Bul letin of the Global Vol can ism Net work Vol ume 28, Num ber 5, May 2003

Ana ta han (Mari ana Is lands) Nearly continu ous ash plumes through May ...... 2

Chiku ra chki (Kur ile Islands) Eruption contin ued through May; long plumes and some ashfall ...... 5

Karym sky (Kamchatk a) Fre quent ash plumes gener ated from Oc tober 2002 through May 2003 ...... 6

Har- Togoo (Mon go lia) Fu ma roles and mi nor seis mic ity since Oc to ber 2002 ...... 7

Mayon (Phil ip pines) Three small ash- and- steam ex plo sions dur ing April- May 2003 ...... 9

Karan ge tang (In do ne sia) Ash explo sions from January through May 2003 ...... 10

Lokon- Empung (In do ne sia) In creased ex plo sive ac tiv ity dur ing January- April 2003; lo cal ash fall...... 11

Ruapehu (New Zea land) Steam plume issued from warm Crater Lake in May, but no eruption ...... 12

Mon owai Sea mount (Ker madec Island s) Volcanic earthquake swarm April-May detec ted by T-wave s ....13

Piton de la Fournaise (Réunion Island) Eruption on 30 May gener ates lava flows within Dolomie u crater ..14

Strom boli (It aly) Lava ef fu sion con tin ues through mid- June; in fra red sat el lite ob ser va tions ...... 15

Nyi ra gongo (DR Congo) 2002-3 lava lake activity, thermal radia tion, and CO2 and SO2 emis sions...... 16

Ro bledo (Ar gen tina) Sat el lite sur veys dur ing May 1996- October 2000 in di cate sub si dence ...... 22

Utu runcu (Bo livia) De for ma tion de tected by sat el lite sur veys; low- level seis mic ity and ac tive fu ma roles...23

Sa ban caya (Perú) Infla tion at Hualca Hualca de tected by satel lite survey s from June 1992 to April 1996 ..23

Santa Maria (Gua te mala) La hars dur ing January- October 2002; ex plo sions and py ro clas tic flows...... 24

Ki lauea (Ha waii) Contin ued lava flows during De cem ber 2002- June 2003 enter the ocean ...... 25

Edi tors: Ed ward Venzke, Jen nifer Fela, Rick Wun derman, and Gari May berry Vol unteer Staff: Jac que lyn Gluck, David Charvo nia, Rob ert An drews, Jessica Ball, and Ashley McCleaf

Global Vol canism Program · National Museum of Natu ral His tory, Room E-421 · Washing ton, DC 20560-011 9 · USA Tele phone: (202) 357- 1511 · Fax: (202) 357- 2476 · Email: gvn@vol cano.si.edu · URL: http://www.vol cano.si.edu/

Subscrip tions are provided by the Ameri can Geo physical Union (see the box on the last page for details). Data are prelimi nary and sub ject to change; contact the original source or the Global Volcan ism Program bef ore using. 2 Ana ta han Smithsonian Institution — Bulletin of the Global

