Brightband has found signal in the noise of weather forecasts three and four weeks out.
Weeks 3 and 4 have long been a predictability desert. They’re too far ahead for today’s atmospheric conditions to help much, and too soon for slow-moving ocean signals like El Niño to dominate. That leaves many people who need to make weather-dependent decisions in weeks 3 and 4 with little to go on.
Forecast skill
AI weather models have started to change this, and ECMWF’s AI Weather Quest competition has showcased exciting progress on sub-seasonal skill. We ran Brightband’s sub-seasonal forecast, BB-SUBS, through the AI Weather Quest evaluation code for 2m temperature and found that it is a significant step beyond every Weather Quest model. In separate evaluations we also found that BB-SUBS outperformed other commercial models.
| Model | Type | Week 3 | Week 4 |
|---|---|---|---|
| BB-SUBS | Brightband | 0.112 | 0.072 |
| LP | AIWP | 0.089 | 0.057 |
| MicroEnsemble | AIWP | 0.086 | 0.055 |
| S2S Multi-Model Mean | NWP | 0.086 | 0.049 |
| CLINT | AIWP | 0.080 | 0.050 |
| AIFS | AIWP | 0.076 | 0.040 |
| CMAandFDU | AIWP | 0.068 | 0.047 |
| EC46 | NWP | 0.062 | 0.036 |
| UWAtmosNVIDIA | AIWP | 0.033 | 0.009 |
| CliMA | AIWP | 0.020 | 0.026 |
BB-SUBS is operational and available now. It is 1.5-2x more skillful than ECMWF’s EC46, the forecast many treat as the sub-seasonal benchmark (even after we debias EC46 to give it its best showing), and 30% more skillful than climatology, the historical average weather for that time of year.
| Week | BB-SUBS v5.3 | EC46 | Improvement |
|---|---|---|---|
| Week 3 | 0.29 | 0.19 | +0.10 ± 0.02 |
| Week 4 | 0.12 | 0.05 | +0.07 ± 0.02 |
If you make decisions that depend on the weather three or four weeks out, BB-SUBS can help. That includes:
- Energy and power trading
- Agriculture
- Water management
- Retail and supply chain
- Insurance and disaster response
See the details and get access to BB-SUBS.
Learn more