ER-2 Doppler Radar (EDOP) Investigations of the Eyewall of Hurricane Bonnie During CAMEX-3 G. M. Heymsfield 1,J. Halverson 2,J. Simpson 1,L. Tian3and P. Bui 4 1NASA Goddard Space Flight Center, Greenbelt, MD 2JCET, University of Maryland Baltimore County (UMBC), Baltimore, MD 3Universities Space Research Associates, Seabrook, MD 4NASA Ames Research Center, Moffut Field, CA August, 2000 For submission to Journal of Applied Meteorology Corresponding author address: Gerald M. Heymsfield, NASA GSFC, Code 912, Greenbelt, MD 20771. ABSTRACT A persistent, mesoscaleregion ofintenseeyewall convection contained within Hurricane Bonnie on23August 1998is examinedfrom multiple observationssynthesizedfrom theNASA ER-2andDC-8 aircraft. The intense convection occurred latein theday asBonniewasattaining its minimum central pressureandduring astagewhen theinner corefeatured amarkedly asymmetric structure. The mainpurpose of thispaper is todescribethe internal structure of aconvectiveburst anditsrelationship to the warm core using a synthesis of high-resolution satellite, aircraft radar, and in situ data. An exceptionally vigorous eyewall tower penetrating to nearly 18 km is described. A second intense eyewall tower, adjacent to the eye, is shown to be associated with a mesoscale subsiding current of air that extends horizontally nearly 20 km into the eye interior. The subsidence occupies a substantial depth within the eye and appears to be a much larger scale feature than the convectively-induced, symmetric overturning which commonly occurs on the upper-level flanks of convective towers in other tropical environments. 1.0 Introduction Factors leading to changes in hurricane intensity, especially at landfall, are of vital importance. Intensity forecasts have little skill and have shown only slight improvement in the past 20 years (DeMaria and Kaplan, 1999). Several factors have been identified which contribute to sudden intensification, such as warmer sea surface temperatures and dynamically forced strengthening of the upper-level outflow (Riehl, 1950). Gray (1998) has hypothesized sudden penetration of convergent wind surges near the surface, leading to unusually intense convection from inflow of high equivalent potential temperature (0E) air. This high energy air is carried aloft by giant cumulonimbus "hot towers" (Malkus and RiehI) leading to warming of the inner core. It is in light of these hot towers that we examine an example of intense, long-lived episodes of deep convection within the eyewall of deepening tropical cyclone Bonnie, often referred to as "convective bursts". Several case studies spanning nearly thirty years, pioneered by Gentry et al. (1970) and more recently by Holliday and Thompson (1979), Steranka et al. (1986), Zehr (1992), Lyons and Keene (1994), and Rodgers et al. (1998), point toward a relationship between the occurrence of convective bursts and sudden intensification. In this paper, we define a convective burst as a mesoscale group or succession of hot towers topped by an anomalously cold anvil cloud mass (seen in geostationary IR imageD" ). The anvil rapidly expands with time, remaining stationary relative to the center location for several hours. Some of the deep cumulonimbi or "hot towers", many of which may be exceptionally vigorous and overshoot their equilibrium level. In addition, recent hurricane intensity change studies have focused on warm core anomalies observed in microwave satellite observations (Kidder et al. 1978;Velden etal.1991).A recentstudy utilizing datafrom the Advanced Microwave Sounding Unit (AMSU) onNOAA-15 (Velden etal. 1999)confirmed that significant warming occurred through deeplayersin theeyesofhurricane Georgesand Bonnieduring 1998.This latter work alink betweenthe magnitude ofthe satellite-derived warm coreand the surfacepressureatthestorm center assumingthestorm wasin hydrostatic balance. What hasnot beenwell understood however, is the details of the mechanismsproducing thesewarm cores. This important andcontroversial issuehasbeenaddressedwith models (Smith 1980;ZhangetaI.2000 ) andobservations (Malkus 1958;Kuo 1959; Shapiro andWilloughby, 1982;Willoughby 1999).Many of theproposed mechanismslinking corewarming with deepeningof the storm invoke an important rolefor subsidencein the eye. In mostconvective burst studies, information onburst morphology and the relationship of thewarm corewas obtained from views obtained from passiveremotesensors. In this paper, we provide detailedobservations onthe internal three-dimensional structure of a convectiveburst, including examplesofvigorous hot towers andthe air motions containedwithin them. Unique tothis study isobservational evidencefor a broad current ofstrong (severalmetersper second)descentinduced on theflank of intenseconvective hot towers, contained within the larger envelope ofa convective burst, andpenetrating deepwithin thewarm coreof thedeveloping eye. TheNASA ER-2and DC-8 participated with threeNOAA aircraft (NOAA-42 andNOAA-43 WP-3D Orion aircraft, and theG-IVSPGulfstream aircraft) on acomprehensivestudy of Hurricane Bonnie on23August 1998 2 during the Convection andMoisture Experiment-3 (CAMEX-3)1.CAMEX-3 