Agriculture

Seven breadbaskets, one scoring rule. Growing-season temperature, wind and precipitation forecasts are resolved against observed truth and ranked by measured error — no crop index is asserted that the evidence cannot carry.

receipts held

1,000

resolved against truth

0%

regions

7

forecast centres

10

Forecast horizon → error

Mean absolute error by lead time, per channel

MEASURED

No channel in this domain has resolved across enough lead times yet.

Which breadbasket is hardest to forecast

Temperature error per region, worst first

MEASURED

No site has enough resolved receipts to rank yet.

Prediction ↔ observation

Daily forecast mean against daily observed mean across the growing regions

MEASURED

No channel has enough resolved days to plot yet.

Which centre is actually right here

Best and worst measured producer per channel

MEASURED

No resolved receipts in this domain yet.

Models trained on agricultural receipts

Stored weights, their measured error and their skill against persistence

MEASURED
residual·agriculture:precipitation +6h0.26+0.01n=51,904
residual·agriculture:precipitation +24h0.26-0.06n=44,334
residual·agriculture:precipitation +72h0.24-0.11n=24,184
residual·agriculture:surface_pressure +6h0.44-0.01n=51,936
residual·agriculture:surface_pressure +24h0.58+0.22n=44,376
residual·agriculture:surface_pressure +72h0.48+0.62n=24,169
residual·agriculture:temperature_2m +6h0.81+0.12n=406,490
residual·agriculture:temperature_2m +24h0.58-0.13n=371,024
residual·agriculture:temperature_2m +72h0.84-0.10n=299,171
residual·agriculture:wind_speed_10m +6h0.74+0.14n=406,427
residual·agriculture:wind_speed_10m +24h0.77+0.23n=370,977
residual·agriculture:wind_speed_10m +72h0.78+0.25n=299,094

columns · MAE · skill vs persistence · training samples