KAJIAN DINAMIKA ERUPSI DARI ENDAPAN PIROKLASTIK 2014 GUNUNG API KELUD, JAWA TIMUR, INDONESIA DAN IMPLIKASINYA TERHADAP BAHAYA DI MASA DEPAN REFERAT Heather, K.H., Louise, R.G., Mirzam, A., Shane, J.C. (2019) Insights Into Eruption Dynamics from the 2014 Pyroclastic Deposits of Kelud Volcano, Java, Indonesia and Implications for Future Hazards, Journal of Volcanology and Geothermal Research, 382, 6-23. Oleh MUHAMMAD RIZKI RADANA NIM : 119150006 PROGRAM STUDI TEKNIK GEOLOGI SJURUSAN TEKNOLOGI INDUSTRI DAN PRODUKSI INSTITUT TEKNOLOGI SUMATERA OKTOBER 2021ABSTRAK Pada tanggal 13-14 Februari 2014 selama kurang lebih 4,5 jam terjadi erupsi Gunung Kelud dengan nilai skala VEI 4. Erupsi tersebut menyebabkan hancurnya kubah lava dan membuat terjadinya hujan abu yang luas yang menyebabkan 7 kematian dan gangguan pada penerbangan di seluruh wilayah Asia-Pasifik. Untuk ‘mengetahui kronologi erupsi dari Gunung Kelud serta mengetahui pola erupsi dan implikasi bahaya yang akan datang maka dilakukanlah penelitian ini. Penelitian dilakukan dengan mengumpulkan sampel endapan piroklastik dari berbagai tahap letusan di sebelas titik yang berbeda. Penelitian menunjukan erupsi terbagi dalam 3 tahap, tahap pertama merupakan fase awal, tahap kedua mengenai_pembentukan kolom plinian, dan tahap ketiga merupakan fase klimaks dari peristiwa plinian dan Jatuhan tephra. Endapan piroklastik tahap pertama memiliki kandungan fragmen litik padat yang tinggi (44 %), endapan tahap kedua kaya akan pumice (70% ) dan tahap ketiga merupakan endapan jatuhan lapili dengan struktur invers yang mengandung pumice (90%) dari gumpalan abu vulkanik setinggi 26 Km. Jika dilihat dari pola erupsi gunung kelud semenjak 1900-an, pola erupsi Kelud yang akan datang kemungkinan akan menunjukan aktivitas pembentukan kubah lava Kembali atau lebih bersifat efusif. Bahaya dari erupsi gunung seperti aliran lahar dan awan panas juga harus diperhatikan karena tetap akan berpotensi terjadi lagi di ‘masa yang akan datang. Kata kunci: Gunung kelud, erupsi, piroklastikABSTRACT On February 13-14th, 2014 for approximately 4.5 hours there was an eruption of Mount Kelud with a VEI scale value of 4. The eruption caused the destruction of the tava dome and made extensive ash rain that caused 7 deaths and disruption to flights throughout the Asia-Pacific region. To know the chronology of eruptions ‘from Mount Kelud and to know the patterns of eruptions and the implications of impending dangers, this study was conducted. The research was conducted by collecting samples of pyroclastic sediment from different stages of the eruption at eleven different points. Research shows the eruptions are divided into 3 stages, the first stage is the initial phase, the second stage regarding the formation of the latian column, and the third stage is the climactic phase of the platian event and the Jangar tephra event. The first stage pyroclastic sediment has a high content of solid lytic fragments (44 %), the second stage sediment is rich in pumice (70 % ) and the third stage is a lapillary precipitate sediment with an inverse structure containing pumice (90%) of voléanic ash lumps as high as 26 Km. When viewed from the eruption pattern of the mountain kelud since the 1900s, the upcoming ‘eruption pattern of Kelud is likely to show the activity of forming a lava dome again or more effusive. The dangers of mountain eruptions such as lava flows and