Visibility and cloud ceiling forecasts are critical for aircraft safety and efficient operations. When visibility or cloud ceilings are low, pilots rely on instruments to navigate instead of visual reconnaissance. The Federal Aviation Administration establishes Instrument Flight Rule (IFR) thresholds—visibility less than three statute miles and/or cloud ceilings at, or below, 1000 feet—for safety.

NWS assesses the quality of IFR threshold forecasts in response to these requirements. Fundamental statistical metrics, specifically Probability of Detection (POD) and False Alarm Ratio (FAR), are used to track IFR forecast performance. Probability of Detection (POD), also known as Accuracy, is a ratio that describes the number of times IFR is correctly forecasted compared to the total number of IFR occurrences. FAR is a ratio that describes the number of IFR forecasts when IFR was not observed compared to the total number of forecast attempts. These two metrics must always be used in conjunction, as one can be improved at the expense of the other. Greater accuracy and a minimized FAR result in safer flights and fewer flight delays; conversely, poorer accuracy and an increased FAR result in a greater incidence of unnecessary flight delays.

Key activities for the development and launch of weather satellites and products are identified and tracked using a project management system.

Metric calculated as the percent of Climate Prediction Center extended range temperature and precipitation outlooks (outlooks out to 14 days) exceeding Heidke Skill Score (HSS) of 10 (10% better than a climatological forecast). To better track results, three individual metrics are tracked – 1) All forecasts combined, 2) Temperature only, and 3) Precipitation only.

Metric calculated as the percent of Climate Prediction Center extended (outlooks out to 14 days) and long range (monthly and seasonal outlooks) temperature and precipitation outlooks exceeding Heidke Skill Score (HSS) of 10 (10% better than a climatological forecast). Due to the scarcity of long range outlooks (six per month), scores are smoothed over a three year period. To better track results, three individual metrics are tracked – 1) All forecasts combined, 2) Temperature only, and 3) Precipitation only.

Temperature outlooks are used by sectors of the U.S. economy, such as energy, agriculture, transportation, etc. as one factor in resource decision making. Temperature outlooks are reported as the probability of temperature being above normal or below normal or, where no definite guidance can be provided, equal chances. This is the cumulative skill calculated for regions where predictions are made.

For each reported flash flood event, the flash flood warning lead-time is the difference in minutes between the issuance of a flash flood warning and the onset of a geographically corresponding flash flood event. The lead times for all flash flood events, within the United States and territories served by the National Weather Service, are averaged to calculate the national average flash flood warning lead-time metric for a given fiscal year. This average includes all warned events with zero lead times and all unwarned events.

Historically, flash flood warnings were issued and verified on a countywide basis. Starting in FY 2010, the storm-based warnings were implemented with verification based solely for the areas impacted by the warning and event.

Seasonal 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.

The measure captures the count of significant and discrete NOAA research and development products that have transitioned to application at operational U.S. weather services and in the U.S. weather commercial sector. In addition, these weather products support NOAA contributions to the Weather Research and Forecasting Innovation Act of 2017.

Proposed new indicator for FY 2022. Measure is under development.

Proposed new indicator for FY 2022. Measure is under development.