CSTAR Update – One Year Since the Kickoff Workshop

On October 28th and 29th, 2010 over 30 meteorologists from the National Weather Service (NWS) as well as faculty and students from NC State University (NCSU) gathered in Raleigh to begin implementation of a 3 year Collaborative Science, Technology, and Applied Research (CSTAR) project. The latest project is entitled “Improving Understanding and Prediction of Hazardous Weather in the Southeastern United States: Landfalling Tropical Cyclones and Convective Storms.”

There have been many accomplishments and changes since the project kicked off at the end of October 2010.  A list of some of the accomplishments for the 5 sub-projects is shown below.

Project wide items –  

  • The collaborative group has expanded with the addition of NWS Huntsville, AL who has joined the High Shear-Low CAPE portion of the CSTAR project
  • Collaborators are from NC State, 11 WFOs, 3 national Centers,  a NOAA laboratory which now includes more than 50 individual collaborators.
  • A blog, entitled Collaboration for Improved Meteorology in the Mid-Atlantic and Southeast (CIMMSE), was developed to facilitate discussion of ongoing events and share research results between PIs, students, and collaborators: https://cimmse.wordpress.com.
  • A new NWS-NCSU Collaborative Research and Training Web Site was developed. The web site will feature a public area where training materials, references, project updates, and project deliverables will be presented.

Improving Inland Gridded Forecasts of Tropical Cyclone Winds and Wind Gusts –

  • A survey of methods and approaches forecasters use to create wind and wind gust grids in GFE for tropical cyclones was conducted when the project started (results)
    Verification of NDFD winds for the limited number of tropical cyclones in the study domain since the NDFD became operation was completed. Initial results consistently indicated an over prediction of wind speeds.
  • Development of wind/wind gust climatologies is nearing completion and should include Weibull plots of wind speed at numerous locations, comparisons of Weibull distributions for climatology vs. landfalling TCs, land decay plots comparing wind speed vs. time after landfall.
  • NHC collaborators described the methodology and tools (or lack thereof) for analysis and forecasting of the quadrant wind radii for tropical cyclones which are used to generate the TCM wind product and to populate wind grids in GFE via the TCMWindTool (details with additional slides)
  • We were fortunate to have Hurricane Irene move through our study area during the project which allowed significant real time and post event examination of the forecasts and the storm’s winds and wind gusts. NC State student Bryce Tyner visited WFO RAH a couple of times to watch the process in which forecasters create wind/wind gust grids during TC’s. An Irene post assessment was completed.
  • This research continues to point to the need for a more systematic and science based approach to adjusting the NHC produced TCM wind guidance to produce wind and wind gust grids at WFOs.

Tropical Cyclone Initial Conditions –

  • A method of improved initial conditions was developed by graduate student Briana Gordon and Dr. Etherton. Initial efforts utilized the GFDL bogus vortex approach which did not perform well. Subsequent efforts used data assimilation approaches and a hybrid ensemble Kalman filter (EnKF) approach was devised and tested for the case of Hurricane Earl (2010).
  • The development of a potentially improved, yet computationally affordable method for obtaining initial conditions for strong hurricanes represents a major accomplishment in the project. Routine implementation of this approach was hampered by the unanticipated loss of a substantial computer resource.
  • Perhaps most important, this effort has resulted in a much better understanding of how the modeled hurricane evolves with and without a bogus vortex. This includes a better understanding of what the bogus vortex is and its limitations.
  • While the bogus vortex may provide a minimum sea level pressure that is closer to the observed, the modeled hurricane does not evolve as realistically as expected. In contrast, a modeled hurricane initialized with the GFS, which typically has a much higher sea level pressure than the observed, will often, within 6-12hours, achieve a sea level pressure and structure that is close to what was achieved with the bogus.
  • These results argue that it is not just an accurate depiction of the hurricane’s central sea level pressure that is important in initial conditions, but rather the whole structure of the pressure field, relative humidity, etc. that needs to be initialized more accurately.

