Event Summary
     National Weather Service, Raleigh NC

Tropical Storm Hanna, September 2008
Updated 2009/06/24


Satellite Image of Tropical Storm Hanna at 1605Z on 2008/09/04 - Click to enlarge


Event Headlines -

...Hanna moved onshore near the border between North and South Carolina a little after midnight on September 6th and then tracked northward across the Coastal Plain of North Carolina near the Interstate 95 corridor...
...The heaviest rain fell in a corridor extending from near Laurinburg, NC north to near Durham, NC and then to near Roxboro, NC. The axis of the heaviest rainfall was skewed to the left of the track with a sharp decrease in rainfall amounts beyond 160 km or 100 miles left of track. Very little precipitation was observed across the Foothills and Mountains of North Carolina...
...Up to 7 inches of rain along with rainfall rates of 2 to 3 inches per hour produced flash flooding across portions of central North Carolina with the most significant flooding in the Triangle area, the Fayetteville area and in Person, Alamance, and Moore Counties...
...No confirmed tornadoes were observed across North Carolina during the period in which Hanna moved across the state...
...Wind Advisories and High Wind Warnings were placed considering the meteorology, and although High Wind Warning criteria based on observations did not verify, there was considerable impact in the warning and advisory areas with downed trees, limbs and power lines...

Event Overview -

The tropical wave that would later develop into Hanna emerged off the coast of Africa on August 19, 2008 and then tracked westward across the Atlantic Ocean. Shower and thunderstorm activity associated with the tropical wave gradually increased as the wave continued westward. On August 26th, the tropical wave developed an area of low pressure around 500 miles east-northeast of the northern Leeward Islands. On August 28th, the low pressure area developed into Tropical Depression Eight at 0000 UTC. Later that day, the system reached Tropical Storm status, and the system was named Hanna. At the time, the low-level center of circulation was partially exposed on the western edge of the convection due to westerly shear from an upper-level low to the northwest of Hanna.

An easterly flow associated with a large subtropical ridge extending over the western Atlantic continued to move Hanna on a westward course. The upper-level low that had been producing wind shear over Hanna for several days took its toll on the system and the low-level center separated from the deep convection, with the satellite appearance briefly resembling that of a subtropical storm on August 30th. The upper-level low shifted southward and gradually dissipated the next day, resulting in a lower shear environment conducive for strengthening. Deep convection began to form over the circulation center on September 1st and the storm began to intensify. This was a very active period for tropical cyclones in the Atlantic with Hurricane Gustav, Tropical Storm Hanna, Tropical Depression 9, and a few other areas of interest.

The development of deep convection which began around 00 UTC on September 1st, continued overnight and into the next morning when a period of rapid intensification began. At around the same time, a building deep-layer ridge over the eastern United States forced Hanna to slow and turn southwestward. Reconnaissance aircraft indicated that Hanna reached Hurricane intensity by 18 UTC on September 1st, while located just north of the Caicos Islands. Hanna reached its maximum intensity at around 00 UTC on September 2nd. Just after the period of rapid intensification, Hanna abruptly began to weaken as strong northerly shear associated with the building ridge over the United States and Hurricane Gustav increased. Hanna weakened to a Tropical Storm by 12 UTC on September 2nd.

Hanna continued to weaken as it tracked west-southwestward near the Southeastern Bahamas. Early on September 3rd, Hanna turned southeastward passing less then 50 miles from the northern coast of Haiti. Hanna turned northward during the afternoon of September 3rd and then turned northwestward completing a counter-clockwise loop. During the loop, the storm briefly exhibited quasi-subtropical appearance as it interacted with an upper-level low over the Bahamas before re-strengthening later on September 3rd.

Hanna continued to struggle with a persistent easterly to northeasterly shear on September 4th that was responsible for an asymmetric appearance on satellite imagery. Hanna’s intensity remained between 55 and 60 kts while the center passed just east of the central and northwestern Bahamas. Water vapor satellite imagery also showed an area of dry air extending south across the Carolinas, Florida, Cuba and into portions of the central Bahamas. These factors inhibited the organization of the system which maintained a broad circulation with some sub-tropical characteristics.

On September 5th, Hanna moved away from the upper-level low and turned north around the western periphery of the subtropical ridge. Hanna slowly intensified during the day with convection initially developing around the perimeter of the storm during the morning before additional convection developed near the circulation center during the evening. The central pressure of 987 hPa fell to 980 hPa by 15 UTC. After rising slightly the pressure fell again to 978 hPa by 03 UTC on September 6th.

Hanna accelerated northward and made landfall as a 60 kt tropical storm near the border between North and South Carolina at 0720 UTC on September 6th. The storm moved north and weakened as it moved across the western Coastal Plain of North Carolina during the morning hours of September 6th. The axis of heaviest rain fell west of the storm track in an arc from Laurinburg to Sanford to Durham to Roxboro. The strongest winds were observed in the immediate coastal region, especially across the southern coast near Wilmington. Despite the reduced winds, scattered wind damage, primarily from falling trees, was observed across much of central and eastern North Carolina.

Hanna exited North Carolina near Roanoke Rapids and turned northeastward. The circulation center passed very close to New York City shortly after 00 UTC on September 7th and then became extratropical when it merged with a cold front over southern New England.

Media reports indicated that over 530 deaths were associated with Hanna. The majority of them were associated with flooding in northern Haiti. This made Hanna the deadliest tropical cyclone in the Atlantic basin since Hurricane Stan in 2005 and the deadliest tropical cyclone of the 2008 season.

Additional Details

Additional details are available from the National Weather Service Wilmington, NC...
National Weather Service Wilmington, NC Hanna Information


Precipitation Totals from Tropical Storm Hanna

The axis of heaviest rain fell west of the storm track in an arc from Laurinburg to Sanford to Durham to Roxboro where rainfall amounts ranged between 3 to 5 inches with a area of very heavy rain totaling 6 to 7 inches across Hoke, Lee, and southern Chatham Counties. Very little if any rain fell across much of the Foothills and the Mountains of North Carolina.


Precipitation from Tropical Storm Hanna


The map below contains a precipitation analysis across the southeastern U.S. from the Hydrologic Prediction Center for the period of September 4th through September 7th, 2008. Click on the map to open a larger image.

Precipitation analysis across the southeastern U.S. from the Hydrologic Prediction Center for Tropical Storm Hanna - click to enlarge



Maximum Wind Gusts from Tropical Storm Hanna

The map below contains the maximum wind gusts in miles per hour (MPH) from Tropical Storm Hanna as it moved across North Carolina on September 5th and 6th, 2008. Wind Advisories and High Wind Warnings were issued across portions of interior North Carolina in advance of the storm. The wind had a considerable impact in the warning and advisory areas with numerous reports of downed trees, limbs and power lines.

Maximum wind gusts from Tropical Storm Hanna



Severe Weather Reports from Tropical Storm Hanna

The map below contains severe weather reports received by the National Weather Service during the period in which Hanna impacted North Carolina on September 5th and 6th, 2008.

Severe Weather reports from Tropical Storm Hanna




KRAX Radar Loops

A Java loop overview of the entire event with images from every hour between 1300 UTC September 5, 2008 through 1800 UTC September 6, 2008 is available here. Note - this loop includes 30 frames

A Java loop overview of the entire event with images from every 15 minutes between 1300 UTC September 5, 2008 through 1800 UTC September 6, 2008 is available here. Note - this loop includes 117 frames

The KRAX reflectivity image below is from 0801 UTC on September 6, 2008 when very heavy rain was falling across much of the Piedmont of North Carolina. One hour precipitation estimates from the KRAX radar ending at 0752 UTC indicated a large area of 0.5 to 1.5 inches of rain extending in a band across the western Piedmont with embedded amounts of 1.5 to 2.5 inches.

