Etna (It aly) Lava flows from mul ti ple vents dur ing 22 Sep tem ber to 4 No vem ber . . . . . . . . . . . . . . . . 2 San Miguel (El Sal va dor) Rest less ness per sists dur ing 2005-6; heavy trop i cal rains trig ger lahars . . . . . . 3 Ubinas (P?ru) New re port ing re veals ashfalls, large bal lis tic blocks, lahar haz ards, and evac u a tions . . . . . 7 Kilauea (USA) PKK lava tube ac tive Au gust-No vem ber 2006; 10 Oc to ber col lapse pit at Pu`u `O`o . . . . . 11 Home Reef (Tonga) More de tails on the new is land and drift ing pum ice rafts, in clud ing sat el lite data . . . . 12 Rabaul (Pa pua New Guinea) Erup tions of vary ing in ten sity at Tavurvur; ex plo sion on 14 No vem ber 2006 . . 15 Likuranga (Pa pua New Guinea) Tragic CO2-gas ac ci dent in open hole at in ac tive vol cano . . . . . . . . . . 17 Mi chael (S Sand wich Is lands) Clear IR sat el lite view on 28 Oc to ber 2006 sug gests lava in side the crater . . 19 Ed i tors: Rick Wunderman, Ed ward Venzke, Sally Kuhn Sennert, and Catherine Galley Vol un teer Staff: Rob ert An drews, Ve ron ica Bemis, Hugh Replogle, Jerome Hudis, Jackie Gluck, Ste phen Bentley, Jeremy Book binder, and Antonia Book binder Global Vol can ism Program ? Na tional Mu seum of Natu ral His tory, Room E-421, PO Box 37012 ? Wash ing ton, DC 20013-7012 ? USA Tele phone: (202) 633-1800 ? Fax: (202) 357- 2476 ? Email: gvn@si.edu ? URL: http://www.vol cano.si.edu/ The text of the Bul le tin is also dis trib uted through the Vol cano Listserv (vol cano@asu.edu). Bul le tin of the Global Vol can ism Net work Vol ume 31, Num ber 10, October 2006 Etna It aly 37.734?N, 15.004?E; sum mit elev. 3,350 m All times are lo cal (= UTC + 1 hours) The fol low ing Etna re port from Sonia Calvari and Boris Behncke is based on daily ob ser va tions by nu mer ous staff mem bers of the I s t i tuto Nazionale d i Geofis ica e Vulcanologia (INGV). As pre vi ously re ported here (BGVN 31:08), a 10-day-long erup tion vented from the base of the South east Crater (SEC) in mid-July 2006. Erup tive ac tiv ity then shifted to the crater?s sum mit vent dur ing 31 Au - gust-15 Sep tem ber, lead ing to lava over flows and re peated col lapse on the SEC cone (BGVN 31:08). This re port dis cusses the time pe riod 22 Sep tem ber to 4 No vem ber, an in ter val with mul ti ple ep i sodes of erup tive ac tiv ity (roughly eight in all, seven of which in volved a re - turn of ac tiv ity at the SEC sum mit). The ac tiv ity typ i cally in cluded lava flows and Strombolian erup tions. In gen eral, the erup tive ep i sodes be came in creas ingly brief and vig or - ous. Erup tions came from the SEC?s sum mit as well as from mul ti ple vents along frac tures on the SEC?s sides or ad ja cent to it that de vel oped dur ing the re port ing in ter val. As men tioned above, in gen eral, dur ing the re port ing in - ter val, the re nowned SEC sum mit area was only ep i sod i - cally ac tive. Since the SEC?s col lapse of Sep tem ber 2006, it has had a breached E wall. Dur ing this re port ing in ter val, lava flows es caped the crater through the breach to form nar row riv u lets down the steep up per SE flank. Ash from SEC fell on Catania on 30 Oc to ber. On 12 Oc to ber a fis sure opened at ~ 2,800-m el e va tion on the ESE base of the SEC cone, ~ 1 km from SEC?s sum - mit. Lava from this vent trav eled SE, and a map show ing the vents and pat tern of flows through 20 No vem ber in di - cated lava ex tend ing ~ 2 km from the 2,800-m vent (Behncke and Neri, 2006). The 2,800-m vent also sits along the path of some of the SEC lavas from the sum mit crater. By late No vem ber, a com plex flow field from both SEC sum mit and the 2,800-m vents lay on the SE side of the SEC. The field ex tended from the sum mit ~ 3 km, and its dis tal ends reached the W wall of the Valle del Bove. Two other im por tant vents be gan erupt ing in late Oc to - ber. One was on the SEC?s up per S flank. The other, at 3,050 m el e va tion, stood ~ 1 km SW of the SEC?s crater and at a spot ~ 0.5 km from the near est mar gin of Bocca Nuova?s crater. Al though lava emis sions from this vent at 3050-m el e va tion stopped, they later re started and by 20 No vem ber the vent had cre ated a large SW-trending field of lava flows roughly the size of those from the SEC sum mit and 2800-m vent. Erup tive be hav ior 22 Sep tem ber-4 No vem ber. Dur ing this time in ter val, the seven ep i sodes de ter mined by erup - tive ac tiv ity at the SEC oc curred as fol lows. The first ep i sode, which was five days long, started late 22 Sep tem ber from the sum mit of the SEC. Ac tiv ity dur ing the first two days was lim ited to mild Strombolian ex plo - sions, but lava be gan to over flow the SEC?s crater on 24 Sep tem ber, spill ing onto the cone?s SE flank. This ac tiv ity ceased some time on 27 Sep tem ber. The 2nd ep i sode be gan late the af ter noon of 3 Oc to ber with Strombolian ex plo sions from the SEC sum mit, which in creased in vigor dur ing the fol low ing hours. Late that eve ning lava be gan to spill down the SE side of the SEC cone ad ja cent to flows of the pre vi ous two ep i sodes. Fol - low ing a sharp de cline in tremor am pli tude on the af ter noon of 5 Oc to ber, the ac tiv ity ended some time be tween mid - night and the early morn ing of 6 Oc to ber. The 3rd erup tive ep i sode oc curred be tween the eve ning of 10 Oc to ber and the eve ning of the fol low ing day. The SEC?s sum mit pro duced vig or ous Strombolian ac tiv ity and lava again de scended the SEC cone?s SE flank. A sharp drop in tremor am pli tude on the af ter noon of 11 Oc to ber in - di cated the erup tions im mi nent ces sa tion. At the tail end of the 3rd ep i sode, a short erup tive fis - sure opened with vents at ~ 2,800-m el e va tion. Mon i tor ing cam eras fixed the start of this ac tiv ity at 2328 on 12 Oc to - ber. The SEC?s sum mit was quiet through out the fol low ing eight days, leav ing this burst to be con sid ered as ac tiv ity late in the 3rd ep i sode, rather than rep re sent ing the start of the 4th ep i sode in the SEC?s spo radic on-and-off behavior. Trending N90?E-N100?E, the new fis sure re sided on the ESE flank at the base of SEC, a spot also on the Valle del Bove?s W wall. For the first few days, lava was emit ted non-ex plo sively, qui etly spread ing in the up per Valle del Bove and ad vanc ing a few hun dred me ters downslope. Mild spat ter ing on 17 Oc to ber re sulted in the growth of three hor ni tos on the up per end of the erup tive fis sure. Summit SEC ac tiv ity marked the 4th ep i sode since 22 Sep tem ber. As lava ef fu sion con tin ued from the fis sure vents at 2,800 m el e va tion, the SEC started a pow er ful erup tion at 0600 on 20 Oc to ber. Ac com pa nied by a rapid in crease in tremor am pli tude, vig or ous Strombolian erup - tions occurred in the cen tral por tion of the SEC?s sum mit. A vent near the E rim of the SEC?s crater, in the notch cre - ated by the col lapse events of early Sep tem ber, produced large ex plo sions ev ery few min utes and quickly built a new pyroclastic cone. Lava once more flowed down the SEC?s SE side, stop ping N of the 2,800-m fis sure. At that fis sure vent, lava emis sion con tin ued but ap peared re duced com - pared to the pre vi ous days. The SEC ceased is su ing lava the same day it be gan, 20 Oc to ber. The 5th ep i sode in volv ing the SEC was pre ceded on 22 Oc to ber with a few iso lated bursts of ash from the SEC. The ep i sode be gan with strong ac tiv ity at 0700 the next day, when the SEC?s sum mit gen er a ted vig or ous Strombolian dis charges and pul sat ing lava foun tains from two vents. The new pyroclastic cone grew rap idly. Lava spilled down the ESE flank of the cone, to the N of the flows formed in the pre vi ous ep i sodes. Co in ci dent with the above erup tions, INGV re search ers noted an in creased lava emis sion from the 2,800-m vents. This led to sev eral lava over flows (in an area ad ja cent to the hor ni tos formed 17 Oc to ber). Al though Strombolian ac tiv ity and fountaining at the SEC di min ished on the af ter noon of 23 Oc to ber, strong ash emis sions be gan at around 1700, pro duc ing an ESE-drift ing plume. Pul sat ing ash emis sions and oc ca sional bursts of glow ing tephra con tin ued and, at about 1750, the SEC cone?s S flank frac tured. Lava es caped from the frac ture?s lower end, form ing two small lobes. The lon ger lobe reached the base of the cone and then trav eled SE, ul ti - mately to reach ~ 1 km from their source at the new fis sure. The smaller lobe took a path down the cone slightly to the W, but halted be fore reach ing the base of the cone. The new fis sure?s lava sup ply di min ished early on 24 Oc to ber, stop - ping around noon. 2 Etna Smithsonian Institution ? Bulletin of the Global Co in ci dent with the above events, ef fu sive ac tiv ity con - tin ued with out sig nif i cant vari a tions at the 2,800-m vents. The far thest flow fronts reached an el e va tion of ~ 2,000 m to the NW of Monte Centenari, and ex tended ~ 2.5 km from their source. Field ob ser va tions made on 24 Oc to ber re vealed that part of the new pyroclastic cone had sub sided and a new col lapse pit, ~ 50 m wide, had opened on the SE flank of the SEC cone, roughly in the cen ter of the largely oblit er ated col lapse pit of 2004-2005. The 6th ep i sode of SEC ac tiv ity be gan in the late af ter - noon on 25 Oc to ber. Ini tially there was an in crease in tremor am pli tude, as well as both ash emis sions and weak Strombolian ac tiv ity from the SEC?s sum mit. Both the tremor and Strombolian dis charges de creased late that eve - ning, but at 0054 on 26 No vem ber lava was emit ted from a new fis sure. This fis sure, on the SEC cone?s SSE flank, was ac tive only for a few hours and pro duced a very small lava flow. As has of ten been the case dur ing the re port ing in ter - val, the 2800-m vents con tin ued to dis charge lava toward the Valle del Bove. What was to later be come an other im por tant ef fu sive vent opened at 0231 on 26 Oc to ber. The vent de vel oped at ~ 3,050 m el e va tion in an area ~ 700 m S of the cen ter of Bocca Nuova?s crater and ~ 500 m SW of the cen ter of SEC?s crater. This spot sits at the S base of the cen tral sum - mit cone be low the Bocca Nuova, and ~ 700 m to the W of the fis sure that had erupted 2 hours earlier. Field work car ried out on 26 Oc to ber by INGV re search - ers re vealed that the vent at 3,050 m el e va tion had formed at the south ern end of a frac ture field. That field ex tended across the SE flank of Et na?s cen tral sum mit cone to the W flank of the SEC cone. Lava ex trud ing at the 3050-m vent poured out at a de creas ing rate be fore a pause be gan on the eve ning of 26 October. The 7th ep i sode, 27 Oc to ber and into early No vem ber, was first as so ci ated with a new in crease in tremor am pli - tude and cor re spond ing SEC ash emis sions on the af ter - noon of the 27th. These emis sions were fol lowed at 0206 on the 28th by the re ac ti va tion of the vent at 3050-m el e va - tion. Ash emis sions and Strombolian ac tiv ity oc curred from the SEC be tween 0830 and 1100, but no lava over flows were pro duced. On the eve ning of 28 Oc to ber, both ef fu - sive vents at 3,050 and 2,800 m were active. 