1692 VOLUME130MONTHLY WEATHER REVIEWq 2002 American Meteorological SocietyDirect Surface Thermodynamic Observations within the Rear-Flank Downdrafts ofNontornadic and Tornadic SupercellsPAULM. MARKOWSKIDepartment of Meteorology, The Pennsylvania State University, University Park, PennsylvaniaJERRYM. STRAKASchool of Meteorology, University of Oklahoma, Norman, OklahomaERIKN. RASMUSSENCooperative Institute for Mesoscale Meteorological Studies, National Severe Storms Laboratory, and University of Oklahoma,Norman, Oklahoma(Manuscript received 19 March 2001, in final form 7 November 2001)ABSTRACTDespite the long-surmised importance of the hook echo and rear-flank downdraft (RFD) in tornadogenesis,only a paucity of direct observations have been obtained at the surface within hook echoes and RFDs. In thispaper, in situ surface observations within hook echoes and RFDs are analyzed. These ‘‘mobile mesonet’’ datahave unprecedented horizontal spatial resolution and were obtained from the Verifications of the Origins ofRotation in Tornadoes Experiment (VORTEX) and additional field experiments conducted since the conclusionof VORTEX. The surface thermodynamic characteristics of hook echoes and RFDs associated with tornadic andnontornadic supercells are investigated to address whether certain types of hook echoes and RFDs are favorable(or unfavorable) for tornadogenesis.Tornadogenesis is more likely and tornado intensity and longevity increase as the surface buoyancy, potentialbuoyancy (as measured by the convective available potential energy), and equivalent potential temperature inthe RFD increase, and as the convective inhibition associated with RFD parcels at the surface decreases. It ishypothesized that evaporative cooling and entrainment of midlevel potentially cold air may play smaller rolesin the development of RFDs associated with tornadic supercells compared to nontornadic supercells. Furthermore,baroclinity at the surface within the hook echo is not a necessary condition for tornadogenesis. It also will beshown that environments characterized by high boundary layer relative humidity (and low cloud base) may bemore conducive to RFDs associated with relatively high buoyancy than environments characterized by lowboundary layer relative humidity (and high cloud base).1. Introduction and motivationNo obvious characteristics capable of discriminatingbetween hook echoes associated with tornadic and non-tornadic supercells are apparent in radar reflectivity data(Fig. 1). Moreover, recent dual-Doppler radar obser-vations from Verification of the Origins of Rotation inTornadoes Experiment (VORTEX) have shown that, atleast kinematically, the differences between tornadicandnontornadic supercells are subtle, if even distinguishablein three-dimensional velocity data (Blanchard andStraka 1998; Wakimoto and Liu 1998; Wakimoto et al.1998; Trapp 1999; Wakimoto and Cai 2000). The imagesin Fig. 1 highlight a major forecasting challenge—howCorresponding author address: Dr. Paul Markowski, 503 WalkerBuilding, University Park, PA 16802.E-mail: [email protected] tornadic supercells be distinguished from nontor-nadic supercells, let alone their tornado potential be an-ticipated in advance?Although the dynamical relationship remains poorlyunderstood, the association among hook echoes, rear-flank downdrafts (RFDs), and tornadoes is well estab-lished (Markowski 2002a). However, direct observa-tions within hook echoes and RFDs have been scarce.A few observations have been mentioned by van Tassell(1955), Beebe (1959), Garrett and Rockney (1962),Browning and Ludlam (1962), Charba and Sasaki(1971), Lemon (1976), Barnes (1978a,b), Brown andKnupp (1980), and Bluestein (1983). Thermodynamicretrievals have been performed (e.g., Brandes 1984a;Hane and Ray 1985), but small-scale details cannot beresolved, buoyancy fields often are noisy, and data with-in the surface layer, which are perhaps most important,are unavailable. Would in situ surface observations col-JULY2002 1693MARKOWSKI ET AL.FIG. 1. A sample of some of the hook echoes associated with both tornadic and nontornadic supercells from which mobile mesonetobservations have been collected. The hook echoes associated with tornadic supercells are as they appeared 5 min or less prior to torna-dogenesis. No obvious, systematic differences are apparent between the hook echoes associated with the tornadic and nontornadic supercells.lected in the RFDs of the nontornadic and tornadic su-percells analyzed by Blanchard and Straka, Wakimotoet al., and Trapp reveal significant differences? Andcould any differences detected by a mobile mesonet everbe inferred from environmental data available routinelyon larger scales?The ‘‘mobile mesonet’’ is an observing system com-prising vehicle-borne sensors that provide direct mea-surements of temperature, moisture, wind velocity, andpressure (Straka et al. 1996). The observing platformwas developed for VORTEX (Rasmussen et al. 1994)and has been heavily relied upon in smaller subsequentstorm intercept projects organized by the National Se-vere Storms Laboratory and University of Oklahoma.In this paper, 5 years of observations of nontornadic andtornadic supercell RFD and hook echo regions are sum-marized. The paper has two main objectives: 1) docu-ment the surface thermodynamic fields in the proximityof tornadic and nontornadic mesocyclones at a resolu-tion not previously possible; 2) determine if differencesexist at the surface in the hook echoes and RFDs oftornadic and nontornadic supercells.a. Documentation of surface thermodynamic fieldsBrandes (1984a) and Hane and Ray (1985) wereamong the first to use the pioneering methods proposedby Gal-Chen (1978) and Hane et al. (1981) to retrieve1694 VOLUME130MONTHLY WEATHER REVIEWthermodynamic (buoyancy and pressure) fields in su-percells from multiple-Doppler-radar-synthesized three-dimensional wind fields. However, the computation ofbuoyancy requires assumptions about lateral boundaryconditions and the forcing for buoyancy involves onemore derivative than the forcing for pressure. Thus,buoyancy fields retrieved in past studies may containsignificant uncertainties and usually contain greaternoise than retrieved pressure fields.The spatial resolution of retrieval results is limitedby the resolution of the dual–Doppler radar data. Usu-ally horizontal resolution is limited to 1–3 km. Fur-thermore, ground clutter contaminates Doppler veloci-ties; therefore,