KRAX reflectivity image - click to load loop



Regional Radar Loop

A Java loop of regional radar imagery from 1158 UTC on September 5, 2008 through 1758 UTC on September 6, 2008 is available here. Note - this loop includes 60 frames.

The regional reflectivity image below is from 0718 UTC September 6, 2008 just as Tropical Storm Hanna was making landfall near the border of North and South Carolina. Note the distribution of precipitation with the most of the significant precipitation falling to the north and especially northwest of the circulation center. Outside of one relatively large band of rain to the northeast/east of the circulation center, there is a relative dearth of precipitation across the eastern half of the storm, including the eastern portion of the circulation center.


Regional reflectivity image - click to load loop



Surface Analysis

The surface analysis from 06 UTC Saturday, September 6, 2008 is shown below. The map depicts the landfall of Tropical Storm Hanna near the border of North and South Carolina which occurred at 0720 UTC on September 6, 2008.

A Java loop of surface analysis imagery from 18 UTC Thursday, September 4, 2008 through 12 UTC Sunday, September 7, 2008 shows the evolution of the event. Note the ridge of high pressure that extends from off the New England coast into the Mid Atlantic and Southeast on Thursday and Friday. A slow moving cold front extended from Great Lakes region south into the Ohio Valley and the Mississippi Valley on Friday and Saturday.

Surface analysis from 06 UTC Saturday, September 6, 2008



Mechanisms Determining the Distribution of Precipitation with Tropical Cyclones

Numerous studies have examined the various factors that play a role in precipitation distribution associated with Tropical Cyclones (TC's). Intuitively, factors such as the tropical cyclone's speed and direction of motion along with its areal extent play a role in the resultant precipitation. Hanna provides an opportunity to examine and review many of the factors that determine the distribution of precipitation associated with a modestly organized tropical storm. A brief discussion of the various mechanisms involved in the distribution of precipitation is included below with more complete discussions contained the subsequent sections of the event summary.

Cote, Bosart, Keyser, and Jurewicz (2008) investigated concentrated areas of heavy rainfall that are observed well upstream from the land falling tropical cyclone and are separate from the cyclone’s primary precipitation shield. They described these events as Predecessor Rainfall Events (PRE's). PRE’s typically form under the following conditions: 1) when persistent, deep meridional flow transports tropical air far from the TC, 2) in favored upslope regions or along synoptic/mesoscale boundaries, 3) along and just west of a low-level theta-e ridge, near the strongest gradient 4) near a midlevel jet-entrance region confluence zone and 5) under favorable upper-level jet dynamics.

Hart (2003) examined cyclone phase space diagrams to plot the life cycle of tropical cyclones based on their barotropic and baroclinic structure. The cyclone phase space diagrams incorporate thermal wind and thermal asymmetry relationships can be used to anticipate and monitor extratropical transitions which have been shown to impact the distribution of precipitation. The cyclone phase space diagrams can also be used to evaluate the structure of subtropical storms and anticipate the evolution between tropical, subtropical, and extratropical systems.

Extratropical Transition (ET) is a gradual process in which a warm-core tropical cyclone loses tropical characteristics and become more extratropical in nature (Jones et al. 2003). During ET, the precipitation area expands poleward of the center and the heaviest precipitation is shifted to the left of the tropical cyclone track (Atallah and Bosart 2003).

Subtropical cyclones are cyclones in tropical or subtropical latitudes that have characteristics of both tropical cyclones and mid-latitude (or extratropical) cyclones. These storms usually have a relatively broad zone of maximum winds located farther from the center, and typically have a less symmetric wind field and distribution of convection.

In-situ Cold Air Damming (CAD) can have a significant impact on the distribution of precipitation with tropical cyclones. The establishment of the in-situ stable layer can lead to the development of a low-level boundary along portions of the perimeter of the CAD region. Several studies have shown that a surface boundary can focus the heavy precipitation in a mesoscale band located along and on the cold side of the surface boundary. Banded mesoscale precipitation structures associated with surface boundaries have been previously documented with landfalling and/or transitioning tropical cyclones (Atallah and Bosart 2003; Colle 2003). Klein (2006) noted that the area of low-level frontogenesis is typically located left and poleward of the storm track with the enhanced precipitation region along or in cold sector of the boundary.

Cline (2003) discussed the role of along track or cross track bulk shear in the distribution of precipitation. He found that tropical cyclones with cross track shear have the precipitation distorted in an asymmetrical fashion while along track shear results in precipitation spreading well in advance of the storm.

Atallah et al. (2007) studied the distribution of precipitation accompanying U.S. landfalling and transitioning tropical cyclones. They argue that storms that have most of their rainfall distributed on the west side of storm’s track or Left of Track (LOT) interact with upper level trough and potential vorticity maximum in a synergistic way. In addition, the axis of precipitation of LOT storms often is aligned parallel to the movement of the storm which increases the period of heavy rainfall. A left of track precipitation distribution is often characterized by a positively tilted mid-latitude trough approaching the tropical cyclone from the northwest. The precipitation stretches out well north of the system due to impulses moving around the upper trough and frontogenesis ahead of the trough.

Atallah et al. (2007) noted that tropical cyclones with their rain distributed to the east of the storm’s track or Right of Track (ROT) typically have the axis of the 850-200-hPa shear oriented more westerly than the more north-south oriented axis associated with LOT cases. The location of the upper level jet streak is much farther removed for ROT cases than for LOT cases. The baroclinic zone as shown by the leading edge of the 1000-500 hPa thickness is also much farther removed from the storm. A right of track precipitation distribution is often characterized by tropical cyclones that are generally steered by shear lines or through a break in the subtropical ridge. Rainfall tends to be concentrated near and right of track.

Klein (2007) noted important differences in the vertical structure across surface boundaries in the LOT and the ROT conceptual cross sections. The warm-frontal boundary that is often associated with LOT cases is characterized by a very deep frontogenesis pattern that tilts toward the cooler air with height and is strongest in the boundary layer and at upper-levels. The ascent pattern tilts toward cooler air with height and is located on the warm side of the frontogenesis maximum. Ascent is greatest in the mid-troposphere above and on the warm side of the frontogenesis region. The warm-frontal boundary associated with the ROT pattern contains strong frontogenesis that is largely confined near the surface. Strong, upright ascent is located over and on the warm side of the frontal boundary with the largest values immediately above the boundary layer.


Predecessor Rain Events (PRE’s) Associated with Tropical Cyclones

A conceptual model of a left of track PRE ahead Of SR Or AR TC - click to enlarge During recent active Atlantic hurricane seasons, observational evidence has suggested that heavy mesoscale rain events can sometimes form unexpectedly well in advance of landfalling and near-coastal tracking tropical cyclones (TC’s). These areas of heavy rain, well in advance of tropical cyclones, are significant in that they can produce flash flooding well ahead of the main precipitation area associated with the TC. In addition, they can produce antecedent conditions that make flash flooding directly associated with the TC more likely (saturate the soil and raise the levels of rivers and streams).