29 Oc to ber ash emis sions from the SEC be came more vig or ous dur ing the early morn ing of the 30th and fine ash fell over in hab ited ar eas to the S, in clud ing Catania (27 km from the SEC). In ter mit tent bursts of glow ing tephra were re corded by INGV-CT sur veil lance cam eras, al though later anal y sis re vealed that most of the tephra was lithic rather than ju ve nile. Ash emis sions grad u ally di min ished and ceased at around 0800 on 29 October. Ash was again emit ted from the SEC shortly be fore 1300 on 31 Oc to ber, and in mi nor quan ti ties at least once per day through 5 No vem ber. No in can des cent ejec tions oc curred from this crater af ter 28 Oc to ber un til the eve ning of 4 No vem ber (dur ing 1830-2005) when weak Strombolian ex plo sions were re corded by the INGV-CT surveillance cameras. The vent at 3050-m el e va tion con tin ued to emit lava on 29 Oc to ber. The ef fu sion rate was es ti mated as 1 to 5 m 3 per sec ond. Emit ted lava de scended SW to ~ 2,400 m elevation. Lava also con tin ued to flow from the 2,800-m vents on the 29th, but the as so ci ated lava flow front ad vanc ing from these vents had moved lit tle since 24 Oc to ber. Lava con tin - ued to flow from both vents dur ing the first days of No vem ber, but the ef fu sion rate had clearly dropped by the 3rd when ac tive flows had re treated upslope from the dis tal fronts. Sim i larly, a he li cop ter over flight on the morn ing of 5 No vem ber dis closed ac tively flow ing lava con fined to the up per most parts of the lava flow fields. Geo logic Sum mary. Mount Etna, tow er ing above Catania, Sic ily?s sec ond larg est city, has one of the world?s lon gest doc u mented re cords of his tor i cal vol ca nism, dat ing back to 1500 BC. His tor i cal lava flows of ba saltic com po si - tion cover much of the sur face of this mas sive vol cano, whose ed i fice is the high est and most vo lu mi nous in It aly. The Mongibello stratovolcano, trun cated by sev eral small cal de ras, was con structed dur ing the late Pleis to cene and Ho lo cene over an older shield vol cano. The most prom i nent mor pho log i cal fea ture of Etna is the Valle del Bove, a 5 x 10 km horse shoe-shaped cal dera open to the E. Two styles of erup tive ac tiv ity typ i cally oc cur at Etna. Per sis tent ex - plo sive erup tions, some times with mi nor lava emis sions, take place from one or more of the three prom i nent sum mit crat ers, the Cen tral Crater, NE Crater, and SE Crater (the lat ter formed in 1978). Flank vents, typ i cally with higher ef fu sion rates, are less fre quently ac tive and orig i nate from fis sures that open pro gres sively down ward from near the sum mit (usu ally ac com pa nied by strombolian erup tions at the up per end). Cin der cones are com monly con structed over the vents of lower-flank lava flows. Lava flows ex tend to the foot of the volcano on all sides and have reached the sea over a broad area on the SE flank. Ref er ences: Behncke, B., and Neri, M., 2006, Mappa delle colate laviche aggiornata al 20 Novembre 2006 (PDF file on the INGV website). In for ma tion Con tacts: Sonia Calvari and Boris Behncke, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Sezione di Catania, Pi azza Roma 2, 95123 Catania, It aly (URL: http://www.ct.ingv.it/). San Miguel El Sal va dor 13.434?N, 88.269?W; sum mit elev. 2,130 m All times are lo cal (= UTC - 6 hours) Ac cord ing to El Sal va dor?s Servicio Nacional de Estudios Territoriales (SNET) ac tiv ity lev els at San Miguel have gen er ally re mained sim i lar to those dur ing Jan u ary 2002 when a mi nor plume rose above the sum mit crater (BGVN 27:02). The vol cano?s vigor con tin ued into at least Oc to ber 2006 at a level slightly at or above the base line of normal activity. Re cent pub li ca tions have dis cussed the vol cano and its lahar-haz ard po ten tial (Escobar, 2003; Chesner and oth ers, 2003; Ma jor and oth ers, 2001). Fig ures 1 and 2 are taken from the lat ter publication. In Jan u ary 2005 ob serv ers saw new fumaroles as well as small land slides on the N and SW wall of the crater. The ac - cu mu la tion of mass-wasted ma te rial in the crater led to a rise in the el e va tion of the crater floor. Dur ing Feb ru ary 2005, weak fumaroles and small rock land slides per sisted in the cen tral crater. Dig i tal sen sors in - Volcanism Network, Volume 31, Number 10, October 2006 San Miguel 3 stalled there re corded fumarolic tem per a tures in real time. On the outer por tions of the cone the ter rain is steeply slop - ing and con tains prom i nent gul lies (figure 3). The SNET re ports for March and April 2005 noted that the crater was struc tur ally weak due to the fumarolic ac tiv - ity, on go ing rock al ter ation, oc ca sional land slides, and frac - tures on the west ern pla teau. Microseismicity had in - creased; but it did not ex ceed typ i cal base-line lev els. Work ers at the Santa Isabel farm (finca) noted N-flank lahars af ter heavy rains dur ing March. The N flank con tains abun dant fine-grained vol ca nic de pos its of the sort eas ily swept away dur ing times of heavy rain. In tense rains dur ing May 2005 were as so ci ated with trop i cal storm Adrian (over an un stated in ter val the me te o - ro log i cal sta tion near the vol cano, San Miguel UES, re - corded 428 mm of rain fall). As a re sult of the de luge, fumarolic ac tiv ity from the crater in creased. The crater walls re mained in tact, but eroded ma te rial pre vi ously de - pos ited in the cen tral crater that was poorly con sol i dated had to some de gree sta bi lized. Sub stan tial fur ther com pac - tion, set tling, or col lapse in the cen tral crater seemed to have ceased by July 2005. Dur ing Au gust 2005 the cri sis at volcan Santa Ana forestalled visits to San Miguel. A spike in seis mic ac tiv ity oc - curred dur ing Au gust 2005, with 7,048 long-pe riod earth quakes, com pared to July 2005, with 2,239 long-pe riod earth quakes. SNET re ports noted that based on mon i tor ing, San Miguel gen er ally re mained within its base-line of nor mal be hav ior dur ing the re - port ing in ter val. Fig ure 4 shows a his to gram of long-pe riod and vol - cano-tec tonic events from the SNET re ports for the in ter val September 2005-June 2006. On 14 Sep tem ber 2005 a vis it - ing group (OIKOS- Soliradaridad Internacional) made a trek to the sum mit and vid eo taped the scene there. SNET said the video dis - closed a lack of s ig nif i cant changes in the crater; how ever, they saw de bris-flow de pos its in sum mit drainages on the vol - cano?s out board flanks. The vis i - tors de scribed both sounds of de - gas sing and mod er ately in tense odors of H2S. Dur ing the course of Sep tem ber the seis mic sys tem re corded several minutes of tremor. 4 San Miguel Smithsonian Institution ? Bulletin of the Global Fig ure 1. In dex map in di cat ing El Sal va dor?s vol ca nic front and the lo ca tion of volcan San Miguel. Ma jor cit ies are also shown (cir cles). From Ma jor and oth ers (2001). Fig ure 2. The lahar haz ard map of San Miguel de picts likely lahar paths, which are shown as col ored or shaded ar eas. The con tour in ter val is 20 m; the ur ban cen ter ~ 11 km NE of the sum mit is San Miguel. From Ma jor and oth ers (2001); their plate 1, crop ped, highly re duced, and ex clud ing the key. Fig ure 3. A photo of San Miguel taken from the N on 22 Feb ru ary 2005 show ing the steep sides of the up per slopes and the in cised drainages there. Al though much of the area on the vol cano is ru ral, haz ards could eas ily af fect 40,000 res i dents liv ing nearby. Cour tesy of Servicio Nacional de Estudios Territoriales (SNET). The Oc to ber 2005 SNET re port noted that work ers at the plan ta tion Santa Isabel noted N-slope lahars as so ci ated with rain fall. The lahars were also de scribed as small de bris flows; they de scended from the high-el e va tion head wa ter ar eas, which are steep sided and nar row. The No vem ber re - port com mented about the quan tity of de bris-flow ma te rial ac cu mu lat ing at the base of some N-flank chan nels. The same re port also men tioned that mod er ate de gas sing was seen in the crater leav ing ar eas of abun dant sul fur, which ap peared as yel low zones in one or more fumarolic areas. The No vem ber 2005 re port of SNET also dis cussed sub stan tial land slides in side the crater that were fol lowed by wid en ing of the fun nel-shaped area of col lapse in the cen tral crater. The land slides had left three dis tinct perched rem nants of the crater floor (small ter races) at var i ous el e - va tions on the crater walls. The crater?s west ern plain (one such ter race of the sort men tioned above) was sta ble but showed ar eas of sub si dence (fig ure 5). Lahar mon i tor ing dur ing De cem ber 2005 dis closed erosion of eas ily mo bi lized cin ders and sco ria ma te rial on the N to NW flanks dur ing the pre vi ous wet sea son. De - cem ber seis mic ity was el e vated, but cracks in the crater changed lit tle com pared to pre vi ous mea sure ments. A field team vis ited the sum mit on 11 Jan u ary 2006 and again in Feb ru ary and found few sub stan tive changes in the crater. On the as cent route dur ing Jan u ary, the team saw a small re - cent ?fall of ma te rial? reach ing 40 cm thick. Some fumaroles dis charged yel low ish gases. Dur ing Feb ru ary the team con ducted mea sure ments of cracks on the west ern plain but found few changes, sug gest ing the headscarp had moved lit tle if at all. Feb ru ary and March tremor episodes were centered at ~ 5 Hz and lasted 1-3 minutes. The March 2006 SNET re port noted small rockslides on the crater?s N and S sides and, with the be gin ning of the rainy sea son in March 2006, there was a po ten tial for the de vel op ment of lahars. Dur ing the March visit the team found abun dant gran u lar ma te rial in the gul lies on the NW flank, judged to be the re sult of de bris flows. Mon i tored cracks remained stable. With the ar rival of the wet sea son in April, lahars