stages of the erupti on, with one ash advi sory indi cat ing ash Ana ta han to 13.4 km alti tude on the 11th. At times multi ple clouds were mov ing in dif fer ent di rec tions at dif fer ent al ti tudes. Mari ana Is lands, cen tral Pa cific Ocean On 13 May Joe Kaipat from the CNMI Emergency 16.35°N, 145.67°E; sum mit elev. 788 m Manage m ent Offic e (EMO) and seism ologist Doug We ins All times are lo cal (= UTC + 10 hours) (Washing ton Uni versit y) flew to Sarigan (6.5 km W of Anata han) to retri eve seis mic data from a broadband instru - The explo sive erupti on that began on 10 May is the first ment insta lled on 6 May. Records from the Sarigan stati on docu mented erup tion from Ana ta han in his tori cal time. showed increa sed seism icit y com mencing at about 1300 on There were no residents on the island due to their evacua- 10 May. The acti vit y re mained very strong for about 36 tion follow ing a shallow earthquake swarm in 1990 (Moore hours be fore de creas ing. and others, 1994), and another in 1993 (Sako and others, Ac tiv ity dur ing 13- 30 May 2003. A helicop ter over- 1995). Anata han is a com posit e volcano that erupts lavas flight on 13 May showed that the is land was still erupting, that are primar ily dac itic in compo si tion. It has the larg est but with less inten sit y than on 11 May. Large vol canic cal dera of the volca noes in the Com monweal th of the bombs were observe d fly ing high in the air over the cra ter Northern Mariana Islands (CNMI). The presence of this region, and the whole W side of the is land was covered cal dera in di cates that large ex plo sive erup tions are pos si ble. with ash, includ ing the seism ograph site. The village ap- Strong acti vit y conti nued over the next few days (Bul le - peared to have 15-30 cm of ash (figure 1). Ash advi so rie s tin v. 28, no. 4), with high ash plumes seen in satel li te im - for 13-14 May reporte d that a dense ash cloud was drifting agery. The area within ~ 55 km of the island was also W away from the island, but that it was not conti nuous and placed off-li mits to all boats and aircraft not approved by varied in size. Ash plumes through 17 May gener all y the CNMI Emergency Man age ment Of fice (EMO). A drifted NW or WNW. The erupti on clouds through May af- smaller but nearly conti nuous erupti on colum n rose from ter the ini tial ac tiv ity were gen er ally be low ~ 6 km. the E cra ter of Ana tahan for the next sev eral weeks. Ac tiv - On 18 May the EMO group took an overfli ght ac com pa- ity was con tinu ing in early July, but at low lev els. nied by David Hil ton (Scripps Insti tuti on of Oceanog ra phy) The EMO in vited USGS sci en tists to pro vide as sis tance and Tobias Fischer (Uni versit y of New Mexico) . They re- in track ing the vol cano’s ac tiv ity and as sess ing po ten tial ported a rising plume com prised of steam and ash. The hazards shortly af ter the erupti on began. USGS sci enti sts cloud was much lighter in color, with a reduced ash com po - first arrive d in Saipan on 30 May to work direct ly with nent com pared to the initi al phases of the erupti on. Bombs, EMO of fi cials to in stall and main tain moni tor ing equip- possi bly up to several meter s in size, were being tossed into ment and inter pret data from overfli ghts and a single seis- the air; most fell back into the E crater. The ash was be ing mome ter op er at ing on Ana ta han. This sta tion be came op- blown W, but most of the ashfal l was still on the E side of erati onal on 5 June. the island. The team landed on the E side of the island and Begin ning of the erupti on, 10-12 May 2003. On 6 May de ployed a PS- 2 seis mometer that ap pears to have record ed re search ers from Wash ing ton Uni ver sity, Scripps In sti tu- earthquakes and some tremor. At that site they found ejecta tion of Oceanog ra phy, and the EMO aboard the MV Super thought to be from the initi al stage of the erupti on. The Em er ald deploy ed a seis mograph on Anata han as part of a ground/vegeta ti on near and under the ejecta was not joint US-Japa n Mariana Subduc - tion Im ag ing Ex peri ment, which is funded by the Nati onal Science Founda ti on. There were no indi - cati ons of an im pending erupti on. During the night of 10-11 May the ship was again approach ing Ana ta han when sci en tists ob- served a trem endous lightning display ahead. As morning broke, they saw a pil lar of steam and ash bil lowing to an al ti tude of 9 km. The ship had to detour around the island to avoid the ashfal l. Ini tial re ports in di cated that the erupti on began around 2100 on 10 May. Satel lite data in ter- preted by the Washing ton Vol- canic Ash Advi sory Center (VAAC) showed that the erupti on appeare d to have started by 1730. An ash plume was clearly visi ble in im agery at 2232, result ing in an advi sory be ing issued to the avia- tion com munit y at 2300 (1300 UTC). Plume heights were re- Figure 1. The village on Anata han covere d with ash, 13 May 2003. The recentl y deploye d seism o graph is barely ported to be 10- 12 km in the early visible in the clearing to the left. Note the ash on the roofs. Courtesy of Doug Weins. Volcanism Network, Volume 28, Number 5, May 2003 Anata han 3 scorched. Most of the mate ria l appeare d to be non- ju ve- with three main layers. The low er most layer con sisted of nile. The largest fragm ents were up to 50 cm across. The ~ 5 cm of fine- grained ash. Next was a layer ~ 15 cm thick team heard “booms” coming from the crater. compr ised of ac cre tion ary lap illi with some fine ash. At the The on go ing ex plo sive ac tiv ity ex ca vated a deep cra ter top was a 5- cm- thick layer that was a mix ture of coarser within Ana ta han’s E cra ter. Sci en tists es ti mated the in ner grained ash and an gu lar clasts of sco ria ceous mate rial. The cra ter was nearly at sea level by about 20 May; be fore the abandoned vil lage, where a team led by Pat rick Shore erupti on, the floor of the E crater was 68 m above sea level. (Washing ton Uni versit y) was working on the seism ic sta- On 20 May the EMO group took an overfli ght and insta lled tion in stalled on 6 May, was similarly cov ered in ash with a telem etered seis mic station. Pres sure waves from detona - many buildings having col lapsed roofs. Two sec tions also tions in the E cra ter were felt on the E flank. From a helicop ter the team also ob served rocks sev eral meter s across be ing thrown up above the E cra ter rim and fal ling back into the cra ter. Ash conti n- ued to fall on the west ern two- thirds of the island and out to sea. The ash cloud size and length was vari able during 17-23 May; it conti nued in general to drift WNW from the is land, at times spread ing over a wide area. On 23-24 May, ty phoon Chan- hom shifted the prevail ing east winds to the S, blowing the erup tion col umn to ward Sai pan and Guam. Light ashfal l re sulted in flight can cel la tions and the clo - sure of the Saipan and Guam in- ter na tional air ports. Resi dents of Sai pan reporte d a rot ten egg smell as so ci ated with the ash fall. The report from Sai pan was that 1-2 mm of ash had fallen on the is land. EMO person nel took an over- flight on 27 May and reporte d that ash cloud heights reached 3 km, signifi cantly lower than during the first few days of the erupti on. The ash cloud was more opaque and laden with ash; the color was closer to that of 10-11 May than more recent plumes. The stream- ing ash cloud, still ex hib iting vari able size and length, drifted NW and NNW through 29 May. Fieldwork on 21 May 2003. Hil ton and Fischer arrive d by ship at Ana ta han at ap proxi mately 0630 on 21 May. The acti vit y level was similar to that on their visit 2 days ear lier. The ship sailed through the ashfal l out to the SW side of the is land, and conti nued along the W coast. The W coast was draped in ash; vege- tation was com pletely cov ered Figure 2. Sum mary of MODIS ther mal alerts detect ed at Anata han, 11-28 May 2003. Each dot defines the giving the is land a gray pallor. geodetic loca tion of the pixels flagged by the MODVOL C algo rit hm (Wright and others, 2002) as con taining They landed at 0815 and spent ~ 4 volcanic hot-spots. However , al though the coor di nate de scribes the center point of each pixel, the hot-spots could have been locate d anywher e in the square boxes (which por tray the nominal 1-km pixel size of the MODIS hours ashore. A trench through instru m ent.) The shaded circles denote the abso lute lim its within which the volcanic hot-spots respon si ble for the the recent depos it s on the beach anoma lies must have been sited (based on a sta tis tical analysis of long- term hot- spot lo cation sta bil ity at other area ex posed a 25- cm sec tion volca noes). The hot-spot loca tions are refer enced to WGS-84 ellip soid. Map coor di nates are in UTM zone 55 from the present erupti ve phase (north). Courtesy of the HIGP MODIS Therm al Alert Team (http://modis.higp.hawaii .edu). 4 Ana ta han Smithsonian Institution — Bulletin of the Global re vealed ini tial ash, cov ered by ac cre tion ary lap illi, then a mix- ture of ash and sco ria ceous mate - rial. Pumice was float ing in water- collection ves sels by the build ings. From the ship the scien tists set up the COSPEC instru m ent and started a traverse through the plume around 1330. The tele- scope was ori ented ver tically and the ship made a N- to-S tran sect through the volcani c plume at a distanc e of ~ 1.5 km from shore. Sul fur di ox ide (SO2) in the plume was re corded im me di ately. The transec t took 50 minutes unti l no SO2 was be ing detec ted. In addi - tion, they sailed through the ash fallout. During the traverse , the volcano erupted every 5 minutes with a deep resonat ing boom. The width of the volcani c plume was ~ 6 km and its direc ti on was to the SW. From the COSPEC measure - Figure 3. Ash plume from Anata han (in dicate d by ar rows) visible in MODIS im agery from the Aqua satel lite , ments and wind speed data pro- 0320 UTC on 30 May. Im age proc essed by NOAA with data from NASA. Cour tesy of NOAA/NASA. vided by NOAA, the SO2 flux was esti mated to be 3,000-4,500 metric tons/day. As the group sailed away from the island around 1430 there was a very on its long axis was center ed ~ 570 km W of Anata han. An large erupti on with a signifi cantly louder “boom” than had SO2 cloud, again displac ed from the ash, extende d ~ 1,100 been heard previ ously , followe d by a dark billow ing ash- km from a point ~ 510 km W of the vol cano to a point ~ 700 laden plume. km SE of it. This cloud contai ned ~ 110 kt of SO2. On 13 MODIS ther mal ob ser va tions. Ther mal sat el lite ob ser - May a data gap covered the Marianas though ash was de- vati ons of the current erupti on of Anata han provided by the tected far ther W, with no sig nificant new SO2 evident. On HIGP MODIS Ther mal Alert Team (http://modis.higp.ha- 14 May a low-leve l SO2 plume appeare d to be drift ing W waii.edu) con firmed that activ ity was heav ily con cen trated from Ana ta han. in the E crater (figure 2). The most re cent hot-spot (as of As of May 30 the Earth Probe TOMS instru m ent conti n - 1700 UTC on 28 May) was observe d on 24 May. The large ued to de tect signifi cant SO2 emissions from Anata han. No amounts of ash produced during the erupti on will have ob - TOMS data were collec ted dur ing 15-23 May due to a tech - scured some therm al anomalies from the MODIS sen sor. ni cal fault on the space craft. There af ter, TOMS detected Plumes were clearly visible on MODIS im agery on 14, 21, SO2 clouds in the region of Ana tahan on 24 May (~ 19 kt 22, 25, 26, 28, and 30 May (figure 3). The persis tent, long SO2), 25 May (~ 23 kt mini mum), 26 May (~ 35 kt), 28 plume from this is land vol cano was frequentl y detec ted in May (~ 70 kt), and 30 May (~ 50-100 kt). Data gaps cov- imager y from a wide va ri ety of sat el lite plat forms. ered the Marianas on other days. Given the persis tent ash SO2 data from TOMS. Sim on Carn reporte d that the plume from the vol cano re ported by the Washing ton Earth Probe Total Ozone Mapping Spectrom eter (EP VAAC, these SO2 clouds are pre sumed to be the prod uct of TOMS) has observe d SO2 and ash emissions from the on- conti nuous emiss ions and not disc rete explo s ive events. going erupti on. No emissions were de tected in the EP Ob serva ti ons from 20 May-8 June 2001. Ana ta han was TOMS overpass at 0116 UTC on 10 May, several hours be- visit ed during 20 May-8 June 2001 as part of fieldwork in fore the reporte d erupti on onset . On May 11 a data gap over the Northern Mari anas (Trusdell and others, 2001), includ - the Mari anas prevente d detec tion of proxim al emissions , ing helicop ter obser va ti ons on 4 June. At that time line though a small ash cloud (no larger than ~ 120 km across) lengths on the Anata han EDM network were meas ured and was de tected ~ 500 km ESE of Ana tahan at 0027 UTC. showed no signifi cant changes. Most line lengths exhib it ed Washing ton VAAC esti mates suggest ed a height of 14-15 small contra cti ons when com pared to the data from the km for this cloud. A weak SO2 cloud was also observe d, 1994 survey. Defor m ati on appeare d to be slowing down displac ed from the ash cloud and center ed ~ 560 km SE of with no sig nifi cant changes. Tempera tures were meas ured Anata han. This cloud contai ned an esti mated SO2 mass of for several boil ing pots and springs on the floor of the E ~ 10 kilotons (kt), but it is suspect ed to be only the distal cra ter. The tempera ture of the ponds as well as fu maroles end of a larger SO2 plume obscure d by the data gap. Diffuse ranged from a mini mum of 96.7°C to a maxim um of ash was also appar ent at least 500 km W of the vol cano at 100.3°C. 0205 UTC, but no meas urable SO2. Back ground. The elongate d, 9-km -long island of Ana- The EP TOMS orbit was better placed on 12 May at tahan in the centra l Mari ana Is lands consis ts of two coa- 0115 UTC. At this time an ash cloud extend ing ~ 560 km lescing vol canoes with a 2.3 x 5 km, E-W- trending sum mit Volcanism Network, Volume 28, Number 5, May 2003 Chiku ra chki 5 de pres sion formed by over lap ping sum mit cal deras. The larger western caldera is 2.3 x 3 km wide and extends east- Chiku ra chki ward from the sum mit of the west ern vol cano, the is land’s 788 m high point. Ponded lava flows overlai n by py rocla s- Kuril Is lands, Rus sia tic de pos its fill the cal dera floor, whose SW side is cut by a 50.325°S, 155.458°E; sum mit elev. 1,816 m fresh-looki ng smaller cra ter. The sum mit of the lower east- All times are lo cal (= UTC + 11 hours) ern cone is cut by a 2- km-wide cal dera with a steep- walled inner cra ter whose floor is only 68 m above sea level. The The erupti on that began on 18 April 2003 (Bul le tin v. sparseness of vegeta ti on on the most re cent lava flows on 28, no. 4) conti nued throughout May and into early June. Ana ta han in di cate their prob able Holo cene age. Accord ing to observ ers, ash fell on the town of Severo- Ref er ences: Moore, R.B., Koyan agi, R.Y., Sako, M.K., Kurilsk (~ 60 km from the vol cano) on 1 May. Observ ers Trusdell, F.A., Kojim a, G., Ellorda, R.L., and Zane, S., from Vasil iev Cape noted weak fum aroli c ac tivit y on 3 1994, Vol cano logic in ves ti ga tions in the Com mon wealth of May and sat ellite data from the USA and Rus sia that day re - the Northern Mari ana Is lands, September-Octobe r 1990: vealed a gas-and- ste am plume more than 150 km long and U.S. Geologi cal Survey Open-Fil e Re port 91-320, 31 p. moving towards the ESE and S. Satel li te data conti nued to Sako, M.K., Trusdell, F.A., Koy anagi, R.Y., Kojim a, show gas- and- steam plumes, pos si bly con tain ing ash, G., and Moore, R.B., 1995, Volcani c inves ti gati ons in the throughout the rem ainder of May (table 1). Sat elli te im ag- Com monweal th of the Northern Mari ana Is lands, April to ing was obscure d by clouds on other days. On 13 May, ash May 1994: U.S. Geologi cal Survey Open-Fil e Re port 94- depos it s were reporte d on the ENE and SSE flanks of the 705, 57 p. volcano and near the sum mit. At 1800 on 15 May, observ - Trusdell, F.A., Sako, M.K., Moore, R.B., Koyan agi, ers on Param ushir Is land reporte d a strong ashfal l at Podg- R.Y., and Schil ling, S., 2001, Preli minary studies of seis- orny set tle ment. mic ity, ground de for ma tion, and ge ol ogy, Com mon wealth During the period 1930 to 2310 on 27 May, Leonid Ko- of the Northern Mari ana Is lands, May 20 to June 8, 2001: tenko on Param ushir Is land reporte d that ash explo sions at- U.S. Geologi cal Survey, prepared for the Offic e of the taining heights of 500 m above the cra ter were observe d Gover nor, the Emergency Manage m ent Offic e, and the Of- from Shelek hov Bay. The ash plume at 0900 on 28 May fice of the Mayor of the Northern Islands , Com monweal th (2200 UTC, 27 May), rose 4,000 m above the cra ter. On 29 of the Northern Mari ana Is lands. May an ash plume rose ~ 1,200 m above the cra ter and ash Wright, R., Flynn, L.P., Gar beil, H., Harris, A.J.L., and fell on the town of Severo-Kuril sk. Pil ger, E., 2002, Autom ated volcani c erupti on detec tion us- Ad di tional in for ma tion about the 2002 erup tion. Pre- ing MODIS: Rem ote Sensing of Envi ron m ent, v. 82, p. vious KVERT reports indi cat ed that the erupti on that began 135- 155. on 25 January 2002 had conti nued through 16 March (Bul- In for ma tion Con tacts: Juan Takai Cama cho and Ra- leti n v. 27, no. 4), but no further reports were made about mon Chong, Com monweal th of the Northern Mari ana Is- that ac tiv ity. How ever, later in for ma tion was re ceived that lands Emergency Manage m ent Offic e, P.O. Box 10007, showed the erupti on conti nuing through at least 22 April. Sai pan, MP 96950 (URL: http://www.cnmiemo.org/, Accord ing to sat elli te data from AVO for 18 March, two Email: juant ca ma cho@hot mail.com); Frank Trusdell , Ha- consec u tive GMS infra red im ages (1732 and 1832 UTC) waiian Volcano Obser va tory , P.O. Box 51, Hawaii Na- showed a narrow, ~ 150-km -long cloud, which extende d tional Park, HI, 96718-0051 (Email: trusdell @usgs.gov, SE from Para mushir Is land. There was no indi ca ti on of ash URL: http://hvo.wr.usgs.gov/cnmi/); Doug Wiens and Pat - based on the split-window technique. On the after noon of rick Shore, Washing ton Uni versit y, St. Louis, McDonnell 20 March, a gas-and- ste am plume with some ash extende d Hall 403 Box 1169, St. Louis, MO 63130 (Email: 200 km SE. Param ushir Is land was obscure d by clouds [email protected]); Al lan Sau ter and David Hilton , Scripps In sti tu tion of Ocean og ra phy, UCSD, 9500 Gil man Es ti mated Drive, La Jolla CA, 92093-0225 (Email: asau- Date Time, UTC Di rec tion Length, km [email protected]); Wash ing ton VAAC, Sat el lite Analy sis Branch (SAB), NOAA/NESDIS E/SP23, NOAA Sci ence 05 May — 50 NW Center Room 401, 5200 Auth Road, Camp Springs, MD 07/08 May — 150 E/SE 20746 USA (URL: http://www.ssd.noaa.gov/); Si mon A. 12 May 0019 20 SE Carn, TOMS Volcani c Emissions Group, Joint Center for 12 May 0449 156 E Earth Sys tems Tech nol ogy (NASA/UMBC), Uni ver sity of Mary land Bal timore County, 1000 Hilltop Cir cle, Bal ti- 13 May 0043 100 E more, MD 21250, USA (Email: [email protected], URL: 13 May 0102 70 SE http://skye.gsfc.nasa.gov/); Rob Wright, Luke Flynn, Har- 13 May 0200 50 E old Garbeil , Andy Har ris, Matt Patrick , Eric Pil ger, and 13 May 0423 178 E/SE Scott Rowland , Hawai’ i Insti tute of Geophys ics and Plane- 13 May 0639 400 E/SE tol ogy, Uni versit y of Ha wai’i, 2525 Correa Road, Hono- 17 May — 50 SW lulu, HI 96822, USA (Email: [email protected] .edu, 18 May — 50 NE URL: http://modis.higp.ha waii.edu); George Stephens , Op - 21 May — 10 NW era tional Sig nifi cant Event Imagery (OSEI) team, World 27 May 0600 100 NE Weather Bldg., 5200 Auth Rd Rm 510 (E/SP 22), 27 May 2200 100 NE NOAA/NESDIS, Camp Springs, MD 20748USA (Email: 29 May AM 15-20 NE [email protected]). Table 1. Sat el lite data re ports of gas- and- steam and ash plumes emanat ing from Chikura chki. Courtesy of KVERT. 6 Karym sky Smithsonian Institution — Bulletin of the Global during the next 2 weeks. On 6 May L. Kotenko (A KVERT plosions to heights of 1,000 m above the volcano (~ 2,500 contac t on the is land) reporte d that hunters had observe d m alti tude) and gas blow-outs. A faint 10-km -long plume fresh ash depos it s on the SW flank on 22 April and that ash- ex tend ing SSE was visible in an AVHRR sat el lite image; fall was also noted in Severo-Kuril sk. no ash was detec ted. Seism icit y on 25-26 Octo ber indi cat ed Back ground. Chikura chki, the highest volcano on Para- pos si ble vig or ous gas emis sions last ing 5-10 min utes, with mushir Is land in the north ern Kur ils, is ac tu ally a rela tively the probabil ity of a lava flow. At 1350 on 31 Octo ber, pi lots small cone construc ted on a high Pleis tocene volcani c edi - reporte d that an ash plume rose 4 km and extende d SE. Ac - fice. Oxi dized scoria depos it s cover ing the upper part of the cording to seism ic data from the Kam chatka Experi mental young cone give it a dis tincti ve red color. Lava flows from and Me thodi cal Seis mol ogi cal De part ment (KEMSD), the 1816-m -high Chikura chki reached the sea and form capes char ac ter of seis mic ity af ter 1400 on 31 Oc to ber in di cated a on the NW coast; several young lava flows also emerge mov ing lava flow. At 1314 on 31 Oc tober, the MODIS sat- from beneat h the scoria blanket on the eastern flank. The ellite im age showed a large bright ther mal anomaly at the Tatari nov group of six volcani c center s is locat ed im medi - volcano and a plume ~ 60 km long that extende d WSW. At ately to the south of Chikura chki. In contra st to the fre- 1100 on 1 Novem ber, pi lots reporte d that an ash plume rose quently ac tive Chikura chki, the Tatarinov volca noes are ex- 4 km and extende d SE. tensive ly modifie d by erosion and have a more com plex Ac tiv ity dur ing No vem ber 2002. Lo cal shal low seis mic structure . Tephro c hronol ogy gives evide nce of only one events total ed ~ 200-250 each day. The charac ter of the erupti on in histori cal time from Tatari nov, al though its seis mic ity in di cated ash- and- gas ex plo sions to heights of southern cone contai ns a sulfur-encrus ted crater with fum a- 1,000-2,000 m above the volcano and vigor ous gas emis- roles that were ac tive along the margin of a cra ter lake un til sions lasting 5-10 minutes . At 1605 on 1 Novem ber, a 50- 1959. km-long plume was observe d extend ing E in sat elli te im- In for ma tion Con tacts: Olga Gir ina, Kamchatka Vol- agery; no ash was de tected. Accord ing to data from canic Erupti ons Re sponse Team (KVERT), a coop era ti ve KEMSD, at 2357 on 20 Novem ber, a seism ic event last ing program of the Insti tute of Vol canic Geol ogy and Geo- 20 minutes indi cat ed that ash explo sions to heights of 1,000 chemis try, Far East Divi sion, Rus sian Acad emy of Sci- m above the cra ter and hot avalanc hes possi bly occurred. ences, Piip Ave. 9, Petropavlovsk-Kam chatskii 683006, On 27 Novem ber, a > 100-km gas-and-st eam plume ex- Russia (Email: girina@ kcs.iks.ru), the Kam chatka Experi - tending ESE from the crater of the volcano was observe d in men tal and Me thodi cal Seis mol ogi cal De part ment MODIS sat el lite im agery. Heli cop ter ob ser va tions by (KEMSD), GS RAS (Russia ), and the Alaska Volcano Ob- KVERT sci enti sts at 1151 on 1 Decem ber identi fied an ash ser va tory (USA); Alaska Vol cano Ob ser va tory (AVO), a plume to ~ 500 m above the cra ter extend ing SE. co op era tive pro gram of (a) U.S. Geo logi cal Sur vey, 4200 Ac tiv ity dur ing De cem ber 2002. Lo cal shal low seis mic Uni ver sity Drive, An chor age, AK 99508- 4667, USA events total ed ~ 190-230 each day. The charac ter of seis- (URL: http://www.avo.alaska.edu/; Email: micit y indi cat ed that ash-gas explo sions to heights of 1,000 [email protected]), (b) Geo physi cal In sti tute, Uni ver sity of m above the volcano (~ 2,500 m altitude) and vigor ous gas Alaska, P.O. Box 757320, Fairbanks, AK 99775- 7320, emis sions last ing 5-10 min utes were pos si bly oc cur ring. USA (Email: [email protected]), and ©) Alaska Di- The top of the volcano and its SE flank were covered with vision of Geologi cal and Geophysi cal Surveys , 794 Univer - re cent ash fall and de bris from con tinu ing Vul ca nian / sity Ave., Suite 200, Fairbanks, AK 99709, USA (Email: Strom boli an erupti ons. The old crater was covered by the [email protected]). new cinder-ash cone. On 12 De cem ber, two sec tors of ash falls extend ing S and SE from the vol cano were noted in a MODIS sat el lite im age. Ac tiv ity dur ing Janu ary 2003. Lo cal shal low seis mic Karym sky events total ed ~ 110-200 each day. The charac ter of seis- micit y indi cat ed that ash-gas explo sions to heights of 1,000 Kam chatka Pen in sula, Rus sia m above the volcano (~ 2,500 m or 8,200 ft. ASL) and vig- 54.05ºN, 159.43ºE; sum mit elev. 1,536 m orous gas emissions lasting 5-10 minutes were pos sibly oc- All times are local (=UTC + 11 hours, + 12 hours after 26 curring. From 1559 unti l 1609 on 8 January , a se ries of Oc to ber) shal low events that possi bly indi cat ed hot avalanc hes were regis ter ed. On 9 January , a ~ 50-km plume extend ing ESE Ac cord ing to the Kam chat kan Vol canic Erup tions Re- from the vol cano was noted. sponse Team (KVERT), the alert level Color Code re- Ac tiv ity dur ing Feb ru ary 2003. The alert level Color mained at Yellow (vol cano is restles s; erupti on may oc cur) Code rem ained at Yellow unti l 27 Febru ary , when it was from Octo ber 2002 to 27 Febru ary 2003, when it was lowered to Green (vol cano is dorm ant; norm al seis micit y dropped to Green (vol cano is dorm ant; norm al seis micit y and fu ma rolic ac tiv ity). Ac cord ing to sat el lite data from and fu marolic ac tiv ity). The level was raised again to Yel - Russia , a weak ther mal anom aly was noted on 3 Febru ary . low in March, lowered to Green on 29 March, and raised to Seism ic acti vit y was at background levels on 20-23 Yellow on 18 April, where it re mained through May. Seis- Feb ru ary. micit y was above background levels unti l 20 Febru ary , af- Ac tiv ity dur ing March 2003. The alert level Color Code ter which it fluc tuate d betwee n at and above background was raised to Yel low as the ac tiv ity of the vol cano slightly levels unti l 16 May, when seism icit y re mained above back- in creased. Seis mic activ ity was at back ground lev els on ground levels . 13-18 March and slightly above background levels on 19 Ac tiv ity dur ing Oc to ber 2002. From 4 to 31 Oc tober, March when seis mic data indi cat ed possi ble hot ava- ~ 200-250 local shal low seism ic events occurred per day. lanches. Weak volcanic earth quakes were also reg istered The charac ter of the seism icity indi cated ash-an d-g as ex- on this day. Ac cord ing to MODIS- satellite data from Rus - Volcanism Network, Volume 28, Number 5, May 2003 Har-T ogoo 7 sia and the USA, ash depos it s extend ing more than 30 km cano on 19 and 21 May were noted. For the week of 24-30 SW from the vol cano on 17-20 March and gas-ste am May, 280-330 local shal low seism ic events occurred per plumes drifting more than 15 km NW and SW on 18 March day. The char ac ter of the seis mic ity in di cated ash- and- gas and on 20 March, respec tively, were noted. Seis mic activ ity explo sions to heights of 1,000 m and gas blow-outs. A ther- dropped to background levels for the week of 20 March. mal anom aly conti nued to be observe d. On 25-26 May, Ac cord ing to sat ellite data from Rus sia, a weak ther mal gas-and- ste am plumes extend ing 15-115 km SSE from the anom aly was observe d on 25 March, and a gas-and- ste am volcano were noted. Ash de posit s on the snow in a differ ent plume extend ing 10 km ESE was noted on 28 March. Ac- direc ti on from the vol cano were noted on 26-27 May. cording to heli copter ob serva ti ons on 31 March by the In- Back ground. Karym sky, the most acti ve volcano of the sti tute of Volcan ol ogy (IV), Far East Di vi sion, Rus sian Kam chat ka’s east ern vol canic zone, is a sym met ri cal strato - Acad emy of Sci ences, the large old ac tive cra ter of the vol - volcano construc ted within a 5- km-wide cal dera that cano and its black ESE flank were noted, but the new formed during the early Holocene . The caldera cuts the cinder-ash cone was not seen. This cone was proba bly de- south side of the Pleisto cene Dvor volcano and is locat ed stroyed and its products formed ash-deposi ts extend ing outside the north margin of the large mid-Ple istocene > 35 km ESE, which were noted on the 17-18 March Polovinka caldera, which contai ns the smaller Akade mia MODIS- satellite im ages. Nauk and Od no boky cal deras. Most seis mic ity pre ced ing Ac tiv ity dur ing April 2003. The alert level Color Code Karym sky erupti ons has origi nated beneat h Akadem ia was dropped to Green during the week of 29 March-4 Nauk caldera, which is locat ed im medi ate ly south of the April, when seism ic acti vit y was at background levels . Karym sky vol cano. The Holocene lava dome La gerny and Seism icit y rose above background levels during the week the SW edge of the Karym sky cal dera is structur ally related of 18-24 April, when ~ 40-100 volcani c earthquakes per to Karym sky. The caldera enclos ing Karym sky vol cano day were record ed, and the haz ard status was raised to Yel- formed about 7,600-7,700 radio car bon years ago; construc - low. The char ac ter of the seis mic ity in di cated ash- and- gas tion of the Karym sky stratovol cano be gan about 2,000 explo sions up to 1,000 m above the cra ter. Accord ing to years later. The latest erupti ve period be gan about 500 sat elli te data from Russia , ash de posit s up to 35 km or years ago, fol lowing a 2,300-yea r quies cenc e. Much of the longer extend ed in differ ent di recti ons on 19-22 April. Ac- cone is mantl ed by lava flows less than 200 years old. His- cording to observ ers from IV, on 18-24 April occa siona l tori cal erupti ons have been Vul canian or Vulcanian- ash-gas explo sions up to 2,500 m above the cra ter occurred Strombolian with mod er ate ex plo sive ac tiv ity and oc ca- each day, and on 21 April, an ash-gas plume rose 1,500 m. sional lava flows from the summit crater. Seism ic acti vit y was above background levels on 24-27 In for ma tion Con tacts: Olga Gir ina, Kamchatka Vol- April and at background levels on 27-30 April. During 24- canic Erupti ons Re sponse Team (KVERT), a coop era ti ve 26 April 50-100 volcani c earthquakes per day were regis - program of the Insti tute of Vol canic Geol ogy and Geo- tered. The char acter of the seis micity indi cated that three chemis try, Far East Divi sion, Rus sian Acad emy of Sci- erup tion events (pos si bly ash- and- gas ex plo sions and rock ences, Piip Ave. 9, Petropavlovsk-Kam chatskii 683006, avalanc hes) oc curred on 24 April. Accord ing to sat elli te Russia (Email: girina@ kcs.iks.ru), the Kam chatka Experi - data from Russia , wide ash de posit s longer than 35 km and men tal and Me thodi cal Seis mol ogi cal De part ment three narrow ash de posit s less than 5 km long extend ing SE (KEMSD), GS RAS (Russia ), and the Alaska Volcano Ob- and W and SW from the vol cano, respec tively, were noted ser va tory (USA); Alaska Vol cano Ob ser va tory (AVO), a on 25 April and 28-29 April. Accord ing to observ ers from co op era tive pro gram of (a) U.S. Geo logi cal Sur vey, 4200 IV FED RAS, on 24 April, an ash-gas plume rose 2,500 m Uni ver sity Drive, An chor age, AK 99508- 4667, USA above the cra ter. (URL: http://www.avo.alaska.edu/; Email: tmiller@ Ac tiv ity dur ing May 2003. The alert level Color Code usgs.gov), (b) Geo physi cal In sti tute, Uni ver sity of Alaska, re mained at Yel low for the month, with in ter mittent ex plo - P.O. Box 757320, Fairbanks, AK 99775- 7320, USA sive erup tions con tinu ing. Oc ca sional ex plo sions up to (Email: [email protected]), and ©) Alaska Divi sion 1,500 m above the volcano, produc ing ash, were consid ered of Geo logi cal and Geo physi cal Sur veys, 794 Uni ver sity to be pos si ble, as well as ash fall within a few tens of kilo - Ave., Suite 200, Fairbanks, AK 99709, USA (Email: meters. Seis mic activ ity was at back ground lev els dur ing [email protected]). 3-16 May. Ac cording to sat elli te data from Russia , the sum - mit of the volcano was black on 4 May. For the week of 10-16 May, seism ic data indi cat ed that 10 ash-and- gas ex- plosions reached heights up to 1,000 m above the cra ter, Har- Togoo and hot avalanc hes possi bly occurred. Accord ing to sat el- lite data from the USA and Rus sia, a weak 1- pixel ther mal cent ral Mongo l ia anom aly on 14 May, and strips of ash depos it s extend ing 48.839°N, 101.402°E; sum mit ele vati on 1,750 m > 10 km to the S, SSE and SE on 14-15 May were noted. Seism icity was above background levels on 16-30 May. In De cem ber 2002 in for ma tion ap peared in Mon go lian During 18-21 May, 150-320 local shal low events oc- and Rus sian newspa pers and on nati onal TV that a vol cano curred per day. The char acter of the seis micity indi cated in Centra l Mongo li a, the Har-Togoo volcano, was produc - ash-and- gas explo sions to heights of 1,000 m above the ing white va pors and consta nt acousti c noise. Becaus e of volcano, gas blow-outs and hot ava lanches. Ac cording to the poten ti al hazard posed to two nearby set tlem ents, sat el lite data from the USA and Rus sia, a 2- 4- pixel ther mal mainly with re gard to poten ti al blocking of rivers, the Di- anom aly was observe d during 18-22 May. Ash depos it s on rector of the Re search Center of Astron om y and Geophys - snow E and SE of the volcano were noted on 18 May. Gas- ics of the Mongo li an Academ y of Science s, Dr. Bekhtur, steam plumes extend ing up to 45 km NE and N of the vol- or gan ized a sci en tific ex pe di tion to the vol cano on 19- 20 8 Har- Togoo Smithsonian Institution — Bulletin of the Global

