainly agree that a longer period of low activity is well correlated with colder climates. I am studying my methodology to see if there are in-built index values that would explain these phenomena without modelling them separately since that modelling would have to be either driven by actual solar activity values or modelled statistically. I have one index value worked out for weak solar cycles but it is based on so few cases (four) that I am not very confident that it means anything in detail. If you follow the link in my signature you'll find the daily normals for the Dalton minimum period and these are generally quite a bit lower especially in winter, than even the long-term averages for 240 years let alone the modern climate period.

Anyway, the first subject to be discussed in detail will be lunar declination. Given that it's a weekend, let's say this will be up and running by Monday for sure, perhaps Sunday if I get enough time. I plan to spend a lot of time on this project during the weeks to come, so once we get going, there should be lots to see. However, I will try to limit it to 2-3 graphs a day with a brief discussion.

  • Like 2

Share this post


Link to post
Share on other sites

As far as the moon is concerned I have noticed from time to time that the weather changes from settled to unsettled, or vice versa at the time of a full moon, though I have never made records.

It seems that with the progression of time more is being learned on how the condition of the stratosphere affects the weather here below, so it would not be so far fetched to visualise the magnetosphere influencing the stratosphere.

Share this post


Link to post
Share on other sites

Thanks for sharing your thoughts and discoveries again Roger. Always an interesting read.

Share this post


Link to post
Share on other sites

Just getting ready to post some graphs on the lunar signals, but this map will illustrate the location of timing line one in eastern North America. Observation over many years has revealed that the timing lines have a tendency to oscillate east-west. The line marked "50" on the map is the equilibrium position of timing line one. The line just to its west marked "45" is an observed position in Jan-Feb 2010 and the loop is a postulated motion of lunar energy events showing a further second-order displacement during the Moon's orbital cycle.

The numbering system is known in the theory as timing number and the convention is that each timing line defines timing sectors, when the timing line is at equilibrium it has timing number 50, and timing number 00 would be the mid-way point west to the next timing line (which in this case is timing line 9). Thus any given location would have both a defined meteo-latitude (for the Great Lakes, due to proximity to the NMP, meteo-latitude is about 7 deg higher than terrestrial) and timing number.

The timing lines do not all oscillate in phase, the system seems to respond to changing positions of field sectors. I will post another map during the discussion of timing line three and its proximity to the UK and Ireland.

post-4238-0-47965500-1357524909_thumb.jp

Edited by Roger J Smith
  • Like 1

Share this post


Link to post
Share on other sites

Discussion of signals in CET and Toronto temperature series

(part A -- lunar signals)

note that part B -- solar system magnetic field signals, will follow in about a week

__________________________________________________

A-1 Lunar declination

As discussed, the Moon has a 27.32(166) day cycle of declination. If the Moon were moving exactly along the ecliptic plane, the annual path of the Sun through the sky, that declination would regularly range from 23.4 deg N to 23.4 deg S every 27.32 days. The "northern max" would be in the position of a 21 Dec full moon (or the 21 June solar position) while the "southern max" would be in the position of a 21 June full moon. These happen to be positions where the ecliptic plane crosses the galactic equator. I have defined the actual positions of N Max and S Max to be where the Moon achieves a position of 6h and 18h R.A. just for anyone checking tables. A good approximation if you have this sort of guidance is where the Moon enters Cancer or Capricorn (some tables are set up that way).

