Description of NCWF Viewer for D2 versison of NCWF-6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Document Written By: James Pinto

System Engineers: Nancy Rehak and Dan Meganhardt

Display Software Developed By: Gary Cunning

Research Applications Laboratory

National Center for Atmospheric Research

First Authored: 20 September 2005

Last Modified : 25 April 2006

 

 


 

 

1. Introduction

 

The NCWF Viewer was written in Java to display1-6 hour probabilistic forecasts of convection with looping capability. The viewer displays the merged probabilistic forecast of convection along with RUC forecasted echo tops. The probabilistic forecasts of convection are based on blending radar-based extrapolation forecasts and RUC-based probabilistic forecasts (RCFP) of convection. The RCFP is a derived product that uses output (e.g., forecasted rainfall rates) from the latest operational version of the RUC model (soon to be RUC13 – see Benjamin et al. 2004) while the radar-based probabilities are produced using combination of extrapolation and trending and includes growth(1) and dissipation(2) of convection. The forecast probabilities indicate the likelihood that aviation disrupting convection (e.g., radar echo exceeding 35 dBZ) will be present at a given location at the given forecast time. The probability maps can also be used to derive the forecasted coverage of storms. For example, a user may be interested in coverages exceeding 40%. This can be determined by contouring the region with P >= 40%. In a perfectly reliable forecast, one would expect 40% of the area within this contour to have aviation-impacting convection. Detailed descriptions of how the probability forecast are generated in the radar-based and RCFP systems are given in Megenhardt et al. (2004) and Weygandt et al. (2004), respectively. The system is currently being evaluated by a select group of users and scientists. Development of the system is funding by the FAA Aviation Weather Research Program.

 

2. Description of viewable fields

 

When the viewer is started it brings up a national scale view with the current base interest and the 2-hr performance polygon (Figure 1).

 


Figure 1. Default configuration displayed when the NCWF viewer is first started shown base interest field and 2 hour performance polygon.

 

The time bar below the map of the CONUS depicts a 10 hour moving time window. It is marked with the present time (red line) and the forecast/performance time that is currently displayed (orange square). The time that is displayed (past, current or future) may be changed by left clicking on the time bar.  The time chosen will remain fixed until either a different time is chosen or a movie loop is started (see below). While the time chosen remains fixed, the forecast valid time or performance polygons will be updated to be valid at the selected valid time until the left end of the time bar is reached by the orange square. Here, the orange square will remain, depicting past performance polygons and base interest that is about 3 hours old, until a new time or movie loop is selected by the user.

 

 

The display field called base interest is actually a derived product given in units of VIP (Video Integration Processor).  which is a WSR-88D Level 3 data product that has been mapped into a national mosaic and made available in realtime by UNISYS.  This field is used along with lightening data in the radar-based forecast algorithm to detect where storms are currently present. In rough terms, a VIP level of 3.5 is typically used as a minimum threshold above which convection is expected to impact aviation both en route and depatures and landings. A VIP level of 3 corresponds to a radar reflectivity of 40 dBZ and a VIL of between 3.5 and 6.9 kg m-2  (Hallowell et al. 1999).

 

The 1,2,3,4,5, and 6 hour probability forecasts (transparent pink shades – Figure 2) and performance polygons (magenta contours) are available for display. The radar-based forecasts are updated every 15 min and overlaid on the latest available base interest data. By default, the 2 hour performance polygon for the radar-based forecast is displayed along with the base interest. The legend/labels at the lower left corner of the radar-based forecast panel indicate the selected time for the base interest observation and either the performance polygon valid time or the forecast valid time. The performance valid times should generally be within 20 min of the base interest time.

 

The RCPF forecasts tend to have a latency of about two hours while the extrapolation forecasts are updated every 15 min. Thus, when blending the two systems, we are merging the 3 hour RCFP forecast with the 1 hour radar-based extrapolation forecast, the 4 hour RCFP forecast with the 2 hour radar-based extrapolation forecast, etc. New RCPF forecasts are generated every 60 min through post processing of the output from successive runs of the operational RUC model.

