ZODOC TROPICAL CYCLONE EXPERIMENT
Prospective Long-Range Timing Forecasts — 2026
The 2026 ZODOC Tropical Cyclone Experiment is a prospective test of whether astronomical timing signals can identify periods of increased tropical cyclone activity weeks in advance.
Forecasts are prepared from astronomical timing data and issued before the forecast periods begin. Once issued, the forecast windows are not revised to fit subsequent weather.
The primary forecast target is timing. ZODOC identifies specific Times of Interest (TOI) during which tropical cyclone development or intensification is expected to become more likely.
Cyclogenesis is defined as the first occurrence of Tropical Depression strength (approximately 25–30 kt). Earlier Invest or disturbance stages do not count as a genesis hit. Rapid Intensification (RI)is evaluated separately when it is specifically forecast.
Each forecast is subsequently compared with observed tropical cyclone activity and classified using predetermined verification rules. Forecast issue dates, lead times, hits and misses are retained so that the experiment can be evaluated objectively.
The purpose is not to replace conventional meteorological forecasting, but to test whether astronomical timing contains useful long-range information about when tropical cyclone activity is more likely to occur.
2026 FORECAST vs. OBSERVED CYCLOGENESIS
Forecast vs. Observed Cyclogenesis
The timeline above compares the prospectively issued ZODOC Time-of-Interest windows with observed tropical cyclone genesis through September 21, 2026. Basin colors identify each genesis event; filled markers occurred within a forecast TOI window and open markers occurred outside the windows.
2026 STATISTICAL VERIFICATION
1. Genesis Concentration
From May 1 through September 21, 2026, ZODOC Time-of-Interest (TOI) windows covered 89 of 144 calendar days (61.8%). During the same period, 33 of 45 observed cyclogenesis events (73.3%) occurred within those forecast windows.
If cyclogenesis were distributed randomly across the observation period, calendar coverage provides a simple null expectation. At 61.8% coverage, approximately 27.8 of the 45 genesis events would be expected to occur within TOI windows. The observed total was 33, approximately 5.2 events above that expectation.
The observed genesis concentration was therefore 11.5 percentage points above calendar coverage. A one-sided exact binomial test gives p ≈ 0.073 for observing 33 or more of 45 genesis events within the TOI windows.
This result is suggestive but does not reach the conventional p < 0.05 threshold. The binomial result should also be interpreted as a preliminary statistical measure because tropical cyclone genesis events are not necessarily completely independent in time.
2. Forecast Window Verification
Through September 21, 2026, 14 of 17 prospectively issued Time-of-Interest (TOI) windows verified, producing an observed window hit rate of 82.4%.
The forecasts were issued before the forecast periods began and were not subsequently revised to fit observed tropical cyclone activity. Average forecast lead time through W17 was approximately 30 days.
The 14 HIT / 3 MISS result is reported descriptively rather than assigning an arbitrary 50% chance baseline. TOI windows differ in duration, and tropical cyclone activity varies substantially through the season; therefore, a simple 50/50 binomial model would not provide a well-defined null expectation for window verification.
Window verification is consequently presented as a measure of forecast performance, while the genesis concentration analysis above provides the experiment's primary preliminary statistical test.
3. Cyclone Activity (CEI)
Cyclone Energy Index (CEI) provides a broader measure of tropical cyclone activity by incorporating daily cyclone intensity rather than counting genesis events alone.
From May 1 through September 21, 2026, total observed cyclone activity was 15,605 kt CEI. Of this total, 10,690 kt (68.5%) occurred within ZODOC TOI windows.
TOI windows occupied 61.8% of the observation period, while containing 68.5% of measured cyclone activity. Cyclone activity was therefore concentrated within TOI periods by 6.7 percentage points above calendar coverage.
A simple binomial significance test is not appropriate for CEI because cyclone intensity persists across consecutive days and observations from the same storm are statistically dependent. A permutation or block-randomization analysis that preserves the temporal structure of cyclone activity is more appropriate for evaluating statistical significance.
CEI is therefore reported here as a descriptive secondary result. A formal significance estimate will require an appropriate time-preserving randomization analysis.
