Shear-Driven Instabilities as the Origin of Multi-Banded Cloud and Precipitation Structures in an Extratropical Cyclone

Gene a ed using he o icial AMS L A T EX empla e 6.1 Shea -D i en Ins abili ies as he O igin o Mul i-Banded Cloud and1 P ecipi a ion S uc u es in an Ex a opical Cyclone2 S ephen R. Guimonda 3 aDepa men o A mosphe ic and Plane a y Sciences and Se e e Wea he Resea ch Cen e , Hamp on Uni e si y, Hamp on, VA, USA 4 5 Co esponding au ho : S ephen R. Guimond, [email p o ec ed]6 1 ABSTRACT: This pape in es iga es he dynamics go e ning mul i-banded cloud and p ecipi a ion in ex a opical cyclones h ough a case s udy om he NASA In es iga ion o Mic ophysics and P ecipi a ion o A lan ic Coas -Th ea ening Snows o ms (IMPACTS) ield campaign. On 1 Feb ua y 2020, a low-p essu e sys em eme ged o he No h Ca olina coas a 1200 UTC, deepening by 7 hPa in six hou s as i accele a ed no heas o e he A lan ic. High- esolu ion GOES isible image y e ealed mul iple bands o high- e lec ance cloud o he no h/no heas o he cen e , along wi h clus e s o con ec i e cells close o he co e. Wa ele analysis iden i ied a dominan mul i-band wa eleng h o 30 km and a seconda y peak a 15-20 km. Ai bo ne ada measu emen s om IMPACTS ligh s showed deep con ec ion nea he cen e and na ow, ele a ed e lec i i y bands linked o he mul i-band ea u es a he ou . Nume ical simula ions ep oduced he mul i-bands, enabling explo a ion o hei dynamical o igin. In insic phase speed calcula ions e ealed ha , con a y o se e al p e ious s udies, he dominan mul i-bands we e no g a i y wa es. Ins ead, he ea u es we e iden i ied as dynamic ins abili ies (Kel in–Helmhol z ins abili y) a ising om e ical wind shea and low Richa dson numbe s nea he uppe -le el ou low. G a i y wa es we e p esen in he low o mid le els (0–6 km) gene a ed by con ec ion, bu hey did no accoun o he s ong pe u ba ions in he mid o uppe le els (6–10 km). This s udy p esen s new scien i ic insigh in o he go e ning dynamics o mul i-banded s uc u es in ex a opical cyclones ha highligh s he ole o shea -d i en ins abili ies. 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 2 SIGNIFICANCE STATEMENT: Ex a opical cyclones along he Uni ed S a es Eas Coas can25 p oduce in ense bands o snow and o he ypes o p ecipi a ion ha cause a el chaos and po en ial26 loss o li e. In his wo k, a s o m o his ype was s udied using NASA ai c a and sa elli e27 da a as well as compu e simula ions. The esul s show ha he s o m bands we e no caused by28 a mosphe ic wa es, as scien is s o en hough , bu by sha p changes in he wind speed wi h heigh 29 called shea . This shea c ea ed uns able laye s ha olled and mixed he ai , o ming o ganized30 cloud and p ecipi a ion pa e ns. Unde s anding hese undamen al physical p ocesses can help31 o ecas e s be e p edic when and whe e hea y snow bands will o m, imp o ing wa nings and32 public sa e y.33 1. In oduc ion34 Banded s uc u es in ex eme wea he sys ems a e pe u ba ions o a balanced, backg ound35 low ha can o ganize and concen a e a iables such as mois u e, momen um, and ene gy. In36 ex a opical cyclones (ETCs), he ocus o his pape , he concen a ion o hese a iables can37 lead o in ense bands o mul i-phase p ecipi a ion a he su ace ha a e di icul o measu e,38 model and p edic wi h signi ican consequences o socie y. Fo example, snow all associa ed39 wi h ETCs in he win e mon hs can o en o ganize in o mul iple bands ha d op la ge amoun s40 o snow in a sho ime causing ehicle c ashes, ligh cancella ions and shu downs o schools and41 businesses. The cu en unde s anding o he dynamical p ocesses con olling he o ma ion and42 e olu ion o hese p ecipi a ion mul i-bands (as opposed o a single, la ge-scale band) and hei 43 ep esen a ion/p edic abili y in nume ical models is e y limi ed. These limi a ions we e pa o he44 mo i a ing ac o s o he ecen ly comple ed In es iga ion o Mic ophysics and P ecipi a ion o 45 A lan ic Coas -Th ea ening Snows o ms (IMPACTS; ?) ield expe imen , which sough o imp o e46 he unde s anding o p ecipi a ion mul i-bands in ETCs om a ious pe spec i es.47 ?used ada obse a ions, soundings, and eanalysis da a o examine he en i onmen s con aining48 banded s uc u es in a la ge se o ETC win e s o ms. They ound ha mul i-bands we e no well49 co ela ed wi h on ogene ical o cing and associa ed de o ma ion zones. This esul is consis en 50 wi h he no ion ha de o ma ion alone canno explain he egula , oscilla o y na u e o he a ious51 ields connec ed o mul i-bands. The au ho s also examined he p esence o condi ional symme ic52 ins abili y (CSI) in he en i onmen o a ious ypes o banded s uc u es. While CSI was p esen 53 3 in he as majo i y o banded cases, he e was no clea sepa a ion be ween single, la ge-scale bands54 and he smalle -scale, mul i-band s uc u es. This sugges s ha CSI may no be he undamen al55 p ope y unde lying he o ma ion and e olu ion o mul i-bands.56 In hei e iew pape , ?no ed ha he banding o clouds and p ecipi a ion in ETCs appea s o57 be