Cote, Bosart, Keyser, and Jurewicz (2008) investigated concentrated areas of heavy rainfall that are observed well upstream from the landfalling tropical cyclone and separate from the cyclone’s primary precipitation shield. They described these events as Predecessor Rainfall Events (PRE's). They investigated Atlantic tropical cyclones from 1998-2006 and identified a PRE if the event met 3 criteria...
  • Coherent area of rain displaced pole ward of the TC
  • Maximum rainfall rates exceeded 100 mm in 24 hours
  • Moisture transport from the TC toward the PRE

    A total of 47 of these predecessor rain events (PRE’s) occurring downstream of 21 TC’s were identified between 1998 and 2006, representing approximately one third of all landfalling TC’s during this time period. The more extreme PRE’s pose an important forecast challenge because they can cause significant flooding with little prior warning, especially in those 25% of cases in which the TC itself later produces additional heavy rainfall across the same region. During the 1998-2006 period, 47 PRE's associated with 21 TC’s were identified. The study found...
  • An average of around 2 PRE’s per TC
  • Nearly 1/3 of all US landfalling TC's produced at least one PRE
  • There were 5 cases where the TC did not make US landfall

    PRE’s generally form...
  • When persistent, deep meridional flow transports tropical air far from the TC
  • In favored upslope regions or along synoptic/mesoscale boundaries (ahead of the boundary, where surface-based convection is favored or immediately behind the boundary, where elevated convection is favored)
  • Along and just west of a low-level theta-e ridge, near the strongest gradient
  • Near a midlevel jet-entrance region confluence zone
  • Under favorable upper-level jet dynamics (position and amplitude)
  • Note that when the trough axis was either in a state of de-amplification, or had already moved to near or east of the longitude of the tropical cyclone, a PRE typically did not occur.

    Loop of regional radar reflectivity of Hanna - click to enlarge Many of the PRE's documented by Cote, Bosart, Keyser, and Jurewicz (2008) occurred at higher latitudes then the Carolinas but they can occur in the Southeast. The diagram above and to the right shows a conceptual model of a PRE that is producing the heaviest precipitation to the left of track from a Southeast recurving or Atlantic recurving tropical cyclone.

    During this event, there was an area of rain that moved westward across coastal and then central portions of the Carolinas and Virginia during the afternoon and evening of September 5th. This area of precipitation met two of the three requirements for a PRE but the rainfall intensity was insufficient. The radar loop to the right shows that the rain that preceded Hanna across North Carolina was a coherent area of rain that displaced poleward of the TC and that it was enhanced because of moisture transport from the TC toward the PRE but the maximum rainfall rates were well below 100 mm (3.94 inches) in 24 hours.

    It is also important to note that the heavy rain prior to a tropical cyclone can establish an in-situ Cold Air Damming (CAD) region. The formation of the CAD can then initiate the development of a surface boundary along the edge of the CAD region. This boundary can then act as a focus for subsequent precipitation. Additional discussion on the development of a CAD region and interaction of the associated boundary on the distribution of precipitation is located a few sections below. Finally, the rain that falls before the main tropical cyclone rain shield can lead to additional wind damage by saturating the soil and promoting falling trees.


  • Cyclone Phase Space Diagrams

    Cyclone phase space diagram for Hanna - click to enlarge The cyclone phase space diagram developed by Hart (2003) provides a mechanism to objectively determine whether a system is warm or cold core and whether the thermal structure will promote asymmetry. The diagram describes the evolution of a cyclone including tropical, extratropical, subtropical, and hybrid structures, and provides a way to visualize the location of the cyclone in the continuum between tropical and extratropical. Three parameters are used to represent the “phase” of a cyclone:
    1. B – Storm-motion-relative 900–600 hPa thickness gradient across the cyclone (graphic)
    2. -VTL – Magnitude of lower-tropospheric cyclone thermal wind (900–600 hPa) (graphic)
    3. -VTU – Magnitude of upper-tropospheric cyclone thermal wind (600–300 hPa) (graphic)
    There are two phase diagram types that are constructed by plotting either 1 vs. 2 - phase diagram #1 construction (B vs. -VTL: thermal asymmetry versus lower-tropospheric thermal wind) or 2 vs. 3 - phase diagram #2 construction (-VTL vs. -VTU: upper vs. lower-tropospheric thermal wind).


    The cyclone phase space diagram #1, thermal asymmetry versus lower-tropospheric thermal wind, for Tropical Storm Hanna produced from the AVN model, is shown above and to the right (click to enlarge). The diagram is provided by the Cyclone Phase Evolution: Analyses & Forecasts web page (http://moe.met.fsu.edu/cyclonephase/). The vertical axis of a graph measures the symmetry of the system (B – Storm-motion-relative 900–600 hPa thickness gradient across cyclone) while the horizontal axis of the graph measures the thermal structure (-VTL – magnitude of lower-tropospheric cyclone thermal wind (900–600 hPa)) and whether the system has warm or cold core structure. The colors of the plotted points indicate the intensity of the storm in pressure. The numbers on the points correspond to the date and the location of the storm in the Atlantic basin is shown in the upper right hand portion of the image. A much more detailed description of the cyclone phase space is available from (Hart 2003).

    Daily cyclone phase space diagrams of thermal asymmetry vs. lower-tropospheric thermal wind from the AVN model for Hanna are available below (all images are from 00 UTC on the given date)
    09/01 | 09/02 | 09/03 | 09/04 | 09/05 | 09/06 | 09/07 | 09/08

    Examining the thermal asymmetry versus lower-tropospheric thermal wind cyclone phase space diagram from the AVN model for Hanna indicates Hanna was flirting with the symmetric warm-core and asymmetric warm-core boundary on September 2nd and 3rd as Hanna was interacting with an upper low near the Bahamas and then again on September 6th and 7th as the storm moved poleward up the East Coast. The transition from symmetric to asymmetric warm-core is indicative of the beginning of the extratropical transition process.

    The cyclone phase space diagram #2, upper vs. lower-tropospheric thermal wind for Tropical Storm Hanna produced from the AVN model are available below. The diagram is provided by the Cyclone Phase Evolution: Analyses & Forecasts web page (http://moe.met.fsu.edu/cyclonephase/). The vertical axis of a graph measures the thermal structure (-VTU – Magnitude of upper-tropospheric cyclone thermal wind (600–300 hPa)) while the horizontal axis of the graph measures the thermal structure (-VTL – Magnitude of lower-tropospheric cyclone thermal wind (900–600 hPa)). The colors of the plotted points indicate the intensity of the storm in pressure, the numbers on the points correspond to the date and the location of the storm in the Atlantic basin is shown in the upper right hand portion of the image.

    Daily cyclone phase space diagrams of upper vs. lower-tropospheric thermal wind from the AVN model for Hanna are available below (all images are from 00 UTC on the given date)
    09/01 | 09/02 | 09/03 | 09/04 | 09/05 | 09/06 | 09/07 | 09/08

    Examining the upper vs. lower-tropospheric thermal wind cyclone phase space diagram from the AVN model placed Hanna in the deep warm-core section of the phase diagram on September 2nd and 3rd. Hanna then flirted with the boundary between deep warm-core and shallow warm-core boundary on September 4th and 5th as the system interacted with a large area of dry air as seen in the visible satellite image from 1315 UTC on September 4th. The 1100 PM National Hurricane Center Discussion on Hanna from September 3rd and 1100 AM Discussion on Hanna from September 4th noted that Hanna displayed some subtropical characteristics and appearance. By September 6th Hanna was again plotted in the deep warm-core section of the phase diagram as deep convection developed near the circulation center as shown in the 1915 UTC visible satellite on September 5th.

    It is important to note that the quality of cyclone phase space diagrams is dependent on the quality of the model analysis and forecasts used to create them. The cyclone phase space diagrams are subject to model analysis errors as well as errors in the forecast of structure, intensity and track of the cyclone (Brennan and Knabb, 2008). In addition, the models used to create the cyclone phase space diagram are relatively coarse (0.5° to 1° for global models) and may have difficulty in predicting small scale changes in the cyclone. Regardless, analyzing and forecasting the phase or phase transition of cyclones is difficult but tools such as the cyclone phase space diagram can provide forecasters with additional insight into this difficult forecast problem.


    Extratropical Transition

    Extratropical Transition (ET) is a gradual process in which a warm-core tropical cyclone loses tropical characteristics and become more extratropical in nature (Jones et al. 2003). During ET, the precipitation expands poleward of the center and the maximum precipitation is shifted to the left of the tropical cyclone track (Atallah and Bosart 2003). In addition, during ET, cyclones begin to tilt back into the cooler air mass with height, and the primary energy source evolves from the release of latent heat from condensation to baroclinic processes. The change in the structure of the precipitation field from the more symmetric distribution in a tropical cyclone to the asymmetric distribution during ET results from increasing shear and baroclinicity.