and en hanced fumarolic out put be came ap par ent. One de bris flow in ter sected a high way. On 23-24 April, 105 mm of rain was re corded at plan ta tion (finca) Santa Isabel. Fig ure 6 shows the re sults of one lahar which left a trail of de bris dur ing the rainy in ter val. Ear lier in the month on the 16th, a tremor or multi-phase ep i sode lasted over an hour. In April 2006, an in crease in fumarole de gas sing within the crater and small land slides con trib uted to the in sta bil ity of the de pos its on the NW flanks of the vol cano. Steam em - a nated from the fumaroles oc ca sion ally form ing a weak col umn that reached the edge of the crater. There was a slight in crease in seis mic ity through out the month. Seis mic ac tiv ity in creased in March and April 2006 (fig ure 4). Rocks in the crater show in tense hy dro ther - mal al ter ation with a yel low ish red dish color. Small rock land - slides were ob served in the N and S zone of the crater. Dur ing June 2006, the tem per - a ture of the fumaroles, open ing of cracks and the gas dis charge by the crater of the vol cano, re - mained sta ble. There was an in - crease of small land slides within the crater. The anal y sis of the seis mic ity in di cates that the vol - cano is slightly above its base line of nor mal be hav ior. New land - slides and cracked rock were ob - Volcanism Network, Volume 31, Number 10, October 2006 San Miguel 5 Fig ure 4. A plot of seis mic ity at San Miguel dur ing Sep tem ber 2005-June 2006. Cour tesy of SNET. Fig ure 5. (top) A photo of San Miguel taken on 16 No vem ber 2005 showing the ?west ern plain? of San Miguel?s crater (a ter race rep re sent ing a rem nant of a for mer crater floor). A con sid er able por tion of the re main ing ter race is in the pro cess of sub si dence (slump ing). (bottom) A photo of the same area taken on 15 Feb ru ary 2006 (look ing S). A zone of lo cal sub si dence, a pit along the head scarp, ap pears in the fore ground but the sub si dence also in cludes the re gion to the left of the large ar cu ate area ex tend ing well be yond the pit and still con spic u ous in the up per left edge of the pho to graph. Cour tesy of SNET. served in the walls of the crater (fig ure 7). Rains have trans - ferred vol ca nic ma te rial down the NW flank. Seis mic ity grad u ally in creased in both fre quency and mag ni tude be - gin ning on 16 June. 47 VT earth quakes and 7,505 LP earth - quakes were re corded, an amount that sur passes those reg - is tered in May; but smaller than those registered in March and April (figure 4). Dur ing July 2006, sta bil ity con tin ued with re spect to fumarole tem per a tures, crack open ings, and gas emis sions around the crater. How ever, the seis mic ity in creased by ~ 70%. Small and spo radic land slides took place in side the crater off the SE to SW walls. In tense hy dro ther mal al ter - ation in the NW wall was also ob served. SNET did not re - port any lahars dur ing July 2006; how ever in tense rains have con tin ued to re move vol ca nic ma te rial from the NW flanks. The fumarolic field gave off weak emissions. In Au gust 2006, the mon i tored pa ram e ters such as fumarole tem per a ture, crack open ing, and vi sual es ti mates of gas dis charge main tained nor mal lev els. The seis mic ity di min ished sig nif i cantly in relation to July. Dur ing Sep tem ber 2006, San Miguel reached a low level of ac tiv ity. There were no sig nif i cant changes in the mor phol ogy of the vol cano as re ported in pre vi ous months. At the S wall, there were ev i dence of small rock slides. A sud den in crease in seis mic ity oc curred on 9 Oc to ber 2006. Con tact was made with other ob ser va to ries and it was de ter mined there were no land slides or rock falls as so ci ated with the event. Seismic in creases such as 9 Oc to ber had previously oc curred, par tic u larly on 19 June 2003 and from 2-6 May 2004. The 9 Oc to ber increaes were at trib uted to gas emission from the crater. Ref er ences: Chesner, C. A., Pullinger, C., Escobar, C. D., 2003, Phys i cal and chem i cal evo lu tion of San Miguel Vol cano, El Sal va dor. GSA Spe cial Pa per 375. Escobar, C.D., 2003, San Miguel Vol cano and its Vol - ca nic Haz ards; MS the sis, Mich i gan Tech no log i cal Uni ver - sity, De cem ber 2003. 163 p. Ma jor, J.J.; Schil ling, S.P., Pullinger, C.R., Escobar, C. D., Chesner, C.A, and Howell, M.M., 2001, Lahar-Haz ard Zonation for San Miguel Vol cano, El Sal va dor: U.S. Geo - log i cal Sur vey Open-File Re port 01-395. (Avail able on-line.) Geo logic Sum mary. The sym met ri cal cone of San Miguel vol cano, one of the most ac tive in El Sal va dor, rises from near sea level to form one of the coun try?s most prom - i nent land marks . The unvegeta ted sum mit of the 2,130-m-high vol cano rises above slopes draped with cof - fee plan ta tions. A broad, deep crater com plex that has been fre quently mod i fied by his tor i cal erup tions (re corded since the early 16th cen tury) caps the trun cated sum mit of the tow er ing vol cano, which is also known lo cally as Chaparrastique. Ra dial fis sures on the flanks of the ba - saltic-andesitic vol cano have fed a se ries of his tor i cal lava flows, in clud ing sev eral erupted dur ing the 17th-19th cen - tu ries that reached be yond the base of the vol cano on the N, NE, and SE sides. The SE flank lava flows are the larg est and form broad, sparsely veg e tated lava fields crossed by high ways and a rail road skirt ing the base of the vol cano. The el e va tion of flank vents has risen dur ing his tor i cal time, and the most re cent ac tiv ity has consisted of minor ash eruptions from the summit crater. In for ma tion Con tacts: Carlos Pullinger, Seccion Vulcanologia, Servicio Geol?gico de El Sal va dor, c/o Servicio Nacional de Estudios Territoriales, Alameda Roo - se velt y 55 Avenida Norte, Edificio Torre El Sal va dor, Quinta Planta, San Sal va dor, El Sal va dor (URL: http:// www.snet.gob.sv/Geologia/Vulcanologia/). 6 San Miguel Smithsonian Institution ? Bulletin of the Global Fig ure 7. San Miguel?s S crater wall ex poses zones of al tered and frac tured rocks. A pla nar zone of struc tural weak ness ap pears to wards the right. Photo taken on 22 June 2006. Cour tesy of SNET. Fig ure 6. A San Miguel photo show ing a part of the freshly scoured upslope chan nel in the Gato ero sional gully. The ma te rial de pos ited in the chan nel con sisted of re worked vol ca nic rocks and must have de scended as a small lahar or de bris flow. Sev eral such flows oc curred dur ing heavy late-April rains at the start of the rainy sea son, a few days be fore this pic ture was taken. Cour tesy of SNET (from their April 2006 re port). Ubinas Per? 16.355?S, 70.903?W; sum mit elev. 5,672 m All times are lo cal (= UTC - 5 hours) Ubinas be gan erupt ing ash on 25 March 2006 (BGVN 31:03 and 31:05); ash erup tions and steam emis sions con - t in ued through at least 31 Oc to ber 2006. Erup tive benchmarks dur ing that pe riod in cluded a lava dome in the crater on 19 April. Ashfall in late April forced the evac u a - tion of Querapi res i dents, who re sided ~ 4.5 km SE of the crater?s ac tive vent, to Anascapa (S of the sum mit). Ash col umns rose to al most 8 km al ti tude dur ing May. This re port dis cusses on go ing erup tions through 31 Oc - to ber 2006 as drawn from Bue nos Ai res Vol ca nic Ash Ad - vi sory Cen ter (VAAC) re ports and es pe cially from an en - light en ing 26-page re port pub lished in P?ru dur ing Sep tem ber 2006 by the Institutio Geol?gico Minero y Metal?rgico?INGEMMET (Salazar and oth ers, 2006). It in cludes a de tailed dig i tal el e va tion map with hazard zones. Back ground. Ubinas lies 90 km N of the city of Moquegua and 65 km E of the city of Arequipa (fig ure 8). The bulk of ad ja cent set tle ments re side to the SE, and gen - er ally at more dis tance, to wards the E. Fig ure 9 shows a shaded re gion where airfall de pos its took place dur ing the span 1550-1969. The zone of de pos its in cludes some mod - ern settlements. The geo logic map on fig ure 10 shows the area of the set tle ments SE of the sum mit in cludes large Ho lo cene de - pos its, in clud ing those from de bris av a lanche(s) at ~ 3.7 ka, and units con tain ing pyroclastic flows. The map also in di - cates de pos its of volcaniclasics, gla cial mo raines, airfall-ash lay ers, and lava flows. Ex ten sive Mio cene de - pos its en ve lope both the NE flanks (Pampa de Para) and SW flanks. The map of haz ard zones (fig ure 11) in di cates a nested, tear-drop shaped set of zones, with com par a tively lower in - ferred haz ard to the NE and NW. The SE-trending, elon gate area of haz ards fol lows the key drain age in that direction. El e vated haz ard zones also fol low many of the roads pass - ing through the re gion. Erup tions dur ing 2006. Salazar and oth ers (2006) re - ported that the cur rent erup tive cri sis could be di vided into three stages. Dur ing July 2005-27 March 2006, the erup tion was pri mar ily gas dis charge ris ing 100-300 m above the crater. Dur ing 27 March-8 April the erup tions con sisted of ash emis sions and gas pro duced by phreatic ac tiv ity (fig ure 12). After a mod er ate ex plo sion on 19 April, Ubinas produced ash and gas, and ex plo sions ejected vol ca nic bombs. Sev eral views into the crater appear on figure 13. Volcanism Network, Volume 31, Number 10, October 2006 Ubinas 7 Fig ure 8. Map in di cat ing the geo graphic set ting of the Pe ru vian vol ca nic front (in set) and the area around Ubinas. From Salazar and oth ers (2006). Fig ure 9. The bound ary of iden ti fied Ubinas ashfall from the years 1550 to 1969 ap pears as a curve across the S por tion of this map, 10-12 km from the sum mit crater. Note the SE-sec tor set tle ments (and their re spec tive dis tances from the sum mit crater) for the dis trict cap i tal Ubinas (6.5 km), Tonohaya (7 km), San Miguel (10 km), and Santa Cruz de Anascapa (~ 11 km), and Huarina (15 km). Map taken from Rivera (1998). Fig ure 10. Geo logic map of Ubinas shown here with out the key, which is avail able in the orig i nal re port. From Salazar and oth ers (2006). On 7 May 2006 a mod er ate ex plo sion sent ash to ~ 3 km above the sum mit. Al though the sit u a tion calmed in the fol - low ing days, an im pres sive bomb fell 200 m from the crater on 24 May 2006 (fig ure 14). Larger out bursts oc curred on 29 May and 2 June, prompt ing the civil de fense de ci sion to evac u ate res i dents in the S-flank Ubinas val ley, in clud ing the set t le ments of Ubinas, Tonohaya, San Miguel, Huatahua, and Escacha. Res i dents evac u ated were lodged in