This diagram shows the ecliptic, the location of some fixed stars (in red) and the current locations of slowly moving planets in green (they are all moving like the Moon, from right to left across the diagram). Notice that planets are not quite on the ecliptic plane. Jupiter is approaching it now from the south and will cross near the northern max position in summer-autumn 2013. Saturn recently reached its northerly latitude max and is slowly returning closer to the ecliptic. Positions where the Moon or planets cross the ecliptic are known as nodes (ascending and descending). An object's inclination measures its orbital angle relative to the ecliptic. Jupiter has a rather small inclination of less than 2 deg, Saturn is closer to 3 deg, Mars, Mercury and Venus achieve inclinations higher than 5 deg and some asteroids move 30 or even 30 deg above and below the plane. The Moon has a range of 5.1 deg and its ascending node, currently located near Spica, moves back along the ecliptic in the opposite direction to prograde motion. This cycle takes 18.6 years to accomplish and it explains a declination cycle of that length, during which maximum declination ranges from 29 deg down to only 19 deg. At present the range is about 21 deg and falling (the minimum is in 2015).

(note for those unfamiliar with astronomical terminology, latitude measures departures from the ecliptic while declination measures departures from the equator. Meanwhile, right ascension goes with declination and longitude goes with latitude in terms of how they relate to the sky -- right ascension is measured along the equator and longitude extends from the ecliptic. But as the ecliptic is close to crossing the equator at the zero point of both R.A. and longitude, they are only marginally different for objects near the ecliptic. Right ascension is measured in hours and when the Moon is at its highest declination (N Max) it is near 6h R.A. -- to further familiarize yourself, on the diagram, the vernal equinox is the point near the right margin of the diagram where the ecliptic rises above the equator. The autumnal equinox is the point where it falls below between "R" for Regulus and "Sp" for Spica. )

post-4238-0-78562300-1357531832_thumb.jp

Edited by Roger J Smith
  • Like 1

Share this post


Link to post
Share on other sites

The blue semi-circle on the diagram above shows the current (7 Jan) location of the Moon as well as, more importantly to our discussion, its location on 1-1-1772. Now bear in mind that the Moon could be at any instant 0-5.1 deg above or below this path, which is scaled to 23.4 deg up and down from the celestial equator. That would be a second-order variation at most, and we will discuss its significance after establishing the nature of the signal for the declination cycle.

Since the Moon was in that position at the start of the CET daily series (1-1-1772) the signal places S Max on day 4 and N Max on day 18 of the 27.32(166) day signal. The data are arranged in groups of 82 days, then days 1-27, 28-55, and 56-82 are separated to create the cycle. About every six years a day is dropped from the longer cycle to keep the numbers aligned. So when you look at these graphs, bear in mind that day 28 is based on about 1/3 as much data as the rest (just like leap year day is based on 1/4 as much data). A large variance on day 28 may not be quite as significant as any other day's anomaly.

The declination cycle will contain any semi-permanent signals from fixed stars, and the galactic equator, which are always encountered in similar timing positions. Over the entire 240 years of the CET (and 172 years of Toronto data) I have found that these signals are rather small but they get a lot larger when you group similar years where moving parts of the model (planets and therefore solar system magnetic field signals) are in similar orientations to the Moon's orbit. There are also three other lunar cycles of roughly equal significance to the declination cycle that are running on different time scales. So the small declination signal is really about one-quarter of a larger lunar signal that is based on declination, phase, latitude, and perigee-apogee or distance. We'll go through these, but first, let's just look at the declination signal in the data.

The first graph below this text is the raw signal for the CET, first showing it on the standard graph (1.0 to -1.0 anomalies) with two historical intervals for comparison (Dalton 1796-1819, and modern 1988-2011), then just the lunar declination signal on a 10x magnified graph (so the range of the graph is +0.1 to -0.1). Notice that the peaks in temperature over the long term have been near S Max and N Max but there is also a historical trend for temperatures to increase in the period after N Max, more so as the climate warms. I will give a possible explanation for that in the next post where I can compare the CET signal to the Toronto signal.

GRAPH 1 Below -- 240-yr CET lunar declination signal, with two historical intervals (see text above)

GRAPH 2 Below -- 240-yr CET lunar declination signal, on 10x mag graph

post-4238-0-07507000-1357527370_thumb.jp

post-4238-0-68453800-1357527420_thumb.jp

Share this post


Link to post
Share on other sites

We return to "While Britain slept" with this comparison of the CET declination signal and the Toronto signal. To achieve a valid comparison, I reduced the CET period to 1841-2011 and that matches the Toronto data (all 2012 data are stored in these files but not incorporated into the signals, although one year makes no visible difference to signals based on this much data).

