17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB CIMMS / University of Oklahoma NWS Meteorological Development Laboratory Decision Assistance Branch Location: National Severe Storms Laboratory, Norman, OK CIMMS / University of Oklahoma NWS Meteorological Development Laboratory Decision Assistance Branch Location: National Severe Storms Laboratory, Norman, OK Gregory J. Stumpf The Four-dimensional Stormcell Investigator (FSI) for AWIPS NOAA NWS Meteorological Development Laboratory Decision Assistance Branch Location: Global Systems Division, Boulder, CO NOAA NWS Meteorological Development Laboratory Decision Assistance Branch Location: Global Systems Division, Boulder, CO M. Thomas Filiaggi

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB To increase warning skill and lead time, and therefore improve public service, 3D/4D visualization will allow forecasters to better analyze the signatures which are useful in diagnosis of severe and tornadic storms. discover new clues and signatures not easily observed using traditional 2D analysis methods. To increase warning skill and lead time, and therefore improve public service, 3D/4D visualization will allow forecasters to better analyze the signatures which are useful in diagnosis of severe and tornadic storms. discover new clues and signatures not easily observed using traditional 2D analysis methods. Four-Dimensional Stormcell Investigator (FSI)

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Why three (and four) dimensions? Because the atmosphere is four-dimensional! Storm researchers have used 3D/4D displays for years to facilitate their understanding of signatures related to severe weather. How can operational forecasters utilize 3D/4D visualization in an effective, and rapid manner? Because the atmosphere is four-dimensional! Storm researchers have used 3D/4D displays for years to facilitate their understanding of signatures related to severe weather. How can operational forecasters utilize 3D/4D visualization in an effective, and rapid manner?

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB The Lemon Technique (1977) Original storm structure analysis methods presented using vertical cross- sections (RHIs) RHIs are not possible with WSR-88D Original storm structure analysis methods presented using vertical cross- sections (RHIs) RHIs are not possible with WSR-88D

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB AWIPS users must infer vertical storm structure using 2D representations of radar data. All Tilts: a series of 2D “horizontal” cross-sections But PPIs are not horizontal – elevation angles are slanted. AWIPS Vertical Cross-sections (OB7.1) Not easy to use Only complete during short period at end of volume scan Data are remapped to Cartesian space Not dynamic (must re-draw to get new cross-section) Warning decision making demands quick and rapid assessment of the 3D structure of storms. AWIPS users must infer vertical storm structure using 2D representations of radar data. All Tilts: a series of 2D “horizontal” cross-sections But PPIs are not horizontal – elevation angles are slanted. AWIPS Vertical Cross-sections (OB7.1) Not easy to use Only complete during short period at end of volume scan Data are remapped to Cartesian space Not dynamic (must re-draw to get new cross-section) Warning decision making demands quick and rapid assessment of the 3D structure of storms. AWIPS 3D storm interrogation shortcomings

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Solution: Four-Dimensional Stormcell Investigator (FSI) Integrate National Severe Storms Laboratory (NSSL) Warning Decision Support System – Integrated Information (WDSSII) 3D/4D display technology into AWIPS Uses Open GL hardware acceleration on standard video graphics cards FSI will be launched via a D2D extension. Once extension is loaded, mouse click will open FSI application centered on storm of interest Integrate National Severe Storms Laboratory (NSSL) Warning Decision Support System – Integrated Information (WDSSII) 3D/4D display technology into AWIPS Uses Open GL hardware acceleration on standard video graphics cards FSI will be launched via a D2D extension. Once extension is loaded, mouse click will open FSI application centered on storm of interest

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB “FSIalpha” Requirements Provide for quick and easy access to the data (15-30 seconds per storm) Provide linkages between 2D representations and 3D data Still difficult for users to overcome tendency to view in traditional 2D manner Radar data must be “native” 8-bit resolution polar grids Z, V, SRM, SW 4D: Animation while slicing and dicing GUI should have similar look and feel to D2D Keyboard shortcuts Processes must be stable and cause minimal impact to AWIPS system resources Provide for quick and easy access to the data (15-30 seconds per storm) Provide linkages between 2D representations and 3D data Still difficult for users to overcome tendency to view in traditional 2D manner Radar data must be “native” 8-bit resolution polar grids Z, V, SRM, SW 4D: Animation while slicing and dicing GUI should have similar look and feel to D2D Keyboard shortcuts Processes must be stable and cause minimal impact to AWIPS system resources

