Base Interest and Hazard Detection
Two fields depicting the areas of precipitation and their intensity are produced by NCWF2: Base Interest and Hazard Detection. The Base Interest field depicts the location and intensity of precipitation ranging from light to extreme and includes precipitation that is both convective and stratiform in nature. The Convective Hazard Detection field is a meant to depict areas of precipitation that are convective in nature and, because of their intensity, pose a significant hazard to aviation. The Base Interest field is available for display to help place the Convective Hazard Detection into context. As such, the user can easily compare the Base Interest field with the Convective Hazard Detection field, by toggling the two fields on and off, to see which storms are being classified as hazardous to aviation by the algorithm.
The Base Interest or Significant Weather field is produced by directly combining UNISYS VIL data with lightning data from National Lightning Detection Network (NLDN) provided by Vaisala. The convective hazard field is determined by following a sequence of three steps: (1) Areas of stratiform precipitation (regardless of intensity) are removed from the UNISYS VIL field; (2) Areas where echo tops are less than 15 Kft are removed to eliminate areas of light precipitation, ground clutter and AP; (3) This modified VIL field is then merged with lightning data to obtain the final product.
(1) The first step in producing the National Convective Hazard Detection involves removing regions of stratiform precipitation. This is done to isolate and track only the most intense convective elements which pose the greatest risk to aviation interests.
UNISYS VIL field given in units of kg m-2
(mouse over image to see impact of stratiform filter – turns most of the VIL regions dark blue).
The image above demonstrates effect of applying the stratiform-convective partitioner (Steiner et al. 1995) to the UNISYS VIL field. The partitioner removes areas of stratiform precipitation based on texture and intensity evident in the VIL. Mouse over the image to toggle between the original field and stratiform-filtered field.
How does removing stratiform precipitation improve the forecasts? Stratiform precipitation tends to trail behind and move more slowly than convection which is typically located along the leading edge of an organized area of storms or mesoscale convective system (MCS). Since we are only interested in accurately forecasting the position of convection, we need to track only the leading edge of the MCS which typically moves much faster and in different direction than the trailing stratiform precipitation (see image below for example of a large area of trailing stratiform).
(2) The second step involves removing areas of VIL that have echo tops less then 15 Kft. This is done to remove any areas of low-topped stratiform precip that made it through the stratiform filter and as a way to remove bad data such as that resulting from Anomolous Propogation (AP) and ground clutter. The echo tops filter is used to remove these artifacts which have echo tops of zero associated with them. Removing AP is very important because looks and moves like real weather and thus can be tracked and erroneously included in the forecasts.
(3) The final step is to merge the NLDN lightning data with the modified VIL field. The NLDN lightning strike data are binned in 10-min intervals to determine lightning rate. The lightning rate is first thresholded then converted to VIL using the relationship given in Table 1 (see below). The lightning rate and VIL (which are now in the same units) are then merged by taking the maximum of the lightning-VIL pair at each grid point.
Mouse over image above to see how threshold is applied to the binned lightning data before it is blending with the Unisys VIL data.
Filtered VIL in g m-2. Mouse over the image above to show the impact of blending lightning data with filtered VIL.
Table 1: Approximate Correlations between various hazard detection fields and VIP levels with corresponding meteorological conditions. |
|||||||||||||||||||||||||||||||||||||
Convective hazard detection field.
The final derived field is then displayed in VIP levels 1-6 that are determined using the relationships given in Table 1 to do the conversion from g m-2 to VIP levels. This field, now called the National Convective Hazard Detection (was NCWD2) is used as an input to the NCWF forecast algorithm. The NCHD was designed to detect areas of aviation-impacting convection. Forecasts produced by the NCWF2 software are based on NCHD. In addition, the NCHD is used to statistically evaluate several operational and developmental convective forecast products (see RTVS web page – http://www-ad.fsl.noaa.gov/fvb/rtvs/index.html).