weakly ela ed o CSI. ?used a iscous o m o he Sawye -Eliassen equa ion o unde s and58 banding mechanisms associa ed wi h on ogene ical o cing. Fo b oad o cing and nega i e mois 59 po en ial o ici y, which is a condi ion ha e lec s he p esence o CSI, mul iple bands de elop60 wi h an in ensi y ha scales wi h he deg ee o ins abili y. Howe e , i is no clea om he heo y61 o ?exac ly how CSI coupled wi h mois u e and li , associa ed wi h on ogenesis, would p oduce62 he egula ly spaced bands o clouds and p ecipi a ion obse ed in he eal a mosphe e. In an63 uns able en i onmen , pa cels a e accele a ed exponen ially in he di ec ion o he displacemen 64 and do no main ain an oscilla o y beha io .65 ?and ?desc ibed se e al condi ions ha should be me o associa e p ecipi a ion mul i-bands66 wi h CSI. One o he key equi emen s is ha he bands ha e no in insic p opaga ion as hey should67 be mo ing di ec ly wi h he en i onmen al low. Howe e , se e al s udies o mul i-bands do show68 in insic p opaga ion, which when coupled wi h he desc ip ion abo e, ques ion he ole o CSI in69 he band dynamics.70 Ano he possible culp i o mul i-bands is he p esence o a wa e phenomenon o o ganize he71 oscilla ions in he s a e a iables and p o ide he li necessa y o elease any ype o ins abili y72 in he en i onmen . Se e al p e ious s udies ha e ocused on he ole o g a i y wa es as he73 mechanism o o ganizing he cloud and p ecipi a ion ields in o mul i-band s uc u es in ETCs74 (e.g., ???). These wa es can be gene a ed om low imbalances (e.g., associa ed wi h uppe -75 le el je s), con ec i e pe u ba ions, and/o low o e opog aphy. ?documen ed la ge-ampli ude76 mesoscale g a i y wa es (wa eleng hs o 200 - 260 km) du ing se e al win e s o ms using ime77 se ies o su ace p essu e and wind measu emen s. These g a i y wa es we e ound o o igina e a 78 uppe le els om je s eak imbalances and he associa ed geos ophic adjus men p ocess (e.g., ?).79 Fo la ge-ampli ude wa es, a su ace e lec ion o he uppe -le el pe u ba ion ene gy was de ec ed80 and hypo hesized o become apped a lowe -le els due o s ong s a ic s abili y (“wa e duc ing”;81 ?). This wa e duc ing allows o a longe esidence ime o he g a i y wa e ene gy in he lowe 82 4 le els o he a mosphe e esul ing in he po en ial o o ganized clouds and p ecipi a ion as shown83 by sa elli e and ada ins umen s.84 ?also highligh ed he ole o mesoscale (∼200 km) g a i y wa es using nume ical simula ions o 85 an ETC using a ious esolu ions down o 4 km spacing. The au ho s p oposed a concep ual model86 o he gene a ion and e olu ion o mesoscale g a i y wa es based on hese simula ions. Fi s , he87 wa es we e ini ia ed in he uppe le els o he sys em om he geos ophic adjus men p ocess.88 Then, con ec ion associa ed wi h he on al ea u es and local ins abili y anspo ed some o he89 g a i y wa e ene gy owa ds he su ace whe e i can be duc ed wi hin a s a ically s able laye . This90 duc ed wa e hen in e ac s wi h he mois u e ield o c ea e a con ec i ely-coupled wa e ha can91 be main ained and possibly ampli ied o e a signi ican amoun o ime.92 Se e al ques ions and unce ain ies come o mind ega ding he applicabili y o hese p e ious93 g a i y wa e s udies o he gene al p oblem o ETC mul i-bands. How does con ec i e ac i i y94 p e e en ially anspo he wa e ene gy ini ia ed a uppe le els (nea he opopause) downwa d95 o low le els and keep he wa e in ac ? The p esence o a s ong s a ic s abili y laye a lowe 96 le els and e lec ing laye abo e o enable wa e duc ing appea s o be a special se o ci cums ances97 and i is no clea how his laye can be o med and main ained in egions o he sys em ha a e98 con ec i ely ac i e.99 The goals o he p esen s udy a e: (1) To de e mine he key s uc u es and spa ial/ empo al scales100 o mul i-banded p ecipi a ion ea u es in ETCs and (2) To de e mine he o igin and dynamical101 p ocesses associa ed wi h hese bands. To add ess he abo e goals and de elop po en ially new102 unde s anding o ETC mul i-bands, a case s udy om he NASA IMPACTS ield campaign is103 analyzed wi h mul i-scale nume ical simula ions and a a ie y o emo e sensing measu emen s.104 Case s udies a e an impo an i s s ep owa ds de eloping a deep, holis ic unde s anding o a105 physical p ocess, which hen allows ha unde s anding o be es ed mo e b oadly on a la ge 106 collec ion o sys ems. The echnical no el ies o he p esen wo k a e in he s udy o mode n107 emo e sensing measu emen s and nume ical models along wi h some new me hods o analysis o108 he p oblem o ETC mul i-bands.109 5 Fig. 1. Synop ic maps o mean sea le el p essu e and on al analysis on 1 Feb ua y, 2020 a (a) 1200 UTC and (b) 1800 UTC. 2. B ie o e iew o ex a opical sys em110 Du ing he win e o 2020, NASA o ganized he i s phase o a mul i-yea e o called IMPACTS111 o s udy he mic ophysics and dynamics o ETCs wi h a ocus on unde s anding banded egions o 112 p ecipi a ion. On 1 Feb ua y 2020, a double su ace low-p essu e sys em along a s a iona y on 113 o med o he coas o he Ca olinas