9.2 ANALYSIS OF SUMMER CONVECTION OVER CENTRAL ALABAMA

Gary D Goggins*, Matthew L Grantham, Scott W Unger, Kevin B Laws, and Kevin J Pence National Weather Service Forecast Office Birmingham, Alabama 35040

Logan Dawson North Carolina State University Raleigh, North Carolina 27695

1. Introduction boundaries that often initiated summertime

Short-term* forecasting throughout thunderstorm development. the southeastern United States during the In this study across Alabama, summer season can be very challenging for numerous boundaries were found to initiate operational forecasters. Often, forecasters convection, including thunderstorm outflow are mistakenly under the impression that boundaries, ridge top or orographic thunderstorm development during the boundaries, sea-breeze fronts, shallow summer months is “random” and can “pop- fronts (defined as fronts without the up” at any location throughout a given area. presence of dynamical upper level support), However, as noted in studies in Colorado deep synoptic fronts, convective horizontal (Wilson and Schreiber 1986), and North rolls, boundaries caused by differential Carolina (Koch and Ray 1997), it is readily heating, and boundaries of unknown origin. acceptable that to produce deep layer A total of 467 boundaries were identified convection, some kind of triggering or lifting during the summer of 2009 over Alabama, mechanism is needed in the lower boundary and statistics were gathered to determine layer to lift surface air parcels to their level the convective nature of these triggering of free convection. mechanisms. Upper-air sounding During the summer of 2009 (June information was also used as a supplement through September), forecasters at the for thunderstorm initiation based on the National Weather Service Office in amount of instability present on a given day, Birmingham, Alabama (WFO BMX) along with in-situ calculations of the LFC attempted to identify several of these and LCL heights. The data and statistics triggers for convective initiation through the from this study were compared to a use of detailed mesoscale analyses using previous study done by Koch and Ray conventional surface observations. These (1997), where they identified different analyses were augmented with other boundaries typically associated with remote sensing operational tools, such as summertime convection in central and NEXRAD, and high resolution visible and eastern North Carolina. infrared imagery from the Geostationary The findings in this presentation will Operational Environmental Satellite present initial statistics of boundary (GOES). The radar and satellite datasets, interactions to aid forecasters in the often along with the surface mesoanalyses, were difficult, short-term forecast process. Also, used daily, in real-time, to detect a variety of additional high resolution satellite imagery and products, such as those created by the * Corresponding author address: Short-term Prediction and Research and Gary D Goggins Transition Center (SPoRT), will be National Weather Service Forecast Office described for the future of determining 465 Weathervane Rd. Calera, Alabama 35040 boundary locations (Goodman, S.J. et al. Email: [email protected] 2004). The experimental products such as those developed at the SPoRT Center, could be utilized throughout many National the radar due to Rayleigh scattering from Weather Service Forecast Offices across insects, most likely. Boundaries identified the southeastern United States to improve farther from an individual radar were most short-term convective initiation and likely from Bragg scattering, where a precipitation forecasts during the summer thermal plume or cloud was present. Each season. radar was analyzed from throughout the coverage area, and boundaries were 2. Methodology and Preliminary Results identified from each separate field of view to

Daily analyses during the summer determine the total boundaries for that months June through September of 2009 particular day. The study that was are attempted daily beginning each performed was very different from Koch’s morning. Mesoanalyses of surface study in 1997, due to the fact that radar observations over Alabama and throughout coverage was much more extensive for the the surrounding region of Mississippi, Alabama study. Typically, only one WSR- Southern Tennessee, Georgia, and the 88D was utilized for the North Carolina Florida panhandle were manually study. performed. Mesoscale Analysis and A sample of the daily detailed Prediction System (MAPS) Surface mesoanalysis performed on 28 June 2009 Assimilation System (MSAS) and the Local at 2100 UTC is shown in figure 1. As the Analysis Prediction System (LAPS) were initial convection formed along previous both utilized as an aid in locating larger outflow in northern Alabama, additional scale features in the analyses. The daily convection initiated along a slow-moving mesoanalysis was performed to determine synoptic front (extreme northern periphery possible locations of synoptic, or possibly of the study area), and along a prefrontal mesoscale boundaries, based on surface shallow boundary and sea breeze data alone. Next, 1200 UTC sounding data boundary, both located across the southern- were analyzed to determine the amount of third of the analyzed region. Subsequent instability present, along with additional deep layer convection also formed along sounding parameters, including the LFC, outflow boundaries across central Alabama which an air parcel must reach for as boundary interactions incurred. Note the thunderstorm formation. GOES 1km visible utilization of remote sensing data to identify resolution satellite imagery was also and label the various boundaries present. employed to compliment the mesoanalysis Once boundaries were identified, due to the course resolution in the various they were categorized using the same surface reporting stations. However, on methods as in Koch and Ray (1997). multiple occasions, the viewing angle from Boundaries were labeled as autoconvective GOES imagery became obstructed as thick if they are able to initiate convection alone, cumulonimbus anvil generated from the with no other boundary interactions. Other initial deep layer convection made boundary interactions were labeled as subsequent boundary identification difficult mergers (one boundary overtaking another at times. This was also noted in Wilson and boundary), intersections (boundaries Schreiber (1986) and Koch and Ray (1997). colliding at an angle of >30°), and collisions NEXRAD radar data throughout the region, (boundaries colliding at an angle <30°). including KBMX, KMXX, KHTX, KGWX and Seven boundary types were subjectively KEOX were then utilized. Due to the identified and documented into a statistical cumulonimbus anvil development, radar spreadsheet. The boundary types identified data became the principal tool used for include convective outflow, sea breeze, mesoscale boundary identification. As was unknown boundaries, horizontal convective found in Koch’s study in North Carolina, rolls, shallow fronts, synoptic fronts, and boundaries were viewed within 50 nm from boundaries caused by differential heating.