ref u gee camps (fig ure 15). On 18 June in stru ments re corded two ex plo sions. Ash clouds dis charged; the sec ond one also ejected in can des - cent blocks ~ 1 km SE of the crater. The early stages of a ris ing plume seen at 0822 on 18 July ap pears on fig ure 16. Sim i lar mag ni tude ash emis sions were noted on 23, 24, and 30 June 2006, and in can des cent rocks fell up to 1.2 km from the sum mit crater. Dur ing 10, 17-19, 22, 27 July, and 7 Au gust 2006 there were var i ous ex plo sions (fig ure 16). Re sult ing ash clouds ex tended more than 70 km SE or SW. 8 Ubinas Smithsonian Institution ? Bulletin of the Global Fig ure 11. The SE cor ner of the Ubinas haz ard map, show ing the cen tral crater, and the haz ard zones that fol low the main drain age (Rio Ubinas) lead ing SE through the most pop u lated re gion close to the vol cano. Map key is omit ted. Mar gins of the map note that its con struc tion was a part ner ship of nu mer ous groups, in clud ing French col lab o ra tors at Blaise Pascal Uni ver sity and IRD. Taken from Salazar and oth ers (2006). Fig ure 12. Ubinas gas emis sions as seen from un stated di rec tion on 4 April 2006. From Salazar and oth ers (2006). Fig ure 13. Views look ing down into the Ubinas crater on 31 March, 19 April, and 26 May 2006. The for mer was taken in com par a tively mild con di tions. The 19 April photo was taken when a 60-m-di am e ter lava body was first seen on the crater floor (the color ver sion of this photo shows faint red in can des cence pen e trat ing the steamy scene). The 26 May cap tured a rel a tively clear view of the steam ing dome on the crater floor. March and April pho tos from and Salazar and oth ers (2006); May photo from the INGEMMET website. In Au gust 2006, ash plumes reached 4.6-7.6 km al ti tude and were oc ca sion ally vis i ble on sat el lite im ag ery. The di - rec tion of drift of the ash var ied widely. On 12 Au gust, ash dis persed more than 100 km to the SE and S. On 14 Au gust an as tro naut on the In ter na tional Space Sta tion took a pic - ture of the ash plume from Ubinas (fig ure 17). The most sig nif i cant ef fect on peo ple and the en vi ron - ment has come from ashfall (fig ure 18). GOES sat el lite im - ages in di cate vis i ble air borne ash for dis tances greater than 60 km from the vent. Fig ure 18 in di cates net ash ac cu mu la - tion through about Au gust 2006, ex trap o lat ing sam pling points with con cen tric cir cles. The re port spe cif i cally noted ash thick nesses of 1.5 cm at ~ 4.5 km SE in Querapi, 0.1-0. 8 cm in Sacoaya, 0.5-0.8 cm in Ubinas, 0.3-0.4 cm in Anascapa, 0.15 mm in Huatahua, and less than 0.1 cm in Chacchag?n. The ac cu mu la tion has ap par ently been due to on go ing ashfalls On 13 April, sev eral mil li me ters of ash dusted all sur faces in Querapi, ~ 4.5 km from the center of the summit crater. Avi a t ion re ports of ash plumes. As sum ma rized in ta ble 1, ash clouds were re ported by the Bue nos Ai res Vol ca nic Ash Ad vi - sory Cen ter (VAAC) on 2 May and then dur ing 2 Au gust through Oc to ber on a nearly daily ba sis. The ob ser va tion sources were usu ally pi lot?s re ports (AIREPs) and/or sat el lite im ages (GOES 12). Af ter 8 Au gust, ash emis - sions were es sen tially con tin u ous to 31 Oc to ber. Dur ing the later in - ter val, the avi a tion color code was gen er ally Red. Plumes rose to 10 km and higher dur ing 23-26 Oc to ber. Ref er ences: Rivera, M., 1998, El volc?n Ubinas (sur del Per?): geolog?a, historia eruptiva y evaluaci?n de las amenazas volc?nicas actuales: Tesis Ge?logo, UNMSM, 132 p. Volcanism Network, Volume 31, Number 10, October 2006 Ubinas 9 Fig ure 14. Ubinas erup tions in May 2006 ejected vol ca nic bombs, seen here in their im pact crat ers. A 2-m-di am e ter bomb (top), struck ~ 200 m from the crater. A crater con tain ing a large, partly bur ied, smooth-faced bomb is seen in the bot tom photo. Nu mer ous bucket-sized an gu lar blocks ap pear on the far side of the im pact crater. Two ge ol o gists stand ad ja cent a ~ 2-m-long block that ended up on the im pact crater?s rim. The bomb frag ments were of andesitic com po si tion. Top photo from Salazar and oth ers (2006); bottom photo from INGEMMET website. Fig ure 15. Set tle ment camp hous ing fam i lies tak ing ref uge from Ubinas ash. This camp, named Chacchagen, housed peo ple from the S-flank set tle ments of Ubinas, Tonohaya, San Miguel, Huatahua, and Escacha. In set shows the ash-dusted face of a lo cal child. Cour tesy of Salazar and oth ers (2006). Fig ure 16. A mod er ate Ubinas ex plo sion on 18 July 2006 gen er ated this ris ing ash plume. Cour tesy of Salazar and oth ers (2006). Rivera, M., Thouret, J.C., Gourgaud, A., 1998, Ubinas, el volc?n mas activo del sur del Per? desde 1550: Geolog?a y evaluaci?n de las amenazas volc?nicas. Boletin de la Sociedad Geol?gica del Per?, v. 88, p. 53-71. Salazar, J.M., Porras, M.R., Lourdes, C.D., and Pauccara, V.C., 2006, Evaluaci?n de seguridad f?sca de ?reas aleda?as al volc?n Ubinas: INGEMMET (Instituto Geol?gico Minero y Metal?rgico Direcci?n de Geolog?a Ambiental, Sep tem ber 2006), 26 p. Thouret, J.C., Rivera, M., Worner, G., Gerbe, M.C., Finizola, A., Fornari, M., and Gon za les, K., 2005, Ubinas: the evo lu tion of the his tor i cally most ac tive vol cano in south ern Peru: Bull. Volc., v. 67, p. 557-589. Geo logic Sum mary. A small, 1.2-km-wide cal dera that cuts the top of Ubinas, Peru?s most ac tive vol cano, gives it a trun cated ap pear ance. Ubinas is the north ern most of three young vol ca noes lo cated along a re gional struc tural lin ea - ment about 50 km be hind the main vol ca nic front of Peru. The up per slopes of the stratovolcano, com posed pri mar ily of Pleis to cene andesitic lava flows, steepen to nearly 45 de - grees. The steep-walled, 150-m-deep sum mit cal dera con - tains an ash cone with a 500-m-wide fun nel-shaped vent that is 200 m deep. De bris-av a lanche de pos its from the col - 10 Ubinas Smithsonian Institution ? Bulletin of the Global Fig ure 17. This im age taken from the In ter na tional Space Sta tion (ISS) cap tures Ubinas dis charg ing a light-col ored ash cloud roughly to the S (N is up on this photo). The cloud had been ob served ear lier on sat el lite im ag ery at 0600 lo cal (1100 UTC) on 14 Au gust 2006. One-hour and 45-min utes later (at 1245 UTC), an ISS as tro naut took this pic ture at non-ver ti cal (oblique) an gle to the Earth. Pum ice and ash blan ket the vol ca nic cone and sur round ing area, giv ing this im age an over all gray ap pear ance. Shad ows on the N flank throw sev eral older lava flows into sharp re lief. (Photograph ISS013-E-66488 ac quired with a Ko dak 760C dig i tal cam era us ing an 800 mm lens). Photo pro vided by the ISS Crew. Fig ure 18. Net ash ac cu mu la tion around Ubinas from start of erup tion in March through about Au gust 2006. Ash has cov ered ag ri cul tural fields in the val ley and pas tures in the high lands, se ri ously af fect ing the two main eco nomic ac tiv i ties in the area, ag ri cul ture and cat tle ranch ing; and has caused re spi ra tory and skin prob lems. Cour tesy of Salazar and oth ers (2006). Ob ser va tion date (2006) Erup tion de tails: VA (Vol ca nic Ash) CLD (Cloud) OBS (Ob served) FL (Flight Level) 02 May VA CLD FL180/200 MOV SE 02 Aug VA CLD DENSE ASH CLD FL160/230 MOV NE. ASH POORLY DE FINED VIS I BLE GOES-12 SAT EL LITE IM AGE 03 Aug-04 Aug VA CLD FL220/240 MOV SW 05 Aug VA CLD OBS FL370 MOV NE 06 Aug-07 Aug VA CLD OBS. AC TIV ITY RE PORTED CON TIN U OUS AND IN CREAS ING EMIS SION FL160/260 SNTR OVER PEAK SPREAD FROM THE SUM MIT IN ALL DI REC TIONS UP TO A DIS TANCE OF 20 KM 07 Aug-08Aug VA CLD OBS FL200 MOV E/NE 10 Aug-14 Aug VA CLD OBS FL180/245 MOV SE. ASH OBS IN SAT EL LITE IM AGE 17 Aug-18 Aug VA CLD FL 160-200 MOV SE/ESE APROX. 60NM 19 Aug VA CLD FL180/250 MOV SW 20 Aug-21 Aug VA CLD FL180/230 MOV ESE/SE APROX. 20NM 22 Aug VA CLD OBS FL180/300 STNR ~ MOV SE 25 Aug-26 Aug VA CLD OBS FL230/235 MOV S. ASH NOT IDEN TI FI ABLE ON SAT EL LITE IM AG ERY 28 Aug-30 Aug VA CLD OBS FL160/250 MOV SE. SAT EL LITE IM AG ERY RE VEALED A LIGHT TRACE OF ASH EX TEND ING TO SE OF THE SUM MIT 31 Aug VA CLD OBS FL 160/250 APROX MOV NE~E 01 Sep-23 Sep VA CLD OBS FL 160/250 MOV NE~E 24 Sep VA CLD FL300 MOV SSE 27 Sep VA CLD OBS FL180/230 and up to FL280 01 Oct-11 Oct VA CLD OBS FL160/180 MOV E~ S 12 Oct-14 Oct Emis sions in ter mit tent VA CLD OBS FL160/220 MOV SE~NE~N 15 Oct-21 Oct VA CLD FL160~ 240 MOV S~ SE 23 Oct-26 Oct VA CLD FL180/350 (Un usu ally high al ti tude) MOV N~E~W 26 Oct-29 Oct VA CLD FL180/240 MOV N~NW swing to S 30 Oct-31 Oct VA CLD FL 280/300 MOV SW Ta ble 1. Com pi la tion of avi a tion re ports (spe cif i cally, 195 Vol ca nic Ash Ad vi so ries, VAAs) on Ubinas and its plumes dur ing May through 31 Oc to ber. The sec ond col umn shows some con trac tions used in the ta ble (eg., ?VA CLD FL 160" means ?Vol ca nic ash cloud at Flight Level 160"). Flight Level is an avi a tion term for al ti tude in feet di vided by 100 (eg., FL 200 = 20,000 feet = ~ 7 km al ti tude). Cour tesy of the Bue nos Ai res VAAC. lapse of the SE flank of Ubinas ex tend 10 km from the vol - cano. Wide spread Plinian pum ice-fall de pos its from Ubinas in clude some of Ho lo cene age. Ho lo cene lava flows are vis - i ble on the vol cano?s flanks, but his tor i cal ac tiv ity, doc u - mented since the 16th cen tury, has con sisted of in ter mit tent mi nor ex plo sive erup tions. In for ma tion Con tacts: Jersy Mari?o Salazar, Marco Rivera Porras, Lourdes Cacya Due?as, Vicentina Cruz Pauccara, Instituto Geol?gico Minero y Metal?rgico (INGEMMET), Av. Canad? No 1470, Lima, Peru (URL: http://www.ingemmet.gob.pe/); Bue nos Ai res Vol ca nic Ash Ad vi sory Cen ter, Servicio Meteorol?gico Nacional, Ar gen - tina (URL: http://www.ssd.noaa.gov/VAAC/OTH/AG/ mes sages.html); ISS Crew, Earth Ob ser va tions Ex per i ment and the Im age Sci ence & Anal y sis Group, NASA John son Space Cen ter, 2101 NASA Park way Hous ton, TX 77058, USA (URL: http://www.nasa.gov/cen ters/john son/home/ in dex.html); Na tional Aero nau tics and Space Ad min is tra - t ion (NASA) Earth Ob ser va tory (URL: ht tp : / / earthobservatory.nasa.gov/NaturalHazards/). Kilauea Ha wai ian Is lands 19.421?N, 155.287?W; sum mit elev. 1,222 m All times are lo cal (= UTC -10 hours) -Lava from Kilauea con tin ued to flow through the PKK lava tube from its source at Pu`u `O`o to the ocean dur ing this re port ing pe riod from late Au gust to the end of No vem - ber 2006. About 1 km S of Pu`u `O`o, the Campout lava flow branches off from the PKK tube. Through No vem ber, the PKK and