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.

The annual number of peer reviewed publications is an indicator of productivity and relevance and is tracked using online resources. Peer review is one of the important procedures used to ensure that the quality of published information meets the standards of the scientific and technical community.
* FY 2019 is the first year that we are reporting on both NOAA-authored publications (previously the only type of publication measured) and NOAA-funded publications.

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.

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

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.

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.

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.

A winter storm warning provides NOAA customers and partners advanced notice of a hazardous winter weather event that endangers life or property, or provides an impediment to commerce. Winter storm warnings are issued for winter weather phenomena like blizzards, ice storms, heavy sleet, and heavy snow. This performance indicator measures the accuracy of winter storm events.

Improving the accuracy and advance warnings of winter storms enables the public to take the necessary steps to prepare for disruptive winter weather conditions.

The public, private sectors, emergency managers, and government institutions at all levels in this country and abroad use NOAA tropical cyclone forecasts to make decisions regarding the protection of life and property. This measure represents the difference between the projected intensity of these storms and the actual intensity in knots (kt) for Atlantic Basin tropical cyclones (i.e., tropical depressions, tropical storms, and hurricanes).

* Annual Hurricane Season begins June 1 and ends November 30. The final values are produced after a verification and validation period.

The measure is validated by computing the average difference (error) for all the 48-hour forecasts occurring during a calendar year. Because tropical cyclones are relatively rare events, this measure can show significant annual volatility. Projecting the long-term trend, and basing out-year goals on that trend, is preferred over making large upward or downward changes to the targets each year.

The lower this measure is the better, so being below the target is exceeded and above the target is not met. 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.

*Please also note that a FAR is the number of false alarms per the total number of warnings. So if a FAR actual is higher than the target, then the target has not been met. If the actual is the same number or less, then it has been met or exceeded.