March 2003. The sci en tific team also in cluded M. Ulzii bat, seis molo gist from the same Research Cen ter, M. Gan zo rig, the Di rector of the Insti tute of Infor m atics , and A. Iva nov from the Insti tute of the Earth’s Crust, Sibe ria n Branch of the Rus sian Acad emy of Sci ences. Geo logi cal set ting. The Miocene Har-Togoo shield vol - cano is situate d on top of a vast volcani c pla teau (figure 4). The 5,000-yea r-old Khorog (Horog) cone in the Taryatu- Chulutu vol canic field is locat ed 135 km SW and the Quaternary Urun-Dush cone in the Khanuy Gol (Hanuy Gol) vol canic field is 95 km ENE. Pliocene and Quater nary volcani c rocks are also abundant in the vi cinit y of the Holo - cene volca noes (Devyat kin and Smelov, 1979; Logatc hev and others, 1982). Analysis of seism ic acti vit y re corded by a net work of seism ic stati ons across Mongo li a shows that earthquakes of magni tude 2-3.5 are scatter ed around the Har-Togoo volcano at a distanc e of 10-15 km. Ob ser va tions dur ing March 2003. The name of the vol- Figure 5. Photo graph of the second hole sam pled at Har-T ogoo, with cano in the Mongo li an language means “black-pot” and ham mer for scale, March 2003. Cour tesy of Alexei Ivanov . through questi oning of the local inhabi tant s, it was learned that there is a local myth that a dragon lived in the vol cano. The local inhabi tant s also menti oned that marm ots, previ - The tem perature of the gas at the first, sec ond, and third ously abundant in the area, began to migrat e westwards five holes was +1.1, +1.4, and +2.7°C, re spec tively, while air years ago; they are now prac tically absent from the area. tem perature was -4.6 to -4.7°C (meas ured on 19 March Acousti c noise and venting of color les s warm gas from 2003). Re peated meas urem ents of the tem perature s on the a small hole near the sum mit were noti ced in Octo ber 2002 next day gave values of +1.1, +0.8, and -6.0°C at the first, by local resi dents. In Decem ber 2002, while snow lay on sec ond, and third holes, re spec tively. Air tem perature was the ground, the hole was clearly visi ble to local visi tors, and -9.4°C. To avoid bias due to di rect heat ing from sunli ght a sec ond hole could be seen a few meters away; it is un clear the meas urem ents were perform ed under shadow. All whether or not white vapors were noti ced on this oc casion. meas urem ents were done with Chechtemp2 digi tal ther- During the inspec tion in March 2003 a third hole was seen. mometer with preci sion of ±0.1°C and accu racy ±0.3°C. The second hole is locat ed within a 3 x 3 m outcrop of cin - Inside the mouth of the first hole was 4-10- cm -thi ck ice der and pum ice (figure 5) whereas the first and the third with suspended gas bubbles (figure 5). The ice and snow holes are lo cated within massive ba salts. When close to the were sampled in plas tic bottles, melted, and tested for pH holes, consta nt noise re sem bled a rapid river heard from and Eh with digi tal meters. The pH- meter was cali brated by afar. The sec ond hole was cov ered with plas tic sheeting Horiba Ltd (Kyoto, Japan) standard solu ti ons 4 and 7. Wa- fixed at the margins, but the plasti c was blown off within ter from melted ice appeare d to be slightly acidic (pH 6.52) 2-3 sec onds. Gas from the second hole was sam pled in a in com pari son to wa ter of melted snow (pH 7.04). Both pH mechani cally pumped glass sam pler. Analysis by gas chro- values were within neutra l solu ti on values. No prominent matog ra phy, perform ed a week later at the Insti tute of the differ ence in Eh (108 and 117 for ice and snow, respec - Earth’s Crust, showed that ni trogen and atm ospheri c air tively) was re vealed. were the major con stituents. Two digi tal short-peri od three-com ponent stati ons were insta lled on top of Har-Togoo, one 50 m from the de gassing holes and one in a re mote area on base ment rocks, for monitor ing during 19-20 March 2003. Every hour 1-3 mi cro seis mic events with magni tude < 2 were re- corded. All seis mic events were vir tu ally iden ti cal and re sem bled A- type volcano- tectonic earth- quakes (fig ure 6). Arri val differ - ence be tween S and P waves were around 0.06-0.3 seconds for the Har-Togoo stati on and 0.1-1.5 sec onds for the rem ote stati on. Assum ing that the Har-Togoo sta- tion was lo cated in the epicen tral zone, the events were lo cated at ~ 1-3 km depth. Seism ic episodes similar to volcanic tremor s were also recorded (figure 6). Figure 4. Photo graph of the Har-T ogoo volcano viewed from west, March 2003. Courtesy of Alexei Ivanov . Volcanism Network, Volume 28, Number 5, May 2003 Mayon 9

life-thre atening vol canic flows might be channele d by gul lies. Ac tiv ity dur ing April 2003. Follow ing a small ash ex- plosion on 17 March 2003 (Bul le tin v.28, no.3), a brief burst of ash and steam occurred at about 0600 on 5 April. The ash col umn rose to ~ 1.5 km above the sum mit crater before being blown SW. The ex plosion was recorded as a low- frequency vol canic earth quake, sig ni fy ing a shal low source. Prior to the explo sion, the volcano’s seism ic net- work had detec ted three small low-frequenc y vol canic earthquakes and three low-frequenc y short-durat ion har- monic tremor s in the past 24 hours. Elec tronic tilt meters in - dicat ed conti nuing slight infla tion of the edific e. The in- Figure 6. Exam ples of an A-type volcano- tectonic earthquake and volcanic tremor epi sodes recorde d at the Har-T ogoo station on 19 March creases in ac tiv ity strongly in di cated the like li hood of 2003. Cour tesy of Alexei Ivanov . sudden ash explo sions . Although no major erupti on was ex - pected im medi ate ly after the explo sion of 5 April, there was growing evi dence that magma was ascend ing the volcano’s Con clu sions. The ab nor mal ther mal and seis mic activi - con duit. ties could be the result of ei ther hydro ther mal or volcani c Ac tiv ity dur ing May 2003. A small explo sion from the proces ses. This acti vit y could have started in the fall of crater at 0721 on 6 May produced a brownish ash-and- 2002 when they were di rectly observe d for the first time, or steam colum n that rose to ~ 450 m above the sum mit crater possi bly up to five years earli er when marm ots started mi- and was blown SW. The ash-and- ste am colum n rose slowly grat ing from the area. Further studies are planned to inves ti - with mini mal noti ceable force and was not detec ted by the gate the cause of the fu marolic and seis mic activi ties. vol cano’s seis mic net work, in di cat ing a very shal low At the end of a sec ond visit in early July, gas vent ing source. No signifi cant seism icit y occurred prior to the ex- had stopped, but seism icit y was conti nuing. In August there plosion. However, electroni c tilt meter s on the N and S will be a workshop on Russian-Mongol ian coop era ti on be- flanks conti nued to show infla tion. Likewise , a pre cise lev - tween Insti tuti ons of the Russia n and Mongo li an Acade- eling survey on 24 April 2003 showed a general infla tion of mies of Sci ences (held in Ulan-Bat or, Mongo li a), where the the N flank. Alert Level 1 re mained in ef fect. work being done on this vol cano will be present ed. At 1813 on 14 May, a small ash puff was emit ted from Ref er ences: Devyatkin, E.V. and Smelov, S.B., 1979, the sum mit crater . This very brief explo sion caused a small Posi ti on of basal ts in sequence of Ce nozoic sedi ments of volum e of ash and steam to rise less than 100 m above the Mon go lia: Iz ves tiya USSR Acad emy of Sci ences, geo logi- crater and to later be blown NW. The Mayon Resthouse and cal serie s, no. 1, p. 16-29. (In Rus sian).Lo gatchev, N.A., Sta Miseri cor dia seis mic stations re corded the ash puff as a Devyatkin, E.V., Malae va, E.M., and others, 1982, Ceno - small-am plitude event. Prior to the ash explo sion, one zoic depos it s of Taryat basin and Chulut u river val ley (Cen - short-durat ion tremor was re corded. Volcani c gas outputs tral Han gai): Iz ves tiya USSR Acad emy of Sci ences, geo- were nota bly moder ate in vol ume, and the sul fur diox ide logi cal serie s, no. 8, p. 76-86. (In Rus sian). emission rates increa sed from the previ ous 1,824 metric In for ma tion Con tacts: Alexei V. Iva nov, In sti tute of the tons per day (t/d) to ~ 3,088 t/d. The seis mic charac ter isti cs Earth Crust SB, Russia n Academ y of Science s, Irkutsk, as so ciated with the ash and steam emission ap peared simi- Rus sia (Email: ai va [email protected]); Bekh tur and M. lar to, though smaller than, pre vious explo sions since 22 Ulzii bat, Resea rch Center of Astron om y and Geophys ics , Octo ber 2002, indi cat ing that this ash puff was very minor. Mon go lian Acad emy of Sci ences, Ulan- Bator, Mon go lia This as sess ment was also con sis tent with the smaller vol- (Email: bekhtur@rc ag.url.mn and [email protected]); M. ume of ash produced. Gan zo rig, In sti tute of In for mat ics MAS, Ulan- Bator, Mon - Back ground. Beau ti fully sym met ri cal Mayon vol cano, go lia (Email: gan zo rig@ar vis.ac.mn). which rises to 2,462 m above the Al bay Gulf, is the Phil ip- pines’ most acti ve volcano. The structur all y sim ple vol cano has steep upper slopes aver ag ing 35-40° that are capped by a small summit crater. The his torical erup tions of this Mayon basaltic- andesitic vol cano date back to 1616 and range from Strom bo lian to ba sal tic Plin ian. Erup tions oc cur pre- Lu zon, Phil ip pines dominate ly from the centra l conduit and have also pro- 13.247°N, 123.685°E; sum mit elev. 2,462 m duced lava flows that travel far down the flanks. Pyro cla sti c All times are lo cal (= UTC + 8 hours) flows and mudflows have com monly swept down many of the approxi m ately 40 ravines that radi ate from the sum mit The Phil ip pine In sti tute of Vol can ol ogy and Seis mol- and have often devas tat ed populat ed low land areas . May- ogy (PHIVOLCS) reporte d small ash and steam explo sions on’s most violent erupti on, in 1814, killed more than 1,200 from the Mayon volcano on 5 April, 6 May, and 14 May peo ple and dev as tated sev eral towns. 2003. The alert status for the area around the volcano re- In for ma tion Con tact: Phil ip pine In sti tute of Vol can ol- mained at Alert Level 1 on a scale of 0-5 (indi cat ing an in - ogy and Seismol ogy (PHIVOLCS), De part ment of Science creased likeli hood for steam-drive n or ash explo sions to oc- and Technol ogy, PHIVOLCS Building, C.P. Gar cia Ave- cur with lit tle or no warning). PHIVOLCS rem inded the nue, Uni versit y of the Philip pines Cam pus, Diliman, Que- publi c to conti nue avoiding entry into the 6- km-radi us Per- zon City, Phil ip pines (URL: http://www.phivolcs.dost. manent Dan ger Zone (PDZ), es pe cially in the sectors where gov.ph/). 10 Karan ge tang [Api Siau] Smithsonian Institution — Bulletin of the Global

blasting sound was heard, and a dark-gray ash colum n Karan ge tang [Api Siau] reached 1,500 m. Ash fell to the E around Dame and Kar- alung village s, and over the sea. Lava avalanc hes from the Siau Isla nd, Indo ne si a S cra ter travele d 1,000 m toward the Batang and Batu riv- 2.47°N, 125.29°E; sum mit elev. 1,784 m ers. On 20 April another explo sion produced a 1,500-m - high ash colum n, and ash fell E over the sea. This explo sion During 6 January-4 May 2003 ex plosions pro duced ash was followed by lava ava lanches and a py ro clastic flow to - that fell on vari ous parts of the cra ter. The S (main) cra ter ward the Batang river that reached as far as 2,500 m. Lava emit ted “white-gray ash” that reached 150-400 m high. On avalanc hes extende d 1,500 m down the S and W slopes. some nights, a red glow was visible reaching 25-50 m over Blast ing noises oc curred for about 3 minutes . the cra ter. The N cra ter emitted a “white- thin ash” plume On 22 April an explo sion ejected ash and glow ing mate - that reached 50-300 m high. Fluctu at ing seism icit y was rial. The ash colum n reached 1,750 m and ash fell on the W dominate d by multi phase earthquakes and emissions (ta ble slope, includ ing Lehi, Mini, Kinali, and Hi ung vil lages, 2). The Alert Level re mained at level 3 (on a scale of 1 to 4) while glow ing mate ria l rose up to 750 m. This explo sion through at least 4 May. was fol lowed by lava avalanc hes towards the W and S that On 11 and 12 January , ash explo sions at the S crater were ac compa nied by a pyro clastic flow to ward the Batang were ac compa nied by glow ing mate rial that reached 200 m river that extende d 2,250 m. On 24 April, an explo sion high and scat tered 500 m toward the E and W parts of the ejected ash to 750 m and ash fell eastward into the sea. crater . An ash colum n rose up to 500 m above the cra ter. Glowing mate ria l from the explo sion travele d toward the W Two explo sions at the S crater on 14 January produced an slope. During late April, the boom ing noises were once ash colum n up to 300 m high; glowing mate ria l rose up to again ac com pa nied by con tinu ous glow ing ava lanches. 50 m and fell around the cra ter. Some of the mate rial en- These decrea sed during the first days of May. tered the Beha River, and ash fell into the sea E of the is- Back ground. Karange tang (Api Siau) vol cano lies at land. Explo sions on 29 January and 6 Febru ary caused ash - the northern end of the is land of Siau, N of Sulaw esi . The fall SW (Beong vil lage) and SSW (Akesem beka village , 1,784- m- high strato vol cano con tains five sum mit cra ters Ta rurane, Ta ta ha deng, Be bali, and Salili), re spec tively. A along a N-S line. Karange tang is one of Indo ne sia ’s most boom ing noise was heard frequentl y throughout the report acti ve volca noes, with more than 40 erupti ons recorded period, and during early Febru ary was someti mes accom pa- since 1675 and many addi ti onal small erupti ons that were nied by thick gray emissions up to 350 m above the cra ter. not docum ented in the his tori cal record (Catalog of Ac tive Begin ning in early March, the boom ing noise was ac- Volca noes of the World). Twentieth- century erupti ons have com panied by glowing lava avalanc hes that travele d from in cluded fre quent ex plo sive ac tiv ity some times ac com pa- the sum mit towards the Kaheta ng (1,250 m), Ba tuawang nied by pyro cla sti c flows and lahars. Lava dome growth has (750 m), Batang (1,000 m), and Beha (750 m) rivers. On 6 occurred in the sum mit crater s; collaps e of lava flow fronts March an explo sion from the S crater ejected ash 750 m has also pro duced py ro clas tic flows. high that fell in the E part of the cra ter. The noises and ava- In for ma tion Con tact: Vol cano logi cal Sur vey of In do- lanches de creased dur ing mid- to- late March. ne sia (VSI), Jalan Dipo ne goro No. 57, Bandung 40122, In- An explo sion on 15 April was followed by lava ava- do ne sia (Email: dali@vsi.dpe.go.id; URL: http://www.vsi. lanches toward the W and S parts of the crater . A loud dpe.go.id).