    Studies by Atallah and Bosart (2003) along with Evans and Hart (2003) have established that the structure, size and orientation of mid-latitude troughs play an important role in determining whether ET will occur. In general, a diffluent upper level trough with a negative tilt and an eastward motion are all conditions favorable for extratropical transition. A study by Hart and Evans (2001) showed that 40-50% of hurricanes in the Atlantic basin that recurved into higher latitudes experience some form of ET. Extratropical transition often results in drastic changes in the distribution of heavy precipitation as tropical moisture can be transported great distances from the cyclone center and the interaction of the cyclone with jets and fronts can result in significant rainfall totals.

    In the cyclone phase space diagram, Extratropical Transition is identified when the system moves from symmetric/warm-core to asymmetric cold-core. For Hanna, this occurred from September 6th through the 7th as the storm moved northeast across coastal North Carolina, Virginia and into the Northeast. The cyclone phase space diagram from the AVN | CMC (Canadian) | NOGAPS are fairly consistent in depicting the ET from September 6th through the 7th.



    Subtropical Cyclones

     - click to enlarge A subtropical cyclone is defined in the AMS Glossary of Meteorology as a cyclone in tropical or subtropical latitudes that has characteristics of both tropical cyclones and mid-latitude (or extratropical) cyclones. They occur in regions of weak to moderate horizontal temperature gradient and extract the associated available potential energy, as do baroclinic cyclones, but they also receive some or most of their energy from convective redistribution of heat acquired from the sea, as do tropical cyclones. In comparison to tropical cyclones, subtropical cyclones have a relatively broad zone of maximum winds located farther from the center, and typically have a less symmetric wind field and distribution of convection. Precipitation associated with subtropical cyclones tends to be removed from the circulation center and is often scattered in bands some distance from the center.

    During the first few days of September, Hanna struggled with strong northerly wind shear while being forced south by an upper level low pressure system that moved into the western Atlantic. The shear from the upper level low and even some of the outflow from Hurricane Gustav (which had already moved inland over Louisiana) inhibited much of Hanna's convection and forced the storm south to near the Haitian coast late on September 2nd. Late on September 3rd the upper low north of Hanna moved further away and convection began to redevelop, especially late on the 4th.

    Water vapor satellite imagery from CIRA on September 5, 2008 at 1445 UTC - click to enlarge Hanna drifted northwest as a large area of dry air across the southeastern U.S. moved into the coastal wasters, portions of the Bahamas and into Cuba. Hanna would eventually ingest much of this drier air on the 4th and into the 5th before convection redeveloped near the circulation center. The upper vs. lower-tropospheric thermal wind cyclone phase space diagram for Hanna placed the cyclone on the boundary between deep warm-core and shallow warm-core at 00 UTC on September 4th and 5th. In the cyclone phase space diagram, subtropical cyclones are identified in the shallow warm-core and/or the asymmetric warm core section. The interaction with the upper level low and especially the large area of dry air shown on the visible satellite image from 1315 UTC on September 4th contributed to the subtropical appearance. The National Hurricane Center Forecast Discussions on Hanna noted that Hanna displayed some subtropical characteristics and appearance (Hanna Discussion from 1100 AM on September 3rd, 1100 PM on September 3rd, 1100 AM on September 4th). The shear, drier air, limited convection, and broad circulation are consistent with subtropical cyclone characteristics.

    By 00 UTC September 6th Hanna was again plotted in the deep warm-core section of the phase diagram as deep convection developed the previous day near the circulation center as shown in the 1515 UTC GOES infrared satellite image from September 5th.



    Cold Air Damming, Surface Boundaries and Precipitation

    As Tropical Storm Hanna approached the Carolina coast on September 5th and 6th, it encountered a dry air ridge over western portions of the Mid-Atlantic States. Outer rainbands ahead of the system moved inland during the day on the 5th, encountering a fairly uniform temperature gradient across NC. However, surface wetbulb temperatures were in the low to mid 60's across portions of central and western NC, while in the mid 70's over eastern NC, indicative of a strong moisture gradient across the state. As precipitation fell into the dry air ridge, evaporative cooling caused temperatures to fall quickly into the low 70's, while readings along the coast remained in the low to mid 80's. This developing thermal lead to the establishment of an in-situ stable layer or cold air damming region across the western Piedmont of North Carolina, and eventually manifest itelf as a low-level boundary in central NC. This is evident in the surface analysis performed by a forecaster on duty overnight (05z on the 6th). Low-level frontogenesis increased dramatically as the boundary interacted with the cooler, more stable airmass to the west. This created a large shield of very heavy rain as warm moist air feeding inland on the forward side of the storm was lifted up and over the stable layer. Comparatively to cool/cold season events, the dry air ridge associated with cold air damming events is a much more subtle feature in the warm season and this was the case with Hanna.

    Several studies have shown that a surface boundary can focus the heaviest precipitation in a mesoscale band located along and on the cold side of the surface boundary. Banded mesoscale precipitation structures associated with surface boundaries have been previously documented with landfalling and/or transitioning tropical cyclones (Atallah and Bosart 2003; Colle 2003). More recently, Srock and Bosart (2009) studied the impacts of dry-air and devloping cold-pool/stable layers on land-falling tropical cyclones, noting that the interaction of coastal fronts and CAD-like boundaries play a very important role in the mesoscale devlopment of the rainfall intensity and distribution (schmatic below). Klein (2006) noted that the area of low-level frontogenesis is typically located left and poleward of the storm track with the enhanced precipitation region along or in cold sector of boundary. In addition, he noted that a slight theta-e gradient commonly occurred.

    Mesoscale influences on rainfall distribution (from Srock and Bosart 2009)

    Additional work is being conducted to evaluate the role that the in-situ cold air damming region and the resultant low-level boundary played on the precipitation distribution. Model simulations are being conducted using the Weather Research and Forecasting (WRF) Model to test the response of the rainfall location and amount to the presence and absence of the boundary.



    Shear and Precipitation Distribution

    Cline (2003) discussed the role of along track or cross track bulk shear in the distribution of precipitation. He found that tropical cyclones with cross track shear typically have the precipitation distorted in an asymmetrical fashion while along track shear results in precipitation spreading well in advance of the storm. Forecasters at the NWS Raleigh use an AWIPS procedure to display the 850-300 hPa bulk shear vector and magnitude to diagnose the potential precipitation distortion. Bulk shear values exceeding 30 kts are typically sufficient to distort the precipitation pattern with the precipitation field displaced down shear and left of the shear vector.

    NAM 6 hour forecast of 850-300 hPa bulk shear valid 06 UTC 09/06- click to enlarge Chen et. al. (2006) studied the relationship between the structure of tropical cyclone rainfall and the environmental flow by computing the rainfall asymmetry relative to the vertical wind shear. They used the Tropical Rainfall Measuring Mission Microwave Imager (TRMM) rainfall estimates over the six oceanic basins across the world. They defined the environmental vertical wind shear as the difference between the mean wind vectors of the 200- and 850-hPa levels over an outer region extending from the radius of 200–800 km around the storm center. They found that the vertical wind shear and storm motion are two of the most important factors contributing to rainfall asymmetries in tropical cyclones. In addition, they noted that rainfall asymmetry decreases with storm intensity and increases with shear strength. This study also found that the rainfall asymmetry maximum is predominantly down shear and to the left for shear values greater then 15 kts.

