The CSTAR Tropical Cyclone QPF group held its first conference call on Thursday, August 4th. There were 6 participants from 2 WFOS’s, the SERFC, HPC, and NCSU.
Over the past year, this project has been split between two studies, though they originated from essentially the same research area, the improvement of rainfall forecasts form tropical cyclones. Drs. Gary Lackmann and Brian Etherton, along with NCSU student Briana Gordon, have worked on an ensemble approach to improving the initialization of tropical cyclones in model forecasts. Initialization is a significant problem for simulating the strength, size, etc. of tropical systems, and ultimately can lead to a poor QPF forecast. Due to the graduation of Briana (journal article in press) and changes in funding at RENCI, emphasis has now shifted to the second part of the project; understanding and forecasting the interaction of tropical cyclones and boundaries.
Most of the call covered the following topics:
– Recent developments and changes at RENCI
* There has a been a shift in funding and emphasis away from hazardous weather related work at RENCI, which has caused personnel changes and resources that there once in place for the tropical cyclone initialization project have been shifted toward a student to work on the tropical cyclone QPF study.
– Update on the tropical cyclone initialization project
*NCSU graduate student Briana Gordon recently defended her thesis on The Sensitivity of Tropical Cyclone Forecast Simulations in the WRF Model to Initial Conditions.
Review of Tropical Storm Hanna
– Heavy rain over the NC/VA Piedmont is, in part, attributed to evaporative cooling from the outer rainbands of TS Hanna (Sept. 2008) and the development of an in-situ frontal zone over central NC. Strong lift up and over the in-situ cool pool (cold-air damming) caused a narrow region of enhanced rainfall (4-7 inches) between Raleigh-Durham and Greensboro, with little to rain further inland in the mountains.
– While previous work has provided case studies and schematics of how TC’s interact with boundaries, there is a need to better understand the predictability of weak, mesoscale, and/or in-situ type boundaries as a tropical cyclone approaches. David Roth noted that models handle large scale, well defined boundaries quite well in the near-term forecast, but struggle with subtle or washed out boundaries.
– Jim Hudgins noted that, with TS Hanna, model forecasts showed the heaviest rainfall along the upslope regions of Virginia. However, very little actually fell along the eastern slopes of the Appalachians, and instead fell over the Piedmont. Discussions suggest that this is likely due to the cool pool (similar to over NC; above) and there is evidence in past cases of a similar scenario where a barrier jet and re-enforcement of the cool pool develops. A question going forward is, “how predictable is this scenario, through model forecasts and forecaster recognition?”
* There are times that models have the boundary in the right place, but QPF may be wrong due to poor forecasts of the tropical cyclone itself.
– New NCSU graduate student Jordan Dale is expected to be working on this project starting this fall (2011).
– A list of cases for study is being compiled, with a focus on tropical storms and weaker hurricanes. Stronger hurricanes will be harder to simulate and complicate results.
– Need to separate a list of cases into categories; for example, boundary interaction type or upper level forcing strength.
– Jonathan Blaes requested that participants try to gather data and feedback if any events occur during this year’s tropical season and a post mortem would be very beneficial to the project.
– An update on the research plan and conference call are expected by late September.