In this comparison below, the first thing you may notice is that the Toronto signal (shown in lighter colours, on a purple-green scheme against the CET red-blue) is somewhat larger in amplitude.

I've always defined amplitude of signals as the departure above or below a mean (not the range from top to bottom) so for the lunar declination CET signal, that was about .06 C deg ... for the Toronto data it is closer to .17 C deg and as the 1841-2011 period is shorter, the CET amplitude is slightly larger but only .11 C deg. The Toronto data started on day 13 of the established CET declination cycle so it was adjusted to the same period as CET. Just a note, when I talk about any given year's declination cycle, that is actually the standard declination periods with the earliest start date in that year, and ending with the end of the data in the early part of the next year. The data tends to be arranged in groups of 14,15,14 cycles on an irregular basis (some declination years are 355 days, some 383 days).

The two signals are varying in phase and look rather similar. The Toronto data do not peak as high at N Max -- this is because in colder than average weather the N Max event almost always runs south of Toronto and reinforces cold weather (there is a tendency for this to happen in coldest UK patterns also). So in warmer months there is often a strong warming peak at N Max. Remember, these are signals of blending of many different weather patterns, showing a background effect that the Moon has. What is much more significant for the purposes of long-range forecasting and climate modelling is how the Moon interacts with different pattern set-ups.

But we have to take this one step at a time, and this step shows a similar forcing (albeit on a small scale overall) by lunar declination, with the similar observation of a warmer presumably zonal flow that develops after N Max and ends with S Max. Here's what I think that means. For both climate zones, a large ocean to the west is a source of mild air masses. The Pacific may be further from Toronto than the Atlantic is from central England, but given the stronger jet streams often found in the mean trough at the base of the geomagnetic trough that contains timing line one, the timing difference may be slight. Northern Max is presumed to be a period of rising heights of subtropical highs in both oceans. The period around N Max + 7d (which happens to correspond roughly to the SpC event) is a time where lunar tidal or gravitational forces are pulling south and presumably the heat discharged by the swelling subtropical highs is then forced into warm sectors of stronger than average lows riding over the crest of the ridges produced. This creates a mean warming signal that is more evident in warmer periods than in colder periods (when the warm peaks tend to retreat to their easier forcing periods of S Max and N Max).

GRAPH below shows the comparison of CET and Toronto lunar declination signals in the period 1841 to 2011, each scaled to a zero-anomaly. The actual mean anomaly of the CET signal in that period was +0.10 C relative to 1772-2011. The Toronto signal was calculated 1841-1994 and had drifted up to +0.03 by 2011. Both were reassigned to zero internal means.

This graph is normal range (+1.0 to -1.0 C)

Toronto data are in the hatched bars with purple and green colour coding, the CET data are as in previous graphs in red and blue bars. The dates of S and N max are normalized.

post-4238-0-70999100-1357528476_thumb.jp

Edited by Roger J Smith

Share this post


Link to post
Share on other sites

The utility of these lunar declination cycles would be very limited if they were always in this small range, but the research has shown two things, first, that the declination signal is embedded in a more complex lunar signal based on other variables as mentioned above (phase, latitude, distance) and also, the declination signal itself looks much larger for groups of similar years. When you study smaller periods such as 24 years or 8 years, the amplitudes increase to about 0.5 C deg for the declination cycle. But for individual years, the typical profile of 14-15 cycles covers a range of over two C deg with amplitudes typically 1.3 to 1.5 C deg.