    Forecast values of bulk shear during Hanna (see examples below) generally suggested a shear vector oriented a small angle across the storm's track. The storm motion based on National Hurricane Center advisories at 03 UTC and 09 UTC was between 360 degrees (north) and 20 degrees (north-northeast). The National Hurricane Center forecast track issued at 06 UTC indicated that Hanna would begin to turn more northeastward shortly. The forecast values of bulk shear valid at 06 UTC 09/06 generally indicated values from the south-southwest or southwest at between 30 and 40 kts. These magnitudes slightly exceed the 30 kt threshold noted by Cline. In addition, the cross track angle is rather modest, generally less then 30 degrees. This would indicate a tendency for the precipitation to be focused along and possibly to the right of the storm track.

    One other item to consider is the actual forecast data and its ability to accurately depict the amount of cross track shear. The examples below use the NAM80 data available in AWIPS but it should be noted that there were differences in the NAM and GFS forecast output. For example, a comparison of the NAM (shown in blue) and GFS (shown in green) 6 hour forecast of 850-300 hPa bulk shear valid 06 UTC 09/06 shows the differences between the two sets of guidance.

    National Hurricane Center advisory location and forecast track issued at 06 UTC
    NAM 6 hour forecast of 850-300 hPa bulk shear valid 06 UTC 09/06
    NAM 18 hour forecast valid 06 UTC 09/06
    NAM 30 hour forecast valid 06 UTC 09/06
    Regional radar image of Tropical Storm Hanna valid 06 UTC 09/06/2008
    HPC precipitation analysis and track of Tropical Storm Hanna

    While the forecast of bulk shear was not a clear indicator of the left of track precipitation distribution associated with Hanna, Tropical Storm Alberto (June 2006) was a much better example of the impact of along track shear. Alberto moved onshore over the Florida panhandle and then moved across Georgia and the Carolinas. The system also interacted with a west to east surface boundary. The deep layer 850-300 hPa bulk shear was generally oriented along the storm track which allowed the precipitation to be distorted downstream and ahead of the cyclone center. The resultant precipitation produced the heaviest rainfall along and especially to the left of the track.
    National Hurricane Center advisory location and forecast track issued at 09 UTC
    NAM 12 hour forecast of 850-300 hPa bulk shear valid 12 UTC 06/14 from the 06/14 00 UTC NAM
    Regional radar image of Tropical Storm Alberto valid 12Z 06/14
    HPC precipitation analysis and track of Tropical Storm Alberto

    While the forecast of bulk shear was not a clear indicator of the left of track precipitation distribution associated with Hanna, Tropical Storm Ernesto (August/September 2006) was a much better example of the impact of cross track shear. Ernesto interacted with a west to east surface boundary as it moved onshore near Cape Fear and then moved north across the Carolinas. The deep layer 850-300 hPa bulk shear was generally across the storm track which allowed the precipitation to be distorted down shear and to the right of cyclone center. The resultant precipitation produced the heaviest rainfall along and especially to the right of the track.
    National Hurricane Center advisory location and forecast track issued at 06 UTC
    NAM 6 hour forecast of 850-300 hPa bulk shear valid 06 UTC 06/14
    Regional radar image of Tropical Storm Ernesto valid 06 UTC 09/01/2006
    HPC precipitation analysis and track of Tropical Storm Ernesto



    Tropical Cyclone Interaction with an Upper Level Trough

    Left of track figure from Atallah et al. (2007) - click to enlarge Atallah et al. (2007) studied the distribution of precipitation accompanying U.S. landfalling and transitioning tropical cyclones. In the two conceptual models from Atallah et al. (2007) shown to the right and below, the curved black lines represent streamlines of the upper-tropospheric flow. Arrows represent motion and deep tropospheric shear with the relative magnitudes given by the length of the arrow. The curved green line represents the trajectory of a parcel starting near the surface in the warm sector and ending in the mid to upper-troposphere in the cool sector. The area inside the light gray circle represents regions of precipitation.

    One portion of the study composited 14 cyclones that had the bulk of their rainfall occur to the left (west) of the track (LOT) of the cyclone circulation. Their composite argues strongly that storms that have most of their rainfall distributed on the west side of storm’s track interact with upper level trough and potential vorticity maximum in a synergistic way. In addition, the axis of precipitation of LOT storms often is aligned parallel to the movement of the storm which increases the period of heavy rainfall. The 850-200 hPa shear in the composite suggests that there was a strong jet streak present to the north or northwest of the cyclone and implies that the storm was located near the right entrance region of a jet streak where frontogenesis is favored. The composite also suggests that the axis of the upper trough shifts from having a slight positive tilt or neutral one to having a negative tilt as the storm undergoes extratropical transition. Storms undergoing a transition to becoming extratropical typically have their heaviest rainfall LOT.

    A left of track precipitation distribution is often characterized by a positively tilted mid-latitude trough approaching the tropical cyclone from the northwest. The precipitation stretches out well north of the system due to impulses moving around the upper trough and frontogenesis ahead of the trough. The trough often transitions from a positive to negative tilt during its interaction with tropical cyclone.

    Recent tropical cyclones with a left of track precipitation distribution include Alberto (2006), Gaston (2004), and Floyd (1999).


    Right of track figure from Atallah et al. (2007) - click to enlarge Another portion of the Atallah et al. (2007) study composited 16 cyclones that had the bulk of their rainfall occur to the right (east) east of the track (ROT) of the cyclone circulation. These composites indicate that the axis of the 850-200-hPa shear is typically oriented more westerly than the more north-south oriented axis associated with LOT cases. The location of the upper level jet streak is much farther removed for ROT cases than for LOT cases. The baroclinic zone as shown by the leading edge of the 1000-500 hPa thickness is also much farther removed from the storm. The 1000 hPa circulation is much weaker, especially on the northern side where there is a much weaker gradient implying that the easterly low-level winds are quite a bit weaker for ROT storms than LOT storms.

    A right of track precipitation distribution is often characterized by tropical cyclones that are generally steered by shear lines or through a break in the subtropical ridge. Rainfall tends to be concentrated near and right of track. When the downstream ridge amplifies in response to diabatic heating ahead of a weak mid-latitude trough, the increased shear shifts precipitation to the right of the storm track.

    Recent tropical cyclones with a right of track precipitation distribution include Ernesto (2006), Isabel (2003), and Bertha (1996).

    LOT and the ROT conceptual cross sections from Klein (2007)- click to enlarge Klein (2007) noted important differences in the vertical structure across the surface boundary in the LOT and the ROT conceptual cross sections. These cross sections, which are constructed parallel to the lower-level (e.g., 925–500-hPa) layer-averaged thermal gradient and across the region of strongest frontogenesis, include the frontogenesis and vertical motion fields. The cross section include frontogenesis which is shaded in green and vertical velocity dashed and contoured in red.

    The warm-frontal boundary associated with LOT cases is characterized by a very deep frontogenesis pattern that tilts toward the cooler air with height and is strongest in the boundary layer and at upper-levels. The ascent pattern tilts toward cooler air with height and is located on the warm side of the frontogenesis maximum. Ascent is greatest in the mid-troposphere above and on the warm side of the frontogenesis region.

    The warm-frontal boundary illustrated in the ROT schematic contains strong frontogenesis that is largely confined near the surface. Strong, upright ascent is located over and on the warm side of the frontal boundary with the largest values immediately above the boundary layer.

    Cross section from the 24 hour NAM80 forecast of Petterssen frontogenesis , potential temperature, and omega across the lower Ohio River Valley and the Carolinas valid at 12 UTC on Saturday September 6, 2009 - click to enlarge Although Hanna does not fit either conceptual model ideally, it is more similar to the left of track conceptual model. Note with Hanna there is a significant upstream upper level trough with the right entrance region of the jet over the Ohio Valley where wind speeds range between 60 and 80 kts. Compare Hanna with the upper level pattern associated with Alberto, which also has a left of track precipitation distribution. With Alberto the upper level trough extends into the Ohio Valley and the right entrance region of the jet extends into the Mid Atlantic. Finally, compare Hanna with the upper level pattern associated with Ernesto which is a right of track event. With Ernesto there is an upper level trough upstream in the Ohio Valley but the heights are not very low. More significantly, the upper level jet is well removed from the Carolinas as the right entrance region of the jet is located in New England.

