Farmer's Almanac 2018-19 Tally

[Maestrobjwa]

Okay, so even if just for pure entertainment purposes...I just wanted to start a thread where I can just tally the times the Almanac gets it right and when it doesn't! (And mods, if this is out of line, I understand) This isn't saying it'll definitely be right or wrong...just a fun observance of their predictions--be it spot on or way off!

(Note: I won't be covering every prediction it makes...primarily the stormy ones!)

[Maestrobjwa]

So, to start...

HIT: For Nov. 16-19: "Some rain, wet snow, then fair/cold"

What actually happened: Rain and snow occurred just a day before this date range (the 15th)...and now we have fair and cold weather on tap for the next several days

[AfewUniversesBelowNormal]

18 hours ago, Maestrobjwa said:

So, to start...

HIT: For Nov. 16-19: "Some rain, wet snow, then fair/cold"

What actually happened: Rain and snow occurred just a day before this date range (the 15th)...and now we have fair and cold weather on tap for the next several days

I'd like to see this too. I tracked it when I was a teenager and it was ~+25% vs random

[Roger Smith]

When I was starting my research many years ago, I looked at some of their forecasts. This would have been around 1981-84. It became obvious to me that their method had something to do with picking an analogue winter and date shifting it to match lunar dates of the current winter. Here's a hypothetical example. You think 1909-10 will be a good analogue for 2018-19. Maybe solar activity and other teleconnections come into the decision too. So you look up the lunar dates for 1910, compare them to 2019, and shift the weather data already recorded for 1910 to the 2019 corresponding dates. It can't be more than 15 days different to shift either forward or backward to match them up. (I have not quoted a real analogue year here, nor have I looked up the difference in lunar dates).

Every 19 years the moon has similar dates, also 8 and 11 years are small corrections. That means little to chances of choosing those years because the moon is only viewed as a sort of modulator of the overall weather pattern, not the creator of it. So if the overall setup looks good in the analogue winter, the number of days you have to shift is not a big concern. 

I'll follow this thread, and perhaps I can tell you from the results which analogue winter they are using and reveal the date shift involved. Then you can predict what the Almanac will predict from looking at that analogue winter and shifting the dates accordingly. Let's hope 2009-2010 is it. If so, the shift will be about -9 days (lunar events in that winter ran 9 days later than this winter). That makes the period Jan 21 to Feb 19 look very enticing (because it snowed several times Jan 30 to Feb 28 of 2010). If I had to use this method alone, I might just choose 2009-2010 and date shift it myself.

The one thing that they might or might not realize is that date shifting requires some phase alterations because some lunar events shift slightly different numbers of days, so just because you have a stellar analogue ten days later than your winter forecast of choice does not mean you can do a seamless date shift of ten days, some parts have to be shifted more like eight or twelve days. And that might concentrate or defuse a storm system to have the two separate energy sources shifting different amounts. 

Just picking a year at random and lunar-date-shifting won't be very useful because you might randomly select a totally different pattern. This only works (to some extent) if you have a good pattern match. Another variable that doesn't shift seamlessly is lunar distance, the perigee cycle is independent of phase and declination. Perigee analogues are found every 9 years back (8.86 if you want to be more precise). And hey lookee there, 2019-9 = 2010. Hmm. 

[Maestrobjwa]

18 hours ago, Roger Smith said:

When I was starting my research many years ago, I looked at some of their forecasts. This would have been around 1981-84. It became obvious to me that their method had something to do with picking an analogue winter and date shifting it to match lunar dates of the current winter. Here's a hypothetical example. You think 1909-10 will be a good analogue for 2018-19. Maybe solar activity and other teleconnections come into the decision too. So you look up the lunar dates for 1910, compare them to 2019, and shift the weather data already recorded for 1910 to the 2019 corresponding dates. It can't be more than 15 days different to shift either forward or backward to match them up. (I have not quoted a real analogue year here, nor have I looked up the difference in lunar dates).

Every 19 years the moon has similar dates, also 8 and 11 years are small corrections. That means little to chances of choosing those years because the moon is only viewed as a sort of modulator of the overall weather pattern, not the creator of it. So if the overall setup looks good in the analogue winter, the number of days you have to shift is not a big concern. 

I'll follow this thread, and perhaps I can tell you from the results which analogue winter they are using and reveal the date shift involved. Then you can predict what the Almanac will predict from looking at that analogue winter and shifting the dates accordingly. Let's hope 2009-2010 is it. If so, the shift will be about -9 days (lunar events in that winter ran 9 days later than this winter). That makes the period Jan 21 to Feb 19 look very enticing (because it snowed several times Jan 30 to Feb 28 of 2010). If I had to use this method alone, I might just choose 2009-2010 and date shift it myself.

The one thing that they might or might not realize is that date shifting requires some phase alterations because some lunar events shift slightly different numbers of days, so just because you have a stellar analogue ten days later than your winter forecast of choice does not mean you can do a seamless date shift of ten days, some parts have to be shifted more like eight or twelve days. And that might concentrate or defuse a storm system to have the two separate energy sources shifting different amounts. 

