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.

Lands End

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)
99%100%96% confident dry (P=99% · run 3–9h) 14:00Z (92%)
100%100%100% confident dry (P=100% · run 6–9h) 15:00Z (99%)
—100%100% confident dry (P=100% · run 8–9h) 15:00Z (98%)
——100% confident dry (P=100% · run 7–9h) 15:00Z (98%)
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
99%97%98% 83% confident dry (P=99%)
99%98%98% 83% confident dry (P=99%)
—98%98% 100% confident dry (P=98%)
——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)
95%98%90% confident dry (P=95% · run 3–9h) 14:00Z (88%)
99%99%98% confident dry (P=99% · run 6–9h) 14:00Z (98%)
—100%100% confident dry (P=100% · run 8–9h) 14:00Z (97%)
——100% confident dry (P=100% · run 7–9h) 14:00Z (96%)
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
97%96%97% 67% confident dry (P=97%)
98%98%99% 83% confident dry (P=98%)
—98%99% 100% confident dry (P=98%)
——99% 100% confident dry (P=99%)
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)
88%95%81% confident dry (P=88% · run 3–9h) 13:00Z (83%)
99%98%95% confident dry (P=99% · run 6–9h) 13:00Z (98%)
—99%99% confident dry (P=99% · run 8–9h) 13:00Z (96%)
——99% confident dry (P=99% · run 7–9h) 13:00Z (95%)
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)
74%88%71% ambiguous (P=74% · run 3–9h) 11:00Z (76%)
96%95%90% confident dry (P=96% · run 6–9h) 12:00Z (95%)
—98%97% confident dry (P=98% · run 8–9h) 12:00Z (95%)
——97% confident dry (P=97% · run 7–9h) 12:00Z (93%)
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
62%83%56% 33% ambiguous (P=62%)
91%90%88% 83% confident dry (P=91%)
—94%92% 100% confident dry (P=94%)
——94% 100% confident dry (P=94%)
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)
62%81%61% ambiguous (P=62% · run 3–9h) 11:00Z (73%)
95%91%85% confident dry (P=95% · run 6–9h) 11:00Z (93%)
—96%94% confident dry (P=96% · run 8–9h) 11:00Z (94%)
——94% confident dry (P=94% · run 7–9h) 11:00Z (91%)
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.