system depends in large measureon the underlying NWP modelrather than on the subsequent ap-plication of analogs.

Pat tern-recognition and forecastanalogs show promise for discern-ing the patterns associated withmajor weather events. Future workrequires regular updating oftheevents database as well as the asso-ciated fingerprints.—PAUL KNIGHT(THE PENNSYLVANIA STATE UNI-VERSITY), B. ROOT, G. YOUNG, S.GREYBUSH. R. GRUMM, R. HOLMES,AND ]. Ross. "A Fingerprinting Tech-nique for Major Weather Events," inthe July Journal of Applied Meteo-rology and Climatology.

COULD OCEAN-MODULATEDW I N D BURSTS LEAD TOBETTER EL NINO FORECASTS?Episodes of strong westerly windsover the tropical Pacific, knownas westerly wind bursts (WWBs),have accompanied every signifi-cant El Nino event ofthe past 25years. In many models, WWBsare treated as a purely stochastic(i.e., random) forcing, consistentwith the description of ENSO as adamped oscillatory system drivenby external noise. Recent observa-tions, however, show that while theprecise timing of a specific WWBevent is unpredictable, the char-acteristics of WWBs, such as thelocation and likelihood of occur-rence, depend upon the large-scalesea surface temperature {SST)field. The WWB characteristicsare therefore "modulated" by theSST and should not be consideredas purely stochastic or external tothe coupled ENSO system. We findthat when the feedback betweenWWBs and SST is included in anumerical model, the amplitudeof ENSO variability is twice aslarge as from a model with purelystochastic WWBs. A new picture

of ENSO seems to emerge: a pos-sibly self-sustained oscillatingsystem maintained by the effectsofthe modulated WWBs.

We examined the consequencesof WWB modulation by the oceanusing an ocean general circula-tion model coupled to a statisticalatmosphere model with the in-clusion of a novel seniistochasticWWB parameterization. Theparameterization was constructedsuch that the likelihood of WWBoccurrence increases as the west-

ern Pacific warm pool extends; asemistochastic formulation thathas both the deterministic andstochastic elements of WWBs seenin observations.

We found that modulation ofWWBs by SST strongly affectsthe characteristics of ENSO. Inparticular, coupled feedbacksbetween SST and WWBs may besufficient to transfer the systemfrom a damped regime to onewith self-sustained oscillations.Modulated WWBs also play a

140E1990 1991 1992 1993 1994 1995 1996 1997

Evidence for westerfy wind burst modulation. Equatorial zonal wind anoma-lies in the TAO buoy data greater than 4 m s~' (blue) are superimposedon the western Pacific warm pool extent (pink) as measured by the 29°CSST isotherm. These data imply that WWBs are 2.8 times more likely tooccur when the warm pool extends past the date line. (GEBBIE ET AL.)

100WM40W run' :^ 1 1 ' J I t •! 1 i .i I I I

Diverse dynamical regimes in an ENSO model. The response of model ENSOvariability to different westerly wind-burst formulations. The left panelsshow a time-longitude section of W W B zonal wind-stress anomaly (T) av-eraged between 5°N and 5°S. The right panels show the resulting Nino 3.4index for three model experiments. The top row is from a purely stochasticW W B representation, the middle from a semistochastic representation inwhich the statistics of WWBs depend upon the SST, and the bottom rowis from a deterministic W W B parameterization where the precise W W Btiming is completely determined by the SST. Observed W W B character-istics seem to be close to the semistochastic formulation. (GEBBIE ET AL.)

1156 BAflJ SEPTEMBER 2007

HflWCMECHOES

With population growth and the decline in the water,there are the elements in the equation which you could call 'theperfect drought."*

—BILL PATZERT, a climatologist at NASA's Jet Propulsion Laboratory inPasadena. California, on conditions in Southern California. Through late

June, Los Angeles was suffering through its driest year since record keepingbegan in 1877. with only 3.21 inches of rainfall over the previous 12 months

and no rain in the forecast before September. Average rainfall during itsJuly-June rainy season is 15.14 inches. Additionally, the Sierra Nevada

mountains had delivered only 20% of their normal rain production to theregion. Combining these conditions with regional population growth overthe past 50 years of 2-4 times the national average. Southern California isfacing a portentous water situation where, as Patzert points out, "Mother

Nature is converging with human nature." (SOURCE: The Guardian)

role in the irregular timing ofwarm episodes and the asym-metry in the size of warm andcold events in this ENSO model.Parameterizing the modula-tion of WWBs by an increaseof the linear air-sea couplingcoefficient seems to miss impor-tant dynamical processes, and apurely stochastic representationof WWBs elicits only a weakocean response.

The new view of wind burstsmay have implications for thepredictability of El Nino. Themodulation of WWBs by theocean suggests that ENSO's pre-dictability limit may be longerthan would be anticipated if theWWBs were purely stochastic.If the observation that WWBsare modulated by the SST canbe translated into a practicalWWB prediction scheme, thereis hope for enhanced skill inENSO forecasts.—GEOFFREYGEBBIE (HARVARD UNIVERSITY),I. EiSENMAN, A. WITTENBERG,AND E. TziPERMAN. "Modulationof Westerly Wind Bursts by SeaSurface Temperature: A Semi-Stochastic Feedback for ENSO,"in the September Journal oftheAtmospheric Sciences.

VORTICITY-BASED DETECTIONOF TROPICAL CYCLOGENESISAlthough tropical cyclogenesis(TCG) is a veryactive area of re-search, it remainsa highly debatableand unresolvedtopic. While con-siderable attentionhas been paid totropical cycloneformation, littleattention has fo-cused on obser-vational studiesof the very earlystages of TCG,otherwise referredto as the genesisstage. In the past,the early stages ofTCG were unveri-fiable in surfaceobservations, dueto the paucity ofmeteorologicaldata over the trop-ical oceans. Usingsatellites, we devel-oped a robust tech-nique for locatingand monitoringtropical distur-

bances from the early genesis stageto the tropical depression stage inthe Atlantic Basin, Overall, thedetection technique proved veryeffective, identifying such tropi-cal disturbances approximately19 to 101 h before classification astropical cyclones by the NationalHurricane Center.

"Ihis new approach used a com-bination of Qu/fcSCAT observa-tions of surface vorticity andGOES imagery of cloud-top tem-peratures to track systems (thosethat did develop into tropicalcyclones) backward through theirdevelopment. Our technique isbased on identification of surfacevorticity and wind-speed signa-tures that exceed certain threshold

QSCAT Vorticity 2132 UTC 2 November 2001Windlegend (ms

0,0

5,0

10.0

15.0

20.0

Rain-Flag

<1 2 3 4 b x10 s

An example of the early stages of TCG identifiedby the vorticity-based detection technique showsTropical Cyclone Noel (2001) 26 h before classifica-tion as a subtropical storm. The background colorrepresents QuikSO^T-derived relative vorticity, withdark green representing all vorticity values greaterthan 5.0 x 10 ^ s '. The black, solid lines signify thelocations where the detection technique's criteriaare met within 75 km from the cloud-cluster centerin the associated GOES infrared image. This vor-ticity signature is associated with the nontropicaloccluded low that spawned Noel, and illustrates anapparent surface circulation. Noel eventually grewinto a weak hurricane. (GIERACH ET AL.)

AMERICAN METEOROLOGICAL SOCIETY SEPTEMBER 2007 BAHS" | 13S7