Referring to the CET data again, the only year I could find that managed to stay almost within the bounds of our standard graph, +1.0 to -1.0 C, was 1923 ... and that graph is below as an illustration -- bear in mind that almost every other year would show a larger variance than this. Day 13 on this graph cuts off a value off-grid (-1.13 C).

GRAPH below is the lunar declination signal for 1923 (anomalies relative to 1772-2011)

and value for day 13 is actually below edge of grid at -1.13.

post-4238-0-56235000-1357529249_thumb.jp

Share this post


Link to post
Share on other sites

The above seems like enough for one day of reading and so I will pause here ... the next data to be illustrated will be the remaining three lunar signals with a discussion of how that part of the research model fits together and how it can be related to the other (and apparently larger) set of signals from the solar system magnetic field. The two sets are related, in fact, when you combine years with similar field positions, and study just the lunar signals, you find that they are often closer to an amplitude of 1 C deg than the background cycles and their modest amplitudes. I will show an example of that near the end of the lunar discussion.

Since we are now into the discussion of lunar cycles, I will mention that the research has uncovered a difference in signal from Toronto to CET, in that for Toronto the various events in the lunar cycle show a peak almost at the same time (lag being a few hours due to the position of the timing line) but for the CET data there is a tendency for a weak signal at event time and a stronger signal later by 3-4 days. This probably illustrates that the system is in fact geomagnetic in its driving mechanisms and therefore the signals are modulated by a lag time from upstream stronger events that may be considerably stronger than the events created directly on the various timing lines. In other words, the system may reflect hemispheric energy processes and not local weather cycles near timing lines.

It should also be kept in mind that this is a study of all data over the full year and you can see slight differences if you look at the stats for seasons or months. However, that is partly because lunar phase always has the same orientation to declination at those times of the year, so separating the two out is the first challenge.

Temperature is not the only element that can show a signal based on declination. I found that there was a 15-mb pressure fall on average within 24-48h of northern and southern max for Malin Head (Ireland) in the period 1973-2008 in the mid-winter months when these declination maxima are near full and new moons. There is also a 5-8 mb pressure wave evident in the Toronto data for some of the lunar events. There is a 3-5 times background precip signal for a full moon - northern max combination in Toronto data from 15 Dec to 5 Jan when these events are well-phased. And bear in mind that not every event hits Toronto or any other observing site, wherever you situate the research, some events will miss north and some will miss south. That latitude variation is probably controlled by non-lunar factors so the research should attempt to track the mobile signals and assess their actual strength, and on that basis, the signals are probably a lot stronger than any static data analysis would show. My estimate is that the average northern max low crossing timing line one is 15-20 mbs deeper than background pressure and extreme cases include (for Toronto in particular) some all-time extreme values. The number of temperature and rainfall records at N Max is clearly non-random and indicates that this is a strong atmospheric event especially in North America. I have seen a strong correlation with severe weather outbreaks with this event.

Share this post


Link to post
Share on other sites

A-2 Lunar Phase signals

With reference to the ecliptic plane diagram, the Moon requires 27.32 days (its sidereal period) to move through that cycle but in that time, the Sun edges forward almost 8% of the distance around the cycle, so the synodic period is 29.53(059) days. This creates an independent signal for lunar phase that once again is presented in a format that is determined by the start date of 1-1-1772. As we mentioned, southern max was on the 4th of January in that year and the new moon on the 5th. So the standard graph period that is generated by this cycle runs from four days before new moon to about ten days after full moon.

Note also this graph extends 30 days compared to the 28 days of the declination cycle graphs.

The letters n and f depict dates of new and full moon in the cycle.

The signal for lunar phase is almost twice as large in amplitude as the declination cycle (for both the CET series and for Toronto) and this makes it less necessary to show a magnified version, so here are the standard-graph views of the CET lunar phase signal and the comparison with Toronto (this time just comparing the data sets and not for period). The signals are almost out of phase which is interesting, perhaps it leads to some research possibility indicating an east-west oscillation over the Atlantic basin.