Deep vol canic Shal low vol canic Date (2003) Ex plo sions Mul ti phase Emis sions Tec tonic Ava lanches (A- type) (B- type)

06 Jan-12 Jan 11 16 2 178 178 28 — 13 Jan-19 Jan 9 16 2 133 42 40 — 20 Jan-26 Jan 12 37 — 189 52 27 — 27 Jan-02 Feb 6 28 1 228 118 22 — 03 Feb- 09 Feb 17 84 1 162 306 23 — 10 Feb- 16 Feb 9 30 1 85 102 16 — 17 Feb- 23 Feb 9 46 — 97 8 32 — 24 Feb- 02 Mar 48 68 — 78 17 34 — 03 Mar-09 Mar 19 29 1 48 9 24 398 10 Mar-16 Mar 14 11 — 27 7 30 125 17 Mar-23 Mar 24 145 — 82 4 23 4 24 Mar-30 Mar 21 68 — 35 1 33 2 31 Mar-06 Apr 8 83 — 30 — 36 — 07 Apr- 13 Apr 18 143 — 116 6 50 — 14 Apr- 20 Apr 12 257 32 226 26 32 7 21 Apr- 27 Apr 13 373 2 93 6 17 309 28 Apr- 04 May 32 255 — 243 1 21 29 Table 2. Seis micity at Karange tang dur ing 6 January-4 May 2003. Courtesy VSI. Volcanism Network, Volume 28, Number 5, May 2003 Lokon- Empung 11

Ash emissions conti nued to thicken and deep and shal- Lokon- Empung low vol canic earthquakes increa sed during early Febru ary . Emis sion earth quakes also in creased, in di cat ing some low northe rn Sula wesi , Indo ne si a ash explo sions . On 8 Febru ary at 0443 an explo sion ejected 1.36°N, 124.79°E; sum mit elev. 1,580 m ash and glowing mate ria l. A boom ing sound was heard for 30 sec onds. A dense ash cloud reached 1,400 m above the During 6 January-4 May 2003, higher-than- normal ac- cra ter. Ash fell over the S part of the crater and around tivit y was dominate d by deep and shallow vol canic earth- Kayau, Tara-tara I and II, and Woloan II and III village s. quakes (ta ble 3), along with gas-and- ash emissions . Several Ashfal l reached thicknesse s of 0.5-1 mm. The Alert Level ex plo sions oc curred dur ing a pe riod of in creased ac tiv ity in was increa sed from 2 to 3 (on a scale of 1-4). late January-earl y April. Throughout the report pe riod, a Explo sions occurred on 10 Febru ary at 1405 and 2219. “white- thick ash” emission rose 25-500 m above Tom- The maxim um am pli tude of the explo sion earthquakes was paluan cra ter. The Vol cano logi cal Sur vey of In do ne sia 50 mm. The height of the ash colum n could not be observe d (VSI) issued a specia l re port during 1-13 Febru ary 2003 due to heavy rain. Explo sion acti vit y con tinued on 12 and that descri bed ac tivit y in 2002 and early 2003 leading up to 16 Febru ary . VSI reporte d that the Alert Level was in- the re cent in crease in ac tiv ity (table 4). creased to 4 on 12 Febru ary at 0800. From that time On 25 January , there was a felt shock (I on the MMI through 1100 on 12 Febru ary , shal low vol canic earth- scale). Dur ing late Janu ary, ash emissions from the cra ter quakes increa sed to a total of 164. An explo sion fol lowed at thickened and emission earthquakes increa sed. On 3 Febru - 1102, but the ash colum n could not be observe d due to ary the number of deep volcani c earthquakes began to in- heavy rain. Tremor was re corded be ginning on 13 Febru ary crease at 0600; by 1000, 35 had occurred. with am pli tudes of 0.5-38 mm.

Date Deep vol canic Shallow volcanic Emis sion Tec tonic Ex plo sion

06 Jan-12 Jan 03 1 6 10 13 — 13 Jan-19 Jan 03 1 3 — 20 — 20 Jan-26 Jan 03 8 6 4 23 — 27 Jan-02 Feb 03 6 4 31 11 — 03 Feb- 09 Feb 03 239 763 4 9 — 10 Feb- 16 Feb 03 32 23 7 14 4 17 Feb- 23 Feb 03 239 763 4 9 1 24 Feb-2 Mar 03 97 353 52 19 12 03 Mar-09 Mar 03 — 3 185 6 2 10 Mar-16 Mar 03 — — 90 14 — 17 Mar-23 Mar 03 2 4 38 17 — 24 Mar-30 Mar 03 49 335 33 7 1 31 Mar-06 Apr 03 7 130 5 18 1 07 Apr- 13 Apr 03 4 15 86 17 — 14 Apr- 20 Apr 03 44 285 — 17 — 21 Apr- 27 Apr 03 46 98 — 14 — 28 Apr- 04 May 03 25 71 — 24 —

Table 3. Seis micity at Lokon during 6 January-4 May 2003. Courtesy VSI.

Date Event

09 Feb 02 An explo sion ejected ash to ~ 1,000 m above the cra ter. Ash fell on Kakaskasen, Telete, and Ruru kan villages in the Tondano Dis trict in thicknesses of 0.5-2 cm. 10 Apr 02 At 2302 volcanic earthquakes be gan to incre ase, reaching a total of 184 events. An explo sion at the same time ejected ash to ~ 1,000 m and glowing mate ria l to 250 m above the cra ter. Ash fell on some villages in thick nesses of 1-3 mm. 12 Apr 02 At 1816 an explo sion ejected ash to 800 m and glowing mate ria l to 150 m. Ash drifted S and fell around Kayawu village. 23 Dec 02 At 0532 an ex plosion at Tom paluan crater produced an 800- m-high ash colum n. Ash drifted S and fell around the edifice. Before the explo sion, an incre ase in seism icity (130 volcanic earthquakes in less than 12 hours) was noted. 03 Feb 03 Volcanic earthquakes be gan to incre ase, with a total of 255 events oc curring through 7 Feb ruary . 08 Feb 03 Tremor was followed by an explo sion at 0443 that ejected ash to 1,400 m above the cra ter. The ash drifted S and was accom panied by glow ing mate rial. Ash fell around Ta ratara, Waloan, and Kayawu vil lages, at thick nesses of 0.5-1 cm. 10 Feb 03 After two days repose, at 2219 an explo sion occurr ed. The height of the ash colum n could not be observe d due to heavy rain near the sum mit. The explo sion was preceded by a boom ing sound. Based on seism o graph record ings, the explo sion was of medium-high inten sity . Explo sion earthquakes stopped at 2335. A phreatic eruption at 1406 lasted for 8 minutes. 12 Feb 03 A signifi cant incre ase in volcanic earthquakes, mainly during 0100-1000. An ex plosion at 1408 was followed by a larger ex plosion at 1102 (based on seism ic data; vis ual obser va tion obscure d by thick fog). At 1133 the explo sion dim inished. At 1225 continu ous tremor began with am plitudes of 13- 55 mm that contin ued until 0046 on 13 Febru ary . Table 4. Sum mary of a special report of ac tivity at Lokon during 2002-2003. Cour tesy VSI. 12 Ruapehu Smithsonian Institution — Bulletin of the Global

VSI reporte d that during 18-20 Febru ary , there were 16 explo sions and a “white-gray ash” colum n rose 500 m. An Ruapehu explo sion on 22 Febru ary was pre ceded by a swarm of 224 shal low vol canic earthquakes . On 21 Febru ary , 29 deep New Zea land volcani c earthquakes occurred. Within two days, the 39.28°S, 175.57°E; sum mit elev. 2,797 m number of vol canic earthquakes decrea sed graduall y and ended with a large explo sion on 23 Febru ary at 1034. The Since the middle of March 2003 the tem perature of ex plo sion was ac com pa nied by thun der ing and a boom ing Ruape hu’s sum mit Crater Lake had been slowly rising. The sound, and a “thick-gray ash” colum n reached 2,500 m lake tem perature rose from 30°C on 5 March (Bul le tin v. above the cra ter. Ash drifted toward the SE. Tremor (with 28, no. 2) to a high of 41.6°C on 15 May (ta ble 5). Similar an am pli tude of 1-20 mm) began soon after the explo sion. values were recorded in January 2003 when the lake tem- Lokon was at Alert Level 3 dur ing 17-23 Febru ary . perature reached 42°C. This is the fourth time that the tem - During 24 February-2 March, 12 explo sions occurred pera ture of the Cra ter Lake has risen above 35°C since the and a “white-gray ash” colum n rose 300 m. An explo sion start of 2001, and the tem perature has been above 30°C on 2 March at 2129 was accom panied by glowing mate ria l since Decem ber 2002. It is not unusual for the tem perature that fell within the crater . A dark gray ash colum n rose to cy cle over peri ods of 6-9 months; minor hydro ther mal 1,500 m above the cra ter and ash fell toward the Tondano activ ity can oc cur in the lake dur ing tempera ture peaks. area (~ 14.5 km from the crater ) with a thickness of ~ 1 mm. Lake tem perature s dropped steadil y from 41ºC after mid- Tremor (with am pli tudes of 0.5-25 mm) began soon after May. However, during the late morning of 26 May a steam the explo sion. The explo sion had been pre ceded by a plume was observe d rising 200-300 m above Crater Lake. swarm of 204 shal low vol canic earthquakes . A total of 77 No seis mic ity ac com pa nied this plume, sug gest ing that it deep vol canic earthquakes occurred dur ing 26 February-1 was gener ate d by atm ospheri c condi ti ons alone (a warm March 2003. Fol lowing the 2 March explo sion, there were lake and a cold, windles s, morning). Steam plumes were 2 medium-int ensity explo sions that produced a ~ 600-m - also observe d on 28 March and 21 April. high “white-gray ash” colum n. Ash explo sions and emission earthquakes ended on 14 March. On 24 March, the Alert Level was low ered to 2. Date (2003) Cra ter Lake Tem pera ture Nor mal ac tiv ity con tin ued, com prised mainly of “white- thick ash” emissions from Tompaluan crater that reached 05 Mar 30°C up to 300 m. Tremor conti nued with am pli tudes of 0.5-12 28 Mar 35°C mm. 11 Apr 38°C On 27 March at 0156, an explo sion produced a 1,500- 29 Apr 39.4°C m-high ash colum n that was accom panied by glowing ma- 15 May 41.6°C teria l. Boom ing and blast ing sounds were heard. Ash 26 May Slightly over 40°C drifted S and some fell around the edific e, while glowing 29 May 36°C mate ria l reached 400 m high before fal ling around the cra - 01 Jun 33°C ter. Acti vit y was low after the explo sion. Tremor conti nued Table 5. Lake wa ter tem pera tures meas ured at Ruape hu’s Cra ter Lake, 5 with am pli tudes of 0.5-24 mm. March-1 June 2003. Cour tesy of IGNS. Fol low ing an other ex plo sion on 1 April, ac tiv ity at Lokon decrea sed. A “white-thi ck ash” plume conti nued to rise 100-450 m above the cra ter. Seism icit y was dominate d Weak in ter mittent seis mic tremor was re corded through by tremor with am pli tudes of 0.5-25 mm. Shal low vol canic early April, then re mained at a con stant moder ate level dur - earthquakes increa sed on 15 April to 106 events. Through ing 12-17 April. The fol lowing week, 18-24 April, there 20 April, the daily number of shallow vol canic earthquakes was an in crease in tremor ac compa nied by discrete vol canic fluc tuate d betwee n 23 and 56 events, but there were no ex- earth quakes. By 2 May vol canic tremor lev els had declined, plosions . Acti vit y re mained low, but above norm al, through but vol canic earthquakes conti nued to oc cur. Levels of vol- at least 4 May. canic tremor fluc tuate d during the week of 3-9 May, with Back ground. The twin vol canoes Lokon and Em pung, several peri ods of enhance d tremor and small volcani c rising about 800 m above the plain of Tondano, are among earthquakes . Tremor had decli ned by 16 May, and seism ic- the most acti ve volca noes of Sulaw esi . Lokon, the higher of ity re mained very low through the 30th. The level of vol- the two peaks (whose sum mits are only 2.2 km apart), has a canic tremor began to increa se slightly in early June, but the flat, crater less top. The morpho logi cally younger Em pung lake tem perature was still decli ning during the week of volcano has a 400-m -wide, 150-m -deep crater that erupted 7-13 June. Very low levels of acti vit y conti nued through last in the 18th century , but all subse quent erupti ons have the 20th. There were no signifi cant changes observe d in the origi nated from Tom paluan, a 150 x 250 m wide double lake wa ter chemis try. The hazard status rem ained un- cra ter situ ated in the sad dle be tween the two peaks. His tori - changed at Alert Level 1. cal erup tions have pri mar ily pro duced small- to- moderate Back ground. Ruapehu, one of New Zea land’s most ac- ash plumes that have occa sion all y dam aged croplands and tive volca noes, is a com plex stratovol cano construc ted dur - houses, but lava-dom e growth and py rocla sti c flows have ing at least 4 cone-buil ding epi sodes. The 110 cu km vol- also oc curred. canic massif is elon gated in a NNE- SSW di rec tion and is In for ma tion Con tact: Vol cano logi cal Sur vey of In do- surrounded by a 100 km3 ring plain of volcani clasti c debris, ne sia (VSI), Jalan Dipo ne goro No. 57, Bandung 40122, In- includ ing the Murim oto debris-aval anche deposit on the do ne sia (Email: [email protected]; URL: http://www.vsi. NW flank. A serie s of subpli nian erupti ons took place at dpe.go.id). Ruapehu betwee n about 22,600 and 10,000 years ago, but Volcanism Network, Volume 28, Number 5, May 2003 Mon owai Sea mount 13 pyro cla sti c flows have been infre quent at Ruapehu. A sin- mount, rises to within 100 m of the sea surface about half- gle his tori cally ac tive vent, Cra ter Lake, is lo cated in the way betwee n the Kerm adec and Tonga island groups. The broad sum mit region, but at least five other vents on the volcano lies at the southern end of the Tonga Ridge and is sum mit and flank have been acti ve during the Holocene . slightly off set from the Ker madec volca noes. Small para- Fre quent mild- to- moderate ex plo sive erup tions have oc- sitic cones oc cur on the north and west flanks of the subm a- curred in his tori cal time from the Cra ter Lake vent, and rine volcano, which rises from a depth of about 1500 m and tephra char acter istics sug gest that the cra ter lake may have was named for one of the New Zealand Navy bathy metric formed as early as 3000 years ago. Lahars produced by survey ships that docum ented its morphol ogy. Num erous phreatic erup tions from the sum mit crater lake are a hazard erupti ons have been de tected from subm arine acous tic sig - to a ski area on the upper flanks and to lower river val leys. nals since it was first recog nize d as a volcano in 1977. A In for ma tion Con tact: In sti tute of Geo logi cal & Nu- shoal that had been reporte d in 1944 may have been a pum - clear Sci ences (IGNS), Private Bag 2000, Wairakei, New ice raft or water distur bance due to de gassing. Surface ob - Zea land (URL: http://www.gns.cri.nz/). ser va tions have in cluded wa ter dis col ora tion, vig or ous gas bubbli ng, and ar eas of upwel li ng water , someti mes accom - panied by rum bling noises. In for ma tion Con tacts: Do mi nique Rey mond and Olivier Hy ver naud, Labo ra toire de Geo phy sique, Mon owai Sea mount CEA/DASE/LDG Ta hiti, PO Box 640, Papeete, French Poly ne sia (Email: rey mond.d@la bo geo.pf, hy ver naud@la- Ker madec Is lands, New Zea land bo geo.pf). 25.887°S, 177.188°W; sum mit elev. -100 m

