U.S. Temperature Forecasts Skill 2.0 (Updated Observational Dataset)
performance.commerce.gov | Last Updated 28 Mar 2024Seasonal outlooks are reported as the probability of temperature being above normal, near normal, below normal or, where no definite seasonal guidance can be provided, equal chances. The seasonal outlooks are the cumulative skill calculated for regions where predictions are made. These forecasts are verified using a 48-month running mean of Heidke Skill score computed for seasonal outlooks for each 3-month seasonal mean (e.g., January-February-March mean; February-March-April mean; March-April-May mean; and so on). The previous dataset used to verify the CPC seasonal temperature outlook has developed a cold bias across a large portion of the western United States during the past 5-10 years. CPC has developed a new dataset which consists of Synoptic and Metar stations only, which is consistent throughout the historical and real time periods. This change will not impact the measure itself, nor how it is reported, but will remove biases identified in the old verification dataset, resulting in a more accurate performance indicator.
This dataset has the following 6 columns:
Column Name | API Column Name | Data Type | Sample Values |
---|---|---|---|
Performance Indicator | performance_indicator | text | U.S. Temperature Forecasts Skill 2.0 (Updated Observational Dataset) view top 100 |
Fiscal Year | fiscal_year | calendar_date | 2025-09-30T00:00:00.000 2024-09-30T00:00:00.000 2023-09-30T00:00:00.000 2022-09-30T00:00:00.000 view top 100 |
Actual | actual | number | 32 27 view top 100 |
Target | target | number | 27 27.5 view top 100 |
Proximity | proximity | number | 1.19 1 view top 100 |
Status | status | text | Exceeded Met view top 100 |