GRAPH 1 below is the CET 1772-2011 lunar phase signal on a grid from +1.0 to -1.0 C,

GRAPH 2 below compares that with the Toronto 1841-2011 signal, which is shown in lighter colours and which has generally a larger amplitude as well as being out of phase

post-4238-0-14269800-1357616610_thumb.jp

post-4238-0-51555500-1357616784_thumb.jp

Edited by Roger J Smith

Share this post


Link to post
Share on other sites

Now, while it is true that lunar phase is independent of declination over the full range of annual data, there is a dependent relationship for any similar interval of the year, especially one defined in a narrow range of dates. If we take the data for a given month, then if we have chosen declination data, the new and full moons will occur in a range of about 2 days in that data, and vice versa, if we take phase data then the northern and southern maxima will occur over a range of about 2 days.

The following chart shows the first declination cycle of the year, mostly January and some early February CET data determined by dates of the first southern max in the period 4 Jan to 31 Jan, and starting three days before that date. This generally places the new moons in the range of days 5 to 7 and the full moons in the range of 19 to 22. For each "month" of the declination data a more month-specific hybrid profile of declination and phase becomes possible, and for more specific research or forecasting interests, you can narrow down even further and get such narrow ranges of start dates that declination-phase are locked into daily ranges.

For the January data shown below, in general the S Max warming is less significant than in the annual average, and it has become evident from my research that S Max is a better "warming" event in spring and summer. I think this may be due to average trajectory of the southern max low pressure event tending to be further south than N Max and giving more frequent cases of a track south of the UK. Conversely, the N Max does better in January (and December) at raising temperatures than in other months of the year. However, it should be added that in a separate study of coldest winters, N Max is only a feeble warming which probably indicates that in the coldest patterns, N Max events go south as well.

The two temperature spikes after N Max are thought to be produced by RC and SpC events generated on either timing lines one or two, since they show a similar pattern to the Toronto data but lagged by 2-3 days.

Note that the amplitude of this January declination-phase hybrid is almost 0.3 C deg, considerably larger than either the annual declination or phase signals. When smaller samples are isolated based on other considerations of analogue, signals begin to get interesting for a long-range forecasting application, up around 1.0 C deg.

The second graph shows this CET signal again with the Toronto signal added for comparison. The main differences are that southern max and more particularly new moon (in January) provide warming signals, and there is a very pronounced cold surge before northern max, a time when a strong arctic high often moves southward to reach the east coast by the time of N Max by which time it has become a warming signal. The same process over the North Atlantic would probably provide a much less obvious cold signal as the air mass trajectory would be from Greenland to Ireland (rather than from northern Alberta to the Great Lakes). So it's possible that some comparisons of lunar signals need to be run through a climatology filter before we conclude that different processes are underway. As stated for the CET analysis, there are regular 3-4 day warmings after the N Max - full moon period that are timed (in Toronto data) for the RC and SpC events. These become a lot stronger if some other signal like a JC event or lunar perigee overlaps.

One other singularity to note in Toronto data, and possibly for CET data although I haven't been watching daily weather in the UK for as many years -- the N Max full moon and S Max new moon combinations decouple around 5-8 January -- before that (from the December solstice or a few days before actually) they are almost simultaneous, by late January they are separated by 2-3 days which is enough time to place the energy peaks on two timing lines -- so around 5-8 January they are forced into a double-low formation that you can expect to become increasingly decoupled moving into January. This may explain why the UK and Ireland climates seem to feature frequent windstorm events that are sometimes in pulses every 24 hours as we saw last month at the N Max (27th) and full moon (28th) events.