Campout sys tems fed two widely sep a rated ocean en tries named East Lae`apuki and East Ka`ili`ili, re - spec tively. Kilauea?s ac tiv ity dur ing this re port ing pe riod in cluded nu mer ous small break outs from the Campout flow, new sky lights along the PKK tube, and vari able ac tiv - ity at the ocean en tries, in clud ing small streams of lava cross ing the coastal bench. Intermittant lava fountaining 15 m in land of the W edge of the East Lae`apuki bench was noted in late Sep tem ber-early Oc to ber. In can des cence was also in ter mit tently vis i ble com ing from the East Pond and Jan u ary vents, the South Wall com plex, and the Drainhole vent in Pu`u `O`o?s crater. In gen eral, dur ing this re port ing pe riod the in fla tion ary trend con tin ued at the sum mit of Kilauea, in ar eas S of Halema`uma`u crater and trem ors remained at a very typical moderate level at Pu`u `O`o. Dur ing 30 Au gust-12 Sep tem ber, crews re ported vis i ble lava streams on the W side of the East Lae?apuki delta and oc ca sion ally from the East Ka?ili?ili en try. On 1 Sep tem ber, the East Lae?apuki lava bench was an es ti mated 22 hect ares (54 acres) and East Ka`ili`ili was an es ti mated 2.3 hect ares (5.8 acres). On 30 Au gust, and 1 and 6 Sep tem ber, the Campout flow es caped from the PKK tube. On 11 Sep tem - ber, Park Ser vice field crews re ported two lava flows vis i - ble down the en tire length of the pali. In can des cence was in ter mit tently vis i ble from the East Pond and Jan u ary vents, the South Wall com plex, and the Drainhole vent in Pu`u `O`o?s crater. Dur ing 13-23 Sep tem ber, lava from the Campout and PKK sys tems con tin ued to flow off of a lava delta into the ocean and break out flows were vis i ble on the pali. On 20 Sep tem ber, a tour pi lot re ported see ing three large lava flows from a break out 10 m in land from the old sea cliff at East Lae`apuki (fig ures 19 and 20). On 23 Sep tem ber, in - can des cence from above Pulama pali in the di rec tion of Pu`u`O`o was likely due to sev eral new and re ac ti vated sky lights on the upper PKK tube. Lit to ral fountaining on 27 Sep tem ber was re ported about 15 m in land of the W edge of the East Lae`apuki bench. Lava jet ted about 30 m in the air ac com pa nied by loud rum bling and jet ting sounds. Ob serv ers re ported ground shak ing. Over the next cou ple of days, 3-4 lava streams were vis i ble on the W side of East Lae`apuki en try, as were in ci dents of tephra jet ting and lava fountaining 15-23 m (50-75 ft) high. Glow had been vis i ble from the East Lae`apuki en try and the Campout flow break out on the pali, but not from the Ka`ili`ili en try. The con sis tent lack of vis i ble glow from the Ka`ili`ili en try was due to the ab sence of a very large bench, forc ing lava to re main hid den at the base of the seacliff. Volcanism Network, Volume 31, Number 10, October 2006 Kilauea 11 Fig ure 19. Aerial view of the lava bench at East Lae`apuki, look ing NE on 20 Sep tem ber 2006. An ac tive lava flow is go ing over the sea cliff in roughly the cen ter of the ar cu ate fault scarp in the wid est part of the lava bench be low it. White steam plumes from the ocean en try were blown to wards along the coast to wards the left. In the col ored ver sion of this shot the ad ja cent sea wa ter con tains a green ish hue. Cour tesy of HVO. Fig ure 20. A lava flow at Kilauea breaks out to the sur face 10 m in land from a sea cliff on 20 Sep tem ber 2006 . The lava pours over the cliff in places as thick cur tains and else where as smaller riv u lets and drip ping falls. Af ter the fall the lava pro ceeded across the up per bench as a se ries of braided streams. To ward the left, some read ers might claim they see a slen der Pel?, danc ing with arms up raised. Courtesy of HVO. Ob serv ers re ported that on 28 Sep tem ber the floor of the Drainhole vent had been re placed by an over turn ing lava pond. As of 29 Sep tem ber, a new tube and flow that formed on the E side of the Campout flow ex tended ~ 180 m. An other flow went W and butted up against the PKK tube. The USGS field crew also found a small stag nant break out of lava at ~ 60 m el e va tion. It flowed E to cover a lit tle more of the long-aban doned Royal Gar dens sub di vi - sion. In the Pu?u O?o vi cin ity, a new col lapse pit pho to - graphed in early Oc to ber had en gulfed pre-ex ist ing spatter cones (figure 21). Dur ing Oc to ber and No vem ber, break out flows were in - ter mit tently vis i ble on the Pulama pali, at the base of the pali, or on the sea cliff and in can des cence from vents in Pu`u`O`o was vis i ble. For ex am ple, on 25 Oc to ber, two sep a rate break-out lava flows were vis i ble on pali. The up - per flow at about 320 m (1,050 ft) el e va tion con sisted of ?a?a and pahoehoe and the lower flow at 114 m (375 ft) was solely pahoehoe. On 3 and 4 No vem ber, tephra jet ted at the tip of the East Lae`apuki bench. On 15 No vem ber, break - outs re sumed on top of the sea cliff af ter a few weeks with - out ac tiv ity. On 18 No vem ber, the Drainhole vent twice ejected spat ter as high as 25 m above its rim. On 19 No vem - ber, ob serv ers saw small ex plo sions at East Lae`apuki ocean en try as well as well-de fined streams of lava en ter ing the ocean. The next eve ning, six rivers of lava flowed over the bench and into the ocean at the W en try. When weather per mit ted, in can des cence was vis i ble from the East Pond, the South Wall complex, the January vents, and Drainhole vent. In for ma tion Con tacts: Ha wai ian Vol cano Ob ser va tory (HVO), U.S. Geo log i cal Sur vey, PO Box 51, Ha waii Na - tional Park, HI 96718, USA (URL: http://hvo.wr.usgs.gov/; Email: hvo-info@hvomail.wr.usgs.gov). Home Reef Tonga Is lands, SW Pa cific 18.992?S, 174.775?W; sum mit elev. -2 m All times are lo cal (= UTC + 13 hours) An erup tion from Home Reef in early Au gust gen er ated large vol umes of pum ice that floated to Fiji (over 700 km away) in the fol low ing two months; an is land was also cre - ated (BGVN 31:09). Sat el lite data and im ag ery have been used to con firm these ob ser va tions and pro vide ad di tional in for ma tion about this event. Nor man Kuring of the MODIS Ocean Color Team iden - ti fied the ear li est clear shot of the pum ice raft in a Terra MODIS on 7 Au gust at 2120 UTC (8 Au gust at 1020 Tonga time). The im age (fig ure 22) shows a cir cu lar patch of pum - ice over the erup tion site with a small vol ca nic plume emerg ing from it. The first in di ca tion of pum ice raft leav ing the erup tion site is in the Aqua MODIS im age at 0132 UTC on 10 Au gust, but the 9 Au gust over pass was cloudy, so it could have hap pened ear lier. Kuring also com piled other im ages show ing the dis per sion of the pum ice through 22 August (figure 23). Kuring also made a pre lim i nary es ti mate of the area of the pum ice raft on 11 Au gust (10 Au gust at 2150 UTC), pre vi ously en coun tered by the Maiken (BGVN 31:09). A mask was cre ated to cover iden ti fi able ar eas of pum ice, re - sult ing in an area of 9,338 pix els. Each pixel in the im age used cov ers an area of 0.0468 km 2 . The cal cu lated to tal area is ap prox i mately 440 km 2 for that time. Note that this es ti - mate does not take into ac count er rors caused by pum ice be ing a high-con trast tar get (al low ing lin ear patches less than the pixel width to be seen), small iso lated patches of pum ice that could not be rec og nized, ma te rial hid den by 12 Home Reef Smithsonian Institution ? Bulletin of the Global Fig ure 21. Two views of Kilauea?s W gap area il lus trat ing morphologic changes there. (top) Ae rial view of Pu`u `O`o taken in July 2006 shows two spat ter cones.. Note he li cop ter above la bel for scale. (bot tom) An ae rial photo taken on 13 Oc to ber 2006 shows a new col lapse pit that grew to en gulf the spat ter cones. The bot tom of the pit, which formed on the night of 10 Oc to ber, is hid den by fume. Cour tesy of HVO. Fig ure 22. Terra MODIS im age taken on 8 Au gust 2006 (lo cal time) show ing the early stages of the erup tion at Home Reef. A steam plume is vis i ble ris ing from the south ern end of a mass of float ing pum ice cov er ing an area larger than Late Is land to the NW. Cour tesy of the NASA Ocean Color Group. clouds, or frag ments sus pended in the wa ter col umn un der the sur face. In the 8 Au gust MODIS im age, the cir cu lar area was de ter mined by Bul le tin ed i tors to be at least 8 km in di am e ter, so the area cov ered was more than 50 km 2 . Si mon Carn (UMBC) used the Ozone Mon i tor ing In - stru ment (OMI) on NASA?s Aura sat el lite to con strain the tim ing of the erup tion. OMI de tected SO2 emis sions from the vi cin ity of Home Reef be gin ning on 8 Au gust. Emis - s ions ap pear to have peaked some time during 8-9 Au gust. The to tal SO2 mass de tected E of Tonga by OMI on 9 Au gust was ~ 25 ki lo tons. By 12 Au gust there were 3.3 ki lo tons of SO2 in the area (fig ure 24). The emis sion ep - i sode was over by 15 Au gust. HYSPLIT for ward tra jec to ries in - di cated that the SO2 re leased on 8 Au gust may have reached al ti - tudes of 5 km or more. Carn also stated that ?To our knowl edge this is the first ex am ple of sat el lite de tec tion of emis sions from a sub ma rine vol cano. Sig nif i cant scrub bing of SO2 and other soluble volcanic gases is likely during such events.? Terra MODIS data from 4 Sep tem ber 2006 pro vided by Alain Ber nard showed pum ice rafts mov ing SE from Home Reef (fig ure 25). Pum ice that pre vi - ously fol lowed a sim i lar path was found on beaches in south ern Vava?u (BGVN 31:09) by 2 September. Is land evo lu tion. No data or re ports are avail able to de ter mine when the is land built by the 1984 erup tion (SEAN 09:02 and 09:04) eroded be low the ocean sur face. Re cent re ports from mar i ners and lo cal fish er men noted that this cur rent erup tion had built a new is land, im ply ing the ab sence of an is land at that lo ca tion. An AS TER im age in spected by Matt Pat rick from 18 No vem ber 2005 did not show an island. An AS TER im age of the new is land taken on 4 Oc to ber 2006 (fig ure 26) has been stud ied by a num ber of sci en tists, in clud ing Greg Vaughan (JPL), Matt Pat - rick (Mich i gan Tech), and Alain Ber nard (Univ. of Brussels). The im age clearly shows the is land (at 18.991?S, 174.762?W) with large ? and NE-di rected anom a lous ar - eas that are likely caused by vol - ca nic ma te rial sus pended in the wa ter. The new is land is warmer than ad ja cent Late is land. Greg Vaughan pro vided an an no tated ver sion of the im age zoomed in on the new is land, which