 

The storm top heights may also be overlaid apon the probability forecasts. The storm top height forecasts are derived from RUC model output and are based on the equilibrium height calculated from the RUC model environmental soundings. The contour levels are 15, 30 and 45 kft going from light to darker shades of blue.

 

Performance polygons (magenta contours) may only be displayed when the probability forecast is not displayed (i.e., fcst time < 0 min). Performance polygons are available for each of the forecast lead times and should be overlay on the base interest field. Any number of performance polygons may be displayed (they will all have the same forecast valid time). The performance valid times should generally be within 20 min of the base interest time.

 

 

3. Description of viewer functions

 

Movie loops may be initiated by clicking in the control box in the lower left of the display. Note: The display will constantly update with new data while the movie is looping bringing in new forecasts and base interest data as it becomes available.

 

The animation speed may be adjusted by dragging the slider with the left mouse button.

 

Zoom in by clicking the upper magnifying glass button in the upper right of the display and then dragging the mouse over the desired region. Go back or zoom out to previous zoom level by clicking the lower magnifying glass.

 

Pan the images by left clicking the hand button, then dragging the image by holding down the left mouse button in the image and sliding the mouse the way you want to pull the image.

 

Map overlays are available under the Maps pulldown menu.

 

 

 


Figure 2. Example of NCWF viewer in forecast mode depicting 2 (left panel) and 5 (right panel) hour merged probability forecast for a strong line of storms in central Oklahoma on 24 April 2006. See above for detailed description of display fields and functions.

 

 

Footnotes

 

(1)   Note: Probabilities of convection are increased and expanded in areal extent in areas where storm cells occur in regions of expected growth based on CAPE and CIN, layer mean RH, Shear and ThetaE advection. Currently growth is only allowed in the first two hours of the forecast and only for forecasts valid between 1630 and 2230 UTC (1230 and 1830 EDT). See Megenhardt et al. (2004) for a detailed description of the algorithm currently employed.

(2)   Regions of dissipation are determined by tracking areas of convection and calculating the change in mean area coverage with time using an 80 km filter. The rate of dissipation is used to reduce the forecast probabilities. Dissipation trending is only applied to forecasts valid before 1630 UTC and after 2230 UTC.

 

References

 

Benjamin, S.G., G.A. Grell, J.M. Brown, and T.G. Smirnova, 2004: Mesoscale weather prediction with the RUC hybrid isentropic-terrain-following coordinate model, Mon. Wea. Rev., 132, 473-494.

 

Hollowell, et al., 1999: The terminal convective weather demonstration. 8th Conference on Aviation, Range, and Aerospace Meteorology, Dallas, TX, Amer. Meteor. Soc., 200-204.

 

Megenhardt, D., et al., 2004: NCWF-2 Probablistic nowcasts. 11th Conference on Aviation, Range and Aerosopace, Hyannis, MA, Amer. Meteor. Soc. 23 pp.

 

Pinto et al., 2005: Fusing observation- and model-based probability forecasts for the short-term prediction of convection. WWRP Symposium on Nowcasting and Very Short range Forecasting, World Meteorological Organization, Toulouse, France, 4 pp.

 

Weygandt, S., and S. Benjamin, 2004: RUC model-based convective probability forecasts. 11th Conference on Aviation, Range and Aerosopace, Hyannis, MA, Amer. Meteor. Soc. 11 pp.

 

 

Relevant web pages

 

 

NCWF Viewer: http://www.rap.ucar.edu/projects/NCWF/ncwf_viewer/NcwfViewer.jnlp

 

Experimental ADDS, NCWF2: http://weather.aero/convection/

 

ADDS: NCWF1: http://adds.aviationweather.noaa.gov/convection/java/

 

RUC Convective Probability Forecasts: http://ruc.fsl.noaa.gov/rcpf.html

 

Realtime Verification System (RTVS) Convective Statistical Tool:

http://www-ad.fsl.noaa.gov/fvb/rtvs/conv/2001/conv_tool/index.html