High Shear / Low CAPE Severe Weather –

  • Two graduate students arrived late this summer to work with Dr. Parker on the High Shear Low Cape (HSLC) project. We are fortunate to have both Keith Sherburn from Oklahoma University, and Jason Davis from Valparaiso join the effort. Both students earned AMS fellowships and their contributions to the project will be partially supported by fellowship funds. This is a great opportunity since both students will be able to work on this project.
  • Nineteen collaborators from ten WFOs worked together to build a 5 year catalog of HSLC severe weather events across the Southeast. The participants examined severe weather events from 2006 through early 2011 within their respective CWA’s. The SPC mesoanalysis archive was consulted to identify those events for which environmental parameters met the threshold for high shear (0-6 km shear >= 35 kt) and low CAPE (<= 500 J/kg). The collaborators identified a total of 106 severe weather episodes that met the HSLC criteria. From these 106 cases, 25 were chosen as “high priority” for more detailed analysis.
  • NC State collaborators have requested gridded SPC mesoanalysis data for the 106 cases from the SPC collaborators. These data will be analyzed and converted to a GIS/KML format and made available to the WFO collaborators for viewing in Google Earth. The WFOs will use these data for detailed mesoanalysis of the 25 high priority cases.
  • Pat Moore and Justin Lane at GSP are working on a case study of the 5 January 2007 north Georgia and Upstate South Carolina severe weather episode. In addition to mesoanalysis, the case studies will include detailed “volume scan-to-volume scan” analysis of radar data, an analysis of storms that prompted issuance of subsequently unverified Tornado Warnings, and more. Once completed, this case study will be critiqued by all WFO collaborators, with the final result used as a template for analysis of all 25 “high priority” cases.
  • WFO forecasters have noted that the SPC meosanalysis products of CAPE did not provide the desired level of detail on the lower end of the CAPE analysis needed for HSLC events. Based on input from collaborators at the WFOs, the SPC made changes to the contour setup for SBCAPE, MLCAPE, and MUCAPE analysis which will result in an additional contour at either 100 J/kg or 250 J/Kg along with the existing contours at 500 J/Kg, 1000 J/Kg, etc.  It is expected that these new contours will provide operational forecasters with a better approximation of the amount of CAPE in these low CAPE severe weather events.

Inland QPF Associated with Tropical Cyclones –

  • With changes to the NC State Principal Investigators, this portion of the project will be receiving more support and is now fortunate to have a graduate student participating. Jordan Dale, a Penn State graduate, arrived in August, and this project will serve as the foundation for his master’s thesis.
  • Previous work examined the role of evaporational cooling and the subsequent development of a surface boundary, on the precipitation distribution associated with Tropical Storm Hanna.  For Hanna, the evaporational cooling and boundary appeared to modulate the QPF, but the change was not extreme. At least in this case, it was noted not to over emphasize CAD and diabatic processes. 
  • Despite the late start, the project is gaining more focus and will be examining the predictability of inland surface boundaries and their impact on tropical cyclone QPF. One particular question will focus on how much does evaporational cooling in the pre-storm environment set the stage for subsequent QPF.
  • Cases are currently being identified and include those without a boundary, weak boundary, and strong boundary cases.

High-Resolution Mid-Atlantic Ensemble (HME) –

  • To support several of the projects noted above, an effort to establish and support a routine High-Resolution Mid-Atlantic Forecasting Ensemble (HME) was including in the project proposal.
  • This effort was already underway before this CSTAR project started but CSTAR provided additional support that allowed a diverse group of 14 partners to create a mesoscale ensemble from independent members.
  • Real time production and distribution of model ensemble products began in 2010 and products were viewed and utilized by operational forecasters at several WFOs before the unanticipated loss of substantial computer resources and the loss of a principal investigator terminated the operational production of the HME.
  • An effort to document the HME, note its successes, and identify lessons learned and suggestions for future federated ensembling efforts is underway.
  • A follow-up activity to this project has been the creation of a web page that allows forecasters the ability to view a single field (model forecast radar reflectivity) from a variety of high resolution model sources, including many that were in the HME. The page is referred to as the High Resolution NWP Reflectivity Comparison web page – http://www.erh.noaa.gov/rah/hiresnwp/


This entry was posted in CSTAR, General Information, High Shear Low Cape Severe Wx, TC and Boundary QPF, TC Improved Initial Conditions, TC Inland and Marine Winds. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s