Monowai is a frequentl y ac tive subm arine volcano, with a volcani c swarm recorded in Novem ber 2002 (Bul le tin v. 28, no. 2) and another during April-May 2003. A major part Sta tion code Lati tude Lon gi tude of its volcani c ac tivit y is detec ted by hydro-acous tic waves (also called T-waves ) gener ate d during the erupti ons, PAE 17.6619°S 149.5800°W through the Ré seau Sis mique Polynésien (RSP), the French PPT 17.5682°S 149.5761°W Poly ne sian seis mic net work (ta ble 6). PPN 17.5308°S 149.4322°W A strong vol canic swarm locat ed on the Monowai sea- TIA 17.5578°S 149.3458°W mount was recorded during April-May 2003 (figure 7). TVO 17.7825°S 149.2517°W This volcani c swarm was very well locat ed around Mon- MEH 17.8753°S 148.0661°W owai, using the inver sion of the arri val times of T-waves re- PMOR 15.0017°S 147.8942°W corded by the network. As an exam ple of the preci sion of VAH 15.2364°S 147.6272°W lo ca tion, with the con tri bu tion of some IRIS sta tions like TBI 23.3489°S 149.4608°W RAR (Cook Island) to enlar ge the array dim ension, the el- RKT 23.1197°S 134.9733°W lipse of error can typi cally be 13 km on the major axis and 2 Table 6. Seis mic station codes and coor di nates of instru m ents in the km on the minor axis, with a Root Mean Squared (RMS) of French Poly nesian seism ic net work. Courtesy of RSP. 0.25 s. This volcanic swarm was com posed of three episodes last- ing 4-5 days each. It started sud- denly on 10 April 2003 with a rate of 100 events per day (about one signal every 10 minutes ) and reached a maxi mum inten sity later that day. The av er age rate over the first four days was 75 events per day (300 signals be- tween 10 and 14 April), but the number of events de tected is thought to be under es ti mated by a fac tor of at least 3 to 5 be cause only the main packets of recorded T- waves were picked. Vol canic acti vit y started again during 19 April, with 120 events recorded in the next five days. The last epi- sode oc curred be tween 3 and 6 May, with ~ 100 vol canic signals re corded. The swarm ended as sud denly as it started. Back ground. Monowai sea- Fig ure 7. T-wave am pli tude versus time for the TVO seis mic sta tion, show ing the three dis tinct and well sepa rated mount, also known as Orion sea- episodes of the Monowai Seam ount swarm. Cour tesy of RSP. 14 Piton de la Fournaise Smithsonian Institution — Bulletin of the Global

Piton de la Fournaise Réunion Island, In dian Ocean 21.229°S, 55.713°E; sum mit elev. 2631 m All times are lo cal (= UTC + 4 hours)

Erupti ons are com mon at Piton de la Fournais e, with the most re cent acti vit y occur ring in January 2002 (Bul le tin v. 26, no. 12) and November-Dece mber 2002 (Bul le tin v. 27, no. 11). At the end of the Novem ber 2002 erupti on, se- imicity beneat h Dolom ieu crater increa sed from 28 Novem - ber to 23 De cem ber. On 22 De cem ber there were 5,700 seism ic events recorded. At 1002 on 23 De cem ber a magni - tude 3 event oc curred and seism icit y stopped. The next day a new crater was ob served in the SW part of the larger Dolo mieu cra ter. Figure 9. Photo graph of the W part of Dolom ieu crater at Pi ton de la Since March 2003, the exten so m eter network and GPS Fournaise at 0850 on 6 June 2003 showing the active vent and most of the recent lava- flow field. View is to wards the SW. Cour tesy of OVPF. meas urem ents had indi cat ed infla tion of Piton de la Four- naise. A new erupti on began on 30 May within Dolom ieu crater . The erupti on procee ded in multi ple phases through On 4 June at 1155 the erupti on started again from the at least 24 June; ac tivit y through 6 June is reporte d below. same site, enlar ging the lava flow in the W part of Dolo- Seism icit y increa sed slightly on 28 May. At 1137 on the mieu crater . Lava fountai ns reached 15 m in height. Steady morning of 30 May a seism ic crisis began that lasted 17 lava emission conti nued into 6 June (figures 8 and 9). Vol- minutes with a total of 34 events. Tremor appeare d at 1155 canic tremor rem ained sta ble unti l the morning of 6 June, beneat h Dolom ieu crater , and an erupti on started within the when a de creas ing ten dency was noted. Af ter a short pit crater formed on 23 Decem ber 2002. Lava fountai ning phreatic erupti on, the sec ond phase of this erupti on stopped was observe d unti l 1400, after which most surface acti vit y on the evening of 6 June. The lava-flow field had grown to stopped. A lava flow ~ 400 m long and 250 m wide ex- ~ 600 x 400 m in size by that time (figure 10). tended into the W part of Dolom ieu. The total volum e of Back ground. The massive Piton de la Fournais e shield lava emitted dur ing the 30 May ac tiv ity was es timated to be volcano on the French is land of Réunion in the western In- 0.2-0.3 x 106 m3. Seis mic ity be neath the cra ter con tin ued, dian Ocean is one of the world’s most acti ve volca noes. with inter mit tent weak tremor being regis ter ed through 3 Much of its > 530,000 year history overlappe d with erup- June. No de fla tion was de tected, and there was strong de- tions of the deeply dissec ted Piton des Neiges shield vol- gassing in the col lapse area. cano to the NW. Three calderas formed at about 250,000, 65,000, and less than 5000 years ago by progres sive eastward slumping of the volcano. Num er- ous pyro cla sti c cones dot the floor of the cal deras and their outer flanks. Most his torical erup tions have originate d from the sum mit and flanks of Dolom ieu, a 400- m-high lava shield that has grown within the youngest cal dera, which is 8 km wide and breached to be low sea level on the east ern side. More than 150 erupti ons, most of which have produced fluid basal tic lava flows, have oc- curred since the 17th century . Only six erupti ons, in 1708, 1774, 1776, 1800, 1977, and 1986, have origi nated from fissures on the outer flanks of the cal dera. The Piton de la Fournais e Vol cano Ob ser va tory, one of sev eral op er- ated by the Insti tut de Phy sique du Globe de Paris, monitors this very ac tive vol cano. In for ma tion Con tacts: Ob- ser va toire vol cano lo gique du Pi- Figure 8. Photo graph of the SW part of Dolom ieu crater at Pi ton de la Fournaise at 0812 on 6 June 2003 showing ton de la Fournaise (OVPF), In- the ac tive vent and part of the re cent lava- flow field. View is to wards the W. Cour tesy of OVPF. sti tut de Phy sique du Globe de Volcanism Network, Volume 28, Number 5, May 2003 Strom boli 15

extrude d from the 600 m vents on the upper Sciara del Fuoco. Spatter ing built up two horni tos, which in a few days reached an esti mated height of over 6 m. This ac tivit y was accom panied by lava flow effu sion along the upper Sci ara del Fuoco, with lava desce nding to 150 m ele vati on. On 1 June, Strom boli an ac tivit y re sumed at Crater 1 (NE crater). It was reveal ed first through helicopter-borne ther mal surveys , and then by direct obser va ti ons from the eastern Sci ara del Fuoco rim. Most of the ejecta fell within the crater , and from the lower slopes of the volcano only pulsat ing dark ash emissions were observe d. Strom boli an acti vit y stopped around 6 June, and oc casiona l lava flows occurred at the horni tos at 600 m ele vati on on 11 June. The sum mit cra ters showed dis con tinu ous ash emis sion un til mid-June, and the SAR fixed cam era at 400 m ele vati on showed a Strom boli an explo sion with abundant ash emis- sion on the night of 15 June. Figure 10. Topo graphi c map of Dolom ieu crater at Pi ton de la Fournaise MODIS Ther mal Alerts. MODIS therm al alerts for showing the extent of the lava-flow field on 30 May and 6 June 2003. Strom boli cover ing the period from the start of MODIS Eleva tions are in meters, and the Gauss- Laborde Pi ton des Nei ges sys tem data acqui si ti on over Europe in May 2000 unti l the pres ent is used for the map coor di nates . Cour tesy of OVPF. were com piled using data avail able at http://modis.higp.ha - waii.edu/. Paris, 14 RN3, le 27Km, 97418 La Plaine des Cafres, La Réunion, France (Email: Thom as.Staudacher@univ- reunion.fr).

Strom boli Aeo lian Is lands, It aly 38.79°N, 15.21°E; sum mit elev. 926 m All times are lo cal (= UTC + 1 hour late September- late March; + 2 hours late March- late Sep tem ber)

The latest erupti ve epi sode from Strom boli began on 28 Decem ber 2002 (Bulle ti n v. 28, no. 1) and include d a sig- nifi cant explo sion on 5 April (Bulle ti n v. 28, no. 4). This re - port include s field obser va ti ons provided by the Isti tuto Na- zion ale di Geofisica e Vulcano lo gia (INGV) through mid-June 2003. Therm al alerts based on infra red satel li te im agery over the course of this erupti on have been com- piled and sum mariz ed by sci enti sts at The Open Uni versit y. Acti vit y during 17 April-16 June 2003. Effu sion of lava from vents locat ed at ~ 600 m ele vati on, on the upper east- ern corner of the Sciara del Fuoco, conti nued unti l 16 June with a gen er ally decreas ing ef fu sion rate. This caused a sig - nifi cant in crease in the thick ness of the lava field formed since 15 Febru ary to over 50 m. Since the 5 April erupti on, the sum mit crater s of the volcano have been blocked by fall out ma te rial ob struct ing the con duit. Small, oc ca sional, short- lived ex plo sions of hot ju ve nile ma te rial were ob- served on 17 April during a heli copter survey with a hand- held therm al cam era, and on 3 May from the SAR fixed cam era lo cated at 400 m eleva tion on the E rim of the Sci ara del Fuoco. The effu sion rate from the 600- m-ele vation vents on the Sci ara del Fuoco showed a sig nifi cant de cline be tween 1 and 4 May, when infla tion of the upper lava flow field was detec ted through daily helicopter-borne therm al surveys . Infla tion stopped on 6 May, when two new vents opened on Fig ure 11. Alert-rat io, number of alert pixels, and summed 4 mm (MODIS the infla ted crust of the flow field, causing drainage and band 21) spectra l ra diance for MODIS therm al alerts on Strom boli spreading new lava flows along the Sci ara del Fuoco. Be- betwee n 1 Novem ber 2002 and 13 May 2003. MODIS data courtesy of the tween the end of May and early June, gas- rich magma was HIGP MODIS Therm al Alert Team. 16 Nyi ra gongo Smithsonian Institution — Bulletin of the Global