    The image to the right is a cross section from the 24 hour NAM80 forecast of Petterssen frontogenesis, potential temperature, and omega across the lower Ohio River Valley and the Carolinas valid at 12 UTC on Saturday September 6, 2008. The frontogenetical zone and the responsive vertical circulation usually slopes with height toward cold air. In addition, the cold air damming region is indicated by a shallow stable layer can be seen in the potential temperature contours just east of the mountains.


    Precipitation Estimates from the NMQ QPE and KRAX

    An accurate analysis of the amount of precipitation that has fallen from a tropical cyclone is extremely important to forecasters. The amount of precipitation is a critical factor in determining the likelihood of flash flooding and eventual river flooding. Because flooding is typically a result of significant precipitation over an area such as a river or stream basin, relying on point observations of precipitation (rain gauges) only provides a portion of the needed information. The estimation of the amount of precipitation that has fallen is often referred to as Quantitative Precipitation Estimation or QPE. Studies have shown that algorithms which combine sensor inputs such as radar, gauge, and satellite yield more accurate precipitation estimates than those which rely on a single sensor. Recent advances in QPE integrates radar data with other data sources such as rain gauges and satellite information in a process called Muti-sensor Precipitation Estimation or MPE.

    The National Mosaic and QPE (NMQ) Web Page provides real time evaluation and display of experimental techniques and applications used for high resolution 3D mosaics of radar reflectivity data and QPE. The NMQ serves as a test bed for research, development, and evaluation of data and methods for the monitoring and warnings of floods and flash floods and in support of comprehensive hydrology and ecosystem modeling. There are 4 major components of the NMQ system: Data ingest, Products, Analysis tools, and Verification. The NMQ system ingests data from a number of sensors and products from various sources: 128 WSR-88D radars (5 min), Gauge data set that includes 5500 gauges from many different networks (hourly), Satellite IR images (15 min), Rapid Update Cycle 20 km resolution (RUC) model analysis variables (hourly), NWS Hydro Estimator (satellite-based) precipitation (hourly), and NWS Stage IV precipitation (1, 6, and 24 hour).


    Q2RAD_HSR_GC

    The Q2RAD_HSR_GC is a local gauge corrected radar QPE field. The local gauge correction is applied onto the one hour Q2RAD_HSR precipitation field. It runs hourly and uses hourly rain gauge observations from the HADS (Hydrometeorological Automated Data System) data sets at NCEP. In the local gauge correction scheme, radar-gauge biases are calculated at each gauge site and then interpolated onto the NMQ grid using an inverse distance weighted (IDW) mean scheme. The two parameters in the IDW scheme, exponent and radius of influence, are determined through a cross-validation procedure. The interpolated radar-gauge bias field is applied back to the Q2RAD_HSR one hour precipitation field and a local gauge bias corrected one hour precipitation field is obtained. Longer-term accumulations are computed by aggregating the one hour local gauge corrected precipitation fields.

    Q2RAD_HSR_GC Precipitation estimate


    KRAX Precipitation Estimate

    The krax precipitation estimate is a radar-based precipitation accumulation produced at the RPG. The precipitation estimate is derived from the Hybrid Scan Reflectivity (HSR) product from the Enhanced Precipitation Preprocessing (EPRE) algorithm using a specific Z-R relationship. The image below is the product from the KRAX RDA displayed in AWIPS (click on the image to enlarge.)

    Precipitation estimate from KRAX


    Comparison of the local gauge corrected radar QPE product with the Stage IV QPE product

    The local gauge corrected radar QPE product or Q2RAD_HSR_GC generally provides the most accurate and detailed precipitation analysis when compared to the various QPE products (Stage IV, radar only QPE, satellite hydro estimator, etc. The comparison below shows the local gauge corrected radar QPE product from the NMQ web site with the Stage IV QPE product from the NWS AHPS web site for the 24 hour period ending at 12 UTC 2009/09/06 (click on the image to enlarge). The two QPE products below appear very similar with the Q2RAD_HSR_GC having greater detail that allows the user to better interpret the maximas and allows some of the storm structure to be visualized (note the overlapping banding features off the NC/SC coast.)

    Comparison of the Local gauge corrected radar QPE product with the Stage IV QPE product - click to enlarge


    Flash Flooding in the Crabtree Valley area of Raleigh

    Rainfall amounts associated with Tropical Storm Hanna in the Crabtree Valley area ( map ) ranged between 5 and 6 inches. The Crabtree Creek has been frequented by flash flooding during the past several decades. The creek runs from its origin near the William B. Umstead State Park in the northwestern portion of Wake County and flows east across north Raleigh and into the Neuse River in eastern Raleigh. The Crabtree Creek collects a tremendous amount of runoff from urban and suburban locations in Raleigh and it is especially prone to flash flooding.

    The Crabtree Creek at Glenwood Avenue (RLHN7) river gauge was reporting a stage of 5.05 feet at noon on Friday, September 5, 2008 just before the precipitation began. As heavy rain moved into the Crabtree Creek basin at around midnight, the creek began to rise very rapidly. At 100 AM EDT, the gauge reported a stage of 5.59 feet, an hour later the creek had risen 2.7 feet and two hours later at 300 AM EST, the creek had risen 5.89 feet and was at a stage of 11.48 feet. The creek continued to rise, albeit at a slightly slower rate during the remainder of the overnight and morning hours on Saturday. The creek crested at a stage of 18.47 feet at 145 PM EDT on Saturday, September 6, 2008.

    Photos of the flooding along Crabtree Creek near Crabtree Valley Mall
    (click on the image to enlarge)


    Photo looking southeast toward Blue Ridge Road of a flooded Crabtree Creek that covers the Crabtree Valley Greenway with several feet of water - photo courtesy of the National Weather Service - Click to enlarge           Photo looking northwest toward Blue Ridge Road of a flooded Crabtree Creek that covers the Crabtree Valley Greenway with several feet of water - photo courtesy of the National Weather Service - Click to enlarge           Photo looking southeast toward Glenwood Avenue of a flooded Crabtree Creek that covers the Crabtree Valley Greenway with several feet of water - photo courtesy of the National Weather Service - Click to enlarge           Photo of water entering the Crabtree Valley Mall lower level parking deck south of Macy's - photo courtesy of the National Weather Service - Click to enlarge           Photo of the flood gates at Macy's in Crabtree Valley Mall near the previous photo - photo courtesy of the National Weather Service - Click to enlarge          


    Video of the flooding along Crabtree Creek near Crabtree Valley Mall
    The below was filmed on September 6th at 1000 AM along the south bank of the Crabtree Creek just west of Blue Ridge Road (map) initially looking east and then west. (click on the image below to view video.)

    video of the flooding along Crabtree Creek near Crabtree Valley Mall
    Note - the video is 10 seconds in length and approximately 3.5 MB in size.


    Hydrograph of Crabtree Creek at Glenwood Avenue in Raleigh showing the change in stage during the flooding on Friday and Saturday, September 5-6, 2008.

    Hydrograph of Crabtree Creek at Glenwood Avenue in Raleigh


    The table below provides a list of the stage and CFS values at the Crabtree Creek at Glenwood Avenue on Saturday, September 6, 2008.