wi h he sou he n low s a ing o a ∼1006 hPa on 1200114 UTC 1 Feb ua y (Fig.1a). Du ing he nex 6 hou s, he sou he n low in ensi ied o 999 hPa a 115 1800 UTC 1 Feb ua y (Fig.1b) and o 998 hPa a 0000 UTC 2 Feb ua y (no shown). Du ing his116 pe iod, mul iple bands o con ec i e clouds and p ecipi a ion o med o he no h and no heas 117 o he sou he n low-p essu e cen e , which allowed de ailed s udy o hei cha ac e is ics. Fligh s118 om he NASA P-3 and ER-2 ai c a occu ed du ing ∼1200 - 1800 UTC 1 Feb ua y sampling119 he en i onmen and banded ea u es associa ed wi h he sou he n low.120 6 3. Da a and p ocessing121 a. EXRAD122 The ER-2 X-band Dopple Rada (EXRAD) is an X-band, downwa d-poin ing ai bo ne ada 123 ha measu es ada e lec i i y and Dopple eloci y om hyd ome eo s wi h wo beams and ∼20124 m ga e spacing. The i s beam is ixed and poin s nominally a nadi , while he second beam scans125 conically a 20 e olu ions pe minu e a a nominal incidence angle o 32°. F om he al i ude o 126 he NASA ER-2 ai c a (∼20 km), he swa h wid h a he i s le el o use ul da a is ∼23 km.127 The EXRAD scanning beam main- and side-lobes in e ac s ongly wi h he ocean su ace and128 con amina e he p ecipi a ion signal below abou 1 km heigh . The sampling o p ecipi a ion om129 he scanning beam is ∼500 m along- ack and ∼2°in azimu h while he nadi beam along- ack130 sampling is ∼50 m.131 Calcula ions o he h ee-dimensional (3D) wind ield using he EXRAD scanning beam a e132 pe o med using he 3D a ia ional algo i hm desc ibed in ?. The e ie al g id is se o 500 m in133 he ho izon al dimensions and 250 m in he e ical dimension. Fine g id spacing in he e ical134 is possible, bu was no deemed necessa y. A wo o h ee g id poin unning mean il e is applied135 o he aw e ie als in pos -p ocessing o emo e nume ical noise. The e ec i e esolu ion o 136 he wind ields p oduced om he algo i hm has been analyzed wi h la ge eddy simula ions in ?.137 Resul s show ha scales o 5 Δ𝑥and la ge a e ully esol ed, which ansla es o 2.5 km o he138 p esen s udy. Scales below his h eshold, down o he g id scale, a e subjec o inc easing kine ic139 ene gy a enua ion.140 Quali y con ol has been pe o med on he wind ields o emo e da a wi h low signal- o-noise141 a ios and high unce ain ies. While a combina ion o pa ame e s ha e been s udied, he bes quali y142 ields we e ound by emo ing da a wi h s anda d de ia ions la ge han 6 m/s (e o s a is ics a e143 compu ed as pa o he wind algo i hm, see ?). The da a p esen ed in his pape ha e his h eshold144 applied. Valida ion o he EXRAD 3D winds wi h ligh le el in-si u da a om he NASA ER-2145 ai c a du ing IMPACTS 2020 ha e been pe o med. This alida ion showed zonal wind oo 146 mean squa e e o s (RMSEs) o 3.99 m/s wi h a co ela ion coe icien o 0.92. Fo he me idional147 wind, he RMSEs a e 4.53 m/s wi h a co ela ion coe icien o 0.89. These s a is ics a e collec ed148 7 a a ious loca ions ac oss he ada swa h and o di e en ime o se s. Fu he in o ma ion abou 149 hese e o s a is ics and he EXRAD scanning beam wind e ie als can be ound in (?).150 Fo scien i ic analysis, he EXRAD scanning beam is used o p esen he ho izon al wind ield,151 while he nadi beam is used o display he e ical wind ield, since i has highe quali y. Se e al152 co ec ions o he nadi beam Dopple eloci ies a e needed be o e he e ical eloci y can be153 analyzed o science. The exp ession o he nadi beam Dopple eloci y (𝑉𝑑) is gi en by154 𝑉𝑑=® 𝑉𝑤·ˆ𝑒−® 𝑉𝑔·ˆ𝑒+𝑉𝑛(1) whe e ˆ𝑒=𝑥ˆ 𝑖+𝑦ˆ 𝑗+𝑧ˆ 𝑘 𝑟and155 ® 𝑉𝑤=𝑢ˆ 𝑖+𝑣ˆ 𝑗+ (𝑤+𝑣𝑡)ˆ 𝑘, (2) ® 𝑉𝑔=(𝐺𝑆ℎ∗𝑠𝑖𝑛𝑇)ˆ 𝑖+ (𝐺𝑆ℎ∗𝑐𝑜𝑠𝑇)ˆ 𝑗+ (𝐺𝑆𝑣)ˆ 𝑘(3) and 𝑉𝑛 ep esen s he e ec s o non-uni o m beam illing.156 In hese equa ions x,y,z a e he Ea h- ela i e coo dina es o he ada pulse olumes (?), is157 he ange, u, ,w a e he componen s o he Ea h- ela i e wind, 𝑣𝑡is he hyd ome eo allspeed,158 𝐺𝑆ℎ,𝐺𝑆𝑣a e he ho izon al and e ical componen s o he ai c a g ound speed and Tis he159 ai c a ack angle.160 Sol ing o he e ical eloci y yields,161 𝑤=(𝑉𝑑−𝑉𝑛+® 𝑉𝑔·ˆ𝑒)𝑟−𝑢𝑥 −𝑣𝑦 𝑧−𝑣𝑡.(4) The calcula ion o hyd ome eo allspeeds om he e lec i i y measu emen s ollows he s udies162 o ?and ?. The non-uni o m beam illing e ec s a e emo ed om he Dopple eloci ies163 ollowing ?. Ai c a mo ion can esul in an enna poin ing angles ha in e cep he ho izon al164 wind ield, con amina ing he e ical winds. The ho izon al wind ields compu ed om he165 EXRAD scanning beam, desc ibed abo e, a e used o emo e his con amina ion om he nadi 166 beam ollowing equa ion 4.167 8 b. GOES-16168 High spa ial and empo al esolu ion image y o clouds om he ad anced baseline image on169 he Geos a iona y Ope a ional En i onmen al Sa elli e (GOES) - 16 sa elli e a isible and in a ed170 wa eleng hs a e used o ack banded s uc u es. Speci ically, e lec ance da a om he isible171 (channel 2) band a 500 m pixel spacing and 60 s ime upda es is u ilized. This ine sampling in172 space and ime p o ides he de ailed s uc u e o he cloud li ecycle as well as he o ganiza ion and173 e olu ion in o mul i-bands.174 c. Nume ical