The cycles, one by one
What each rhythm is, what it touches, and how far ahead it can honestly be predicted. Figures marked in bold come from this almanac's own records and update with the place and date above.
Clockwork. Astronomy and cicada brood years can be calculated years ahead, to the day or better.
A real driver, like El Niño, or a steady seasonal habit, like migration weeks. It tilts the dice toward one outcome without deciding it.
A rhythm exists, but it's irregular or its effect here is small. Treat it as background.
Tested and no better than a coin flip, or only forecastable about two weeks out.
From a pattern to a prediction
A correlation (r, from −1 to +1) says two records move together. Squared, it says how much of the year-to-year swing one explains: r = 0.3 explains about 9%. That's useful for odds, and far from certainty.
What makes a forecast possible
A forecast needs a lead: the driver has to change before the effect (El Niño in fall, then the winter). It also needs a mechanism and a test on years it wasn't fitted to. Long-term trends, like shrinking ice, carry forward more reliably than cycles do.
Cycle Clock
Each ring is one cycle, shown one full period wide. Now is at the top; time runs clockwise, so the right half is what's coming and the left half is what just passed. Marks sitting under the needle together show cycles that line up.
How to read it
Fast cycles sit inside and slow cycles outside. Each ring's colour shows that cycle's value at that moment (see the ring key). Filled dots are high points (full moon, solar maximum), hollow dots are low points, and small dots are events like meteor showers or median frost dates. Point at any ring to read it.
What it says about the future
The right half of the sky rings (day, moon, distance, year, lunar nodes) is exact. Migration and frost marks are typical timing from past records, not this year's forecast. The QBO, ENSO and PDO rings go blank ahead because nobody can calculate them forward; their odds are in Projections.
Ring key (inside → out)
Where each cycle is now
| Cycle | Now | Next |
|---|
Location
Click anywhere in the region. The place sets the day length, the nearest weather station, the county's cicada brood, the nearest lake, and which nations' calendars come from that area.
Compounding: the next 24 months
How close each cycle is to one of its extremes, month by month, for the chosen place. The alignment index weights each cycle by how many established effects it has in the interaction graph.
How to read it
Rows are cycles and columns are months, starting now. The darker a cell, the closer that cycle is to a peak or a low. The jade bars at the bottom add the rows up, so tall bars are months when several cycles sit near an extreme together.
What it says about the future
The timing is a projection: exact for the sky rows, typical for migration, forecast for sunspots. What a line-up does is not. Most line-ups are coincidences in timing; only links in the Influences section push on each other.
Strongest alignments ahead
Calendar Comparison Strip
One year, with each nation's published calendar shown separately. Under each calendar, observed timing of the events its moons are named for (USA-NPN volunteer data, station records) is drawn on the same axis.
How to read it
The top rows are the Gregorian months and the real lunations (new moon to new moon, dots = full moons). Each nation's calendar follows in its own row. Moss-green bars underneath are observed timing: the bar is the usual range, the dot the typical date, and the ring this year.
What it says about the future
Comparing a named marker (sap, ice, berries) with when it's actually observed shows drift. If the marker keeps arriving earlier, that's a trend you can extend into coming years; a cycle can't explain it.
Braid Timeline
Each band is one record, coloured by its own value. Scroll or pinch to zoom; drag to pan. Hover to read every band at the same moment; click near the now line for the latest values.
How to read it
Three colour families (see the key): plum ↔ ochre for above or below normal, jade for amounts (ice, birds, moose), and amber for activity (sunspots, storms, wolves). Strong stripes repeating at even spacing are a rhythm; a slow colour change is a trend.
What it says about the future
Long records show which rhythms are steady enough to extend: sunspots (about 11 years) are; ENSO wanders. Notice that several bands drift steadily (temperature up, ice down). Those trends are the most reliable thing to project.
Projections
What's coming for the chosen place. Sky and moon events are exact; dates marked ≈ are expectations from past records; the winter outlook gives probabilities, with each model's track record shown. Every forecast is logged before the season and scored afterwards.
How to read it
The calendar lists the coming year; tags say what kind of event each is. Expected dates show the usual range (bar = middle half of years, line = 10–90%). Brown-grey is all years; jade is this year's projection, adjusted for the trend or the forecast ENSO.
How the predictions are made
Exact events come from astronomy. Expected dates come from the station's long-term trend. Winter odds come from the ENSO forecast and from "analog" past winters, and an analog is kept only if it beat the plain trend on past winters (HSS = skill score; 0 = no better than chance).
The coming year here
Expected dates
Winter outlook
Testable-claims register
Sayings, hypotheses and forecasts written down with a test, then scored against the records. A claim "went the claimed way" in a year if the outcome landed on the claimed side of its own long-term (detrended) median, so chance alone gives about 50%.
Interaction Graph
How the cycles affect each other, with direction, lag, mechanism and how strong the evidence is. Hover a link for its mechanism, citation, and what the Great Lakes records here show. Click a node to light up what drives it and what it affects; click again to clear.
How to read it
Arrows run from cause to effect. Line style is the evidence grade (key, bottom left); node colour is the layer (bottom right). Longer links have longer delays.
What it says about the future
A link with a lag is what makes a prediction possible. If acorns lead ticks by two years, today's acorns give odds for ticks in two years. Grade A and B links with a lag are the ones worth using; C and D links are not.
All links as a table
Analysis
The comparisons behind the graph, run on the records in this app. Every series is detrended first so long-term warming doesn't pass as a cycle. A guard against chance correlations (false-discovery rate) is applied throughout.
How to read it
Wavelet power: time runs left to right and rhythm length top to bottom; dark outlined patches are rhythms that are really there. Coherence: where two records keep in step. Lag matrix: each square is the strongest link between a driver and a response. Epoch: the average of the years around an event, against what random years give.
What it says about the future
A significant, persistent rhythm can be extended forward. A coherent pair with a steady lead lets one record act as an early warning for the other. Results that don't survive the chance test shouldn't be used to forecast, however good they look.
Which periods are strong in a record (wavelet power)
Do two records vary together (wavelet coherence)
Lagged correlations, every driver × response
What happens around key events (superposed epoch analysis)
Average anomaly of each record in the years around the events, in standard deviations. The shaded band is what random sets of years give, so a line inside the band is consistent with chance.