With the excep ti on of single-pixel alerts on 8 July and ta nia, It aly (URL: http://www.ct.ingv.it/, Email: cal- 19 Septem ber 2000 (with alert rati os of -0.798 and -0.794, [email protected]); David A Rothery and Di ego Cop pola, De- both barely above the -0.800 autom atic detec tion threshold partment of Earth Sci ences, The Open Uni ver sity, Mil ton of the ther mal alerts al go rithm), ac tiv ity at Strombo li re- Keynes, MK7 6AA, United Kingdom (Email: d.a.roth- mained below the autom atic detec tion threshold unti l No- [email protected], di ego.coppola@ unito.it). MODIS data vem ber 2002 (figure 11). In that month there were two courtes y of the HIGP MODIS Therm al Alert Team. single-pixel alerts, barely above detec tion threshold (-0.790 on 12 Novem ber and -0.792 on 28 Novem ber). Therm al in - fra red radi ance was higher than ever be fore at the time of the MODIS overpass on 20 Decem ber 2002, when there Nyi ra gongo was a two-pixel alert, with alert rati os of -0.667 and -0.749. These five dates were the only MODIS ther mal alerts cen tral Af rica prior to the start of effu sive acti vit y on 28 De cem ber 2002 1.52°S, 29.25°E; sum mit elev. 3,469 m (Bul le tin v. 27, no. 12 and v. 28, no. 1). The source of the All times are lo cal (= UTC + 2 hours) ra di ance to trig ger these alerts was evi dently in can des cence at one or more of the ac tive vents. In the case of a volcano Nyi ragongo, locat ed along the East Afri can Rift (figure such as Strom boli, prior to Decem ber 2002, isolat ed ther- 12), ceased gener at ing flank lava flows follow ing its Janu- mal alerts are more likely to rep re sent the chance co in ci- ary 2002 erupti on, but re mained acti ve inside its sum mit dence of a short-li ved peak of incan des cenc e with the time cra ter where it hosts a rest less lava lake. Ob ser va tions made of MODIS overpass, rather than a sustai ned emission of in - by staff from the Goma Volcano Obser va tory (GVO) in fra red ra dia tion. How ever the November- December 2002 August 2002 include d the opening of a new sinkhole, and ther mal alerts can with hindsight be seen to have been indi - meas ure ments of CO2 and O2 gas con cen tra tions at three fu - cators of enhance d acti vit y in the lead-up to the 28 Decem - marolic ar eas (lo cally termed mazu kus). For con text, hand - ber ef fu sive erup tion. book values for CO2 con cen tra tions and their re sult ing On 28 Decem ber 2002 MODIS recorded its highest ever alert ra tio at Strom boli (+0.419) and highest summed radi ance at 4.0 µm (MODIS band 21) in a seven-pixel alert, corre spond ing to the daily MODIS overpass at 2115 UTC. This is a record of radi ance from 300-m -wide lava flows from the NE crater (Bul le tin v. 27, no. 12). Subse quent to that date, ther mal alerts have occurr ed per sis tently at Strom boli, and evi dently re flect on go ing lava ef fu sion. The general trend of the highest alert ra tio on each date, the number of alert pixels, and the summed 4.0 µm radi ance for all alert pixels on each date shows an expo nen ti al de cline. There are no therm al alerts for 3-7 April 2003 inclu sive , which could be becaus e of cloud cover. There is thus no di- rect record of the explo sion on the morning of 5 April that com pletely covered the upper 200 m of the volcano with bombs. How ever, the mild in ten si fi ca tion of sub se quent thermal-alerts in di cates slight re- invigoration of the on- going lava ef fu sion. Back ground. Spec tacu lar in can des cent night time ex- plosions at Strom boli volcano have long attra cted visi tors to the “Light house of the Mediter ranean.” Strom boli, the NE-m ost of the Aeoli an Islands , has lent its name to the fre- quent mild ex plo sive ac tiv ity that has char ac ter ized its erupti ons throughout much of histori cal time. The small, 926-m -high island of Strom boli is the emergent sum mit of a volcano that grew in two main erupti ve cy cles, the last of which formed the west ern porti on of the island. The Neo- strom boli erupti ve period from about 13,000 to 5000 years ago was followe d by form ati on of the modern Strom boli edi fice. The ac tive summ it vents are lo cated at the head of the Sciara del Fuoco, a prominent horseshoe-shape d scarp formed about 5000 years ago as a result of the most re cent Fig ure 12. Sche matic map il lus trating the trend of the East Af rican rift. of a se ries of slope fail ures that ex tend to below sea level. The rift’s overall shape is curved, concave to wards the E, and it contains a The modern vol cano has been construc ted within this scarp, centra l segm ent com posed of two branches passing on the E and W sides which funnels pyro cla sti c ejecta and lava flows to the NW. of Lake Vic to ria (V). The over lap ping tri an gles la beled N at the N end of Es sen tially con tinu ous mild strom bo lian ex plo sions, some- the rift’s Western segm ent identify the approxi m ate loca tion of Nya muragira and Nyi ragongo volca noes N of Lake Kivu. The latter times ac compa nied by lava flows, have been re corded at volcano sits to the E and closer Lake Kivu. This figure is based on one in Strom boli for more than a mil lennium . an online book by W.J. Klius and R.I. Tilling of the US Geologi cal Survey . In for ma tion Con tacts: Sonia Calvari , Is ti tuto Na zi on - A smaller scale map show ing some of ten men tioned lo cal fea tures ale di Geofisica e Vul cano lo gia, Pi azza Roma 2, 95123 Ca - ap peared in Bul le tin v. 26, no. 3 (Nyam uragira report) . Volcanism Network, Volume 28, Number 5, May 2003 Nyir agongo 17 symptom s in hum ans are discusse d. The GVO has also vealed that the release of cer tain gases could con ceiva bly brought to light re ports from local resi dents of abnor m ally have played an other role as well. Both acetylene (C2H2, a rapid ripen ing of picked ba nanas (and in some cases yams) col or less, flamma ble gas with an odor similar to gar lic and prior to the January 2002 erupti on. slightly less dense than air) and C2H4 (eth yl ene, a col or less, This report also discusse s GVO and resident vol canolo - faintly odorous gas less dense than air) speed up the ripen - gist sum mit crater visit s during late Novem ber 2002-earl y ing pro cess in many ag ri cul tural prod ucts (in clud ing ba- May 2003. In all cases the lava lake within the summ it cra- nanas and yams). Eth ylene can cause ba nana peels to shift ter rem ained dynam ic, with one or more windows on the from green to yel low at low (ppm) concen tra tions. These crater floor expos ing agi tated molten lava. Dur ing this in- gases oc cur natu rally and may form or es cape in as so ciation ter val, degassi ng conti nued and tephra fell on the upper with heat ing or ganic mate rial. In con trast, CO2 gen er ally flanks. A sum mary of some ancil lary obser va ti ons such as slows the ripen ing proces s. For the inter val prior to the seism icit y meas ured on the GVO net work is also provided. Janu ary 2002 erup tion, ob serv ers lack docu men ta tion of in- A later sec tion discusses ash plumes as described in creases in degassi ng or heat ing. avia tion reports. Ash clouds extende d as visible swaths on Seis mic ity and cra ter vis its, No vem ber 2002- May sat elli te im agery for up to ~ 100 km from the vol cano. 2003. Mul tiple GVO cra ter vis its were docu mented: 23- 25 These reports include some as recent as 15 May 2003. The Novem ber 2002; 9-10 and 21-22 January 2003; 4-5 and fi nal sec tion dis cusses MODIS ther mal im agery dur ing late 25-26 Febru ary 2003; 18-19 March 2003; 22- 24 April 2002 through early 2003. The 2003 MODIS data refle ct the 2003; 6 May 2003. GVO also sent out occa siona l update s lava lake seen deep within the summ it crater. Fi nally, sat el- dis cuss ing seis mic ity and other ob ser va tions. lite data show at mospheri c SO2 burdens for the During 23-25 Novem ber 2002, GVO team mem bers Nyiragongo- Nyamuragira re gion dur ing 13 De cem ber Kas ereka Mahinda, Ciraba Mateso, Ar naud Le mar chand, 2002 to 15 June 2003. and Jac ques Durieux watched the active lava lake on the GVO’s August 2002 field obser va ti ons. On 12 August crater floor. The lake was then locat ed within the southern 2002 GVO was called to Bugarura village upslope from crater in the 16 Novem ber col lapsed area. Two broad open- Munigi on the S flank. A new sinkhole had devel oped that ings lay at the bottom of this new de pression; both per mit - morn ing, leaving a steam ing opening ~ 3-4 m in di ameter . ted view ers to see the lava lake’s sur face. A third, smaller Sci enti sts could not see the opening’s bot tom through the opening ejected only high-tem perature gases. The great steam, but they timed falling stones and es timated the sink - quanti ty of gas oc cupy ing the bottom of the crater thwarted hole’s depth at ~ 15 m. The odorles s gas be ing emit ted led ef forts to carry out a pre cise laser- based measure ment of them to be lieve that the steam chiefly rep re sented vapor ized the depth to the lava- lake sur face. The vis ual es tim ate for ground wa ter. this depth from the sum mit was ~ 700 m. GVO staff and col labora tors hoped to advance gas The lava lake was very ac tive, as it was be fore 1977. moni tor ing ef forts by meas ur ing CO2 and other escap ing The lava surface was disturbe d by the rise of abundant large gases at multi ple sites in the region. They conti nued to gas bubbles . Breaking bubbles threw molten fragm ents make spot-chec ks with hand-held de vices, but also sought a onto the margins of the two openings. Consis tent with the more- nearly con tinu ous rec ord from dedi cated moni tor ing bubbles and consta nt degassi ng, a gas plume was visi ble at instru m ents. Although noxious gases are a fa mil iar problem night from Goma. Occa sion all y, light dustings of tephra in vol canic areas , some of the gas concen tra tions in the rift and Pe le’s hair came from the cra ter and fell on the sur- are sur pris ingly high for ar eas ad ja cent hu man habi ta tion. rounding ar eas. Although the current lake was im pressive , The Swa hili word mazuku al legedly connotes places asso - the observ ers pointed out that the crater has contai ned a dy - ci ated with “evil winds,” and the term is cur rently used to namic lava lake for nearly 50 years. The ear lier lake’s sur - de scribe fu marolic ar eas, which have also been described as face was much larger and stood nearly 500 m higher. dry gas vents. Jean- Christophe Ko mor owski ac com pa nied GVO staff Pos si ble pre cur sors to Janu ary 2002 erup tion. In the on a climb up Nyi ragongo on 9-10 January 2003. While on weeks before the 17 January 2002 erupti on, there were the upper slopes, the climbers heard a few detona ti ons asso - widesprea d reports of picked crops ripen ing at unusu all y ciated with more en er getic gas plumes. From the rim they rapid rates. From the settl em ents of Rusay o (8 km SW of saw a deep pit in the SW part of the in ner crater. There were the sum mit) and Katal e (~ 18 km NNE of the sum mit and two vents on the cra ter floor separat ed by a thin rocky ~ 10 km NE of Nyam uragi ra’s sum mit) people reporte d in ridge. The SW vent (vent A) was char acter ized by a high- early January that the nor mal 5- day ripen ing proces ses of pressure fluctu at ing gas jet that gave off very loud roaring bananas placed in the ground decrea sed to only 2 days. noises, along with flames of incan des cent and com busting From Rusay o, people also re ported that sweet pota toes , gases. Condens ing steam clouds here were dense, render ing which are norm ally sun-drie d on the ground surface , dried visual obser va ti ons diffi cult . The other ac tive vent (vent B) even without sun. GVO observ ers saw this first-hand and, was just to the NE and consis ted of an area of stable incan - as a result sought funds to hire porter s and observe Nyi ra- des cence at least 100 m in di ameter with an ac tive lava gongo direct ly, but the erupti on began before the expe di - fountai n. Projec ti ons of lava spatter there took place every tion started. 30-60 sec onds and typi cally reached 40-60 m in height. Although ra diant or conduc ti ve heat may have been a The large area of in can des cence in di cated that a small factor (since heat speeds up the ripen ing proces s), heat’s lava lake must have been present deep in the pit, al though transport to broad areas on the surface by conduc ti on the ob serv ers never saw the mov ing lava sur face. Peak through rocks would be com parati vely slow. Heat at depth high-press ure degassi ng in vent A did not neces sar il y cor re- may have more rap idly reached the sur face in the form of late with peak lava foun tain ing ac tiv ity at vent B. Ob ser va - heated, lib er ated gases (such as steam). Dis cus sions with tions were conducte d for several hours at night and during gas chemist Vern Brown and a scan of the lit era ture also re - the day. La ser bin ocu lar meas ure ments es tab lished the cra - 18 Nyi ra gongo Smithsonian Institution — Bulletin of the Global ter floor’s depth at ~ 800 m. Very light ash consis ting of Pe - cantly. Pe ri odi cally more vig or ous lava foun tains sent le’s hair and tears, and millimeter- sized vitric sco ria frag- smaller fragm ents to higher eleva ti on that cooled to black ments fell conti nuously on the rim. Condi ti ons were made sco ria frag ments. A small sco ria cone had started to build dif fi cult at times when the SO2-rich gas plume blew to- around the ac tive vent. Re cent small lake overflows formed wards the W. thin lobate lava sheets around the vent. The ascent veloc it y Acid rain that flushed the vol cano’s SO2 gas plume, of indi vid ual gas plumes within the crater varied be tween 7 sam pled at eleva ti on 2,600 m, had a pH of 2.26. In contra st, and 12 m/s. rain collec ted in Kibati (below 2,000 m on the SSE flank) A serie s of incan des cent pits extende d to the SE of the on 6 January had a pH of 6.15. Dam age to about two-thi rds acti ve pit along a line that cor responds to a major pre- of the vegeta ti on by acid plume conden sat es was evi dent existing fault-fract ure system trending N25°W. This system above 2,900 m on the SW and S flanks. transec ted the cra ter from NW to SE and linked with the up- Com pared to the last visit by GVO staff, 30-31 Decem - per Sha heru frac ture and 1977 vent network that re acti vated ber 2002, de gassing had increa sed signifi cantly. However in 2002. A hot fracture zone trended N10°E-N20°E in the the level of the lava in the cra ter and/or lake had not risen NE part of the crater wall. This zone had extende d into the and might have dropped lower in the conduit . The gas- acti ve deep crater form ing a conspic uous, elongate , plume height, meas ured regular ly by the GVO, reached vertical- walled can yon. Ob serv ers fre quently heard and 4,500-5,000 m alti tude. At times, although the very loud saw rockfal ls, and noted that those events often gener ate d roaring sound re mained unchange d, the enti re cra ter be- plumes that spread and depos it ed ash over local vegeta ti on. came gas- filled to an ex tent that in can des cence was en tirely Intra- crater ash reached 5 mm in thick ness. The gas plume blocked, even from the vantage of surround ing village s. In- re mained rich in SO2. Rain wa ter col lected at the top of form ati on brought regular ly to the at tenti on of the GVO by Nyi ragongo had a pH of 2.84. the popula ti ons of Kibati , Mudja, Mutaho, and Rusay o vil - The late-Janua ry plume height es ti mated during fa vor- lages at tested to their expo sure to the gas and ash plumes able atm ospheri c condi ti ons by GVO mem bers varied from from Nyi ragongo. Through at least early May 2003 the vol - 4,500 to 5,500 m alti tude. Often, the prevail ing wind car- cano’s haz ard status rem ained at yel low (“vigi lance,” the ried ash, cinder, and Pele’ s hair S towards Ki bati, Rusay o, sec ond low est level on a 4- step scale). Mudja, and Mutaho village s. Another climb enable d observ ers to peer into the cra ter A 13 Febru ary GVO report said that for four consec u- during 21-22 January 2003 (figure 13). Com pared to the tive days, Pe le’s hair fell in Goma, 17 km SSW. Although 9-10 January obser va ti ons, only one opening re mained ac- cloudy and foggy due to the start of the rainy season, Nyira - tive inside the crater . The former vent A probably disap - gongo’s plume reached at least 5 km above the cra ter. Be - peared follow ing a collaps e. The acti ve opening had about tween Goma and the Nyira gongo stood heavy gray-to- blac k the same diame ter and its lava foun tain at tained similar ash-ric h clouds. The fall of Pele’ s hair was due to lava heights com pared to earli er vent B obser va ti ons. The level fountai ns inside the crater . of the lava had not changed in the crater , rem aining deep in The same report noted that seism icit y was probably the vol canic conduit . Degassi ng had increa sed signifi - lower than the previ ous week and consis ted of low tremor, few long-peri od earthquakes , and alm ost no tectoni c earthquakes . Very small-amplitude seis mic noise (small earthquakes ) oc- curred, presum ably due to col- lapses and per haps intra- crater ex plo sions. GVO went on to say that one side ef fect of the ash falls was that vil lages around Goma had seri ous wa ter short ages, since they rely on col lect ing rain fall. All UN agen cies and NGOs were in- formed and asked to start pota ble wa ter dis tri bu tion around Goma. A few more physical problem s might arise because of the Pe le’s hair, includ ing stress on people’ s eyes and breathing. Crops around the vol cano in some cases have been burned by acid rains and ash, while cat tle might also suf fer from inges ti on of ash-poll uted grass. The 25- 26 Feb ru ary as cent re- vealed more robust acti vit y than observ ers had seen on their 4-5 Figure 13. A photo of Nyira gon go’s cra ter and the one opening in the lava lake visible on 22 January 2003. Feb ru ary visit. By the lat ter date, Copyright ed photo used with perm ission of GVO. all vegeta ti on had died near the Volcanism Network, Volume 28, Number 5, May 2003 Nyi ra gongo 19 main crater . Approach ing the rim in the upper 220 m of the and lava splashing 50-60 m high. Oc casiona l waves of lava as cent, tephra falls had ac cu mu lated to form de pos its sev - rolled across porti ons of the crater floor and walls. Excur - eral cen time ters thick; those, along with acidic plumes, had sion mem bers also wit nessed crater- wall col lapses tak ing killed plants. The flora and fauna at lower eleva ti ons were place along the NW and S frac ture zones. still surviv ing, although they showed signs of se rious stress. Widely felt earthquakes also conti nued in the region, Loud sounds were audi ble several kilom eter s from the cen- presum ably relat ed to exten sion along the massive East Af- tral cra ter. Intra- crater ac tiv ity seemed in tense, but thick ri can rift sys tem. For ex am ple, three C- type events occurr ed fumes in the cra ter area thwarted day- time visi bil ity. On 25 on 23 April below Nyira gongo at a depth of ~ 15 km. Dur- Feb ru ary views from the W rim re vealed that a spatter cone ing the whole day of 24 April, sustai ned tremor plus C-ty pe had begun to grow on the cra ter floor. Lava fountai ning oc- events reg istered. On 25 April a few seis mic events oc- curred all night; discharg ing lava proba bly rose more than curred amid sus tained tremor. A main volcano-tec tonic 100 m high, but it was diffi cult to as sess the maxim um rise shock had been re corded and later a se ries of A- type events height. Lava fountai ns chiefly came out at one spot, al- in the Nyira gongo field, be tween the S flank and Lake though a second, much smaller point of emission gave off Kivu. Increa sing tremor followe d. For the rest of the week, mainly flames and sometim es sco ria. Pe le’s hair fell all the seis mic network recorded a concen tra tion of volcani c night long. events to the NW and the S of the volcano, along the prefer - An up date dis semi nated on 27 Febru ary 2003 noted that en tial frac ture axis. com pared to pre vi ous weeks, dur ing 21- 27 Feb ru ary Nyi ra- On 2-3 May unusu all y dense ash plumes were visible gon go’s ac tiv ity had de creased, al though seis mic ity meas- from Goma. Conti nuous ashfal l occurred in many village s ured on the S flanks conti nued to contai n low-am plitude close to the volcano, and perm anent tremor and long-peri od tremor. S- flank seis mic ity also con tained com para tively earth quakes were record ed. SO2 emission rates were rela - few long-peri od (LP) earthquakes . The update also said that tively high during 1-6 May, with the largest emission on 3 local winds had begun to blow predom inantl y from the May (~ 50,000 tons, see TOMS data below). UN peace ENE, thus sweeping plumes and asso ci ate d tephra falls keepers provided a 3 May heli copter flight that gave vol- clear of Goma. A 22 Febru ary visit to the SW- flank set tle- cano lo gists clear views of the crater. The lava lake’s molten ment of Rusay o reveal ed conspic uous tephra depos it s on surface appeare d slightly larger than during a visit to the roofs and trapped in the crevice s of banana trees. cra ter rim on 22- 24 April. At that time a sig nificant plume During a visit to Nyira gongo on 18-19 March, GVO sci- contai ning gas and ash rose high above the volcano. enti sts observe d a thick plume engulf ing the cra ter. Two On 6 May GVO climbers enter ed the vil lage of Kibati , possi ble emission points were noted; one was re lated to the usual de parture point for the ascent , ~ 8 km from the power ful lava and ash emissions , and the other was re lated cra ter rim. Ki bati resi dents told how ash falls and acid rains to a much weaker white- pink plume. An inner ac tive cone had nega tively affected lo cal crops. For ex am ple, bean was visi ble in the cra ter and was at least 200 m in di ameter. leaves had been burnt in many places. Along the ascent , at Lava fountai ns rose to maxim um heights of 150-200 m and 2,260 m ele vati on, Pele’ s hair was found, in clud ing some as low as 50 m. Scoria ejec tion made obser va ti ons diffi cult intac t indi vid ual strands 30 cm long. At 2,700 m ele vati on, at times. Sev eral per manent fu maroles, also ob served dur- thin ash grains com pletely covered the vege tati on. At 3,200 ing the previ ous visit, were seen in the cra ter. m ele vati on on the S flank (~ 270 m below the sum mit), all Dario Tede sco noted that the cone morphol ogy seemed vege ta tion had died. slightly differ ent from the trip 3 weeks earli er. He observe d At mos pheric con di tions ini tially al lowed quite clear that on the N side of the crater a new platform had been views from the crater rim. The lava lake un der went vio lent formed, proba bly due to the con tinu ous ac cu mu la tion of outbursts from bursting of gas bubbles esti mated at up to 40 ejecta, sco ria, and ash. The team saw a huge lava foun tain m wide. The re sult ing projec ti ons of spatter s and surges of at least 150-200 m in height. In contra st, when viewed in splashed on the walls of the pit. The lake had regaine d its late Feb ruary , fountai ns seemed to rem ain be low ~ 100 m in former di mensions (~ 60 m across). The wider lake, re- height. The lava fountai ns gener ate d abundant falli ng ash cently seen from heli copter , had shrunken, leaving a solid of milli me ter size at the ob ser va tion point, a pro cess that plat form on its side. Pres sure of the es cap ing gases seemed lasted all night long. very high and yielded a conti nuous roaring. GVO climbers Stronger and higher lava fountai ns, reaching al most 300 again wit nessed in ter mittent pale yellow- green flames hurl - m high, were witnesse d at 0230 on 19 March. The erupti ve ing from the vents up to 50 m high. vigor as well as the in ten sity of the fal ling tephra de clined At 0644 on 6 May a seism ic shock was felt by the team at 0530. The last wit nessed acti vit y was of 50-m -high foun- on top of the volcano. It was recorded by the whole network tains. A second pit was noted on the E side of the cra ter that as a low-am plitude long-peri od earthquake. Then, fog and had been hidden during the night by the very thick plume. gases halted further sight ings into the cra ter. The fog lifted For many days prior to visit s on 22-24 April the seis mic around 0100 on 7 May; at this time viewers saw a small sta tions con sid ered most rep re sen ta tive of the Nyi ra gongo narrow lava flow in the southern inner wall adja cent the ac- acti vit y only regis ter ed very weak and steady conti nuous tive pit’s margin ~ 200 m above the cra ter floor. The lava tremor. Although other types of seism icit y were absent in escape d out of what looked like a tunnel or tube. Although the, A-ty pe and C-ty pe earthquakes occurred near the vol- the lava de scended at a steep an gle and ap peared to es cape cano. De spite the compara tive seis mic quiet, a prominent from the tube at a con stant rate, its rate of ad vance rem ained gas plume rose from the vol cano. When weather condi ti ons slow. The lava front had not made it to the crater cen ter. Be - perm it ted, the plume top was meas ured at 5-6 km alti tude. low the tube, however, intri cate indi vid ual lava flows had The 22-24 April field excur sion noted five dis tinct vents formed a long delta. on the crater floor, alm ost conti nuous emissions of tephra, Avia tion re ports. A Volcani c Ash Advi sory (VAA) for an agi tated molten-lake surface that include d emerging gas, Nyi ragongo was issued by the Toulouse Volcani c Ash Ad- 20 Nyi ra gongo Smithsonian Institution — Bulletin of the Global visory Center (VAAC) on 6 March 2003. That advi sory pixels ) occurred near the sum mit for the rem ainder of 2002 stated, “A cloud probably contai ning ash can be seen on and into early 2003, consis tent with the kind of lava-lake [visi ble wave length] ME TEO SAT im agery ex tend ing 100 ac tiv ity de scribed above. NM [(nauti cal miles, 185 km)] westward from the vol- At mos pheric SO2. The Earth Probe Total Ozone Map- cano.” Several hours later the ash cloud was no longer visi- ping Spectrom eter (EP TOMS) SO2 data present ed in figure ble. Advi so rie s were also issue on 9, 12, 14, and 15 May 15 are pre liminary. The bars in di cated as “TOMS SO2“ 2003. The one for 9 May noted, “Renewed acti vit y since plot ted on the lower axis of the chart repre sent EP TOMS early May. Ash plume wit nessed during a heli copter flight around early May up to 5-6 km above sea level. Many ash falls and acid rains all around the vol cano.” No cloud was observ able due to con - vec tive weather clouds. The re- ports on 14 and 15 May stated, “Ac cord ing to Goma ob ser va tory [GVO], a plume of steam and ash is of ten emit ted since early May. It may rise 1,500-2,500 m above the vol cano’s sum mit. No new mes sage from Goma ob ser va tory since early May.” Me te oro logi cal sat el lite (ME TEO SAT) im agery was un able to de tect an ash cloud on 14 May due to weather clouds around the volcano. MODIS sat el lite data. Dur ing early 2002 to early 2003 Nyira - gongo was monitore d on a daily ba sis with ther mal sat el lite im- agery (1-km pixel size). Inves ti - ga tors Matt Pat rick, Luke Flynn, Figure 14. A plot il lustra ting MODIS data for Nyi ragongo with the sum for short-wave (4 micron, band 21) Harold Garbeil , Andy Harris, Eric radi ance as well as the sum for long- wave (12 micron, band 32) radi ance for all anomalous pixels in each im age. Pil ger, Glyn Williams- Jones, and The x- axis (time axis) starts be fore the eruption in Decem ber 2001 and ends in early 2003. Courtesy of Hawaii Rob Wright used NASA’s Mod- In sti tute of Geo phys ics and Plane tol ogy, Uni ver sity of Ha waii, Ma noa. er ate Reso lu tion Im ag ing Spec- tro ra di ome ter (MODIS) in stru- ment and processe d these data using the autom ated MODIS ther- mal alert sys tem at the Univer sity of Hawaii, Ma noa. Prior to the January 2002 erup tion, Nyi ra gongo ac tiv ity ap- peared in sig nifi cant; anoma lies were ab sent from the start of the MODIS-based alert system in April 2000, and through all of 2001. Anomalous pixels re- mained absent during 24 February-12 June 2002. The ab- sence of anomalies could be ex- plained ei ther by a lack of expo - sure of the lava lake or by cloud cover obscur ing the heat source from the sat el lite’s view. Nyi ra gon go’s ma jor ef fu sive erupti on in mid-Janua ry 2002 caused strong initial ther mal anomali es (figure 14). Lava flows Fig ure 15. Pre limi nary at mos pheric SO2 data taken from satel lite meas ure ments of the Nyiragongo- Nyamuragira extend ing down the S flank to region during 13 Decem ber 2002 to 15 June 2003. The data along the lower axis are from the EP TOMS instru m ent; the data on the upper axis are from the GOME in strum ent on the European satel lite ERS-2. Only the Lake Kivu re sulted in anomali es data de scribed as “TOMS SO2“ are quan ti ta tive (see text). Blank spaces for cer tain days and time in tervals on the as large as 45 pix els. Af ter wards, chart im ply that ei ther a data gap occurr ed over the region, or that no SO2 was de tected. One of these blank the anoma lies di min ished quickly. inter vals in the EP TOMS data took place during 15-23 May 2003, in this case due to the one in strum ent shut down Small in ter mit tent anoma lies (1-3 during the data- collection pe riod. Courtesy of Sim on Carn. Volcanism Network, Volume 28, Number 5, May 2003 Nyi ra gongo 21 meas urem ents on days when the signal was large enough to Vol ume % CO2 Physical Symptoms al low re trieval of the SO2 mass. The height of these bars corre sponds with the y-axis scale. Note that these val ues 2% 50% incre ase in breathing rate rep re sent the SO2 mass in a satel li te ‘snapshot’ of the vol- 3% 10-m inute expo sure limit; 100% incre ase in canic cloud taken around lo cal noon, and not an SO2 flux. breath ing rate The bars in di cated as “In ferred SO2” on the lower axis de- 5% 300% in crease in breath ing rate, head ache and note days on which the presence of SO2 could be inferred sweating may be gin after about an hour. from EP TOMS data, but the signal was too weak to allow 8% Short- term ex po sure limit re trieval of an atm os pheric SO2 mass. Hence these bars are 8-10% Head ache af ter 10 or 15 min utes. Diz zi ness, non-quanti tative, but they indi cat e that non-tri vial SO2 buzzing in the ears, blood- pressure in crease, emis sions proba bly oc curred. high pulse rate, ex cita tion, and nausea. More, non-quanti tative data appear as bars indi cat ed as 10- 18% Af ter a few min utes, cramps similar to epi lep tic “GOME detec tion” along the upper axis of figure 15; in this fits, loss of conscious ness, and a sharp drop in blood pressure. The victim s recover very case, showing dates when another instru m ent de tected SO2 quickly in fresh air. emissions in the region. These emission dates denote SO2 18- 20% Symp toms similar to those of a stroke de tection over cen tral Af rica by the Euro pean GOME (Global Ozone Moni tor ing Ex peri ment) in stru ment aboard Table 7. The AGA Gas Handbook included these CO2 gas con cen tra tions the ERS-2 satel lite. GOME measure ments are based on (in vol ume per cent) and accom pany ing symptom s for adults in good scans by a visi ble- and ultraviolet-wavel ength spectrom e- health (after Ahlber g, 1985). ter. GOME has in fe rior spa tial and tempo ral reso lu tion to EP TOMS, but is more sensi ti ve to atm ospheri c SO2. TOMS SO2 mass re triev als are de pend ent on the alti - tude of the vol canic plume and are also af fected by mete - 1.2- km-wide sum mit crater mark lev els of former lava orologi cal cloud cover, and therefore may be adjuste d as lakes, which have been observe d since the late-19th cen- more infor m ati on becom es avail able. The largest of these tury. Two older stratovol ca noes, Baruta and Sha heru, are pre limi nary es ti mates dur ing this in ter val was in ex cess of parti ally overlappe d by Nyira gongo on the north and south. 50 kilotons (kt) SO2. These peaks in the first half of May About 100 parasit ic cones are locat ed pri maril y along radial 2003 were truncat ed by an instru m ent shutdown during fissures south of Shaheru, east of the sum mit, and along a 15-23 May. Given the crater and plume obser va ti ons by NE-SW zone extend ing as far as Lake Kivu. Many cones GVO, and other data dis cussed above, the vast major it y of are buried by volu m inous lava flows that extend long dis- the SO2 shown on figure 15 was probably emit ted by tances down the flanks of the volcano. The extre mely fluid Nyi ra gongo. 1977 lava flows caused many fatal iti es, as did lava flows CO2 gas concen tra tions at three mazu kus on the flanks that inun date d porti ons of the major city of Goma in Janu- of Nyi ragongo in vi cinit y of Lac Vert at the ground surface ary 2002. meas ured up to ~ 40% by volum e, but concen tra tions of the Ref er ence: Ahlberg, K., 1985, AGA Gas Handbook: heavier-than- air gas dropped quickly with height above the Proper ti es & Uses of Indus tri al Gases, AB, Lidingo/ Swe- ground sur face. Spot measure ments were made with a Geo - den, ISBN 91-970061- 1-4 (out of print). tech ni cal In stru ments multi- gas land fill ana lyzer. Field In for ma tion Con tacts: Ce les tin Kas ereka Mahinda, notes reporte d CH4 con cen tra tions con sis tently at zero and Kavotha Kalendi Sadaka, Jean-Pie rre Bajope, Ciraba O2 concen tra tions at only one site where it was 22 vol. % at Mateso,and Mathieu Yali re, Goma Vol cano Ob ser va tory the ground surface and 16-17 vol. % nearby. The 15 August (GVO), De par te ment de Geo phy sique, Cen tre de Re cher- 2002 field excur sion was led by GVO scien ti sts Mathieu che en Sci ences Na turelles, Lwiro, D.S. Bu kavu, D.R. Ya lire, Ciraba Mateso, and Kas ereka Mahinda, with Chris Congo; Dario Tede sco, Jacques Durieux, Jean- Christophe Newhall pres ent. Komor owsk i, Jack Lockwood, Chris Newhall, Paolo Pa- Ef fects of car bon di ox ide. People in the region appar - pale, Ar naud Le Mar chand, and Orlando Vaselli, UN- ently under sta nd the hazard of es caping CO2 gas, and in the OCHA resident vol canolo gist s, c/o UN Offic e for the Coor - past sev eral years CO2 gas expo sure has not led to reporte d di na tion of Hu mani tar ian Af fairs, United Na tions Ge neva , hu man fa tali ties. CO2 gas, which is more dense than air at Palais des Na tions,1211 Geneva 10, Swit zerland (URL: equiva lent tem perature and pres sure, can be lethal to hu- http://www.unog.ch); Tol ouse Vol canic Ash Ad vi sory Cen- mans at 9-12 vol. % concen tra tions in as littl e as 5 minutes . ter (VAAC), Toulouse , Météo- France, 42 Avenue G. Corio - The US standards for indoor air qual ity suggest that long- lis, 31057 Toulouse Cedex, France (Email: vaac@meteo. fr; term hum an expo sures rem ain be low 0.1-0.2 vol. %, and URL: http://www.meteo. fr/aeroweb/info/vaac/homepage/ that short-ter m (10- to 15-m inute) expo sures rem ain be low ein dex.html); Matt Patric k, Luke Flynn, Harold Garbeil , 3 vol. %. The odor of CO2 is too weak to warn of danger ous Andy Har ris, Eric Pil ger, Glyn Williams- Jones, and Rob con cen tra tions. Ta ble 7 lists some symp toms as so ci ated Wright, Ha waii In sti tute of Geo phys ics and Plane tol ogy, with the in ha la tion of air con tain ing pro gres sively higher Univer sity of Ha waii, Ma noa (URL: http://modis.higp.ha- lev els of CO2. waii.edu/); Vern Brown , Presi dent, EN MET Cor po ra tion, Back ground. One of Afri ca’s most nota ble vol canoes, P.O. Box 979, Ann Arbor, Michigan 48106-0979 (Email: Nyi ragongo contai ned a lava lake in its deep sum mit crater info@en met.com; URL: http://www.en met.com/); Si mon that was ac tive for half a cen tury be fore drain ing cata- A. Carn, TOMS Volcani c Emissions Group, Joint Center strophi cally through its outer flanks in 1977. In contra st to for Earth System s Tech nology (NASA/UMBC), Univer sit y the low profil e of its neighbor ing shield vol cano, Nya- of Mary land Bal timore County, 1000 Hilltop Cir cle, Bal ti - muragira, 3470- m- high Nyi ra gongo dis plays the steep more, MD 21250 USA (Email: [email protected]; URL: slopes of a stratovol cano. Benches in the steep-wall ed, http://skye.gsfc.nasa.gov/). 22 Ro bledo Smithsonian Institution — Bulletin of the Global