Simula ions175 The Wea he Resea ch and Fo ecas ing (WRF) model e sion 4.2 wi h he ad anced esea ch176 dynamic co e is u ilized o p o ide con ex o he obse a ions. Fou domains a e u ilized wi h177 a la ge, pa en domain a 2 km g id spacing (domain 1) co e ing he ull mo emen o he low-178 p essu e sys em and banded ea u es on 1 Feb ua y 2020. Th ee nes ed domains a 0.667 km179 (domain 2), 0.222 km (domain 3) and 0.074 km (domain 4) g id spacing we e placed o he No h180 and No heas o he low cen e o y and cap u e he ine scales o he mul i-bands. All domains181 u ilize 121 s e ched e ical le els wi h a spacing o ∼75 m a he su ace, ∼200 m a 10 km182 heigh and ∼900 m nea he model op a 20 km heigh .183 Po ions o he mul i-bands we e cap u ed in domains 2 and 3, bu no domain 4. The ocus o 184 he pape is on domain 1 o a ew easons. Fi s , his domain cap u es he en i e y o he sys em,185 which allows bo h la ge-scale and mesoscale ea u es o be analyzed. Second, he wa eleng hs186 o dominan ea u es ound in he obse a ions a e 15-20 km and ∼30 km, which should be well187 esol ed by he 2 km domain gi en he ∼7 - 8 Δ𝑥nume ical dissipa ion ange o WRF (e.g., ?).188 The highe esolu ion domains we e analyzed as pa o his esea ch and will be no ed whe e189 app op ia e.190 The se up o he model is as ollows. The NCEP Global Da a Assimila ion Sys em (GDAS)/Final191 (FNL) ope a ional global analyses a 0.25 °spacing and 6 h empo al spacing a e used as he ini ial192 and bounda y condi ions o he simula ion. This widely used sys em combines he Global Fo ecas 193 Sys em (GFS) model wi h a ious synop ic-scale obse a ions o achie e an op imal s a e o he194 a mosphe e. Fo he 2 km pa en domain he ollowing sub-g id physics schemes we e chosen:195 Thompson o mic ophysics, YSU o bounda y laye ( e ical di usion), Smago insky-2D o 196 9 Two hou s la e a 1600 UTC, he same ields ha e changed signi ican ly. The low-le el e lec i i y310 (Fig. 8a) e eals h ee bands o ien ed a di e en angles and deno ed wi h dashed, black ci cles.311 The mid-le el e lec i i y (Fig. 8b) shows some signa u es o he low-le el bands, such as he312 sou he nmos ci cled egion, bu o he bands o he no h do no ha e he same o ien a ion. These313 bands a e o ien ed app oxima ely pe pendicula o he along- ack axis ha deno es he axis o 314 band p opaga ion. The uppe -le el e lec i i y (Fig. 8c) shows na ow signa u es o mul i-bands315 ha ex end om nea he sys em cen e ou o app oxima ely 300 km adius. The wa eleng hs316 o he e lec i i y bands in Fig. 8 a e ∼30 km, which is e y simila o hose documen ed om317 he GOES da a, shown by he wa ele analysis in Fig. 5. The mul i-bands p esen in he model318 e lec i i y appea simila o hose obse ed in he GOES da a (Fig. 2).319 Figu e 9 shows he absolu e e ical o ici y ield a 1600 UTC. The low-le el (Fig. 9a) o ici y328 shows simila s uc u e o he ield a 1400 UTC. The mid-le el (Fig. 9b) o ici y shows no iceable329 cu a u e in he bands wi h posi i e/nega i e oscilla ions mo e p ominen o he eas o he along-330 ack axis. The uppe -le el (Fig. 9c) o ici y ield shows a ib an mul i-banded s uc u e o he331 no heas o he sys em cen e wi h o ici y oscilla ions up o ±1×10−3𝑠−1and wa eleng hs o ∼332 30 km. This s uc u e is consis en wi h he e lec i i y ield shown in Fig. 8c.333 To examine he mul i-bands mo e closely, he model da a is ou pu a wo minu e in e als and344 in e pola ed o a ack- ela i e g id, cen e ed on he black line in he p eceding igu es, wi h a g id345 spacing o 2 km in he ac oss- ack and along- ack dimension while keeping he na i e model346 e ical spacing. In addi ion, he along- ack (𝑈𝑎) and ac oss- ack (𝑈𝑥) eloci ies we e compu ed347 on his g id,348 𝑈𝑎=𝑢cos(T) + 𝑣sin(T)(5) 𝑈𝑥=𝑢sin(T) − 𝑣cos(T)(6) whe e 𝑢and 𝑣a e he zonal and me idional eloci ies and 𝑇is he g id ack angle o ∼75 °.349 A posi i e along- ack eloci y is mo ing owa ds inc easing along- ack alues (looking down350 he ack), while a posi i e ac oss- ack eloci y is mo ing owa ds inc easing ac oss- ack alues351 ( om le o igh looking down he ack). One snapsho o he bands a di e en le els is shown in352 16 Fig. 6. Ho izon al c oss sec ions o he simula ed e lec i i y (dBZ) in domain 1 (2.0 km g id spacing) on 1 Feb ua y, 2020 a 1400 UTC. Panels (a), (b) and (c) show he 0 - 3 km, 4 - 7 km and 7 - 10 km heigh a e aged ields, espec i ely. The black line in all panels deno es he along- ack axis (no mal o he phase lines) o mul i-bands obse ed in subsequen igu es. 320 321 322 323 he nex wo igu es. This snapsho is ep esen a i e o he mul i-band s uc u e unde in es iga ion353 and hus, mul iple snapsho s a e no shown.354 Figu e 10 shows he absolu e o ici y and e ical eloci y a e aged o e he 6 - 10 km heigh 355 ange on 1 Feb ua y, 2020 a 1530 UTC. In hese igu es, he aw model ields a e il e ed wi h a ∼356 20 km ac oss- ack unning mean and a ∼10 km along- ack unning mean o educe small-scale357 a iabili y and pull ou he la ge scale ea u es ha we e documen ed in he obse a ions. The358 sensi i i y