hot clouds must also be observed because they will still potentially occur again in the future, Keywords: Mount Kelud, eruption, pyroclasticKATA PENGANTAR Puji syukur saya panjatkan kepada Tuhan Yang Maha Esa yang telah melimpahkan rahmat, nikmat kesehatan, dan kasih sayang kepada hamba-Nya, sehingga saya dapat menyelesaikan makalah referat yang berjudul “Wawasan Tentang Dinamika Erupsi dari Endapan Piroklastik 2014 Gunung Api Kelud, Jawa Timur, Indonesia dan Implikasinya Terhadap Bahaya di Masa Depan” yang disusun untuk memenuhi tugas mata kuliah Referat Saya mengucapkan terima kasih yang sebesar-besarnya kepada Bapak Mochamad Igbal S.T, M.T dan Ibu Bella Restu Juliarka S.T, M.Eng selaku dosen pengampu mata kuliah Referat yang telah membimbing proses penyusunan makalah referat ini, sehingga dapat selesai tepat waktu. Saya juga berterima kasih kepada kedua orang tua dan teman-teman yang selalu memberi dukungan dan doa sampai saat ini. Saya berharap makalah referat ini dapat memberikan manfaat yang sebesar- besarnya kepada pembaca untuk pengembangan ilmu kegeologian. Saya juga menyadari ada banyak kekurangan dalam penulisan makalah referat ini, baik dari segi isi penulisan maupun tata tulis. Oleh karena itu, saya sangat terbuka untuk menerima saran, masukan, dan kritik yang tentunya sangat berguna untuk perbaikan makalah kedepannya. Lampung Selatan, 7 Oktober 2021 Penulis, ‘Muhammad Rizki Radana NIM. 119150006 iliDAFTAR ISI ABSTRAK ABSTRAK. KATA PENGANTAR.. DAFTAR ISI. DAFTAR GAMBAR. DAFTAR TABEL BAB I PENDAHULUAN. 1.1. Latar Belakang 1.2. Rumusan Masalah .... 1.3. Maksud dan Tujuan 1.4. Batasan Masalah BAB II TINJAUAN PUSTAKA.... 2.1. Kondisi Geologi 2.2. Brupsi Kelud pra-2014 .... 2.2.1, Erupsi Tahun 1990... 2.2.2. Erupsi Tahun 2007-2008... 2.3. Erupsi Februari 2014... BAB III METODOLOGI PENELITIAN BAB IV HASIL DAN PEMBAHASAN .. 4.1, Stratigrafi dan Dinamika Erupsi Gunung Kelud Tahun 2014. a) 4.1.1. Tahap 1: Fase Awal. i 4.1.2. Tahap 2: Pembentukan kolom plinian... ul 4.1.3. Tahap 3: Fase Klimaks. ul 42. Pola Erupsi Gunung Kelud dan Implikasi Bahaya Erupsi 12 BAB V KESIMPULAN .. DAFTAR PUSTAKA. ivDAFTAR GAMBAR 3 Gambar 2.1. Lokasi Gunung Kelud dan sekitarnya diambil dari citra satelit Gambar 2.2. Citra satelit menunjukkan gunung berapi Kelut dan daerah sekitarnya Kawah sebelum (a) dan setelah (b) letusan Februari 2014.. Gambar 3.1. Titik lokasi pengambilan sampe! Gambar 3.2. Foto endapan vulkanik di titik-titik lokasi pengambil Gambar 4.1. Gambar endapan piroklastik Gunung Kelud 2014 Gambar 4.2. Urutan stratigrafi endapan piroklastik untuk letusan Kelud 2014.. 10 2 Gambar 4.3. Diagram skema dinamika yang terjadi selama Tahap 1-3 erupsiDAFTAR TABEL ‘Tabel 2.1. Tabel sejarah erupsi gunung api Kelud 4 viBABI PENDAHULUAN 1.1. Latar Belakang Gunung Kelud yang terletak di provinsi Jawa Timur merupakan salah satu gunung berapi yang paling mematikan di Indonesia. Gunung Kelud dikenal memiliki sifat cerupsi yang sangat dinamis, seringkali mengalami erupsi eksplosif yang merusak dan berbahaya dan beberapa kali juga mengalami erupsi efusif. Di tubuh Gunung Kelud banyak ditemukan danau kawah ha il dari erupsi-erupsi sebelumnya yang, tentunya sangat berbahaya jika bercampur dengan magama ketika erupsi yang menyebabkan terjadinya banjir lahar. Dalam sejarahnya, besarnya erupsi bervariasi dalam skala Viudkanik Explosivity Index (VEI) berkisar antara 2 sampai 5 VEL. Sejak