In the Toronto data, I have noticed that N Max and full moon events seem to lose intensity after that decoupling, then hit another peak around 21st to 25th of January. This may be related to (as cause rather than effect) of the much-discussed January thaw phenomenon. However, I think it's more to do with timing line separation and perhaps a secondary energy peak from a lunar-stellar event in Gemini that is otherwise too weak to generate much of a signal. There's another energy peak for full moons near 20 Feb due to superimposition of the full moon and RC event. This got very active a few years ago when other events were also superimposed there. The general conclusion I've reached is that the phase-declination combination drives a lot of the active day to day weather but the background pattern is still from another source, namely solar system magnetic field sectors -- and this makes the Moon important but not definitive in its role in creating weather variations.

If the thread remains active, I will post other monthly segments of this hybrid data, the other "months" don't all look like January, and there are some interesting singularities, such as a warm southern max peak for May, and a cold spell around the full moon (before N Max) in Novembers.

GRAPH 1 below shows the January hybrid of declination and lunar phase for the CET data 1772-2011

GRAPH 2 below shows the Toronto hybrid of declination and lunar phase for 1841-2011

post-4238-0-97156200-1357617314_thumb.jp

post-4238-0-94489800-1357620062_thumb.jp

Edited by Roger J Smith

Share this post


Link to post
Share on other sites

A-3 .. The lunar distance (perigee-apogee) signal

We have investigated signals for lunar declination (27.32 days) and phase (29.53 days), but another variable of the lunar orbit is distance from earth, which averages 384,000 km but varies by about 25,000 km either side of that average, every "lunation" (a term meaning lunar orbital period). The perigee occurs 0.233 days later in each sidereal period, on average, which gives us the anomalistic period of 27.55(455) days. That forward motion takes it around the ecliptic once every 8.87 years, or close to 8 and 8/9 years, so that similar years of lunar perigee (relative to declination) occur every nine years, with a more precise schedule 0,9,18,27,35-36,44,53,62,71,80 years. While not exact, this 80-year long cycle is the basis for the perigee data study in the CET data, and makes 1772 a similar year to 2012, namely, with perigee moving from about 6 days before southern max to 3 days before.

The location of perigee in the declination cycle seems to produce a 4.43 and 2.22 year modulation in temperature and therefore could be part of the QBO forcing. When perigee is close to N Max or S Max, there are years that on the average are 0.5 to 1.0 C deg warmer, and when it falls mid-way between them a secondary warming is observed (this is true of either Toronto or the UK data).

Perigee over the whole period of 240 years has produced a very regular cycle where a warming takes place from about 1-2 days before lasting about a week, and boosts the temperature about 0.12 C deg for the CET data and about 0.20 C for Toronto (scaled to climate, these are about the same range).

This is illustrated below -- the first perigee in the data set was on 27 Jan 1772 so the letter P is placed on the graph (once again full sized, no magnification) to show where perigee occurs, the warming is mostly in the first few days of the period which is on a 28-day graph since the period is 27.55 days.

I have not shown Toronto data this time because except for the higher amplitude, that looks very similar. Instead, I have added the second half of the CET data signal to the graph, that's the fainter lines above the bar graph segment. Note that all we've seen happen to this signal in the second half of the data (1892-2011) is a general warming of about 0.3 C deg.

The use of the perigee signal in actual forecasting is more limited to case studies of similar orientation to both phase and declination -- that provides some stronger signals and begins to amplify the case studies towards 0.5 to 0.7 C deg amplitudes. But it should be noted that perigee seems to act as a general "height sweller" and if there is strong northern blocking then a perigee signal can begin to reverse -- the entire data set probably contains a small percentage of these reversals and so the "normal pattern" perigee signal is probably bigger than this composite.

The physical explanation of warming at perigee, if it is indeed a global signal as I suspect, is fairly basic -- closer to the earth, the Moon can create stronger waves, these travel faster too, and so their ability to transfer heat is greater. The reduction ad absurdum would be a much larger orbital range where the Moon came within say 100,000 km of the earth. This would create enormous tidal forces and from my theoretical perspective, intense storms beyond what we've ever seen, therefore no doubt some very prolific warming signals. So this current rather sedate 7% variation around an orbtial distance mean is just a shadow of what that might look like.