focusedonhurricane objectivesincluding synoptic flow andinner core dynamics. Themission goalsfor 23August were to discerneyewall structure ofa moderate,nearly-stationary hurricane, andtoprovide improved initial conditions for trackmodels2.A variety of remote sensing instruments including radars, microwave, visible to infrared range radiometers, and lidar were flown on the ER-2 and DC-8. In addition, the DC-8 carried in situ sensors for meteorological parameters and microphysics measurements. The ER-2 was instrumented with the ER-2 Doppler Radar (EDOP) which is an X-band (9.6 GHz) Doppler radar with dual 3°beam width antennas fixed at nadir and 30° forward of nadir (Heymsfield et al. 1996). Data sets from EDOP will be a principle focus in this paper, corroborated by in situ flight level measurements and dropsondes deployed within the eye and in the environment of intense convection. In Section 2 of this paper, we present a concise description of the various observational platforms utilized throughout the study. In Section 3, the morphology of the convective burst associated with the subsidence is discussed in terms of its larger-scale, dynamic evolution as viewed from satellite. In Section 4, the fine-scale structure of hot towers and subsident region contained within the convective burst is presented from the vantage of the EDOP. These findings are summarized in Section 5 along with discussion of possible mechanisms leading to broad-scale sinking motion in the eye region and its importance in the formation of Bom_ie's warm anomaly. LCamex-3 web page, http:llghrc.msfc.nasa.gov/camex3. 2Mission Summary for flight 980823H, obtained from NOAA Hurricane Research Division (HRD), 2.0 Data and methods 2.1 EDOP processing and analysis Details of the EDOP instrument may be found in Heymsfield et al. (1996). During the 23 August 1998 Bonnie flight, data were collected on the EDOP nadir beam with a 37.5 m gate spacing, and a pulse-repetition-frequency of 4400 Hz providing a Nyquist velocity of about 34 msL Processed reflectivities and Doppler velocity were obtained every 0.5 s, which corresponds to approximately 100 m of aircraft translation (aircraft ground speed - 200 - 220 msl). This oversamples precipitation but is performed to maximize resolution near cloud top and to allow for better aircraft motion corrections. The reflectJvity data have been calibrated to within about I dBZ. These reflectivities are corrected for attenuation using the surface reference approach (Iguchi and Meneghini 1994). The accuracy of this correction was slightly compromised since EDOP's nadir "surface" channel was not available and the surface return saturated the receiver in some of the lighter rain (weaker attenuation) periods. The mean Doppler velocity measurements have a standard deviation of about 0.1 ms _ for typical rain situations and excluding aircraft motions which can produce significantly larger uncertainties. Calculation of vertical air motions, w, from the ED©P-measured Doppler velocities have been described in Heymsfield et al. (1999). A number of steps are involved in going from the measured nadir-beam Doppler velocities vo, to gridded vertical velocity fields. Aircraft motions are first removed from the vD using flight parameters from the ER-2 inertial navigation system (INS) and the antenna tilt angles (the nadir antenna is not exactly pointed at nadir). The Doppler velocities with aircraft motion removed are vertical hydrometeor 4 motions (G)from which thevertical air motion w= vh + vt can be obtained with a hydrometeor fallspeed (vt) assumption based on the reflectivity. The vhand reflectivity measurements are interpolated to two-dimensional grids with identical sampling as the nadir measurements (100 m horizontal, 37.5 m vertical) to correct for pitch variations of the aircraft. Minor filtering of the grids was performed both horizontally and vertically. Calculation of w requires estimation of the vt at each grid point. The vt estimation is the most critical assumption in obtaining w since vt depends on many factors such as particle phase, size distributions, etc. The approach is similar to that used by Marks and Houze (1987) and Black et al. (1996) which uses reflectivity -vt relations relations for the snow, rain, transition (melting), and convection regions. Stratiform regions are separated vertically into three regions: rain, snow, and transition region corresponding to the melting layer. Their approach was modified slightly for EDOP observations by using a more realistic rain reflectivity -vt relation derived for a gamma distribution and also a parabolic profile is assumed in the transition region instead of a linear profile (Heymsfield et al. 1999). Details of the partitioning are described in the aforementioned papers. Difficulties in fallspeed estimation occur in mixed phase regions associated with convection where strong updrafts can loft liquid water, frozen rain, and graupel several kilometers above the melting level. Fallspeeds in these regions are usually between those of snow and rain but can be larger than rain for small hail detected by EDOP in a mid-latitude squall line (Heymsfield et al. 1999). Thus, w estimates are subject to errors mainly from improper aircraft motion removal andv, estimation, where larger errors occur in the convective region. 