Just picking a year at random and lunar-date-shifting won't be very useful because you might randomly select a totally different pattern. This only works (to some extent) if you have a good pattern match. Another variable that doesn't shift seamlessly is lunar distance, the perigee cycle is independent of phase and declination. Perigee analogues are found every 9 years back (8.86 if you want to be more precise). And hey lookee there, 2019-9 = 2010. Hmm. 

Huh...that is a very interesting theory! (Certainly would crack the code a bit...Now I'm an amateur enthusiast, so a little of your post went over my head a bit, lol But I think I kinda get what you're saying: they somehow choose an analogue year (how, btw? Similar lunar or solar records?)

And after that, you're saying the do the date-shifting thing to match the current season?

Interesting...such would imply that things in weather can repeat themselves in a certain way or something...Either way, it always impressed me when they do nail something down to the very day!

[Roger Smith]

I may be describing what they were doing 30-40 years ago and perhaps now they have other methods worked out. But if it's what I think it is, the choice of the analogue year is based on whatever considerations may be available just like you see other LRFs on this site being developed from analogue years, chosen for some reason such as similar El Nino, similar Pacific etc. 

The only difference would be the date shifting for lunar dates. But while some researchers have gone down that lunar road to a conclusion where they think the Moon is a big player, my own research points to the idea that the Moon is just strong enough to modulate the ingredients it is forced to handle by the teleconnections. So the date shifting times the highs and lows, then hopefully the storm tracks are accurate because of the analogue year being a good one. 

If you choose the wrong analogue year and date shift, you'll still see some ghostly correlations because the timing of highs and lows somewhere close to your region will be the same. But the temperature field will be all wrong. That would perhaps happen if it's going to be a cold winter and they choose 2012 as their analogue. 

[Maestrobjwa]

ALMANAC WATCH

So, I'll say this if the almanac has something for a particular date range AND current LR guidance gives even a hint at something in the same time frame! (But again...NOT a forecast or anything...just stating what it says!) 

So, for late November and December:

Nov 28-30: "Stormy; rain/snow/sleet"

Dec 1-3 "Fast-moving storm from the Great Lakes spreads gusty winds, moderate-to-heavy precipitation into New England, the Northeast."

**Could this be the clipper showing up on some models out in fantasy range? (Though the Nov. 28-30 is where it shows up...so we shall see!)

EDIT: Seems even the more conservative Old Farmer's Almanac also predicts snow during the first week of Dec...gonna be interesting to see how the LR evolves!

[Maestrobjwa]

So...since we've had some very early hinting around in the long range for some potential between the 8th and 11th...I still have my Almanac Watch in effect  

Old Farmer's Almanac says snow between 4-7th

Regular Farmer's merely says "Snow showers and scattered flurries, then clearing" for the same period; while predicting a coastal system (p-type not specified) for December 16-19th...

So both say "flakes"...with the Old Farmer's seeming to predict more of a storm (although it did not say "snowstorm" just "Snow") So we shall see! 

[Roger Smith]

So does the Almanac make any bold predictions for Jan or Feb? 

If they went with a 2010 analogue, watch for increasing snowstorm risks in their forecasts after Jan 19th. Roughly on the 2010 schedule moved forward ten days. 

[Maestrobjwa]

On 1/1/2019 at 8:03 PM, Roger Smith said:

So does the Almanac make any bold predictions for Jan or Feb? 

If they went with a 2010 analogue, watch for increasing snowstorm risks in their forecasts after Jan 19th. Roughly on the 2010 schedule moved forward ten days. 

Whoa! I have really been neglecting this this thread lately...sorry about that!

Interestingly enough...the regular doesnt outright say "big storm" for their winter predictions (although they did say snow for Jan 12-15...which is very close to what we're tracking now, lol). But overall, they're a little more conservative and less "big storm!" and non-descript than the last few years for our area. They do, however, forecast a coastal storm the first week of February.

[Roger Smith]

Maybe 1978 is their analogue then. The lunar dates in 1978 were only 3 days later than in 2019. That means an analogue to the east coast blizzard would occur around Feb 3-4. The Great Lakes superstorm timing would be Jan 22-23. If that was the analogue in play. (for me it's not in play for other reasons) ...

When did DC region have a snowstorm midway from new moon to full moon in January? That might provide a clue as to the analogue year (new moon was 6th, full moon is 21st using z time, both back up to near end of previous day using EST to 5th and 20th). I know it wouldn't be 1996 as that storm was just a day or two after full moon. 

[Maestrobjwa]

29 minutes ago, Roger Smith said:

Maybe 1978 is their analogue then. The lunar dates in 1978 were only 3 days later than in 2019. That means an analogue to the east coast blizzard would occur around Feb 3-4. The Great Lakes superstorm timing would be Jan 22-23. If that was the analogue in play. (for me it's not in play for other reasons) ...

When did DC region have a snowstorm midway from new moon to full moon in January? That might provide a clue as to the analogue year (new moon was 6th, full moon is 21st using z time, both back up to near end of previous day using EST to 5th and 20th). I know it wouldn't be 1996 as that storm was just a day or two after full moon. 

Good question! I don't really know where to look for that info, though...I have monthly totals but not specific dates.