GRAPH below shows the signal for lunar distance with day 27 shown as date of lunar perigee

This is for the CET period of record 1772-2011 with the second half of data 1892-2011 shown in the upper section of the graph. Scale is normal (+1.0 to -1.0 anomalies)

post-4238-0-55792800-1357621282_thumb.jp

Share this post


Link to post
Share on other sites

A-4 .. Lunar latitude signals

This is the final section on lunar signals -- will post a summary and research directions the following day.

Just an aside, the Moon's current position is at the "JO" event and in about 24h (less for you reading this) the Moon will pass Venus and reach S Max. New moon is on Friday. As we move ahead with this discussion, I will continue to update on lunar events.

Now, readers may recall that lunar perigee moves forward gradually around the ecliptic every sideral month, resulting in a cycle of 8.87 years for lunar perigee. Meanwhile, the ascending node of the Moon's orbit moves slowly backward along the ecliptic and takes 18.6 years to complete a circuit. As noted earlier, the Moon goes around the ecliptic and not the equatorial plane (in space) and also has a latitude range of 5.1 deg above and below that plane. In Jan 1772, the Moon was approaching a minimum of declination range, and was at its highest latitude about 2 days after Southern Max. This year we are about 1.5 years earlier in the cycle, so fairly similar, the highest latitude occurs about 5-6 days after Southern Max. Next year (2014) will be more similar to 1772 and by 2015 into early 2016 we will reach the declination minimum.

This cycle of 27.21(221) days is known as the draconic period (cycle). Similar to perigee, there is a warming associated with the northern latitude maximum, and it seems to lag behind the peak in the CET hitting about when the Moon is returning to the ecliptic from its 5 deg latitude max. The amplitude of this signal is similar to declination and a bit smaller than perigee or phase signals, at 0.08 C. On the graph, I have shown four key points in the 27.21 day cycle, the latitude max (+) and min (--) and the two points where the orbit intersects the ecliptic (e) -- these are the descending and ascending nodes in the order shown.

The Toronto signal from 1841 to 2011 seems to have less lag and is a direct response to the latitude cycle. It has an amplitude of almost 0.2 C deg. This may indicate that the CET warming is not created in the local climatic system(s) but is imported from a signal upstream. More research on intermediate points such as St John's Newfoundland might help to clarify this process.

Another unexplained detail is that the CET latitude signal is stronger in the period April to November and it virtually disappears from the data in the winter season including March. Perhaps this indicates a preferred path for dispersion of the signal upstream because there is no great difference seasonally in the Toronto signal.

The graphs below show the CET signal and then the comparison of that with the Toronto signal, using the same colour codes as previously.

Despite two attempts, the graphs have decided to perform their own little left-right switch. They may appear in either order by the time you read this post, so the CET signal is the one with only a red-blue colour code and the comparison has the second (Toronto) colour code added.

post-4238-0-14685400-1357702309_thumb.jp

post-4238-0-76758800-1357702571_thumb.jp

Edited by Roger J Smith

Share this post


Link to post
Share on other sites

The interplay of latitude and declination produces the 18.6 year cycle -- even some ancient peoples noticed this and found it significant. Around 1950 to 1970 some prominent climatologists were studying the relationships between the cycle and weather patterns. Nothing very definitive was produced except for Reid Bryson's observation that at maximum declination range the atmosphere seemed to be more prone to meridional blocking patterns. The two coldest winters in modern times in western North America occurred near this declination max (1950, 1969). Here's a list of all declination max years since the CET began (going back to the pre-daily period) and the decimal points add some precision, these are interpreted as a portion of the year so 1950.2 would mean March 1950, etc.