tahun 1919, letusan Kelud telah menunjukkan pola siklik perulangan VEI4 yang bersifat eksplosif (menghancurkan kubah lava) dan periode efusi (pembentukan kubah lava). Pola perulangan serupa juga umum erjadi secara regional dan global di gunung api lainnya seperti Gunung Merapi, Indonesia. Setelah letusan plinian 1990 (VEI 4) ada periode tenang selama 17 tahun, diikuti oleh masa ekstrusi kubah hasil letusan efusif pada 2007-2008. ‘Tumbuh kubah secara bertahap yang membuat sebagian besar bagian danau kawah tertutupi, tanpa menghasilkan lahar. Pada 13 Februari 2014, letusan VET 4 plinian lainnya terjadi, menghancurkan kubah lava hasil erupsi 2007-2008, dengan dua titik erupsi yang terjadi selama kurang lebih 4,5 jam. Untuk memahami dinamika dan mekanisme erupsi serta endapan piroklastik erupsi gunung berapi Kelud secara lebih rinci, maka perlu dilakukan analisis tentang endapan gunung berapi kelud baik dari stratigrafi, litologi, distribusi_maupun sejarah ery ya. Dalam makalah ini akan disajikan ringkasan dari laporan pengamatan, data sedimentologi lapangan, dan komponen dan analisis ukuran butir dari endapan piroklastik 2014 yang berguna untuk merekonstruksi kronologi Jetusan, pola erupsi, serta bahaya dari erupsi Gunung Kelud yang akan datang, 1.2. Rumusan Masalah Berdasarkan hasil kajian permasalahan yang ada, didapatkan rumusan masalah yang akan dibahas yaitu mengenai_dinamika erupsi Gunung Kelud dari endapan material piroklastik hasil erupsi tahun 2014.1.3. Maksud dan Tujuan Maksud dari penelitian kali ini adalah untuk mengkaji dinamika erupsi Gunung Kelud dari endapan hasil erupsi 2014. Adapun tujuan dari penelitian ini adalah: 1. Mengetahui kronologi erupsi dari Gunung Kelud dengan merekontruksi endapan piroklastik erupsi Gunung Kelud 2014. 2. Mengetahui pola erupsi gunung Kelud dan implikasi bahaya yang akan datang 1.4, Batasan Masalah Penelitian pada makalah ini hanya terbatas pada analisis endapan erupsi Gunung Kelud tahun 2014.BABII ‘TINJAUAN PUSTAKA, 2.1. Kondisi Geologi Gunung Kelud merupakan gunung api Kuarter yang merupakan produk proses tumbukan antara lempeng Indo-Australia yang menunjam ke bawah lempeng Asia, tepatnya di sebelah selatan Jawa. Sebagai gunung api muda yang tumbuh pada zaman Kuarter Muda (Holosen), Gunung Kelud merupakan salah satu gunung api dalam deretan gunung api yang tumbuh dan berkembang di dalam Subzona Blitar dari Zona Solo; dimulai dari daerah bagian selatan Jawa bagian tengah (Gunung Lawu) hingga Jawa bagian timur (Gunung Raung) , yang dibatas gawir sesar Pegunungan Selatan . Gunung api ini merupakan gunung api strato, akan tetapi tidak seperti gunung api strato yang lainnya seperti Gunung Merapi atau Semeru, Hal ini karena kerucutnya tidak begitu jelas, puncak tidak teratur, tajam dan terjal, serta kerucut yang rendah. Keadaan puncak-puncak tersebut disebabkan oleh sifat erupsi yang sangat merusak (eksplosif) disertai dengan pertumbuhan sumbat- sumbat lava, seperti puncak Sumbing, Gajah Mungkur, dan puncak Kelud (Zaenuddin, 2007). ny PIT) co Gambar 2.1. Lokasi Gunung Kelud dan sekitarnya yang diambil dari citra satelit, (Jurnal Geologi Indonesia, Vol 6 No. 4 Desember 2011;227-23)2.2.Erupsi Kelud pra-2014 Dalam sejarahnya sebelum terjadi erupsi gunung kelud 2014, Kelud sudah mengalami puluhan kali eupsi yang bisa di lihat dari tabel di bawah (tabel. 