Back ground. The small Robledo cal dera is lo cated 80 Ro bledo km SW of the much larger Cerro Galán cal dera in NW Ar - gen tina. The Holo cene Cerro Blanco lava dome is located Argen t ina on the southern rim of the Robledo caldera. Well-prese rved 26.77°S, 67.72°W; sum mit elev. 4,400 m pyroclastic-flow de posit s from Cerro Blanco are exposed on the floor of the caldera (de Silva and Francis , 1991) and A satellite- based in ter fero met ric syn thetic ap er ture ra- rep re sent its most re cent ac tiv ity. dar (InSAR) survey of the rem ote centra l Andes vol canic Ref er ences: de Silva, S.L., and Francis , P.W., 1991, arc (Prit chard and Sim ons, 2002) reveal ed defor m ati on in Volca noes of the Centra l Andes: Springer-Verl ag, Berli n, the Robledo caldera be tween May 1992 and Octo ber 2000 216 p. (figure 16). Subsi dence was de tected, with a maxim um de- Prit chard, M., and Sim ons, M., 2002, A satel li te geo- form ati on rate in the radar line- of-sight of 2-2.5 cm/year. detic sur vey of large- scale de for mation of vol canic cen tres The sub sidence rate seemed to be decreas ing with time. The in the Centra l Andes: Nature , v. 418, p. 167-170. inferred source depth was 4.5-6 km below sea level. Addi - In for ma tion Con tacts: Mat thew Prit chard and Mark tional de tails about the study and analy sis are available in Si mons, Divi sion of Geologi cal and Planetar y Science s, Prit chard and Sim ons (2002). Cali for nia In sti tute of Tech nol ogy, Pasa dena, CA 91125,

Figure 16. Shaded relief topo graphic map of the centra l Andes with insets showing ar eas of de form ation detect ed by Prit chard and Sim ons (2002). Inter ferogram s (draped over shaded relief) indi cate active de form ation; each color cy cle corre sponds to 5 cm of defor m ation in the ra dar line- of-sight (LOS). The LOS direc tion from ground to spacecra ft (black ar row) is incline d 23° from the ver tical. Black squares indi cate radar frames, and black trian gles show po ten tial vol canic edi fices. Cour tesy of Mat thew Prit chard. Volcanism Network, Volume 28, Number 5, May 2003 Saban c aya 23

USA (Email: [email protected] tech.edu, URL: http://www.gps. The sum mit region of Uturuncu has two acti ve fum arole cal tech.edu/). fields with substa nti al sulfur produc ti on and areas of clay- silica hy dro ther mal al tera tion. Maxi mum tem pera tures in four fum arole s were meas ured at 79-80°C. A hot spring on the NW flanks had a tem perature of 20°C. Utu runcu Uturuncu is a stratovol cano com posed of hy persthene an des ites, hypersthene- biotite dac ites, and biotite- Bo livia hornblende dac ites. Al most all the exposed products are ex - 22.27°S, 67.22°W; sum mit elev. 6,008 m tensive coulée-ty pe la vas and domes; no pyro cla sti c depos - its were ob served. Flow features are well-preser ved on the A large- scale con cen tric pat tern of de for ma tion was de - youngest lavas. A wide vari ety of xeno li ths were found in tected be tween May 1996 and Decem ber 2000 center ed on most la vas, in clud ing ma fic mag matic in clu sions, cu mu- Uturuncu vol cano, Boli via (figure 17), based on satel li te lates, mi cro crys tal line ig ne ous in clu sions, and horn fels of geodeti c surveys (Prit chard and Sim ons, 2002). The ob- pos si ble base ment rocks in clud ing sand stones and cal care - served defor m ati on is pri maril y surface upli ft with a maxi- ous rock types. mum rate at the up lift cen ter of 1-2 cm/year in the ra dar Lavas around the sum mit area appear to be the most re - line- of- sight di rec tion (fig ure 16). A re con nais sance in ves - cent products , but have been affect ed by glacia tion; there is ti gati on by a team com posed of scien ti sts from Boli via, however no present-day ice. There is thus no evi dence yet Chile, the USA, and the UK, took place during 1-6 April for Holocene acti vit y. The recent unrest manifest ed by sub- 2003 to identi fy any other signs of volcani c unrest and as - stan tial ground de for ma tion and re con nais sance seis mic ity sess past volcani c behav ior. in di cate, how ever, that a magmatic system is still pres ent A single- component ver ti cal one- second seis mome ter and therefore further monitor ing is warrant ed. was placed at five loca ti ons for peri ods of up to 14 hours. Ref er ence: Prit chard, M., and Sim ons, M., 2002, A sat- Data were record ed at a rate of 100 samples per sec ond on a el lite geo detic sur vey of large- scale de for ma tion of vol- lap top com puter. Per sis tent low- level seis mic ity was ob- canic centre s in the Centra l Andes: Nature , v. 418, p. 167- served mainly from one source loca ti on on the NW flank, 170. close to the center of defor m ati on observe d by satel li te sur- In for ma tion Con tacts: Mayel Sunagua and Ru ben veys. Two other sources within the vol canic edi fice could Muranca, Geo logi cal Sur vey of Bo livia, SER GEOMIN, not be lo cated with the available data. The rate of vol canic Ca silla 2729, La Paz, Bo livia (Email: mayel suco@hot- earthquakes was up to 15 per hour, and the magni tudes mail.com); Jorge Clav ero, Geologi cal Survey of Chile, were in the 0.5-1.5 range based on coda length. The sources Servi cio Na cional de Ge ología y Min ería (SER GEOMIN), were consid ered to be within 3-4 km of the surface (much Aven ida Santa María 0104, Casil la 10465, Santi ago, Chile shal lower than the de for mation source); more ac cu rate in- (Email: jclav [email protected]); Steve McNutt, Alaska for mation will be available when the data are analyzed Vol cano Ob ser va tory and Geo physi cal In sti tute, Uni ver sity fur ther. of Alaska Fairbanks, 903 Koyukuk Drive, P.O. Box 757320, Fairbanks, AK 99775- 7320 (Email: [email protected]. edu, URL: http://www.avo. alaska.edu/); Mat thew Prit chard, Di vi sion of Geo logi cal and Plane- tary Sci ences, Cali for nia In sti tute of Tech nol ogy, Pasa dena, CA 91125, USA (Email: matt@gps. cal tech.edu, URL: http://www. gps.cal tech.edu/); C. Annen , M. Hum phreys, A. le Friant , and R.S.J. Sparks, De part ment of Earth Sci ences, Uni ver sity of Bris tol, Bris tol BS8 1RJ, UK (Email: Steve.Sparks@bris. ac.uk).