o he mul i-bands o model g id spacing was analyzed by compa ing he s uc u es359 in domain 1 (2 km) and domain 2 (0.67 km). The aw ields, be o e il e ing, clea ly show360 mo e oscilla ions and la ge magni udes in se e al a iables on he highe esolu ion g id, which361 17 Fig. 7. Ho izon al c oss sec ions o he simula ed absolu e e ical o ici y (s−1) in domain 1 (2.0 km g id spacing) on 1 Feb ua y, 2020 a 1400 UTC. Panels (a), (b) and (c) show he 0 - 3 km, 4 - 7 km and 7 - 10 km heigh a e aged ields, espec i ely. The black line in all panels deno es he along- ack axis (no mal o he phase lines) o mul i-bands obse ed in subsequen igu es. 324 325 326 327 is expec ed. Howe e , a e il e ing he ields o he same scales no ed abo e, he mul i-band362 s uc u e looks simila in bo h domains wi h dominan ho izon al wa eleng hs o ∼30 km. Since363 he obse a ions also show dominan wa eleng hs o ∼30 km, he ields simula ed in domain 1 a e364 deemed su icien o examine he dynamics.365 The e ical eloci y (Fig. 10b) displays a simila mul i-banded s uc u e as he o ici y ield370 (Fig. 10a), al hough he o ici y pe u ba ions a e mo e ib an and con inuous when compa ed371 o he e ical eloci y ield. The message om Fig. 10 is ha he posi i e phase lines o o ici y372 in he mul i-bands a e la gely unco ela ed wi h he posi i e phase lines o e ical eloci y. To373 cla i y, he dashed, whi e lines in Fig. 10a a e gene ally ou o phase wi h he peaks in e ical374 18 Fig. 8. Ho izon al c oss sec ions o he simula ed e lec i i y (dBZ) in domain 1 (2.0 km g id spacing) on 1 Feb ua y, 2020 a 1600 UTC showing he sou he n low p essu e sys em wi h embedded mul i-bands. Panels (a), (b) and (c) show he 0 - 3 km, 4 - 7 km and 7 - 10 km heigh a e aged ields, espec i ely. The black, dashed ci cles highligh he loca ions o se e al bands iden i ied in panel (a). These ci cles a e copied on o he panels in (b) and (c). The black line in all panels deno es he along- ack axis (no mal o he phase lines) o mul i-bands obse ed in panels (b) and (c). 334 335 336 337 338 339 eloci y shown in Fig. 10b, which indica es ha he s e ching o p e-exis ing o ici y is no he375 p ima y d i e o he o ici y oscilla ions.376 Figu e 11a highligh s a se ies o o ici y oscilla ions in he mid-le els (3 - 6 km heigh a e age)377 a he same ime as Fig. 10 ha ha e some simila i ies o he uppe -le el ea u es, such as he378 s ong anomaly a ∼75 km along- ack. Howe e , he o ici y oscilla ions in he mid-le els379 a e subs an ially weake in magni ude and no as well-de ined as hose in he uppe le els. The380 co esponding e ical eloci y ield (Fig. 11b) con inues o be ou o phase wi h he o ici y381 19 Fig. 9. Ho izon al c oss sec ions o he simula ed absolu e e ical o ici y (s−1) in domain 1 (2.0 km g id spacing) showing he sou he n low p essu e sys em wi h embedded mul i-bands. Panels (a), (b) and (c) show he 0 - 3 km, 4 - 7 km and 7 - 10 km heigh a e aged ields, espec i ely. The black line in all panels deno es he along- ack axis (no mal o he phase lines) o mul i-bands shown in Figu e ??. 340 341 342 343 wi h he excep ion o he s ong anomaly a ∼75 km along- ack, which is a band o con ec ion382 close o he sys em cen e . While he e a e se e al bands o e ical eloci y be ween 50 - 250383 km along- ack, hese bands do no seem o be as o ganized in he ac oss- ack di ec ion as hose384 shown in he uppe -le els (Fig. 10b).385 Figu e 12 shows e ical c oss sec ions o he mul i-bands a e aged be ween ±50 km ac oss-386 ack. Dis inc oscilla ions in o ici y (Fig. 12a) a e isible in he 6 - 10 km laye ha ex end387 om ∼50 - 300 km along ack. Below 6 km, he o ici y pa e n is no ob ious, al hough some388 oscilla ions ha a e il ed down he ack wi h heigh a e isible in he ∼2 - 6 km laye om ∼389 125 - 200 km along- ack. The e ical eloci y ield (Fig. 12b) shows simila oscilla ions o he390 20 Fig. 10. Ho izon al c oss sec ions, a e aged o e he 6 - 10 km heigh ange, o simula ed da a on he ack- ela i e g id e ealing p ope ies o mul i-bands. Panels (a) and (b) show he simula ed absolu e o ici y (s−1) and e ical eloci y (m s−1), espec i ely. The whi e, dashed lines highligh he posi i e pe u ba ions in o ici y, which a e copied on o he e ical eloci y plo . 366 367 368 369 o ici y in he 6 - 10 km laye , mos appa en in he nega i e pe u ba ions, bu he e ical eloci y391 and o ici y a e mos ly ou o phase as p e iously discussed. The main excep ion is a 75 km392 along- ack, whe e a s ong, posi i e e ical eloci y anomaly is colloca ed wi h a s ong, posi i e393 o ici y anomaly. This ea u e is pa o he main o a ing con ec i e band loca ed jus no heas 394 o he sys em cen e highligh ed in Fig. 8 and Fig. 9. Posi i e co ela ions in o ici y and e ical395 eloci y a e also appa en a ∼125 km along- ack below 6 km heigh .396 Wha a e he dynamics go e ning he oscilla ions obse ed in he uppe -le els o he sys em?397 One migh suspec he pe u ba ions a e g a i y wa es and we e alua e ha po en ial he e. The398 o al phase speed o he bands is calcula