1671.2

1689.8

1708.4

1727.0

1745.6

1764.2

1782.8

1801.4

1820.0

1838.6

1857.2

1875.8

1894.4

1913.0

1931.6

1950.2

1968.8

1987.4

2006.0

There are some trends in the 18.6 year cycle that may be a result of different signals, or may just be coincidence. In North America, hot dry summers are more frequent about 3-6 years into the cycle, after the period of high blocking potential. Colder winters occur more frequently there at declination max and min.

In the UK, I've noticed in the records that average annual temperatures are a bit higher in years with low declination range (the years about 8-10 years after those listed). Also there is often one notable cold winter in the phase between min and max, for example, all of these with their declination year added (calculated from the decimals in the list, and using Year.0 as the event time except for 1947 which uses Year.1 ...

1684 was 12.8

1740 was 13.0

1795 was 12.2

1814 was 12.6

1830 was 10.0

1891 was 15.2

1907 was 12.6

1929 was 16.0

1945 was 13.4

1947 was 15.5

1963 was 12.8

1982 was 13.2

There is another less impressive clustering near max declination (e.g., 1709, 1784, 1820, 1838, 1895, 1969, 1987).

This is not a direct technique that can be used to generate long-range forecasts but it may show that direct methods are buried in the signals -- however, the more significant temperature signals are yet to be explored in this thread -- the solar system magnetic field sectors.

All for today -- a brief summary and discussion of the lunar signals will follow tomorrow about the same time.

Share this post


Link to post
Share on other sites

Wow - time to slow down. I'm at work at the moment, and going to need some serious time to get to grips and digest a lot of this. Thanks Roger... When I have got my head around it I'll tip you off!

Share this post


Link to post
Share on other sites

Yep, I could use a break too. I will resume the discussion with the solar system magnetic field signals over this coming weekend or no later than Monday your time, that will give readers some time to catch up. The only real point we can make so far is that some valid but rather small lunar signals appear to exist and that there is some degree of correspondence between signals observed in two different climate regions. Readers should appreciate that these signals alone would not be sufficient to generate the sort of robust seasonal forecasts that have been presented, and the rest of the story is the more important portion. What is a bit less clear to me as well, is how much overlap exists between lunar signals and solar system magnetic field signals, once you start to gather the second group, and I continue to look for that overlap and to understand it although in raw prediction terms understanding it is not a requirement (the overlap, that is to say, can be small medium or large but the technique remains the same).

  • Like 1

Share this post


Link to post
Share on other sites

RJS index model still shows a warming ahead before the real cold blocking structure and has and does remain a big thorn in my side re the outlook during Jan.

So with models still not agreeing I'm thinking we have some further hurdles to clear before we settle in to maintained/uninterrupted cold...

BFTP

That is true Blast - and I am a big RJS fan... but his index also showed a cold snap in middle December precisely at the time that it warmed up and got very wet. I am getting to grips with his methodology slowly on his separate thread - but it certainly isnt bulletproof.

Edited by Catacol_Highlander

Share this post


Link to post
Share on other sites

That is true Blast - and I am a big RJS fan... but his index also showed a cold snap in middle December precisely at the time that it warmed up and got very wet. I am getting to grips with his methodology slowly on his separate thread - but it certainly isnt bulletproof.

Well I am not a fan of anyone who predicts the weather by plotting invisible J rays

from Jupiter that only they can see.

06z way to progressive but then it is the 06z so not surprised by this.

What is more of a concern is the ECM still not on board but considering its

a very complex pattern unfolding with the models trying to work out how much

energy undercuts and how much travels n/ne.

I am definitely still in the UKMO/GEM camp partly because that is the evolution

I think is right and partly because this to me ties in more with the warming that

is propagating down through the atmosphere.

Share this post


Link to post
Share on other sites

Well I am not a fan of anyone who predicts the weather by plotting invisible J rays

from Jupiter that only they can see.

.

I'm sure he will lose so much sleep over that, I'll call a counsellor for him now.

BFTP

  • Like 8

Share this post


Link to post
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

Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×