he com - puted then had an area of 0.245 km 2 . Vaughan also noted that the Volcanism Network, Volume 31, Number 10, October 2006 Home Reef 13 Fig ure 23. Terra and Aqua MODIS sat el lite im ages show ing the dis per sion of the pum ice raft gen er ated by the erup tion at Home Reef dur ing 7-8 Au gust. By 10 Au gust a large raft was NE of Late Is land. Most of the ma te rial stayed in that area through 12 Au gust be fore break ing up into elon gate pieces that be gan mov ing W to wards Fiji. Cour tesy of the NASA Ocean Color Group. ?day time im age shows con sid er able ac tiv ity in the wa ter around the new Home Reef is land [and] a ther mal plume in the same shape as the pink col ored area in the at tached VNIR im ages (AS TER chan nels 3-2-1 as R-G-B).? Work by Alain Ber nard based on the AS TER ther mal bands de - ter mined that the hot lake on the is land had a max i mum tem per a ture of 64.7?C on 4 Oc to ber. Ber nard cal cu lated the is land area to be 0.230 km 2 on 4 Oc to ber. Com par i son with an other AS TER im age from 12 No vem ber showed that the is land had changed shape and cov ered an area of 0.146 km 2 , a de crease of 0.84 km 2 (figure 27). Float ing pum ice ob ser va tions. Ad di tional pum ice sight ings have been re ported that sup ple ment those de - scribed ear lier (BGVN 31:09). Ar eas known to have been im pacted by the pum ice now in clude Suva Point (where the cap i tal of Fiji, Suva, is lo cated) and Yasawa Is land (N of Viti Levu and E of Vanua Levu). By early No vem ber pum - ice from Home Reef had reached Efate Island in Vanuatu. Crew on the SV Sand piper en coun tered pum ice dur ing tran sit from Tonga to north ern Fiji on 12 Sep tem ber. They went through ?large patches? of pum ice ?all af ter noon? while trav el ing about 200 km over the course of the day. The next eve ning, af ter sun set on 13 Sep tem ber, the boat sud denly slowed and the wa ter ?looked like a thick choc o - late shake.? Lights shin ing down from the rig ging (spreader lights) showed that they were sur rounded by pum ice. The crew ob served pum ice again on 23 Sep tem ber at the southern end of Vanua Levu. Wally John son was fly ing from Suva to Taveuni on 19 Sep tem ber and ob served large amounts of pum ice in the Koro Sea, drawn out into nu mer ous par al lel strings in the di rec tion of the pre vail ing wind and head ing to wards Taveuni. A fair bit of the pum ice had been washed up into ridges on beaches on the NW coast of Taveuni, and up and into pock ets on some of the re cent ba saltic lava flows to the SW. Bernie Joyce for warded ad di tional re ports from Fiji. 14 Home Reef Smithsonian Institution ? Bulletin of the Global Fig ure 24. Sul fur-di ox ide emis sions in the vi cin ity of Home Reef, 12 Au gust 2006 at 0140 UTC. Data ob tained from the Ozone Mon i tor ing In stru ment (OMI) on NASA?s Aura sat el lite. Cour tesy of Si mon Carn. Fig ure 25. Terra MODIS data from 4 Sep tem ber 2006 show ing pum ice rafts mov ing SE from Home Reef. Data was ob tained with a sim ple pro cess ing of bands 1 and 2; pixel size is 250 me ters. Cour tesy of Alain Ber nard. Fig ure 26. AS TER VNIR im age show ing the new is land at Home Reef on 4 Oc to ber 2006. A vol ca nic lake is vis i ble on the is land, as are sub ma rine plumes orig i nat ing from the is land. Some pos si ble small pum ice rafts can also be iden ti fied in this 15-m im ag ery. Mod i fied from orig i nal pro vided cour tesy of Greg Vaughan. Fig ure 27. Com par i son of the is land at Home Reef on 4 Oc to ber (left) and 12 No vem ber 2006 (right) us ing AS TER im ag ery. The size of the is land de creased ap prox i mately 0.84 km 2 over that time pe riod. Cour tesy of Alain Ber nard. On 27 Oc to ber 2006, Rebekah Mue-Soko re ported that the Suva Point area of Viti Levu was filled with pum ice as of 27 Oc to ber, and that it had ap peared some time be fore 8 Oc - to ber. About 6 No vem ber 2006 Lyn and Darcy Smith were on the Fijian is land of Yasawa, N of Viti Levu, and re - ported ?a heap of pum ice on the beach? which ap par ently arrived during their one-week visit. While pum ice has per sisted in Fiji, some reached Vanuatu. Sandrine Wallez re ported pum ice on the W coast of Efate Is land dur ing the night of 4-5 No vem ber. A de posit around 10 cm thick was ob served along 40 km of coast line. The larg est pum ice frag ments were the size of a ten nis ball. Pum ice was still on the beaches in early De cem ber (fig ure 28). Shane Cronin was in Vanuatu in early Oc to ber when a new batch of fresh pum ice washed up on north ern Efate beaches. Pum ice is com monly be ing de pos ited on beaches around Vanuatu, and lo cal res i dents told Cronin that they thought it was com ing from up around the Ambrym-Lopevi area. Douglas Char ley (DGMWR - Vanuatu) re corded ex - plo sion earth quakes on a por ta ble geo phone from south Epi (BGVN 29:04) at the beginning of September. Geo logic Sum mary. Home Reef, a sub ma rine vol cano mid way be tween Me tis Shoal and Late Is land in the cen tral Tonga is lands, was first re ported ac tive in the mid-19th cen tury, when an ephem eral is land formed. An erup tion in 1984 pro duced a 12-km-high erup tion plume, co pi ous amounts of float ing pum ice, and an ephem eral is land 500 x 1500 m wide, with cliffs 30-50 m high that en closed a water-filled crater. In for ma tion Con tacts: Si mon Carn, Joint Cen ter for Earth Sys tems Tech nol ogy, Uni ver sity of Mary land-Bal ti - more County (UMBC), 1000 Hill top Cir cle, Bal ti more, MD 21250, USA (Email: scarn@umbc.edu, URL: http://toms. umbc.edu/); Nor man Kuring, NASA/Goddard Space Flight Cen ter, Code 970.2, Greenbelt, MD 20771, USA (Email: nor man@seawifs.gsfc.nasa.gov, URL: http://oceancolor. gsfc.nasa.gov/); Greg Vaughan, NASA Jet Pro pul sion Lab - o ra tory, Cal i for nia In sti tute of Tech nol ogy, 4800 Oak Grove Dr., Pas a dena, CA 91109-8099, USA (Email: Greg. Vaughan@jpl.nasa.gov); Alain Ber nard, IAVCEI Com mis - sion on Vol ca nic Lakes (CVL), Universit? Li bre de Bruxelles (ULB), CP160/02, av e nue F.D. Roo se velt 50, Brussels, Bel gium (Email: abernard@ulb.ac.be, URL: http://www.ulb.ac.be/sci ences/cvl/homereef/homereef. html); Sandrine Wallez, De part ment of Ge ol ogy, Mines, and Wa ter Re sources (DGMWR), Port-Vila, Vanuatu (Email: sandrine@vanuatu.com.vu); R. Wally John son, 45 Alroy Cir cuit, Hawker, ACT 2614, Australia (Email: wallyjohnson@grape vine.com.au); Shane Cronin, In sti tute of Nat u ral Re sources, Massey Uni ver sity, Palmerston, New Zea land (Email: S.J.Cronin@massey.ac.nz); Tom and Amy Larson, SV Sand piper (URL: http://sandpiper38.blogspot. com/2006_09_01_sandpiper38_ar chive.html, Email: wdc5572@sailmail.com); E.B. Joyce, School of Earth Sci - ences, The Uni ver sity of Mel bourne, VIC 3010, Aus tra lia (Email: ebj@unimelb.edu.au, URL: http://www.ge ol ogy. au.com/) Rabaul New Brit ain, SW Pa cific 4.271?S, 152.203?E; sum mit elev. 688 m All times are lo cal (= UTC +10 hours) The Rabaul Vol cano Ob ser va tory (RVO) re ported that a large, sus tained Vulcanian erup tion be gan at Rabaul at about 0845 on 7 Oc to ber 2006 (BGVN 31:09). A fur ther point re gard ing that erup tion, ab sent from our pre vi ous re - port, was that some mem bers of the Vol ca nic Clouds Group (a listserv dis cus sion group) con ducted sig nif i cant ob ser va - tions and ini tial mod el ing of the 7 Oc to ber erup tion clouds, in clud ing map ping the cloud?s sul fur di ox ide con tent and mak ing fore casts of their dis per sion. In the Volcaniccloud listserv dis cus sions of the 7 Oc to ber clouds, An drew Tupper noted the fol low ing: ?The cloud was at 16 km (up - per tro po sphere/lower strato sphere) when it passed over Manus on its way NW . . . . How ever, the north/north east - ern parts were ini tially higher . . . , with the east ward bit clearly strato spheric. There were mul ti ple flights un der the cloud over Mi cro ne sia for [sic] that re ported that was no ash or smell?this puts a lower bound ary (~ 10 km) on the cloud, con sis tent with our view that the bits at cruising levels had gone to the SE.? Since that event and in ref er ence to the time in ter val for this re port, 4 No vem ber to early De cem ber 2006, RVO has noted that ac tiv ity con tin ued at Tavurvur at vary ing in ten si - ties. The larg est event in the re port ing in ter val took place at 0715 on 14 No vem ber 2006; Tavurvur produced a large ex - plo sion that rose sev eral kilometers above the cone. Volcanism Network, Volume 31, Number 10, October 2006 Rabaul 15 Fig ure 28. Pho to graph show ing beach de pos its of pum ice from the Au gust erup tion at Home Reef on west ern Efate, Vanuatu, on 3 De cem ber 2006. Cour tesy of Sandrine Wallez. Dur ing 4-13 No vem ber, mild erup tive ac tiv ity con tin - ued at Tavurvur, with oc ca sional small-to-mod er ate ash emis sions con tin u ing and blow ing to the SE. An emis sion on 11 No vem ber con sisted of thick white va por ac com pa - nied by oc ca sional small-to-mod er ate ash clouds that drifted vari ably to the SE, S, and NW and re sult ing in fine ash fall down wind. On 12 No vem ber the emis sion was blown W and NW, and on the morn ing of 13 No vem ber the ash cloud drifted N of the volcano. An ex plo sion oc curred at Tavurvur at 0715 on 14 No - vem ber 2006, ac com pa nied by a thick ash cloud that rose to about 2 km above the sum mit be fore drift ing NW. The ex - plo sion show ered the flanks of the vol cano with lava frag - ments, some of which fell into the sea. Fine ash fall oc - curred at Rabaul Town ar eas and down wind to the Ratavul and Nonga ar eas. Con tin u ous ash emis sion fol lowed the ex - plo sion. Seis mic ac tiv ity con tin ued at low lev els; however, high-fre quency earth quakes con tin ued to oc cur within the Rabaul cal dera. Af ter the large ex plo sion on 14 No vem ber, mild erup tive ac tiv ity con tin ued at Tavurvur, con sist ing of con tin u ous thick white va por ac com pa nied by pale gray to gray ash clouds that rose ~ 1.5 km above the sum mit be fore drift ing vari ably S and E of Tavurvur. Dur ing 16-17 No - vem ber, con tin u ous thick white va por ac com pa nied by pale gray ash clouds rose to about 2.5 km above the sum mit be - fore drift ing vari ably to the NW and E with fine ash fall ing on set tle ments down wind, including Rabaul Town. One high-frequency earthquake occurred on 16 November. Mild erup tive ac tiv ity con tin ued at Tavurvur dur ing 18-20 No vem ber. On 18 No vem ber and on the morn ing of 20 No vem ber con tin u ous gray ash clouds rose less than 200 m above the sum mit be fore be ing blown N and NW. Fine ash con tin ued to