ed by using he wo-minu e model ou pu o ack he lines399 o cons an phase using he ac oss- ack eloci y ield (shown la e in Fig. 14b) a e aged o e ±400 21 Fig. 11. The same as in Figu e 10, only o da a a e aged o e he 3 - 6 km heigh ange. No e he colo ba has been expanded on he e ical eloci y panel compa ed o Fig. 10b. 50 km ac oss- ack and 6 - 10 km heigh . T acking hese phase lines esul ed in mean o al phase401 speeds o 26.67 m s−1. In o de o es ima e he possibili y o in insic p opaga ion, he mean low402 in he di ec ion o wa e p opaga ion mus be emo ed om he o al phase speed. The e ical403 c oss sec ions o he bands (Fig. 12) clea ly show ha hey a e p esen wi hin he 6 - 10 km laye .404 Sensi i i y es s in he ac oss- ack a e aging in e al we e pe o med o 100 km, 200 km and405 300 km ack- ela i e g ids. These es s showed ha he mean along- ack eloci y, a e aged o e 406 he app op ia e ac oss- ack dis ance and he 6 - 10 km laye we e ∼27.0 ±0.5 m s−1. Thus, he407 measu ed in insic phase speed o he mul i-bands in his laye is ∼0ms−1.408 Fo comple eness, he heo e ical g a i y wa e speed o his en i onmen is also calcula ed. The409 dispe sion ela ion o in e nal g a i y wa es is410 22 Fig. 12. Ve ical c oss sec ions o simula ed da a on 1 Feb ua y, 2020 a 1530 UTC, a e aged o e he ±50 km ac oss- ack ange showing (a) absolu e o ici y and (b) e ical eloci y. (𝜔−¯𝑢𝑘)2𝑘2+𝑚2−𝑁2𝑘2=0 (7) whe e 𝜔is he angula equency, ¯𝑢is he la ge-scale, a e aged ho izon al windspeed in he411 di ec ion o wa e p opaga ion, 𝑁is he B un –V¨ ais¨ al¨ a equency, 𝑘is he ho izon al wa enumbe ,412 and 𝑚is he e ical wa enumbe .413 Rea anging Eq. (7) o he in insic ho izon al phase speed on he le -hand-side yields414 𝜔/𝑘−¯𝑢=𝑁/√︁𝑘2+𝑚2.(8) The igh -hand-side o Eq. (8) is e alua ed using he da a om he simula ion. No e ha he415 buoyancy equency was es ima ed o be 50 imes la ge han he Co iolis equency and hus, he416 po en ial g a i y wa es a e no signi ican ly a ec ed by he Ea h’s o a ion. Taking an a e age417 o da a o e he sys em a wo di e en ime pe iods and in he 6 - 10 km laye p oduced alues418 o 𝑁a ound 10−2s−1. The dominan ho izon al wa eleng h in he model is ∼30 km and he419 23 Fig. 13. Ve ical c oss sec ions o simula ed da a on 1 Feb ua y, 2020 a 1530 UTC, a e aged o e he ±50 km ac oss- ack ange showing (a) pe u ba ion e ical eloci y (m s−1) and (b) pe u ba ion e lec i i y (dBZ). The black a ows in panel (a) deno e ea u es discussed in he ex . 427 428 429 e ical wa eleng h is aken o be wice he dep h o he pe u ba ions, which equa es o 8 km.420 En e ing hese numbe s p oduces a ho izon al phase speed o 12.30 m s−1, which is ep esen a i e421 o he bands p opaga ing down he ack- ela i e g id o he no heas o he sys em cen e . I is422 clea om his calcula ion ha he measu ed in insic phase speeds ha e a la ge misma ch wi h423 he heo e ical in insic phase speeds o in e nal g a i y wa es, e en wi h signi ican unce ain y424 bounds in a ious pa ame e s. Thus, he obse ed mul i-bands in he uppe -le els canno be g a i y425 wa es.426 Howe e , he e a e mo e sub le oscilla ions in se e al a iables in he low o middle le els o 430 he sys em ha a e e ealed h ough examining pe u ba ion ields. In his analysis, pe u ba ions431 a e de ined as de ia ions o he o al a iables om he 50 km along- ack il e ed a iables. The432 pe u ba ion e ical eloci y a 1530 UTC (Fig. 13a) shows he p ominen mul i-bands in he 6433 24 - 10 km laye along wi h weake pe u ba ions in he ∼0 - 6 km laye deno ed by black a ows.434 The weake oscilla ions exhibi ampli udes app oxima ely h ee o ou imes lowe han hose in435 he uppe laye , making hem clea ly dis inguishable despi e he simila wa eleng hs o ∼30 km.436 Anima ions o e ical eloci y (no shown) appea o show hese wa es emana ing om he deep437 con ec ion p esen a ∼75 km along- ack. The pe u ba ion e lec i i y (Fig. 13b) is la gely438 consis en wi h he e ical eloci y excep he low o middle le el oscilla ions a e abou en imes439 lowe in ampli ude han he uppe laye wa es. This is why he mul i-bands a e no as isible in440 he low-le el p ecipi a ion ield as shown by he e lec i i y (Fig. 8a).441 The low o middle le el wa es (0 - 6 km laye ) we e acked in he pe u ba ion e ical eloci y442 ield wi h he wo minu e model ou pu and he o al phase speeds we e measu ed. This p ocedu e443 p oduced o al wa e phase speeds o ∼30 ±3 m/s. The along- ack eloci y was a e aged ac oss-444 ack (±50 km), along- ack (0 - 342 km) and heigh (0 - 6 km) o ep esen he mean low mo ing445 he wa es, which esul ed in alues o ∼14 ±0.5 m/s. Thus, he measu ed in insic phase speeds o 446 he wa es we e ∼16 m/s. The heo e ical in insic phase speeds o g a i y wa es we e compu ed447 using he same inpu s as be o e wi h he excep ion o a 12 km e ical wa eleng h ( wo imes he448 6 km dep h o he e ical eloci y pe u ba ions). These inpu s p oduced alues o 17.73 m/s,449 which is close o he measu ed in insic phase speed o ∼16 m/s. Thus, hese low o middle le el450 wa es can be iden i ied as in e