WeatherBlend

Multi-model forecast blending for Bonehill Rocks, Dartmoor

Dry-window forecast

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

Dartmoor Nr Hexworthy

2-hour dry window

Phase 3p — Gaussian copula MC over Phase 3o hourly P(wet)

Gaussian copula MC over the rich+orographic Phase 3o's hourly P(wet) marginals. Single empirical Σ per station, fit on train-split observed daytime wet/dry binary sequences (daytime-shape autocorrelation is lead-independent for truth, so per-lead Σs pool into one). Captures within-day wet/dry autocorrelation that an iid sampler misses. Bake-off Brier 0.1064 aggregate on the 9 (station, lead) Bonehill cells.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
89%87%92% confident dry (P=89% · run 1–9h) 15:00Z (88%)
100%100%99% confident dry (P=100% · run 5–9h) 09:00Z (99%)
—100%100% confident dry (P=100% · run 9–9h) 15:00Z (99%)
——100% confident dry (P=100% · run 9–9h) 15:00Z (99%)
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
64%56%62% 50% confident dry (P=64% · τ=39%)
97%95%95% 83% confident dry (P=97% · τ=39%)
—95%96% 100% confident dry (P=95% · τ=43%)
——96% 100% confident dry (P=96% · τ=41%)
Phase 3p — Gaussian copula MC over Phase 3o hourly P(wet)

Gaussian copula MC over the rich+orographic Phase 3o's hourly P(wet) marginals. Single empirical Σ per station, fit on train-split observed daytime wet/dry binary sequences (daytime-shape autocorrelation is lead-independent for truth, so per-lead Σs pool into one). Captures within-day wet/dry autocorrelation that an iid sampler misses. Bake-off Brier 0.1064 aggregate on the 9 (station, lead) Bonehill cells.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
77%76%83% ambiguous (P=77% · run 1–9h) 14:00Z (83%)
99%98%97% confident dry (P=99% · run 5–9h) 08:00Z (98%)
—100%100% confident dry (P=100% · run 9–9h) 14:00Z (98%)
——100% confident dry (P=100% · run 9–9h) 12:00Z (99%)
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
28%45%50% 50% confident wet (P=28% · τ=44%)
87%91%72% 83% confident dry (P=87% · τ=44%)
—95%94% 100% confident dry (P=95% · τ=44%)
——93% 100% confident dry (P=93% · τ=47%)
Phase 3p — Gaussian copula MC over Phase 3o hourly P(wet)

Gaussian copula MC over the rich+orographic Phase 3o's hourly P(wet) marginals. Single empirical Σ per station, fit on train-split observed daytime wet/dry binary sequences (daytime-shape autocorrelation is lead-independent for truth, so per-lead Σs pool into one). Captures within-day wet/dry autocorrelation that an iid sampler misses. Bake-off Brier 0.1064 aggregate on the 9 (station, lead) Bonehill cells.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
65%64%73% ambiguous (P=65% · run 1–9h) 13:00Z (78%)
98%96%94% confident dry (P=98% · run 5–9h) 08:00Z (97%)
—99%99% confident dry (P=99% · run 9–9h) 11:00Z (96%)
——100% confident dry (P=100% · run 9–9h) 11:00Z (99%)
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 3p — Gaussian copula MC over Phase 3o hourly P(wet)

Gaussian copula MC over the rich+orographic Phase 3o's hourly P(wet) marginals. Single empirical Σ per station, fit on train-split observed daytime wet/dry binary sequences (daytime-shape autocorrelation is lead-independent for truth, so per-lead Σs pool into one). Captures within-day wet/dry autocorrelation that an iid sampler misses. Bake-off Brier 0.1064 aggregate on the 9 (station, lead) Bonehill cells.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
56%52%63% ambiguous (P=56% · run 1–9h) 12:00Z (68%)
93%91%87% confident dry (P=93% · run 5–9h) 08:00Z (95%)
—98%99% confident dry (P=98% · run 9–9h) 11:00Z (96%)
——99% confident dry (P=99% · run 9–9h) 10:00Z (98%)
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
16%13%15% 33% confident wet (P=16% · τ=47%)
73%66%67% 67% confident dry (P=73% · τ=47%)
—94%92% 100% confident dry (P=94% · τ=47%)
——82% 83% confident dry (P=82% · τ=54%)
Phase 3p — Gaussian copula MC over Phase 3o hourly P(wet)

Gaussian copula MC over the rich+orographic Phase 3o's hourly P(wet) marginals. Single empirical Σ per station, fit on train-split observed daytime wet/dry binary sequences (daytime-shape autocorrelation is lead-independent for truth, so per-lead Σs pool into one). Captures within-day wet/dry autocorrelation that an iid sampler misses. Bake-off Brier 0.1064 aggregate on the 9 (station, lead) Bonehill cells.

Target date (UTC) +24h +48h +72h Confidence Best start (UTC, P dry block)
44%43%52% ambiguous (P=44% · run 1–9h) 11:00Z (58%)
88%86%82% confident dry (P=88% · run 5–9h) 08:00Z (92%)
—97%97% confident dry (P=97% · run 9–9h) 11:00Z (95%)
——98% confident dry (P=98% · run 9–9h) 09:00Z (98%)
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.