    Date   Time (UTC)/(EDT)    Stage    CFS
    
    09/06  02:00Z/1000 PM EDT  5.39ft   0.097 kcfs
    09/06  03:00Z/1100 PM EDT  5.58ft   0.12 kcfs 
    09/06  04:00Z/1200 AM EDT  5.41ft   0.099 kcfs 
    09/06  05:00Z/100 AM EDT   5.59ft   0.12 kcfs 
    09/06  06:00Z/200 AM EDT   8.29ft   0.66 kcfs  
    09/06  07:00Z/300 AM EDT   11.48ft  1.68 kcfs  
    09/06  08:00Z/400 AM EDT   12.19ft  1.96 kcfs  
    09/06  09:00Z/500 AM EDT   12.86ft  2.24 kcfs  
    09/06  10:00Z/600 AM EDT   14.55ft  3.00 kcfs   
    09/06  11:00Z/700 AM EDT   15.21ft  3.33 kcfs
    09/06  12:00Z/800 AM EDT   15.88ft  3.68 kcfs
    09/06  13:00Z/900 AM EDT   16.96ft  4.26 kcfs 
    09/06  14:00Z/1000 AM EDT  17.45ft  4.54 kcfs 
    09/06  15:00Z/1100 AM EDT  18.01ft  4.87 kcfs  
    09/06  16:00Z/1200 PM EDT  18.22ft  4.99 kcfs  
    09/06  17:00Z/100 PM EDT   18.38ft  5.08 kcfs  
    09/06  18:00Z/200 PM EDT   18.34ft  5.06 kcfs  
    09/06  19:00Z/300 PM EDT   18.26ft  5.01 kcfs  
    09/06  20:00Z/400 PM EDT   17.96ft  4.84 kcfs 
    09/06  21:00Z/500 PM EDT   17.33ft  4.47 kcfs  
    09/06  22:00Z/600 PM EDT   16.59ft  4.06 kcfs  
    09/06  23:00Z/700 PM EDT   15.86ft  3.67 kcfs  
    

    The table below includes impacts at various stages of the creek at Glenwood Ave and a few of the historical crests.

    3 - 5 feet - Typical stages at low flows
    14.00 feet - Water covers the lower parking lot at Crabtree Valley Mall
    18.00 feet - Water enters the lower entrance to the Macy's Department Store
    18.47 feet - Tropical Storm Hanna (September 6, 2008)
    21.50 feet - Hurricane Floyd (September 16, 1999)
    20.00 feet - Most parking lots at Crabtree Valley Mall flood, including the Shell gas station
    22.00 feet - Glenwood Ave and the lower level Crabtree Valley Mall parking lot flood
    23.00 feet - The Glenwood Ave (U.S. 70) bridge floods. Travel on Glenwood Ave and the area around Crabtree Valley Mall is severely hampered
    23.00 feet - Hurricane Fran (September 6, 1996)
    23.77 feet - Tropical Storm Alberto (June 14, 2006)
    24.00 feet - Glenwood Ave (U.S. 70) becomes impassible
    26.00 feet - Water reaches the top of the Glenwood Ave (U.S. 70) bridge railing
    27.69 feet - All time record high (June 29, 1973




    Hydrologic recharge from Tropical Storm Hanna

    Heavy rain from Tropical Storm Hanna produced flooding and significant hydrologic recharge across much central North Carolina. Minor to moderate mainstem river flooding was observed along portions of the Neuse and Cape Fear River basins. Tropical Storm Hanna had a significant role in a 2 class improvement in the Drought Monitor across central North Carolina from early August through early November.

    Drought Monitor change - Click to enlarge




    Viewing Tropical Cyclone Data in Google Earth

    Plot of reconnaissance observations - click to enlarge The National Weather Service in Raleigh has begun to use Google Earth and its GIS capabilities to view tropical cyclone model data and near real-time aircraft reconnaissance data prior to a potential tropical cyclone landfall. These tools were utilized the day of Tropical Storm Hanna’s landfall and were shown to add some benefit.

    One useful product was the near real-time plot of observations from the aircraft reconnaissance missions (Hurricane Hunters). In the image to the right, plots of all the reconnaissance observations are shown. The multicolored wind barbs indicate varying strength of winds measured at flight level (ranging from 1,000 ft for weak storms to 10,000 ft in strong hurricanes.) These winds are utilized by the National Hurricane Center in estimating the maximum sustained winds and determining the wind field. The white filled circles with the numbers inside represents the actual center of the tropical cyclone and the minimum central pressure recorded at the time the aircraft fixed the center.

    Plot of tropical cyclone forecast positions - click to enlarge By overlaying several of these fixes, one can gather a better sense of the track and any deviations the storm might make prior to landfall. In Hanna’s case, the tropical storm shifted slightly west prior to landfall, resulting in a landfall near North Myrtle Beach instead of just east of Wilmington as was forecasted. Such a shift is not a large forecast error and well within the NHC’s cone of uncertainty, but given the shape of the coast and the storm’s parallel movement to the coast, subtle shifts can be critical to rain, wind, and storm surge forecasts. The shift to left before landfall resulted in the heaviest rain band developing just to the west of the Triangle and tropical storm force wind gusts across the southern Coastal Plain.

    Model forecast tracks can also be plotted in Google Earth. A quick overview, as shown in the second image above, can give a general sense if the models are in good agreement (tightly clustered tracks) or in poor agreement (a large spread in potential tracks). It is important to note, that many of the model plots available are not from guidance that is considered reliable in forecasting tropical cyclone movement. Even after looking at this brief overview, forecasters at the NHC and the NWS Raleigh look in-depth at several models and apply their meteorological expertise to generate a forecast. Models, are only a tool, and are not to be used verbatim.

    In addition to tropical cyclone model data and near real-time aircraft reconnaissance data, there are numerous other Google Earth overlays available such as satellite imagery, radar, sea surface temperatures, and more. Note that Google Earth map imagery used under license.



    Mesoscale Data

    Analyzed surface pressure and wind barbs from SPC at 07 UTC on Saturday, September 6, 2008
    The circulation center of Hanna can be seen just off the coast of Myrtle Beach, South Carolina. The analyzed minimum surface pressure of 991 MB was higher then the 983 MB pressure reported in the 06Z National Hurricane Center advisory. This is likely a result of using the RUC for the analysis. A notable wind shift can be seen across the Coastal Plain of North Carolina where easterly winds meet northeasterly winds near the I-95 corridor.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    Analyzed surface temperatures (red/purple), dew points (brown/green), and wind barbs from SPC at 07 UTC on Saturday, September 6, 2008
    A west to east gradient in surface dew points and temperatures can be seen across central and eastern North Carolina. Dew points were as low as the upper 60s across the Northwest Piedmont. A boundary can be seen in the wind direction across the Coastal Plain of North Carolina where easterly winds meet northeasterly winds near the I-95 corridor.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    Analyzed surface theta-e (green) and theta-e advection (purple) from SPC at 07 UTC on Saturday, September 6, 2008
    A significant theta-e gradient can be seen across central and especially eastern North Carolina. The theta-e gradient also corresponds with the surface boundary across the Coastal Plain.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    850 hPa heights, temperatures (red/blue), dew points (green), and wind barbs (black) from SPC at 07 UTC on Saturday, September 6, 2008
    The 850 hPa circulation associated with Hanna is shown below. Note the 19 deg C isotherms and the modest warm advection to the east of the storm center. In addition the strongest 850 hPa winds reached 70 kts well to the east of the circulation center. Much drier air is present along/near the Appalachians where dew points were less than 6 deg C.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    Analyzed surface based convective available potential energy (SBCAPE) (red) and surface based convective inhibition (blue lines - shaded) from SPC at 07 UTC on Saturday, September 6, 2008
    Surface based instability was generally confined to the coastal regions with a large area of negative CAPE present across the Piedmont of North Carolina, mainly along and to the west of the surface boundary.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    Analyzed precipitable water (green) and wind barbs from SPC at 07 UTC on Saturday, September 6, 2008
    Deep moisture is shown across much of central and eastern North Carolina with precipitable water values greater than 2.5 inches. A sharp gradient indicating much drier air can be seen across western North Carolina.