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB “FSIalpha” design WDSSII GUI (wg) Earth-centric coordinates (earth center is origin) Radar data represented by 2D textures u Elevation PPI data on conical surfaces u Vertical and horizontal Xsections on 2D planes 2D textures represented in 3D space Zoom, pan, pitch, and yaw u “Fly” around and through 3D data WDSSII GUI (wg) Earth-centric coordinates (earth center is origin) Radar data represented by 2D textures u Elevation PPI data on conical surfaces u Vertical and horizontal Xsections on 2D planes 2D textures represented in 3D space Zoom, pan, pitch, and yaw u “Fly” around and through 3D data

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Earth-Relative Coordinates Data are plotted using earth center as domain origin All data plotted at lat/lon/ht-MSL Continuous pan and zoom Data are plotted using earth center as domain origin All data plotted at lat/lon/ht-MSL Continuous pan and zoom

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Earth-Relative Coordinates PPI data shows as 2D textures on conical surfaces Can “fly through” data PPI data shows as 2D textures on conical surfaces Can “fly through” data

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Vertical Cross- Sections Click, draw, and drag concept similar to legacy WDSS and WATADS Vertical 2D texture is drawn with PPI textures Can fly around both textures in 3D space Vertical data accurately represented in 3D space 1:1 aspect ratio 1  beamwidth also represented in the vertical Virtual Volume: Cross sections always have full volume of data. Click, draw, and drag concept similar to legacy WDSS and WATADS Vertical 2D texture is drawn with PPI textures Can fly around both textures in 3D space Vertical data accurately represented in 3D space 1:1 aspect ratio 1  beamwidth also represented in the vertical Virtual Volume: Cross sections always have full volume of data.

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Dynamic Cross- Sections Can update X-Section line, either by Dragging entire line Resetting endpoints 2D and 3D pictures are linked Other representations update on-the-fly Can update X-Section line, either by Dragging entire line Resetting endpoints 2D and 3D pictures are linked Other representations update on-the-fly The Lemon Technique

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB FSI Configuration PPI: Plan Position Indicator VDX: Vertical Dynamic X-Section PPI: Plan Position Indicator VDX: Vertical Dynamic X-Section CAPPI: Constant Altitude PPI 3D Flier: Three- Dimensio nal Flier FSIalpha will have a linked 4-panel design:

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB FSIalpha Configuration CAPPI Constant Altitude PPI 3D Flier Three- Dimensional Flier Linked 4-panel design The Lemon Technique The Lemon Technique PPI Plan Position Indicator VDX Vertical Dynamic X-Section PPI Plan Position Indicator VDX Vertical Dynamic X-Section

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Summary of Benefits of FSI Improved vertical cross-sections Dynamic placement and re-position of a cross-section reference line showing real-time updates to the cross-section data. Cross-sections are no longer a one-time requested RPG product, and are instead generated on-the-fly using 8-bit data. Constant Altitude cross-sections (CAPPIs) 8-bit radar data plotted at constant altitude eliminates the need to sample elevation scan data for altitude or reset elevation angle choices in four-panel displays. Cross-section control is also dynamic, showing real- time updates to the cross-section data. 3D visualization 8-bit radar data from elevation scans, vertical cross-sections, and CAPPIs are plotted as 2D textures in 3D space. A forecaster can then zoom, pan, pitch, yaw, and fly about the data in 3D. Virtual Volumes No volume scan is incomplete. As new elevation scans are updated, they replace the old elevation scans in the virtual volume one-by-one. This means that there are full volumes of data available at all times for cross-sections and data perusal. Access to entire data inventory An “All-Tilts” product only allows the forecaster to peruse a sequential order of elevation scan frames. For a 32 frame limit in VCP12, that only comprises 2 ½ volume scans. The FSI allows the user to access any elevation scan in the radar data inventory RPS list, up to 1 or 2 hours of 8-bit data for all elevation scans. Improved vertical cross-sections Dynamic placement and re-position of a cross-section reference line showing real-time updates to the cross-section data. Cross-sections are no longer a one-time requested RPG product, and are instead generated on-the-fly using 8-bit data. Constant Altitude cross-sections (CAPPIs) 8-bit radar data plotted at constant altitude eliminates the need to sample elevation scan data for altitude or reset elevation angle choices in four-panel displays. Cross-section control is also dynamic, showing real- time updates to the cross-section data. 3D visualization 8-bit radar data from elevation scans, vertical cross-sections, and CAPPIs are plotted as 2D textures in 3D space. A forecaster can then zoom, pan, pitch, yaw, and fly about the data in 3D. Virtual Volumes No volume scan is incomplete. As new elevation scans are updated, they replace the old elevation scans in the virtual volume one-by-one. This means that there are full volumes of data available at all times for cross-sections and data perusal. Access to entire data inventory An “All-Tilts” product only allows the forecaster to peruse a sequential order of elevation scan frames. For a 32 frame limit in VCP12, that only comprises 2 ½ volume scans. The FSI allows the user to access any elevation scan in the radar data inventory RPS list, up to 1 or 2 hours of 8-bit data for all elevation scans.