Figure 1. Detailed mesoanalysis (top) from 28 June, 2009 at 2100 UTC, with areas of autoconvection highlighted in orange and additional convection initiated from boundary interactions highlighted in yellow. Composite base reflectivity (bottom-left) and 1km visible GOES- East imagery (bottom-right) are also shown.

The boundary statistics were then synoptic flows helped the sea breeze front compared with the previous work done by form early in the day and pushed it well Koch and Ray (1997). inland, whereas offshore synoptic flows Boundary identifications were impeded the development and propagation performed for a total of 71 days over the of the front. Additionally, this collision-like study period. Overall, 467 boundaries were interaction between the offshore synoptic subjectively identified by operational flow and the boundary triggered strong deep forecasters. Sixty-six percent (66.4%) of layer convection along the boundary. This those boundaries were identified as relationship between the background thunderstorm outflow boundaries, 15.4% synoptic-scale flow and sea breeze fronts sea breeze, 7.7% were unknown, 4.9% suggested that such boundaries observed horizontal convective rolls, 2.1% shallow within the Birmingham CWA (central fronts, 1.7% differential heating, and only Alabama) have to propagate around 150km 1.7% synoptic fronts due to the exclusion of inland, and are often weak signatures that most synoptic frontal events that occurred are not as likely to be convective. Because throughout the period of study. Of the the immediate coastal region was included different boundary types, 90% of the outflow in the observational study in Koch and Ray boundaries were labeled as being (1997), the active sea breeze fronts that are autoconvective, while 84% of all fronts isolated near the coast were observed and (shallow and synoptic), 80% of the documented, while they were not frequently topographical boundaries, and 55% of the analyzed across central Alabama. unknown boundaries were also autoconvective. Combining the shallow and synoptic fronts into one category, fronts are convective 94% of the time, while 85% of the outflow boundaries were convective, 49% of the topographical boundaries (which also included sea breeze fronts), and 28% of the unknown boundaries were also associated with convection. The North Carolina study showed thunderstorm outflows were convective 90% of the time, 84% for fronts, 80% for topographical and sea breeze, and 55% for unknown boundaries. Figure 2 depicts the Figure 2. Frequency of boundaries comparisons between the present study producing new convection as observed across over Alabama and the eastern North Alabama and eastern North Carolina (after Carolina study by Koch and Ray (1997). Koch and Ray 1997). The most significant difference between the two studies was seen in the 3. Future Summer Convection Studies percentage of topographical boundaries that As technology continues to advance were convective. The number of convective throughout the weather forecasting science, topographical boundaries observed in North new tools for short-term forecasting are Carolina was more than double the number becoming available. As these new tools identified in Alabama, likely due to the become readily accessible and operational frequency of sea breeze fronts. Past at NWS Weather Forecast Offices, research indicated that mean low-level flow collaboration between offices can be made affected the distance the sea breeze to improve short-term summertime propagated inland as well as the potential convective forecasting on a region-wide for the boundary to become convective. As scale. The entire region, not just the state noted in Koch and Ray (1997), onshore of Alabama, can therefore benefit from this new technology. A great method of gradients existed along and near the achieving this type of collaboration and leading outflow boundary in north central implementing new forecast tools can be Alabama, and coincident to the prefrontal done in conjunction with the SPoRT Center trough present in east central Alabama into housed at the NASA Marshall Space Flight central Georgia (refer to Figure 1 for Center and the University of Alabama in placement of these features). Figure 4 also Huntsville. The SPoRT program was highlights where the initial deep layer created to accelerate the implementation of convection was forming along these same NASA earth science operations, data gradients. assimilation, and modeling research into Furthermore, additional datasets National Weather Service Forecast Offices from the MODIS satellite can provide details and decision making operations (Goodman, through high spatial resolution products and S.J. et al. 2004). The experimental imagery that will become readily available products utilized through SPoRT are on the next generation of GOES satellites. designed to focus on the regional scale with Lifted Index, Convective Available Potential emphasis on high impact weather forecast Energy, and total column precipitable water improvements on a time scale from 0-24 are just a few of the derived products that hours. The SPoRT program is an excellent will be deemed useful in determining resource that WFO’s can use to improve favorable areas for convective initiation. short-term forecasting for their respective CWA’s throughout the southeast. One 4. Conclusion particular product that the SPoRT program Summer weather patterns is beginning to implement is the Land throughout the southeastern United States Information System (LIS). The LIS can be very complex and extremely difficult integrates satellite-derived datasets, to forecast. As stated in Koch and Ray ground-based observations, and model (1997), the misconception that pop-up reanalysis data to form Land Surface thunderstorms are random, from the days of Models (LSM’s) that characterize surface pre-WSR-88D radars, has unfortunately states and moisture fluxes on the earth’s been carried over into today’s forecast surface (Case, J.L. et al. 2007). Figure 3 offices. However, these so called “random” presents two of the parameters (soil type summertime thunderstorms are almost and vegetation type) that LIS takes into always associated with identifiable account. The LIS can be implemented into boundaries. These boundaries can be local Weather Research and Forecasting identified through careful mesoanalyses of (WRF) models at each office, or become surface observations and the use of radar available within each forecast workstation and satellite datasets. The SPoRT imagery database. Due to the complex terrain and and products will add a layer of additional soil types within each office’s CWA, the LIS information that may help forecasters better can pinpoint soil moisture and vegetation understand the boundaries of unknown type differences that could possibly lead to origin. differential heating and associated Current methods of forecasting boundaries, or moisture boundaries that can probability of precipitation (POP) and possibly initiate thunderstorm development. quantitative precipitation forecasts (QPF) Forecasters will then be able to add this fail to give the general public the best information, which was previously unknown, forecast possible during the summer into their arsenal for short-term prediction of months. Multiple products are now summertime thunderstorm initiation. Figure available, and new tools will become 4 shows an example of the LIS-Noah Land available in the future, to help determine Surface Model (LSM) latent heat flux from more precisely the locations of 1800 UTC on 28 June, 2009. As noted from thunderstorm initiation. A more diligent the figure, significant latent heat flux