2.2 In situ meteorological parameters The DC-8 in situ measurements were comprised of standard flight-level measurements using the Data Acquisition and Display (DADS), plus special higher accuracy humidity and wind measuring instruments. Temperature, pressure, and humidity measurements were provided by the DADS system. Humidity measurements were measured by four separate instruments, but this study uses data only from the JPL Laser Hygrometer which is the best documented with high accuracy (May 1998). This instrument is based upon a near infrared tunable diode laser source operating near 1.37 _m. Originally designed for stratospheric measurements, its precision is 0.05 ppmv (parts per million volume) in the stratosphere, and also very good at lower altitudes. The temperature, pressure, and humidity measurements were used to calculate 0F using the method described by Bolton (1980). The Meteorological Measurement System (MMS) on the DC-8 (Bui 1997) provided high-accuracy 5 Hz wind measurements in addition to independent temperature and pressure measurements. Accuracies of the MMS are: pressure (-T-0.3hPa), temperature (-T-0.3 °C), horizontal wind vector (;1 ms-_), and vertical wind vector (<1 ms-l). 3.0 Evolution of the Convective Burst: Mesoscale Structure 3.1. Satellite IR evolution. The initial intensification of Bonnie from tropical depression to huricane status commenced late in the day on August 20, 1998. (Fig. 1). Hurricane Bonnie first developed an eye early on 22 August, with the deepening trend continuing until a surface minimum pressure of 954 hPa was reached near 0000 UTC on 24 August. During the intensification period, convective bursts were noted in both the satellite IR observations and also by the NASA aircraft pilots. An initial burst formed adjacent to the evolving eye late on 21 August. Its hot towers observed during the aircraft mission were overshooting the tropopause to at least 17 km Within 6 h of this burst, a closed eye appeared in the satellite imagery. In addition, on 22 August, the TRMM Precipitation Radar (PR) again observed an exceedingly tall hot tower or "chimney" cloud in the inner eyewall that topped 18 kin. The maximum surface winds of about 52 ms -1were reached and maintained after these episodes of intense convective activity and near the time of the minimum surface pressure (0000 UTC, 24 August). The intensification of Bonnie is shown in Fig. 2 with a larger scale sequence of GOES IR satellite imagery from 1645 on 23 August (panel A) to 0215 on 24 August (panel F). The coldest IR temperatures with Tb < 200 K are depicted by white regions embedded in the overall cloud shield (mostly red colors), and defines the mesoscale extent of the burst. This sequence indicates a strongly asymmetric region of coldest cloud top temperatures undergoing a pronounced expansion between 1915-0015, and with the most active convection concentrated in the north and east quadrants of Bonnie. The asymmetric mesoscale burst in the inner eyewall (most evident after 23/2015) persists for about 10 h in the same location relative to the center of the vortex. Persistent formation of intense convective cells in the same quadrant of the eyewall is often observedin tropical cyclones(Simpson,1966;Gentry, 1970),andin thecaseof Bonnie,may havebeencausedby aconvergent region establishedby alow-level jet identified in theNOAA-42 WP-3D aircraft flight level wind analysis(see footnote 2).Other interesting featuresof the temperatures areepisodesoferosion ofthe cloud shield on the westhalf (forward quadrants) ofBonnie (e.g.,August 23at 1645UTC [panel A], August 24at0215UTC [panel F]etc.)presumably due to intrusion of drier air on the forward quadrants of the storm. Focussingonthe finer scalestructure of this burst revealsconsiderable evolution in the coldestIR areas(Fig.3). The GOESIRtemperatures from the inner coreregion areshown in Fig.3 for aperiod covering the aircraft flights (1846to2145UTC). During this sequence,GOESwas in ashort interval mode andimageswere collectedatapproximately 7.5min intervals. Theinner core region andapproximate inner eyewall location is about 100km in diameter and isshown on the imagesassumingno changein diameter over the3-hour period. Theconvectiveactivity is strong during the sequence,andthe coldest temperaturesareeastandnorth of the circulation centerin the left rearquadrant. Severalhot towerslabeled CellsA, B,C,and D persist for 30min or morewith minimum temperatures 190-195K, andarepresumed tobeassociatedwith tall overshooting convection along thenortheasteyewall of Bonnie. Thecells develop in successionnearthesamegenesisregion before2100UTC, then weaken,andfinally advectcyclonically andoutward from the eyewall. After 2100,convective activity increasedalong theeasterneyewall. Of particular interest isthe factthat aportion of the cirrus outflow from the vigorous cells advectsover theeasternportions of the inner corewhich causesthe eyetobe obscuredby acirrostratus layer. This cirrostratus is alsovery significant in that