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Proof-of-concept testing FSI will be alpha tested on AWIPS OB7.2 in 2007 (tentative): SR: Gulf Coast WFO (TBD; mid-winter) SR: Norman, OK - as part of the National Hazardous Weather Testbed (spring) CR: Omaha, NE (spring, early summer) ER and WR: to be determined (summer) Similar in concept to WDSS and WDSSII testbeds Application developer staffing during severe weather operations Feedback via surveys, etc. FSI is primary driver for choice of video graphics card (to support 3D visualization) on LX workstation technology refresh, due in Winter Delayed FSI testing until new hardware can be delivered to the field. FSI will be alpha tested on AWIPS OB7.2 in 2007 (tentative): SR: Gulf Coast WFO (TBD; mid-winter) SR: Norman, OK - as part of the National Hazardous Weather Testbed (spring) CR: Omaha, NE (spring, early summer) ER and WR: to be determined (summer) Similar in concept to WDSS and WDSSII testbeds Application developer staffing during severe weather operations Feedback via surveys, etc. FSI is primary driver for choice of video graphics card (to support 3D visualization) on LX workstation technology refresh, due in Winter Delayed FSI testing until new hardware can be delivered to the field.

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Operational release FSI 1.0 operational release: AWIPS OB8.2 (Fall 2007) FSI 2.0 (tentative – OB9.1): Polarimetric variables, Terminal Doppler Weather Radar (TDWR) Beyond: 3D volume rendering and isosurfaces Display of single radars from multiple locations concurrently 3D grids of mosaicked multiple-radar data Integration with near-storm environment data from numerical models FSI 1.0 operational release: AWIPS OB8.2 (Fall 2007) FSI 2.0 (tentative – OB9.1): Polarimetric variables, Terminal Doppler Weather Radar (TDWR) Beyond: 3D volume rendering and isosurfaces Display of single radars from multiple locations concurrently 3D grids of mosaicked multiple-radar data Integration with near-storm environment data from numerical models

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Operational Considerations Training on the understanding of storm signatures in 3D representations to be developed by WDTB. Workload management/human factors Partnering with Klein Associates, a leading decision science company, to conduct Cognitive Task Analysis. 3D visualization is reaching WFOs via other applications 3D experience is gaining in the field. New NWS Concept of Operations (ConOps), and the clustered peers. Multiple-radar applications. Next-generation AWIPS. Should require fully-integrated support for 3D visualization. Training on the understanding of storm signatures in 3D representations to be developed by WDTB. Workload management/human factors Partnering with Klein Associates, a leading decision science company, to conduct Cognitive Task Analysis. 3D visualization is reaching WFOs via other applications 3D experience is gaining in the field. New NWS Concept of Operations (ConOps), and the clustered peers. Multiple-radar applications. Next-generation AWIPS. Should require fully-integrated support for 3D visualization.

17 October 2006NWA Annual Meeting, Cleveland, OHTom Filiaggi – MDL/DAB Questions? NWS Meteorological Development Laboratory Decision Assistance Branch The views expressed in this presentation are those of the authors and do not necessarily represent those of the NWS, NOAA, or CIMMS. The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an official endorsement or approval by the NWS, NOAA, or CIMMS of any product or service to the exclusion of others that may be suitable. NWS Meteorological Development Laboratory Decision Assistance Branch The views expressed in this presentation are those of the authors and do not necessarily represent those of the NWS, NOAA, or CIMMS. The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an official endorsement or approval by the NWS, NOAA, or CIMMS of any product or service to the exclusion of others that may be suitable.