nal g a i y wa es ha a e being gene a ed by he con ec i e ac i i y451 close o he sys em cen e .452 I he dominan , uppe -le el mul i-bands a e no g a i y wa es, wha is d i ing hei dynamics?456 Figu e 14 shows e ical c oss sec ions o he along- ack eloci y and ac oss- ack eloci y. The457 along- ack eloci y (Fig. 14a) shows a je cen e ed a ∼11 km heigh wi h la ge alues o e ical458 shea (maximum alues o 0.01 s−1) loca ed be ween 8 - 10 km heigh and mos no ably be ween459 50 - 225 km along- ack. The loca ions o he la ge e ical wind shea alues ma ch well wi h460 he loca ions o he wa e mo ions associa ed wi h he mul i-bands. Fo example, he ac oss- ack461 eloci y (Fig. 14b) e eals clea wa e mo ions in he 6 - 10 km laye and be ween 50 - 250 km462 along- ack. The e a e pe haps ex ensions o he wa e mo ions down o ∼4 - 5 km in some egions,463 bu o e all, he oscilla ions become less de ec able below 6 km heigh .464 The along- ack and ac oss- ack eloci y ields shown in Fig. 14 ma ch well wi h he EXRAD465 obse a ions shown in Fig. 4, despi e he di e ences in spa ial/ empo al co e age (also no e he466 25 Acknowledgmen s. The IMPACTS ield campaign and he wo k o au ho Guimond a e suppo ed590 by he NASA Ea h Ven u e Subo bi al (EVS) p og am and NASA g an 80NSSC24K0302. Au ho 591 Guimond hanks he IMPACTS eam membe s o o ganizing he campaign, p o iding o ecas ing592 suppo , coo dina ing he ligh s and collec ing he da a. Many hanks a e gi en o D . Paul Reaso 593 o mul iple, leng hy discussions on he wo k ha imp o ed his pape conside ably. Feedback on594 he g a i y wa e analysis om D . S e en Koch is also g ea ly app ecia ed.595 Da a a ailabili y s a emen . The da a se s needed o ep oduce he co e esul s o his pape can596 be ound a he ollowing link: h ps://doi.o g/10.5281/zenodo.17253490597 Re e ences598 Ba nes, H. C., J. P. Zag odnik, L. A. McMu die, A. K. Rowe, and J. Houze, R. A., 2018:599 Kel in–helmhol z wa es in p ecipi a ing midla i ude cyclones. Jou nal o he A mosphe ic Sci-600 ences,75 (8), 2763–2785, h ps://doi.o g/10.1175/JAS-D-17-0365.1.601 Bosa , L. F., W. E. B acken, and A. Seimon, 1998: A s udy o cyclone mesoscale s uc u e wi h em-602 phasis on a la ge-ampli ude ine ia–g a i y wa e. Mon hly Wea he Re iew,126 (6), 1497–1527,603 h ps://doi.o g/10.1175/1520-0493(1998)126⟨1497:ASOCMS⟩2.0.CO;2, URL h ps://jou nals.604 ame soc.o g/ iew/jou nals/mw e/126/6/1520-0493 1998 126 1497 asocms 2.0.co 2.xml.605 F ank, W. M., and E. A. Ri chie, 2001: E ec s o e ical wind shea on he in ensi y and s uc u e606 o nume ically simula ed hu icanes. Mon hly Wea he Re iew,129, 2249–2269.607 Galpe in, B., S. Suko iansky, and P. S. Ande son, 2007: On he c i ical icha dson numbe in s ably608 s a i ied u bulence. A mosphe ic Science Le e s,8 (3), 65–69, h ps://doi.o g/10.1002/asl.153,609 URL h ps://doi.o g/10.1002/asl.153.610 Gane is, S. A., B. A. Colle, S. E. Yu e , and N. P. Hoban, 2018: En i onmen al condi ions611 associa ed wi h obse ed snowband s uc u es wi hin no heas u.s. win e s o ms. Mon. Wea.612 Re .,146, 3675–3690, h ps://doi.o g/h ps://doi.o g/10.1175/MWR-D-18-0054.1.613 Guimond, S., 2023: E -2 x-band ada (ex ad) 3d winds impac s. NASA Global Hyd ome eo ol-614 ogy Resou ce Cen e DAAC, Hun s ille, Alabama, U.S.A., URL h p://dx.doi.o g/10.5067/615 IMPACTS/EXRAD/DATA201, da ase a ailable online, h ps://doi.o g/10.5067/IMPACTS/616 EXRAD/DATA201.617 32 Guimond, S. R., 2007: A diagnos ic s udy o he e ec s o ough in e ac ions on opical cyclone618 qp . Mas e ’s hesis, Flo ida S a e Uni e si y, Tallahassee, FL.619 Guimond, S. R., J. M. Reisne , S. Ma as, and F. X. Gi aldo, 2016: The impac s o d y dynamic620 co es on asymme ic hu icane in ensi ica ion. Jou nal o he A mosphe ic Sciences,73 (12),621 4661–4684, h ps://doi.o g/10.1175/JAS-D-16-0055.1.622 Guimond, S. R., S. S oka, and D. P o zko, 2018: A la ge eddy simula ion o hu icane in en-623 si ica ion. 33 d Con e ence on Hu icanes and T opical Me eo ology, Ame ican Me eo olog-624 ical Socie y, 17, URL h ps://science.gs c.nasa.go /sed/con en /uploadFiles/publica ion iles/625 la ge eddy 33hu guimond.pd , a ailable online.626 Guimond, S. R., L. Tian, G. M. Heyms ield, and S. J. F asie , 2014: Wind Re ie al Algo-627 i hms o he IWRAP and HIWRAP Ai bo ne Dopple Rada s wi h Applica ions o Hu i-628 canes. Jou nal o A mosphe ic and Oceanic Technology,31 (6), 1189–1215, h ps://doi.o g/629 10.1175/JTECH-D-13-00140.1.630 Hanley, D., J. Molina i, and D. Keyse , 2001: A composi e s udy o he in e ac ions be ween631 opical cyclones and uppe - oposphe ic oughs. Mon hly Wea he Re iew,129 (10), 2570–632 2584, h ps://doi.o g/10.1175/1520-0493(2001)129⟨2570:ACSOTI⟩2.0.CO;2.633 Heyms ield, G. M., L. Tian, A. J. Heyms ield, L. Li, and S. Guimond, 2009: Cha ac e is ics o deep634 opical and sub opical con ec ion om nadi - iewing high-al i ude ai bo ne dopple ada .635 Jou nal o he A mosphe ic Sciences,67, 285–308, h ps://doi.o g/10.1175/2009JAS3132.1,636 submi ed Ma ch 2009, e ised June 2009.637 