NPROVS Statistic Plots
Collocated sounding data from NPROVS are used as input into routines that generate a variety of statistics that can be used to evaluate the performance of every collocated data platform. This is done by comparing temperature and moisture data from each platform to a baseline system, typically radiosondes and dropsondes. The resulting statistical data can be downloaded and viewed using graphical applications. They are also used to create the charts that are available from this page.
Long-Term Performance Trends
Long-term performance trend plots are available from the Long-Term Trends page. Hovering over the blue menu bar just above the graph will allow for the selection of platform grouping, terrain, data type, statistic, and pressure level.
To reduce the number of plots on each graph, the platforms are divided into groups:
- NUCAPS IR + MW
- NUCAPS MW only
- AIRS and MIRS
- Forecast/GPRSO: ECMWF Analysis, HRRR Analysis and COSMIC-2
The graphs have been separated into three terrain options: Sea only, Land only, and Land and Sea combined.
Statistics are available for temperature and water vapor mixing ratio percent error.
The available statistic options are bias (platform minus radiosonde), standard deviation, and root mean square.
NPROVS typically generates statistics at 100 pre-defined pressure levels. A subset of the 100 levels are available.
Although NPROVS is capable of generating statistics from 2008 to the present, the plots shown here show the most recent 5 year period. The plots show monthly statistical averages. They are usually updated during the first week of each month.
Statistical Heat Maps
Long-term performance trend plots showing statistics displayed as heat maps are available from the Heat Maps page. Hovering over the blue menu bar just above the graph will allow for the selection of a specific platform profile, terrain, data type, and statistic.
The NPROVS heat maps show long-term trends for a chosen profile at every pressure level over the most recent five year period.
The heat map plots have similar options as the performance trend plots do. In addition to selecting individual profiles, it is also possible to select terrain type, data type and statistic type.