Figure 3. Sample depictions of soil type and vegetation type classes across Alabama. Images provided by the NASA SPoRT Center (Goodman, S.J. et al. 2004) in Huntsville, Alabama.

Figure 4. Land Information System-Noah Land Surface Model latent heat flux output from 28 June, 2009 at 1800 UTC. The black contours indicate regions of convective initiation at the current hour. effort and initiative to utilize these procedures can better serve the public by generating improved forecasts for specific locations, instead of the “shotgun approach” References to POP and QPF forecasts, which currently Case, J. L., K. M. LaCasse, J.A. Santanello blanket an entire area. Jr., W. M.Lapenta, and C. D. Petars- Summertime convection is a Lidard, 2007: Improved Modeling of regional phenomenon, not just an Land-Atmosphere Interactions Using a occurrence that happens throughout central Coupled Version of WRF with the Land Alabama. Collaborations will be possible Information System. 21st Conf. on with other forecast offices throughout the Hydrology, AMS, San Antonio, TX, southeast to develop similar summertime P5A.4. convective forecast methods. Many of the boundaries identifiable in this study crossed Goodman, S. J. et al, W. M. Lapenta, G. J. into other states and CWAs. The end Jedlovec, J. C. Dodge, and J. T. results of careful and diligent analysis are Bradshaw, 2004: The NASA Short-term expected to continue to improve short-term Prediction Research and Transition POP and QPF forecasts for summer (SPoRT) Center: A Collaborative Model convection. for Accelerating Research into Operations. 20th Conf. on Interactive Acknowledgments Information Processing Systems (IIPS) The authors of this study would like for Meteorology, Oceanography, and to thank Jonathan Case, Kevin Fuell, and Hydrology, AMS, Seattle, WA, P1.34. Dr. Geoffrey Stano of the NASA Short-term Prediction Research and Transition Center Koch, S. E. and C. A. Ray, 1997: in Huntsville, Alabama for their assistance Mesoanalysis of Summertime with Land Information System figures, and Convergence Zones in Central and for clarification of datasets related to this Eastern North Carolina. Wea. research. We look forward to future Forecasting, 12, 56-77. collaborations. We would also like to thank Dr. Wilson, J.W., and W. E. Schreiber, 1986: Steven Koch, the director of the NOAA Initiation of Convective Storms by Radar- ESRL Global Systems Division in Boulder, Observed Boundary-Layer Convergent CO, for his insights and guidance with this Lines. Mon Wea. Rev., 114, 2516-2536. project.