Sa ban caya Perú 15.78°S, 71.85°W sum mit elev. 5,967 m

A satellite- based in ter fero met- ric syn thetic ap er ture ra dar (In- SAR) sur vey of the re mote cen tral Figure 17. Photo graph of Uturuncu viewed from the south, April 2003. Courtesy of Steve Sparks. An des volcanic arc (Pritchard and 24 Santa María Smithsonian Institution — Bulletin of the Global

Si mons, 2002) re vealed de for ma tion in the Sa ban caya area flow from July 1999 had stopped its advance in the channel during June 1992-m id 1997. Infla tion was de tected ~ 2.5 of the Nimá II river as of April 2002. km E of the Hualc a Hualc a cone and 7 km N of Saban cay a Since renewal of acti vit y in April and May 2002, a new (figure 16), with the maxim um defor m ati on rate in the radar lava flow had been advanc ing on top of the high part of the line-of- sight being ~ 2 cm/year. While not tem poral ly exist ing lava flow, in front of the Santi aguit o view point. well-const rained, this infla tion seems to have stopped in This consta nt movem ent was fill ing up the ravine that di- 1997, perhaps re lated to the large erupti on of Saban cay a in vided the lava flow from the El Faro farm. The new lava May 1997 (Bul le tin v. 22, no. 7). No de form ati on was ob- flow quickly built a small lobe reaching ~ 300 m high. It ad- served betwee n mid 1997-Dece mber 2001. The inferred vanced in a fan shape toward the S and W flanks, with con- source depth was 11-13 km below sea level. Addi ti onal de - tinuous col lapses from the front. tails about the study and analy sis are available in Prit chard A volcani c ash advi sory issued on 16 August was based and Si mons (2002). on a report from INSI VU MEH about a dome col lapse with Back ground. Sa ban caya, lo cated on the sad dle be tween some near-sum mit ash. However, no ash was evi dent in 6288- m- high Am pato and 6025- m- high Hu alca Hu alca vol- GOES-8 sat ellit e im agery. After 29 August there were fre- canoes, is the youngest of these Holocene volcani c center s. quent col lapses from the crater rim of the Caliente cone, Saban cay a is the only one to have erupted in histori cal time. gener at ing py rocla sti c flows that extende d to the base of the The 5967-m -high Sa bancay a, meaning “tongue of fire” in domes. The greatest col lapse oc curred on 3 Oc to ber, ac- the Quechua Indian language, first appeare d in records in com pa nied by a strong ex plo sion and sev eral py ro clas tic 1595, suggest ing ac tivit y prior to that date. Holocene acti v - flows that de scended all flanks of the volcano at high ity has consis ted of plinian erupti ons followed by emission speeds, cover ing the volcano com pletely in a few minutes of volu m inous tra chytic and dacit ic lava flows, which form and produc ing abundant ashfal l on the SW flank. During an exten sive 15-km -wide apron around the volcano. Re- Oc to ber there were con tin ued collapses of the cra ter rim. cords of histori cal erupti ons of Saban cay a date back to In the early hours of 17 Octo ber the inhabi tant s of the El 1750. Faro and La Flor ida farms, and ar eas such as Pal mar Nuevo Ref er ence: Prit chard, M., and Sim ons, M., 2002, A sat- and part of San Fe lipe Re tal huleu, heard a strong explo sion. el lite geo detic sur vey of large- scale de for ma tion of vol- At OVSAN (Vul canológico Ob ser va tory of San tia guito canic centre s in the Centra l Andes: Nature , v. 418, p. 167- Volcano), this ac tivit y was felt, and a col lapse of the dome 170. from the edge of the cra ter was seen. Af ter 19 Octo ber In for ma tion Con tacts: Mat thew Prit chard and Mark moder ate and strong explo sions occurred at a rate of 3-5 per Si mons, Divi sion of Geologi cal and Planetar y Science s, hour, some ac com panied by rum blings. There was also an Cali for nia In sti tute of Tech nol ogy, Pasa dena, CA 91125, increa se in the number of phreatom agmatic ash explo sions USA (Email: [email protected] tech.edu, URL: http://www.gps. that sent abundant gray ash 800-1,200 m high, disperse d Cal tech.edu/). mainly on the SW flank. In Novem ber observ ers re ported consta nt col lapses of the SE and E lava flows. On the morn- ing of 11 Novem ber there was a se ries of col lapses from the S lava flow, and heavy ash fall on the seism ic station Santa María hous ing. Back ground. Symmetrical, forest-covered Santa María Gua te mala volcano is one of the most prominent of a chain of large 14.756°N, 91.552°W; sum mit elev. 3,772 m stratovolcanoes that rises dramatically above the Pacific coastal plain of Guatemala. The 3772-m-high stratovolcano At Santi aguit o, the acti ve lava-flow front conti nued to has a sharp-topped, conical profile that is cut on the SW gener ate ash plumes through early 2002 (Bul le tin v. 27, no. flank by a large, 1.5-km-wide crater. The oval-shaped 5). IN SI VU MEH re ported that dur ing January- October crater extends from just below the summit of Volcán Santa 2002, ac tiv ity at San tia guito in cluded la hars, ex plo sions, María to the lower flank and was formed during a growth of the lava dome, and collaps es from the Cali ente catastrophic eruption in 1902. The renowned plinian dome. The main lahar during that pe riod oc curred on 8 eruption of 1902 that devastated much of SW Guatemala January 2002. Farm ers in the Monte Claro area heard rock - followed a long repose period after construction of the large falls on the W flank. Field inspec tions near the San Isidro basaltic-andesite stratovolcano. The massive dacitic ravine showed an abundance of mate ria l depos it ed by mud - Santiaguito lava-dome complex has been growing at the flows and other volcani c debris, mainly fine ash. These de- base of the 1902 crater since 1922. Compound dome posit s formed ash knolls called “hum mocks.” The San Is- growth at Santiaguito has occurred episodically from four idro ra vine be gins at the Nimá II river, now covered by the westward-younging vents, accompanied by almost SW lava flow, which cre ated a dam ~ 200-300 m high. A continuous minor explosions and periodic lava extrusion, rupture of the dam in the high part of the Brujo dome con- larger explosions, pyroclastic flows, and lahars. tributed fine mate ria l and blocks to the high-veloc ity lahar, In for ma tion Con tacts: Otoniel Ma tias and Gus tavo which travele d ~ 4 km unti l it was stopped by old landsli de Chi gna, Unit of Volcan ol ogy, Geologi c Depart ment of In- de pos its. ves ti ga tion and Serv ices, In sti tuto Na cional de Sis molo gia, At the height of the Propert y Florida, there are old lahar Vul cano lo gia, Me te oro lo gia e Hi drolo gia (IN SI VU MEH), depos it s, possi bly from the erupti ons of Santa Maria in 7a Av. 14-57, Zona 13, Guate m ala City, Guate m ala; Wash- 1902 and/or Santi aguit o in 1929, with blocks of 1, 2, 3, and ing ton VAAC, Sat el lite Analy sis Branch (SAB), 5 m in di ameter . With the arri val of the rainy sea son, San NOAA/NESDIS E/SP23, NOAA Science Center Room Is idro, which becam e a new chan nel for lahars from May to 401, 5200 Auth Road, Camp Springs, MD 20746, USA Oc to ber, had at least six “strong” la hars. The ac tive lava (URL: http://www.ssd.noaa.gov/). Volcanism Network, Volume 28, Number 5, May 2003 Ki lauea 25

Tongues of lava were visi ble travel ing down Pulam a pali, Ki lauea part of the acti vit y that be gan on 12 May 2002 (named the Mother’s Day flow). Ha waii, USA Through April 2003, Ki lauea conti nued to erupt, send- 19.425ºN, 155.292ºW; sum mit elev. 1,222 m ing lava down its SE flank either travel ing over the land All times are lo cal (= UTC - 10 hours) sur face or through tubes. Lava en tered the sea at the West High cas tle en try; ac tiv ity there was some times weak, From Decem ber 2002 through June 2003, lava from Ki- though one or more glowing ar eas were typi cally seen. On lauea conti nued to flow down the S flanks and into the 16 April a large tract of land not over-run by sur rounding ocean at sev eral points. Seis mic ity gen er ally con tin ued at lava (a kipuka or ahu in the local parlanc e) rem ained within norm al (background) levels . The Mother’s Day flow, which the Kohola lava flow, still ~ 30 cm above the top of infla ted began erupt ing 12 May 2002, conti nued through June 2003 lavas that surround it. On the eastern margin of the swath of (fig ure 18). lava flows going down the steep slopes of Pulam a pali, one Lava flows. During De cem ber 2002, lava conti nued to partly crusted- over lava stream was highly visi ble. The cra - flow into the sea at entry points from two lava deltas . ter of Pu’u ‘O’o was dark and obscure d by fumes, but erup- Moderate- to- large lit to ral ex plo sions tossed spat ter onto the tive ac tivit y at Pu’u ‘O’o conti nued unabat ed. The flows on front of the West High cas tle delta. Sur face lava flows were Pu lama pali were fre quently visi ble at night as streams of visi ble on the coastal flat. On 15 Decem ber, shortly after incan des cenc e from the top of the pali down to the coastal 0700, the Wili pe’ a lava delta par tially col lapsed, los ing flats. Late in April, the E arm of the Mother’s Day flow about 1/3 of its area. The tip of the delta retre ated shore- split in two with the W segm ent be ing more ac tive. A new ward about 260 m and most of the col lapse was in the cen - ocean en try near Lae’apuki only lasted a day be fore the tral part of the delta. Around 15 and 16 Decem ber a sub- flow stag nated. Scattered sur face break outs were seen stan tial collapse oc curred at the West High cas tle delta. On throughout the infla ting Kohola flow, espe cia lly on its W 28 De cem ber mod er ate col lapses oc curred at the Wi li pe’a side. As of 24 April, lava enter ed the ocean at two points lava delta, ap par ently in the area of the 15 Decem ber col- along the West High cas tle delta. lapse. Sur face lava flows were visi ble on the coastal flat In early May, lava flows conti nued to de scend the S and upslope on Pulam a pali. flanks and pour into the sea. On 12 May lava be gan to enter During January and Febru ary 2003, lava conti nued to the sea again at the West High cas tle lava delta. Sur face lava flow into the sea at the West High cas tle en try. Sur face lava flows were visi ble on the coastal flat and the Pu lama Pali flows were visible on the coastal flat and upslope of it on fault scarp. During June, lava conti nued to flow down Ki- Paliuli. Most of the sur face lava flows on the coastal flat lauea’s SE flank, with sur face lava flows oc casion ally visi- crusted over, so that less in can des cence was visible than ble on the coastal flat and upslope at Pulam a pali, and pre vi ously. Rela tively large sur face lava flows were visi ble Paliuli. Small amounts of lava conti nued to flow into the starting on 21 January around 2035. Around 28 January a sea at High cas tle beach. large lava break out oc curred from the West High cas tle lava Geo physi cal ac tiv ity. During De cem ber 2002 and Janu- tube about 170 m inland from the old sea cliff. As of 2 Feb - ary 2003, seis micity was gen er ally at nor mal lev els. The ruary the area of the new breakout was about 6.15 hectar es swarm of long-peri od earthquakes and tremor beneat h Ki- (6.15 x 104 m2), and surface flows and lava in lava tubes lauea’s cal dera, oc ca sion ally seis mi cally ac tive since June travele d down the Pulam a pali fault scarp. The Chain of 2002, conti nued to show some short bursts of tremor inter - Cra ters road was closed due to a wild fire that was started by lava flows. Surface lava flows conti nued to travel through vegeta ti on, ignit - ing fires and caus ing methane ex- plo sions. Rang ers’ of fice huts, restrooms, and signs were moved out of the path of the lava flow, which reached the Chain of Craters Road on 19 Febru ary at 1005. Be- ginning 15 Febru ary and going into March, lava flowed into the sea at the Ko hala en try. Fresh lava oozed out of the cool ing Kohala lava flow, both within the body of the flow and along its E margin. During 26 Febru ary to 3 March lava con tin ued to en ter the sea at the West High cas tle en try, but the lava- flow rate was re duced to a small trickle at the Ko hala en try. Small sur face flows oc curred along the W edge of the Kohala lava flow and sur face lava flows were visi ble Figure 18. Map of lava flows erupted during 1983 through 16 May 2003 from Pu’u ‘O’o and Kupaiana ha. The above the Pu lama pali fault scarp. most re cently ac tive flows are on the SW side of the flow- field. Cour tesy of HVO. 26 Ki lauea Smithsonian Institution — Bulletin of the Global spersed with small earthquakes . Small infla tion and defla - During 30 April to 6 May, distanc es meas ured across tion events occurred at Pu’u’ O’o and Uweka huna tilt me- Ki lauea cal dera between two points ~ 10 km apart, re- ters. The Pu’u ‘O’o tiltm eter showed de fla tion for about mained sta ble as they have since early 2003. There had one week from 10 to 17 De cem ber. During 27-28 Decem - been con sis tent pro gres sive length en ing of this dis tance ber, slight defla tion occurred at the Uweka huna and Pu’u during late 2001 through mid-2002, and some minor fluc- ‘O’o tilt me ters. tua tions af ter that. In gen eral, tilt dur ing late April through Kilaue a’s sum mit began to de flate on 20 January 2003 2 May changed littl e at Uwekahuna stati on (W side of the at 1710, and Pu’u ‘O’o be gan to de flate a few tens of min- cal dera), and showed a progres sive decli ne at Pu’u ‘O’o utes later. Both ar eas de flated well into the next day. On the sta tion (E of the cal dera). In the first few days of May slight 21st at 1610 rapid, brief infla tion began at the sum mit. The in fla tion ary tilt ap peared at both sta tions. infla tion and preced ing defla tion were center ed near the NE Seis micity at Ki lauea’s summit was at moderate- to- high cor ner of Hale mau mau Cra ter, the nor mal cen ter of small levels from about 1 June through 14 June, with many small, de for ma tion events. Seis mic ity in creased with the de for ma- low-frequenc y earthquakes occur ring at shallow depths be - tion events, return ing to norm al levels after wards. By 22 neath the sum mit caldera. The tiny earth quakes oc curred at Janu ary seis micity had re turned to its nor mal level, with the the nota bly high rate of 2-4 per minute. Littl e or no vol canic long-las ting swarm of long-peri od earthquakes and tremor tremor accom pa nied the swarm, how ever. Vol canic tremor at Ki lauea’s sum mit con tinu ing at weak- to- moderate lev els. at Pu’u ‘O’o rem ained moder ate to high, as is the norm. A Dur ing Feb ru ary and March, seis micity was at back- quasi- cyclic in fla tion and de fla tion oc curred at Ki lauea’s ground levels . The long-las ting swarm of long-peri od sum mit and at Pu’u ‘O’o during the week of 6-13 June, but earthquakes and tremor at Kilaue a’s sum mit conti nued at did not cul minate in sig nificant over all tilt. low-to- m oderate levels . On 9 and 10 Febru ary , short peri - Back ground. Kilauea volcano, which overlaps the east ods of defla tion and infla tion occurred at the Uweka huna flank of the massive Mauna Loa shield volcano, has been and Pu’u ‘O’o tiltm eter s. Moder ate tremor was re corded by Hawaii's most active volcano during historical time. the nearest seism ometer to Pu’u ‘O’o unti l the seism ometer Eruptions of Kilauea are prominent in Polynesian legends; broke on 5 March. Moder ate defla tion occurred on 8 written documentation extending back to only 1820 records March, first at the Uweka huna tilt meter and then at the Pu’u frequent summit and flank lava flow eruptions that were ‘O’o tilt meter . Accord ing to a news report, a mem ber of a interspersed with periods of long-term lava lake activity tour group suf fered burns on 10 March when he fell on hot that lasted until 1924 at Halemaumau crater, within the lava while hiking near Chain of Crater s road. summit caldera. The 3 x 5 km caldera was formed in several For about a week in early April, volcani c tremor at Pu’u stages about 1500 years ago and during the 18th century; ‘O’o was relati vely high and small defor m ati on changes oc- eruptions have also originated from the lengthy East and curred, mostly at Pu’u ‘O’o. During 16-17 April, the SW rift zones, which extend to the sea on both sides of the Uweka huna tilt me ter at Ki lauea’s sum mit re corded three volcano. About 90% of the surface of Kilauea is formed of small in fla tions, the last ap par ently right at its crest. Pu’u lava flows less than about 1100 years old; 70% of the ‘O’o has gener all y fol lowed suit, though in this case show - volcano's surface is younger than 600 years. A longterm ing only two of the in fla tions very well. These tilts are not eruption from the East rift zone that began in 1983 has major but conti nue to illus tra te the clear connec ti on be- produced lava flows covering more than 100 sq km, tween Ki lauea’s summit, where most tilt events start, and destroying nearly 200 houses and adding new coastline to Pu’u ‘O’o, 20 km away, where the tilt events fol low a few the island. minutes later. Seis micity dur ing the week was at low to nor - In for ma tion Con tact: Ha wai ian Vol cano Ob ser va tory mal lev els. In stru ments con tin ued to reg is ter the sum mit (HVO), U.S. Geo logi cal Sur vey, Ha waii Vol ca noes Na- swarm of long-peri od earthquakes and tremor, which began tional Park, P.O. Box 51, Hilo, HI 96718, USA (URL: last June. Vol canic tremor at Pu’u ‘O’o re mained ele vated, http://hvo.wr.usgs.gov/kileaua; Email: hvo-info@hvo- as has been the norm for more than a week. mail.wr.usgs.gov).

Sub scrip tions: The Bul le tin of the Global Vol can ism Net work (ISSN: 1050- 4818) is avail able by subscrip tion from the Ameri can Geo physical Un ion (2000 Flor ida Ave nue NW, Wash ing ton, DC 20009, phone 202:462- 6900 or 800:966- 2481, fax 202:328- 0566, Email: [email protected]). An nual sub scrip tion price is $22 to US ad dresses and $39 to all other coun tries. Back is sues can be or dered through AGU Sepa rates; con tact AGU for cur rent pric ing in for ma tion. Or ders must be pre paid; make checks pay able to AGU; VISA/MC are ac cepted. The Smith sonian does not han dle any Bul le tin or ders.

Inter net Ac cess: The com plete text of the GVN Bul le tin is dis trib uted by electronic mail through the VOL CANO List serv mail ing list (vol [email protected]) maintained by Jona than Fink at Ari zona State Uni ver sity (subscribe via [email protected]). Text, figures, and photo graphs can also be ac cessed via the World Wide Web at URL: http://www.vol cano.si.edu/.