    SPC Analysis at 07 UTC on Saturday, September 6, 2008



    NWS Composite reflectivity imagery from 0730 UTC on Saturday, September 6, 2008
    The composite reflectivity imagery below is from the approximate time in which the analysis imagery above is valid. Note the large precipitation band extending north to south across the Piedmont of North Carolina with scattered areas of mainly light precipitation shown across eastern portions of North Carolina. This radar image shows that the heaviest precipitation is falling to the left of the storm track.

    Composite Reflectivity Imagery from 0730 UTC on Saturday, September 6, 2008



    Archived Text Data from the Event

    Select the desired product along with the date and click "Get Archive Data."
    Date and time should be selected based on issuance time in GMT (Greenwich Mean Time which equals EDT time + 4 hours).


    Product ID information for the most frequently used products...

    RDUAFDRAH - Area Forecast Discussion
    RDUZFPRAH - Zone Forecast Products
    RDUAFMRAH - Area Forecast Matrices
    RDUPFMRAH - Point Forecast Matrices
    RDUHWORAH - Hazardous Weather Outlook
    RDUNOWRAH - Short Term Forecast
    RDUSPSRAH - Special Weather Statement
    RDULSRRAH - Local Storm Reports (reports of severe weather)
    RDUSVRRAH - Severe Thunderstorm Warning
    RDUSVSRAH - Severe Weather Statement
    RDUTORRAH - Tornado Warning


     from 



    Lessons Learned

    Since Hanna was expected to move across central North Carolina as a strong Tropical Storm or a minimal Hurricane, heavy rain and flooding were the primary anticipated threat. Forecasters anticipated a rain distribution with the heaviest rain falling to the left of the storm track.

    FFMP was invaluable for issuing timely and accurate Flash Flood Warnings. Situational awareness was also critical with excellent identification that the main hazard would be flash flooding and minor wind damage.

    Forecasters issued Flash Flood Warnings with fairly large polygons covering several counties to make it easier to manage and provide details about the flooding threat. The warnings were issued for 6 hour durations to ensure that warnings would not have to be reissued and extended.

    Crabtree Creek information should be included in Flash Flood Warnings and follow-up statements rather than River Flood Warnings. If needed, an entire warning or statement to Crabtree Creek alone to highlight this high impact location.

    Analyzing and forecasting the phase or phase transition of cyclones is a difficult but tools such as the cyclone phase space diagram can provide forecasters with additional insight into this difficult forecast problem.

    Wind Advisories and High Wind Warnings were placed considering meteorological features and although High Wind Warning criteria winds were not observed, there was a considerable impact in the warning and advisory areas with downed trees, limbs and power lines. It is recommended that generous areal coverage and extent of High Wind Warnings and Wind Advisories be used in tropical cyclone events to account for local gusts that may occur between observations especially in combination with what is typically a very wet ground.

    Adding a special section on the top of the web page that contained links to important imagery and information was well received. Numerous customers and partners passed along their appreciation for the information.

    Real time web based power outage maps were used to get some information on the impact of the winds.



    References


    Ashley, W.S., A.J. Krmenec, and R. Schwantes, 2008: Vulnerability due to Nocturnal Tornadoes. Wea. Forecasting, 23, 795–807.

    Atallah, E., L. F. Bosart, and A. Aiyyer 2007: Precipitation Distribution Associated with Landfalling Tropical Cyclones over the Eastern United States. Mon. Wea. Rev., 135, 2185-2206.

    Atallah, E and L. F. Bosart, 2003: Extratropical transition and precipitation distribution: A case study of Floyd (1999). Mon. Wea. Rev., 131, 1063-1081. Brennan, M. and R. Knabb, 2008: Extratropical and Subtropical Cyclones: NHC Operational Challenges and Forecast Tools.

    Chen, S.S., J.A. Knaff, and F.D. Marks, 2006: Effects of Vertical Wind Shear and Storm Motion on Tropical Cyclone Rainfall Asymmetries Deduced from TRMM. Mon. Wea. Rev., 134, 3190–3208.

    Cline, J. W., 2003: Recent tropical cyclones affecting North Carolina. M.S. Thesis, University of Miami, 43-47 pp.

    Colle, B. A., 2003: Numerical simulations of the extratropical transition of Floyd (1999): Structural evolution and responsible mechanisms for the heavy rainfall over the northeast United States. Mon. Wea. Rev., 131, 2905–2926. Cote, Matthew R, and L. F. Bosart, D. Keyser, and M. L. Jurewicz, 2008: Predecessor rain events in tropical cyclones. 28th Conference on Hurricanes and Tropical Meteorology.

    DeLuca, D. P., 2004: The distribution of precipitation over the Northeast accompanying landfalling and transitioning tropical cyclones. M.S. Thesis, Department of Earth and Atmospheric Sciences, University at Albany-SUNY, 178 pp.

    Hart, R. E., 2003: A cyclone phase space derived from thermal wind and thermal asymmetry. Mon. Wea. Rev., 131, 546-564.

    Hart and Evans, 2001: A climatology of the extratropical transition of Atlantic tropical cyclones. J. Climate, 14, 546– 564.

    Hart, and Evans, 2003: Analyzing and forecasting cyclone structural transition using cyclone phase diagrams. Proc. Second Int.

    Jones, S.C., P.A. Harr, J. Abraham, L.F. Bosart, P.J. Bowyer, J.L. Evans, D.E. Hanley, B.N. Hanstrum, R.E. Hart, F. Lalaurette, M.R. Sinclair, R.K. Smith, and C. Thorncroft, 2003: The Extratropical Transition of Tropical Cyclones: Forecast Challenges, Current Understanding, and Future Directions. Wea. Forecasting, 18, 1052–1092.

    Klein, J. R., 2007: Mesoscale Precipitation Structures Accompanying Landfalling and Transitioning Tropical Cyclones over the Northeast United States. M.S. thesis, University at Albany.

    Klein, J. R., 2006: Mesoscale Precipitation Structures Accompanying Landfalling and Transitioning Tropical Cyclones over the Northeast United States. Northeast Regional Operational Workshop 8, Albany, New York.

    Srock AF, Bosart LF (2009) Heavy precipitation associated with southern Appalachian cold-air damming and Carolina coastal frontogenesis in advance of weak landfalling Tropical Storm Marco (1990). Monthly Weather Review: In Press

    Wes Junker. HPC Heavy Rainfall Forecasting Training Manual.



    Acknowledgements

    Many of the images and graphics used in this review were provided by parties outside of WFO RAH. The surface analysis graphic and regional Hanna precipitation analysis were obtained from the Hydrometeorological Prediction Center. The upper air analysis images were obtained from the University of Wyoming. GOES satellite data was obtained from National Environmental Satellite, Data, and Information Service. Other satellite imagery was obtained from CIRA and the Naval Research Laboratory Monterey Marine Meteorology Division. Storm Relative 16km Geostationary Water Vapor Imagery of Hanna provided by the NOAA Regional and Mesoscale Meteorology Branch and the Cooperative Institue for Research in the Atmosphere. Tropical cyclone phase diagram background images provided by Dr. Michael Brennan. SPC meso-analysis graphics provided by the Storm Prediction Center. The tropical cyclone model data and near real-time aircraft reconnaissance data for Google Earth was provided by http://www.tropicalatlantic.com. Google Earth map imagery used under license. The Crabtree Creek hydrograph provided by the NWS AHPS web page. The various QPE products were provided by the National Mosaic and QPE (NMQ) Web Page. Web based power outage maps provided by Progress Energy. Photos courtesy of the National Weather Service in Raleigh.



    Case study team -
    Barrett Smith
    Jason Beaman
    Michael Moneypenny
    Phillip Badgett
    Darin Figurskey
    Jonathan Blaes

    For questions regarding the web site, please contact Jonathan Blaes.


  • NWS Disclaimer.