House , J. L., and H. B. Blues ein, 2011: Pola ime ic dopple ada obse a ions o kel in–638 helmhol z wa es in a win e s o m. Jou nal o he A mosphe ic Sciences,68 (7), 1676–1702,639 h ps://doi.o g/10.1175/2011JAS3566.1.640 Koch, S. E., and L. M. Siedla z, 1999: Mesoscale g a i y wa es and hei en i onmen in he641 cen al uni ed s a es du ing s o m- es . Mon. Wea. Re .,127, 2854– 2879, h ps://doi.o g/h ps:642 //doi.o g/10.1175/1520-0493(1999)127%3C2854:MGWATE%3E2.0.CO;2.643 33 Lindzen, R., and K. Tung, 1976: Banded con ec i e ac i i y and duc ed g a i y wa es. Mon. Wea.644 Re .,104, 1602– 1617, h ps://doi.o g/10.1175/1520-0493(1976)104⟨1602:BCAADG⟩2.0.CO;645 2.646 Liu, Y., X. San Liang, and R. H. Weisbe g, 2007: Rec i ica ion o he bias in he wa ele powe 647 spec um. Jou nal o A mosphe ic and Oceanic Technology,24 (12), 2093–2102, h ps://doi.o g/648 10.1175/2007JTECHO511.1.649 Lon a , M., F. D. Ma ks, and S. S. Chen, 2004: P ecipi a ion dis ibu ion in opical cyclones650 using he opical ain all measu ing mission ( mm) mic owa e image : A global pe spec i e.651 Mon hly Wea he Re iew,132, 1645–1660.652 McMu die, L., and Coau ho s, 2022: Chasing snows o ms: The in es iga ion o mic ophysics and653 p ecipi a ion o a lan ic coas - h ea ening snows o ms (impac s) campaign. Bull. Ame . Me eo .654 Soc.,103, E1243–E1269, h ps://doi.o g/h ps://doi.o g/10.1175/BAMS-D-20-0246.1.655 Medina, S., and J. Houze, R. A., 2016: Kel in-helmhol z wa es in ex a opical cyclones passing656 o e moun ain anges. Qua e ly Jou nal o he Royal Me eo ological Socie y,142 (697), 1311–657 1319, h ps://doi.o g/10.1002/qj.2755.658 Me ill, R. T., 1988: Cha ac e is ics o he uppe oposphe ic en i onmen al low a ound659 hu icanes. Jou nal o he A mosphe ic Sciences,45, 1665–1677, h ps://doi.o g/10.1175/660 1520-0469(1988)045⟨1665:COTUTE⟩2.0.CO;2.661 Molina i, J., and D. Volla o, 1989: Ex e nal in luences on hu icane in ensi y. pa i: Ou low662 laye eddy angula momen um luxes. Jou nal o he A mosphe ic Sciences,46 (8), 1093–1105,663 h ps://doi.o g/10.1175/1520-0469(1989)046⟨1093:EIOHIP⟩2.0.CO;2.664 Nolan, D. S., and M. J. Onde linde, 2022: The ep esen a ion o spi al g a i y wa es in a mesoscale665 model wi h inc easing ho izon al and e ical esolu ion. Jou nal o Ad ances in Modeling666 Ea h Sys ems,14 (10), e2022MS003 067, h ps://doi.o g/10.1029/2022MS003067, URL h ps:667 //doi.o g/10.1029/2022MS003067.668 P o zko, D. E., S. R. Guimond, C. R. Jackson, J. W. Sapp, Z. Jelenak, and P. S. Chang, 2023:669 Documen ing cohe en u bulen s uc u es in he bounda y laye o in ense hu icanes h ough670 34 wa ele analysis on iw ap and sa da a. IEEE T ansac ions on Geoscience and Remo e Sensing,671 61, 1–16, h ps://doi.o g/10.1109/TGRS.2023.3305998.672 Ruppe , J. H., S. E. Koch, X. Chen, Y. Du, A. Seimon, Y. Q. Sun, J. Wei, and L. F.673 Bosa , 2022: Mesoscale g a i y wa es and midla i ude wea he : A ibu e o uqing674 zhang. Bulle in o he Ame ican Me eo ological Socie y,103 (1), E129–E156, h ps://doi.o g/675 10.1175/BAMS-D-20-0005.1.676 Schul z, D. M., and P. N. Schumache , 1999: The use and misuse o condi ional symme ic ins abil-677 i y. Mon. Wea. Re .,127, 2709–2732, h ps://doi.o g/h ps://doi.o g/10.1175/1520-0493(1999)678 127,2709:TUAMOC.2.0.CO;2.679 Sel ze , M., R. Passa elli, and K. Emanuel, 1985: The possible ole o symme ic ins abili y680 in he o ma ion o p ecipi a ion bands. J. A mos. Sci.,42, 2207–2219, h ps://doi.o g/h ps:681 //doi.o g/10.1175/1520-0469(1985)042%3C2207:TPROSI%3E2.0.CO;2.682 To ence, C., and G. P. Compo, 1998: A p ac ical guide o wa ele analysis. Bulle in o he Ame i-683 can Me eo ological Socie y,79 (1), 61–78, h ps://doi.o g/10.1175/1520-0477(1998)079⟨0061:684 APGTWA⟩2.0.CO;2.685 Uccellini, L. W., and S. E. Koch, 1987: The synop ic se ing and possible ene gy sou ces o 686 mesoscale wa e dis u bances. Mon. Wea. Re .,115, 721–729, h ps://doi.o g/h ps://doi.o g/10.687 1175/1520-0493(1987)115%3C0721:TSSAPE%3E2.0.CO;2.688 Ulb ich, C. W., and P. B. Chilson, 1994: E ec s o a ia ions in p ecipi a ion size dis ibu ion689 and allspeed law pa ame e s on ela ions be ween mean dopple allspeed and e lec i i y690 ac o . Jou nal o A mosphe ic and Oceanic Technology,11, 1656–1663, h ps://doi.o g/10.691 1175/1520-0426(1994)011⟨1656:EOVIPS⟩2.0.CO;2.692 Walke McLinden, M. L. W., A. M. Lo us, L. Li, and G. M. Heyms ield, 2021: Ap-693 plica ion o nonuni o m beam illing (nub ) dopple eloci y e o co ec ion on ai bo ne694 ada measu emen s. IEEE Geoscience and Remo e Sensing Le e s,19, 1–5, h ps://doi.o g/695 10.1109/LGRS.2021.3120647.696 Xu, Q., 1989: F on al ci cula ions in he p esence o small iscous mois symme ic s abili y and697 weak o cing. Qua . J. Roy. Me eo . Soc.,115, 1325–1353.698 35 Zhang, F., C. Da is, M. Kaplan, and S. Koch, 2001: Wa ele analysis and he go e ning dynamics699 o a la ge-ampli ude mesoscale g a i y-wa e e en along he eas coas o he uni ed s a es. Q.J.R.700 Me eo ol. Soc.,127, 2209–2245, h ps://doi.o g/10.1002/qj.49712757702.701 36