WeatherBlend

Multi-model forecast blending for Sennen, Cornwall

Dry-window forecast

P(at least one N-hour dry block in 09:00–18:00 local). 3b champion plus Monte-Carlo challengers.

Trengwainton

2-hour dry window

Phase 3q — Gaussian copula MC over Phase 3c hourly P(wet)

Gaussian copula MC over Phase 3c's hourly P(wet) marginals — the 3p engine bound to the rich (non-orographic) 3c instead of 3o, for coastal locations without a 3o model. Single empirical Σ per station from train-split observed daytime wet/dry sequences. Captures within-day wet/dry autocorrelation an iid sampler misses; scored against the 3b champion at Sennen.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
97%99%94% confident dry (P=97% · run 2–9h) 15:00Z (82%)
100%100%99% confident dry (P=100% · run 4–9h) 13:00Z (95%)
—100%100% confident dry (P=100% · run 6–9h) 15:00Z (94%)
——100% confident dry (P=100% · run 5–9h) 15:00Z (95%)
Monte-Carlo P(an 2-hour dry block runs from each start hour) — one line per forecast horizon, soonest target day shown. The peak is the best time to set off; the height is how likely it stays dry. Each hour is its own probability, so the lines need not sum to the daily "any dry window" figure.

3-hour dry window

Phase 3b — lean (53 features)

53-feature LightGBM. Champion.

Target date (UTC) +24h +48h +72h Model agreement Confidence
97%97%96% 83% confident dry (P=97%)
99%97%96% 83% confident dry (P=99%)
—97%96% 100% confident dry (P=97%)
——96% 100% confident dry (P=96%)
Phase 3q — Gaussian copula MC over Phase 3c hourly P(wet)

Gaussian copula MC over Phase 3c's hourly P(wet) marginals — the 3p engine bound to the rich (non-orographic) 3c instead of 3o, for coastal locations without a 3o model. Single empirical Σ per station from train-split observed daytime wet/dry sequences. Captures within-day wet/dry autocorrelation an iid sampler misses; scored against the 3b champion at Sennen.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
89%95%85% confident dry (P=89% · run 2–9h) 14:00Z (75%)
97%99%96% confident dry (P=97% · run 4–9h) 12:00Z (93%)
—99%98% confident dry (P=99% · run 6–9h) 12:00Z (92%)
——99% confident dry (P=99% · run 5–9h) 14:00Z (91%)
Monte-Carlo P(an 3-hour dry block runs from each start hour) — one line per forecast horizon, soonest target day shown. The peak is the best time to set off; the height is how likely it stays dry. Each hour is its own probability, so the lines need not sum to the daily "any dry window" figure.

4-hour dry window

Phase 3b — lean (53 features)

53-feature LightGBM. Champion.

Target date (UTC) +24h +48h +72h Model agreement Confidence
95%95%96% 67% confident dry (P=95%)
97%97%98% 83% confident dry (P=97%)
—97%98% 100% confident dry (P=97%)
——98% 100% confident dry (P=98%)
Phase 3q — Gaussian copula MC over Phase 3c hourly P(wet)

Gaussian copula MC over Phase 3c's hourly P(wet) marginals — the 3p engine bound to the rich (non-orographic) 3c instead of 3o, for coastal locations without a 3o model. Single empirical Σ per station from train-split observed daytime wet/dry sequences. Captures within-day wet/dry autocorrelation an iid sampler misses; scored against the 3b champion at Sennen.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
81%90%74% confident dry (P=81% · run 2–9h) 13:00Z (70%)
95%97%91% confident dry (P=95% · run 4–9h) 11:00Z (91%)
—98%95% confident dry (P=98% · run 6–9h) 11:00Z (90%)
——96% confident dry (P=96% · run 5–9h) 13:00Z (87%)
Monte-Carlo P(an 4-hour dry block runs from each start hour) — one line per forecast horizon, soonest target day shown. The peak is the best time to set off; the height is how likely it stays dry. Each hour is its own probability, so the lines need not sum to the daily "any dry window" figure.

5-hour dry window

Phase 3q — Gaussian copula MC over Phase 3c hourly P(wet)

Gaussian copula MC over Phase 3c's hourly P(wet) marginals — the 3p engine bound to the rich (non-orographic) 3c instead of 3o, for coastal locations without a 3o model. Single empirical Σ per station from train-split observed daytime wet/dry sequences. Captures within-day wet/dry autocorrelation an iid sampler misses; scored against the 3b champion at Sennen.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
66%79%63% ambiguous (P=66% · run 2–9h) 12:00Z (63%)
87%92%84% confident dry (P=87% · run 4–9h) 11:00Z (86%)
—95%91% confident dry (P=95% · run 6–9h) 11:00Z (88%)
——91% confident dry (P=91% · run 5–9h) 12:00Z (84%)
Monte-Carlo P(an 5-hour dry block runs from each start hour) — one line per forecast horizon, soonest target day shown. The peak is the best time to set off; the height is how likely it stays dry. Each hour is its own probability, so the lines need not sum to the daily "any dry window" figure.

6-hour dry window

Phase 3b — lean (53 features)

53-feature LightGBM. Champion.

Target date (UTC) +24h +48h +72h Model agreement Confidence
53%64%68% 33% ambiguous (P=53%)
84%85%79% 83% confident dry (P=84%)
—92%88% 100% confident dry (P=92%)
——91% 100% confident dry (P=91%)
Phase 3q — Gaussian copula MC over Phase 3c hourly P(wet)

Gaussian copula MC over Phase 3c's hourly P(wet) marginals — the 3p engine bound to the rich (non-orographic) 3c instead of 3o, for coastal locations without a 3o model. Single empirical Σ per station from train-split observed daytime wet/dry sequences. Captures within-day wet/dry autocorrelation an iid sampler misses; scored against the 3b champion at Sennen.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
57%72%53% ambiguous (P=57% · run 2–9h) 11:00Z (58%)
85%87%77% confident dry (P=85% · run 4–9h) 11:00Z (83%)
—91%85% confident dry (P=91% · run 6–9h) 11:00Z (87%)
——86% confident dry (P=86% · run 5–9h) 11:00Z (82%)
Monte-Carlo P(an 6-hour dry block runs from each start hour) — one line per forecast horizon, soonest target day shown. The peak is the best time to set off; the height is how likely it stays dry. Each hour is its own probability, so the lines need not sum to the daily "any dry window" figure.

Dry hour = all four 15-min readings ≤ 0.1 mm. Search bounded to 09:00–18:00 local.