The performance score was developed as an aid to help evaluate NWS upper-air station performance nationwide and improve data availability, quantity, and quality.
The following equation is used to rank station performance, which includes appropriate weighting factors:
Score = A/2 + B + C + D/2 - E/5 - F/5 where,
|
A = Percentage of monthly soundings arriving at NCEP within 3 hours after balloon release (by 14:00 or 02:00 UTC). This is based on NCEP and FMH#3 requirements. To score in this category, all coded messages (with or without data) must be received. This percentage is divided by 2. |
|
B = Percentage of monthly soundings achieving the minimum 400 hPa pressure level requirement or higher. |
|
C = Percentage of monthly soundings achieving 50 hPa or higher. |
|
D = Percentage of monthly soundings achieving 20 hPa or higher. This percentage is divided by 2. |
|
E = Percentage of monthly temperature levels between surface & 400 hPa rejected by NCEP. This percentage is divided by 5. |
| F = Percentage of monthly wind levels between 400 and 100 hPa rejected by NCEP. This percentage is divided by 5. |
Starting with the August, 1999, data, element "A" above now includes soundings with AND without data. Prior to this change, points were only given in this element if the coded messages contained data. Messages without data (i.e., no sounding taken) were not included in this element. This change will result in slightly higher average scores.
Elements B through F are compiled 6 hours or more after balloon release and before NCEP data deletions are applied to the soundings. The maximum score a station can achieve with this equation is 300. Histograms are available that show station performance for some of the above elements
Data used for generating the scores come from data quality reports provided by NCEP each month.
Upper-air station managers should use the scores as only one of several tools for assessing how well their station is performing.
The performance score at an office does not solely reflect station staff performance. Factors such as equipment and communication failures, maintenance personnel performance, and support from NWS Regional office and Headquarters management, must also be taken into consideration when evaluating station scores. While it is common that a low score is the result of a system failure (e.g., lightning strike on the tracking antenna) beyond the control of station staff, scores are not adjusted manually to take this into account. The 92 NWS upper air stations launch a total of up to 5,700 radiosondes per month and thus generation of the scores is automated. To manually go through this large amount of data each month and adjust the scores higher to take into consideration factors beyond the control of station staff would be an excessive workload and is beyond the scope of this performance measure. This policy has been in place since development of the scoring system in 1998.
(1) Stations must not under inflate balloons in an attempt to reach higher sounding termination heights and a higher monthly score. Likewise, over inflation of the balloon must be avoided as this wastes gas and can lower balloon burst heights. Routine under inflation and over inflation of the balloon is unacceptable. Stations should review balloon ascension rate graphics based on the National Climatic Data Center (NCDC) archive and adjust the balloon gas amounts accordingly. NWS stations should have no more than 5 soundings a month ascend outside the range of 275 to 350 meters/minute (average from surface to flight termination)
(2) Stations must not disseminate poor quality soundings that should have been terminated early or not even transmitted. Dissemination of such soundings can erroneously increase the monthly score and is unacceptable for NWS operations. Station staff should review the monthly NCEP Super-adiabatic Lapse Rate Report and other data to help identify poor quality soundings and take corrective action.
Scores are now posted here: https://sites.google.com/a/noaa.gov/nws-radiosonde-observations/performance-scores