fall on vil lages down wind in clud ing Rabaul Town. Ac tiv ity on 19 No vem ber con sisted of emis - sion of thick white va por only, ac com pa nied by roar ing noises heard between 1130 and 1400. Quiet gen er ally pre vailed at Tavurvur dur ing 20-23 No - vem ber. Emis sions then con sisted of thick white va por ac - com pa nied by a small amount of pale gray ash clouds. On 21 No vem ber the emis sions ac cu mu lated in the at mo sphere around the cal dera caus ing haze, and on 22 No vem ber the emis sions rose less than 1,000 m above the sum mit be fore drift ing W. Fine ash fell on vil lages down wind. On the morn ing of 23 No vem ber the emis sion con sisted of white va por ris ing more than a ki lo me ter above the summit before drifting E. On 26 and 27 No vem ber the ac tiv ity con sisted of gen tle spo radic emis sion of subcontinuous, gray to pale gray ash clouds of vary ing thick ness. The ash clouds drifted NW to W re sult ing in fine ash fall down wind. From No vem ber to 1 De cem ber the emis sion con sisted of pale gray to dark gray ash clouds be ing re leased more force fully. The ash clouds rose less than 200 m above the sum mit be fore drift ing E. On the morn ing of 2 De cem ber the emis sion con sisted thick white va por and pale gray ash clouds that rose about 2 km be fore be ing blown ENE. On 3 De cem ber thick pale gray ash clouds that rose about 1 km above the sum mit were emit ted. The ash clouds drifted NE in the morn ing and then slightly to the W in the af ter noon. On the morn ing of 4 De - cem ber the ash cloud rose about 2 km be fore drift ing E. Fine ash fall oc curred in down wind ar eas. There was no glow from the vol cano vis i ble at night. From late morn ing to the af ter noon of 4 De cem ber the ac tiv ity con sisted of emis sion of thick pale gray ash clouds that rose about 500m above the sum mit be fore drift ing NW. In the morn ing of 5 De cem ber the ash cloud rose 200 m be fore drift ing E. By mid-morn ing the ash clouds were ris ing about 1 km above the sum mit be fore drift ing NNW, and dur ing the early af - ter noon the ash clouds drifted briefly to the E and then S be fore go ing back to the E by late af ter noon. On the morn - ing of 6 De cem ber the ash cloud rose about a km be fore drift ing N-NW. The emis sion was ac com pa nied by loud roaring noises. Fine ash fall occurred in downwind areas including Rabaul.. There was no sig nif i cant de for ma tion un til 10 De cem - ber. The RVO re ported that loud and con tin ual roar ing was pres ent from 8 De cem ber 2006 un til the morn ing of 9 De - cem ber, when the roar ing be came in ter mit tent. The roar ing ceased on 10 De cem ber and at that time parts of the cal dera un der went a rapid ~ 1 cm up lift. On 11 De cem ber the vol - cano was quiet with very lit tle fume. At 0400 on 12 De cem - ber, a loud ex plo sion oc curred with an airwave which shook houses in Rabaul. This event gen er ated a bil low ing gray col umn that rose to a max i mum of 1,000 m be fore be - ing blown to the E. Fol low ing the 12 De cem ber ex plo sion sub si dence re turned the site?s level to that of 9 De cem ber. Seis mic ac tiv ity con tin ued at low levels. No high frequency earthquake was recorded. Ta ble 2 shows the MODIS ther mal anom a lies ob served dur ing 22 Oc to ber-12 De cem ber 2006 (see BGVN 31:09 for ear lier Oc to ber anomalies). Geo logic Sum mary. The low-ly ing Rabaul cal dera on the tip of the Ga zelle Pen in sula at the NE end of New Brit - ain forms a broad shel tered har bor uti lized by what was the is land?s larg est city prior to a ma jor erup tion in 1994. The outer flanks of the 688-m-high asym met ri cal pyroclastic shield vol cano are formed by thick pyroclastic-flow de pos - its. The 8 x 14 km cal dera is widely breached on the east, where its floor is flooded by Blanche Bay and was formed about 1400 years ago. An ear lier cal dera-form ing erup tion about 7100 years ago is now con sid ered to have orig i nated from Tavui cal dera, off shore to the north. Three small stratovolcanoes lie out side the north ern and NE cal dera r ims of Rabaul . Post-cal dera erup t ions bui lt ba - saltic-to-dacitic pyroclastic cones on the cal dera floor near the NE and west ern cal dera walls. Sev eral of these, in clud - ing Vul can cone, which was formed dur ing a large erup tion in 1878, have pro duced ma jor ex plo sive ac tiv ity dur ing his - tor i cal time. A pow er ful ex plo sive erup tion in 1994 oc - curred si mul ta neously from Vul can and Tavurvur volcanoes and forced the temporary abandonment of Rabaul city. In for ma tion Con tacts: Steve Saunders and Herman Patia, Rabaul Volcanological Ob ser va tory (RVO), De part - ment of Min ing, Pri vate Mail Bag, Port Moresby Post Of - fice, Na tional Capitol Dis trict, Pa pua, New Guinea (Email: 16 Rabaul Smithsonian Institution ? Bulletin of the Global Date Time (UTC) Num ber of Pix els Sat el lite 22 Oct 2006 1220 2 Terra 22 Oct 2006 1520 1 Aqua 27 Oct 2006 1250 1 Terra 16 Nov 2006 1230 1 Terra Ta ble 2. MODIS ther mal Anom a lies for Rabaul vol cano for 24 Oc to ber through 12 De cem ber 2006. Cour tesy of the Hawai?i In sti tute of Geo phys ics and Plan e tol ogy. hguria@global.net.pg); An drew Tupper, Dar win Vol ca nic Ash Ad vi sory Cen tre (VAAC), Bu reau of Me te o rol ogy, Dar win, Aus tra lia (Email: A.Tupper@bom.gov.au); Na - tional Aero nau tics and Space Ad min is tra tion Earth Ob ser - va tory (URL: ht tp : / /ear thobservatory.nasa .gov/ NaturalHazards/); RSAM def i ni tion (URL: http://vul can. wr.usgs.gov/Mon i tor ing/De scrip tions/de scrip tion _RSAM_SSAM.html); HIGP MODIS Ther mal Alert Sys - tem, Hawai?i In sti tute of Geo phys ics and Plan e tol ogy (HIGP), Uni ver sity of Ha waii at Manoa, 168 East-West Road, Post 602, Ho no lulu, HI 96822, USA (URL: http:// modis.higp.ha waii.edu/); Vol ca nic Clouds Group (Email: volcanicclouds@yahoogroups.com; URL: http://groups.ya - hoo.com/group/volcanicclouds/). Likuranga New Brit ain Is land, Pa pua New Guinea 4.95?S, 151.38?E; sum mit elev. 922 m Al though Likuranga vol cano in West New Brit ain is thought to be of Pleis to cene age (i.e. with out ev i dence of erup tion in the past 10,000 years; John son, 1971, 1970a, b), a boy died of car bon-di ox ide (CO2) as phyx i a tion in a hole at Bakada vil lage on the vol cano?s N flank on 21 Sep tem - ber 2006. De tails of the fol low-up in ves ti ga tion came out in a re port of the Rabaul Vol cano Ob ser va tory (Mulina and Taranu, 2006). This re port is a con den sa tion of that work. The event serves as a re minder of threats from gas re lease in vol ca nic re gions, even those ar eas in re pose or un likely to erupt again. In this case, the link age to biogenic ver sus volcanogenic or i gins of the gas re mains equiv o cal. Likuranga?s sum mit is ~ 13 km NNE of Ulawun?s sum mit (fig ure 29). The vol cano and Bakada vil lage ap pear in several Google Earth images (figures 30 and 31). In 2004, a log ging com pany dug a num ber of holes to build la trines but ceased af ter find ing wa ter at shal low depths. The com pany ul ti mately left the area with out re fill - ing the holes, which are be hind some of the re main ing build ings (fig ure 32). A con spic u ous dis turbed area cor re - sponded with the re ported co or di nates of the hole on the zoomed-in im age of the village (?hole,? figure 32). Back ground on gas haz ards. Nat u ral sources of CO2 in clude vol ca nic outgassing, the com bus tion of or ganic mat ter, and the res pi ra tion pro cesses of liv ing aer o bic or - gan isms. CO2 gas is ~ 1.5 times heavier than air at the same tem per a ture and can col lect in de pres sions, and con fined spaces such as caves and build ings. With out wind to ven ti - late an area, the denser CO2 dis places the typ i cal at mo - Volcanism Network, Volume 31, Number 10, October 2006 Likuranga 17 Fig ure 29. Likuranga sits along the N coast of New Brit ain is land (in set map). The larger fig ure com prises a sketch map of im por tant fea tures along the coast from Likuranga to the Sulu Range. By far the most fre quently ac tive and re ported-on vol cano on the map is Ulawun, al though re cent re ports have dis cussed un rest at both Bamus and the Sulu Range (BGVN 31:09) and re gional seis mic ity has been high in 2006. This fig ure was scanned from John son (1970b) and modified. Fig ure 30. The coastal vil lage Bakada on Likuranga?s N flanks. The vil lage, which has few per ma nent res i dents but is used as a safe ha ven by nearby coastal vil lag ers when Ulawun be comes rest less. The linked line seg ments across the im age crudely ap prox i mate the bound ary be tween East and West New Brit ain. Cour tesy of Google Earth. Fig ure 31. A closer view of Likuranga?s N-flank vil lage Bakada. Cour tesy of Google Earth. sphere, caus ing an ox y gen de fi ciency. For adult oc cu pa - tional ex po sure, one US agency rec om mends a ceil ing limit of 3 per cent CO2 for up to 10 min utes. Watanabe and Moritea (1998) stud ied re sponses of rats to var i ous gases, in clud ing CO2. They dis cuss var i ous types of asphyxia and the related diagnoses of causes of death. Al though the main com po nent of vol ca nic gas is usu ally wa ter va por, other com mon vol ca nic gases can en dan ger life and prop erty. These can in clude, as in this case, car bon di ox ide (CO2); and, in an el e vated tem per a ture en vi ron - ment, a mul ti tude of other gas ses such as sul fur di ox ide (SO2), hy dro gen (H2), hy dro gen sul fide (H2S), car bon mon - ox ide (CO), and hy dro gen flu o ride (HF). The main dan gers to health and life re sults from the ef fects of the ac ids and am mo nia com pounds on eyes and re spi ra tory sys tems. The vol ca nic gases that pose the great est po ten tial haz ard to peo ple, an i mals, ag ri cul ture, and prop erty are sulfur dioxide, carbon dioxide, and hydrogen fluoride. Trag edy at Bakada vil lage. On 21 Sep tem ber 2006, an 8-year-old boy, with his mother nearby, went down a large (2.6 m deep and 3.4 m wide) hole. He en tered the hole try - ing to res cue his dog, which had fallen in. Wit nesses re - called that the boy soon started shak ing and screamed for help. A nearby woman went down the hole to res cue the boy and she, too, fell un con scious. Both were pulled from the hole by peo ple at the rim with the aid of a long stick and knot ted rope. The boy was dead; his hands a pale color. The woman was still breath ing but vom ited blood. She was rushed to a health cen ter where she soon re cov ered. It was es ti mated that the woman was in the hole for 15 min utes and the boy some what lon ger (though this es ti ma tion re - mains crude as it could not be con firmed by any one with a watch during the incident). The fol low ing day, 22 Sep tem ber, vil lag ers threw five small an i mals into the hole and noted that they all died im - me di ately. The vil lag ers also re called that the pre vi ous year an em ployee of the log ging com pany at tempted to burn dried veg e ta tion in the same hole and failed, even af ter add - ing waste die sel fuel to assist the process. In ves ti ga tion and con clu sions. On 25 Sep tem ber RVO sci en tists ar rived in Bakada to in ves ti gate the in ci dent. It should be noted that for two days be fore their ar rival there was mod er ate rain fall in the area. The sci en tists found that a frog and a dog were mov ing freely in the hole along side the re mains of the orig i nal five an i mals. The RVO re port did not in di cate when the frog and dog were put into the hole. In ad di tion, a burn ing pa per low ered into the hole con tin - ued to burn on the bottom surface. On 27 Sep tem ber a re turn visit by RVO with in stru - ments per mit ted the mea sure ment of CO2 emis sions from ad ja cent soil. The re sults listed in ta bles 3 and 4 show that the rate of CO2 emis sion var ied, but gen er ally in creased as they ap proached the hole. The in ves ti ga tors con cluded that CO2 in the 2.6-m-deep hole caused the boy to die of as phyx i a tion and the woman at tempt ing to res cue him to en ter a semi-con scious state. It was also noted that whereas air cur rents may keep CO2 con - cen tra tions ac cept ably low on the land sur face, the same 18 Likuranga Smithsonian Institution ? Bulletin of the Global Fig ure 32. Al though some what fuzzy, this zoomed-in view of Bakada shows the Likuranga hole, which was la beled based on co or di nates pro vided in the RVO re port . Cour tesy of Google Earth. Du ra tion (min utes) Con cen tra tion of CO2 (ppm) over pe riod of time (in min utes) as mea sured from vary ing dis tance away from the hole. 1 m E Soil temp=28?C 5 m E Soil temp=28?C 100 m E Soil temp=27.2?C 100 m W Soil temp=29.5?C 250 m W Soil temp=27?C 0.0 2000 1380 1130 910 850 0.5 220 1490 1140 930 860 1.0 2350 1530 1160 950 870 1.5 2400 1570 1190 980 900 2.0 2550 1700 1220 990 920 2.5 2725 1730 1250 1010 940 3.0 ? 1860 1290 ? 970 3.5 ? ? ? ? 990 Ta ble 3. Pre lim i nary CO2 soil flux made with ap proach to the hole where the child died at Bakada vil lage, Likuranga vol cano. The mea sure ments were made ~ 6 days af ter the trag edy, on 27 Sep tem ber 2006. Af ter Mulina and Taranu (2006). Dis tance from the hole Rate of soil CO2 gas flux emis sion Rate at ppm per min ute Rate at ppm per sec ond 1 m 270 5.4 5 m 128 2.14 100 m E 54.2 0.9 100 m W 40.6 0.676 250 m W 41.9 0.698 Ta ble 4. Pre lim i nary CO2 soil flux anal y ses at var i ous dis tances from the hole; as mea sured on 27 Sep tem ber 2006. Af ter Mulina and Taranu (2006). does not hold true for deep holes. A fi nal con clu sion was that ex ter nal fac tors such as rain may be able to wash out trapped CO2 from the air, but the con tin u ing emis sion of the gas from the soil may lead to fur ther ac cu mu la tions during dry spells. The au thors rec om mended that the log ging com pany re - fill all the holes and that knowl edge of this trag edy be made more-widely known to cope with the dan gers of toxic gases in vol ca nic ar eas. The au thors also sug gests that car bon iso - to pic anal y ses be car ried out on the CO2 re leased at Bakada to de ter mine if it is of mag matic or biogenic origin. Ref er ences: John son, R.W., 1971, Bamus Vol cano, Lake Hargy Area, and Sulu Range, New Brit ain: Vol ca nic Ge ol ogy and Pe trol ogy: Bur. Miner. Resour. Aust. Rec. 1971/55. John son, R.W., 1970a, Ulawan Vol cano, New Brit ain: ge ol ogy,pe trol ogy and erup tive his tory be tween 1915 and 1967: Bur. Miner. Resour. Aust. Rec. 1970/21. John son, R.W., 1970b, Likuruanga vol cano, Lolobau Is land, and as so ci ated vol ca nic cen tres, New Brit ain: ge ol - ogy and pe trol ogy: Bur. Miner. Resour. Aust. Rec. 1970/ 42. Mulina, K., and Taranu, F., 2006, Gas re lated deaths at Bakada vil lage in side Likuranga vol cano, West New Brit - ain on 21st Sep tem ber 2006, re port of Rabaul Vol cano Ob ser va tory. Watanabe, T. and Morita, M., 1998, As phyxia due to ox y gen de fi ciency by gas eous sub stances: Fo ren sic Sci ence In ter na tional, v. 96, no. 1, p. 47-59. In for ma tion Con tacts: Rabaul Volcanological Ob ser - va tory (RVO) (see Rabaul). Mi chael Saunders Is land, South Sand wich Is lands 57.78?S, 26.45?W; sum mit elev. 990 m All times are lo cal (= UTC - 2 hours) Matt Pat rick sent a new Ad vanced Spaceborne Ther mal Emis sion and Re flec tion Ra di om e ter (AS TER) im age, col - lected 28 Oc to ber 2006 over Saunders Is land . In his opin - ion this is the best im age col lected to date ow ing to the lack of a plume ob scur ing the sum mit crater, which was a prob - lem in all pre vi ous im ages. The im proved im age pro vides a clear view of the crater (figures 33 and 34). An a lyz ing the VNIR, SWIR, and Ther mal In fra red (TIR) (not shown in fig ures 33 or 34) im ages to gether shows that the outer crater is 500-600 m wide, with a 180m high-tem per a ture crater in te rior. The lat ter shows up as an SWIR anom aly and may in di cate the rough ex tent of ac tive lava flow be ing ~ 180 m wide. Matt Pat rick chose Villarrica vol - cano in Chile for com par i son to Mt. Mi chael (fig ure 35) since it pres ents a po ten tially good an a - logue in terms of mor phol ogy and ac tiv ity style. Max i mum ra di ant heat flux val ues were sim i lar (up to ~ 150 MW), sug gest ing that the max i mum in ten sity of ac tiv ity may be sim i lar . Mt. Mi chael shows a much lower fre quency of ther mal alerts, which may be the re sult of more fre quent cloud cover in the South Sand wich Is - lands or a greater depth to molten lava in the Mt. Michael crater. Ta ble 5 shows a sum mary of ther mal anom a lies and pos si ble erup tions from Mod er ate Res o lu - tion Imag ine Spectroradiometer (MODIS) sat el lites since No vem - ber 2005. The last re ported ac tiv - ity of Mount Mi chael was noted in the SI/USGS (Smith so nian In - sti tu tion/U.S. Geo log i cal Sur vey) Weekly Vol ca nic Ac tiv ity Re port of 12-18 Oc to ber 2005 (see BGVN 31:04). At that time the first MODVOLC alerts for the vol cano since May 2003 in di cated an in creased level of ac tiv ity in the is land?s sum mit crater and a pre sumed semi-per ma nent lava lake that ap peared con fined to the sum mit crater . Those aler ts occurred on 3, 5, and 6 October 2005. Volcanism Network, Volume 31, Number 10, October 2006 Mi chael 19 Fig ure 33. An AS TER im age of Mt. Mi chael cre ated us ing en ergy in the vis i ble near-in fra red wave length (?VNIR?; bands 3-2-1, RGB), with the in set show ing a closer view of the sum mit crater. There are two small near-IR anom a lies (band 3, 0.807 mi crons wave length) in the oth er wise dark cen ter of the crater, shown as red spots in the col ored im age. The two anom a lies sug gest very high tem per a tures and sup port the idea that fresh lava may re side at the sur face or a shal low level in the crater. Cour tesy of Matt Pat rick. Ref er ences: Lach lan-Cope, T., Smellie, J.L., and Ladkin, R., 2001, Dis cov ery of a re cur rent lava lake on Saunders is land (South Sand wich Is lands) us ing AVHRR im ag ery: Jour nal of Vol ca nol ogy and Geo ther mal Re - search, vol. 112, no. 1-4, p. 105-116 (au thors are mem bers of the Brit ish Antarctic Survey). LeMasurier, W.E., and Thomson, J.W. (eds), 1990, Vol ca noes of the Ant arc tic Plate and South ern Oceans: Amer i can Geo phys i cal Un ion, Wash ing ton, D.C., AGU Mono graph, Ant arc tic Re search Series, v. 48. Geo logic Sum mary. The young con struc tional Mount Mi chael stratovolcano dom i nates gla cier-cov ered Saunders Is land. Sym met ri cal 990-m-high Mount Mi chael has a 700-m-wide sum mit crater and a rem nant of a somma rim to the SE. Tephra lay ers vis i ble in ice cliffs sur round ing the is - land are ev i dence of re cent erup tions. Ash clouds were re - ported from the sum mit crater in 1819, and an ef fu sive erup tion was in ferred to have oc curred from a north-flank fis sure around the end of the 19th cen tury and be gin ning of the 20th cen tury. A low ice-free lava plat form, Blackstone Plain, is lo cated on the north coast, sur round ing a group of for mer sea stacks. A clus ter of par a sitic cones on the SE flank, the Ashen Hills, ap pear to have been mod i fied since 1820 (LeMasurier and Thomson, 1990). Va por emis sion is fre quently re ported from the sum mit crater. Re cent AVHRR and MODIS sat el lite im - ag ery has re vealed ev i dence for lava lake activity in the summit crater of Mount Michael. In for ma tion Con tacts: Matt Pat rick, Mich i gan Tech no log i cal Uni ver sity, Hought on, MI (Email: mpatrick@mtu.edu); Ther mal Alerts Team, Hawai?i In sti tute of Geo phys ics and Plan e tol ogy (HIGP), School of Ocean and Earth Sci ence and Tech nol ogy (SOEST), Uni ver sity of Hawai?i, 2525 Cor rea Road, Ho no lulu, HI 96822, USA (URL: http :/ / hotspot.higp.ha waii.edu/ ); John Smellie, Brit ish Ant arc tic Sur vey, Nat u ral En vi ron ment Re search Coun cil, High Cross, Madingly Road, Cam br idge CB3 0ET, United King dom (URL: http:// www.antarctica.ac.uk/, Email: jlsm@pcmail.nerc-bas.ac.uk); AS TER Sci ence Pro ject Teams, United States and Ja pan (URL: http://asterweb.jpl.nasa.gov/team. asp and http://www.science.aster. ersdac.or.jp/en/science_info/). 20 Mi chael Smithsonian Institution ? Bulletin of the Global Fig ure 34. The AS TER Short Wave In fra red (SWIR; band 9, 2.4 mi crons) im age with a con spic u ous anom aly at the sum mit, with nu mer ous sat u rated pix els. Cour tesy of Matt Pat rick. Fig ure 35. The real-time sat el lite ther mal mon i tor ing (MODVOLC) ra di ant heat flux val ues for Mi chael and Villarrica vol ca noes dur ing the pe riod 2000-11 No vem ber 2006. Cour tesy of Matt Pat rick. Date Time (UTC) Num ber of pix els Sat el lite 01 Nov 2006 0125 1 Terra 31 Oct 2006 1600 1 Aqua 21 Oct 2006 1120 1 Terra 20 Oct 2006 0250 2 Aqua 20 Oct 2006 0100 3 Terra 21 Jul 2006 0120 1 Terra 09 Jun 2006 0920 2 Aqua 21 Jan 2006 0100 1 Terra 20 Dec 2005 0100 1 Terra 06 Oct 2005 0115 1 Terra 05 Oct 2005 0220 1 Aqua 03 Oct 2005 0045 1 Terra Ta ble 5. Ther mal anom a lies mea sured by MODIS sat el lites for Mount Mi chael for the pe riod 3 Oc to ber 2005 to 1 No vem ber 2006. All of the anom a lies ap peared on the SW side of the vol cano. Cour tesy of Hawai?i In sti tute of Geo phys ics and Plan e tol ogy (HIGP) Ther mal Alerts Team.