2.1), adapun erupsi gunung Kelud yang paling dekat sebelum erupsi 2014 adalah erupsi tahun 1990 dan tahun 2007-2008, Tabel 2.1. Tabel sejarah erupsi gunung api Kelud Tagine Vane) pean Copan ——_— le hewenibd rary Ko ? 1 Snide a > roy esata 7 : ; too esol? ' Fememinumred 4 Os TAR Rawucadene 1.578 Femenitunrod 4 tt inna Neaicadee U1 2.2.1. Erupsi Tahun 1990 Pada tanggal 10 Februari 1990 terjadi erupsi Gunung Kelud setelah 24 tahun mengalami dormansi. Volume air danau yang hannya sekitar 1,8 jutam’ merupakan faktor yang membuat tidak terjadinya lahar panas pada letusan kali ini, Letusan terjadi secara beruntun mulai pukul 11.41 - 12.21 WIB dengan skala VEI 4. Tahap awal dari letusan merupakan fase freatomagmatik yang mengakibatkan sebaran abu tipis di sekitar puncak, sedangkan letusan berikutnya lebih besar dengan lemparan pasir, lapilli, dan batu yang tersebar pada radius 3,5 km? Jarak jangkau 1,5 km ke arah timur dan sekitar 5 km ke arah barat, barat laut dan barat daya. Letusan utamanya berupa letusan plinian dengan awan panas menyusuri lembah di baratdaya sejauh 5 km dari kawah, Letusan tersebut berintensitas sedang dengan tephra sekitar 130 juta meter 3. Daerah yang rusak tidak terlalu Iwas, hanya dalam jangkauan radius sekitar 2 km dari kawah, namun demikian sebaran abu letusan jauh lebih luas dan diperkirakan meneapai luasan sekitar 1700 km? (Bourdier, 1997) 2.2.2. Erupsi Tahun 2007-2008 Mulai November 2007 terjadi peningkatan aktifitas vulkanik Gunung Kelud, sebuah kubah lava diekstrusi di danau kawah, yang membuat hampir sepenuhnya menutupi danau (gambar 2.2. a). Gumpalan uap naik 2 km dari aktivitas freatik, meskipun tidak ada aktivitas magmatik eksplosif yang terkait dengan letusan ini. Kubah tumbuh dengan cepat Desember 2007, ketika aktivitas seismik dan laju pertumbuhan menurun, dan pertumbuban berhenti sekitar Mei 2008. Batuan kubah andesit basaltik yang dominan terdapat zonasi, dan secara tekstur bervariasi plagioklas, orthopyroxene, clinopyroxene, Ti-magnetite dan apatit kecil. Kubah tersebut diawasi dengan ketat untuk antisipasi akan aktivitas erupsi baru secara tiba- tiba. Namun, hanya gumpalan putih kecil yang muncul dari kubah setelah tahun 2008, hingga benar-benar hancur pada letusan Februari 2014 (gambar 2.2.b) Gambar 2.2. Citra satelit menunjukkan gunung berapi Kelut dan daerah sekitamya kawah sebelum (a) dan setelah (b) letusan Februari 2014. (Goode, Handley, Cronin, & Abdurrachman, 2019)2.3. Erupsi Februari 2014 Letusan pertama terjadi pada 13 Februari 2014 Pukul 22.00 WIB, diikuti oleh fase plinian utama kedua pukul 23.30, berlangsung total 44,5 jam. Citra satelit (CALIPSO) diambil pukul 01.12 menunjukkan bahwa semburan letusan mencapai ketinggian 18-19 km. Dicitrakan setiap jam oleh MTSAT-2 IR setelah pukul 23.32 waktu setempat, gumpalan dengan cepat diperluas, dengan bagian dari plumes naik ke ketinggian maksimum 26 km. Serangkaian foto dari saksi mata di Blitar diambil sekitar pukul 00.38 hingga 01.00 waktu setempat pada tanggal 14 Februari 2014, menunjukkan letusan secara bertahap meningkat intensitasnya, dengan awal gumpalan abu-abu gelap yang menjadi semakin pucat warnanya. plumes tersebar terutama di sebelah barat gunung berapi, di atas Samudra Hindia, dengan abu berwarna coklat pucat jatuh sejauh 240 km dari gunung berapi ke timur laut, barat laut, barat daya, dan barat. Pada 18 Februari, hujan lebat menghasilkan lahar yang ‘mencapai Kediri (35 km barat laut barat laut), Blitar (20 km barat daya), dan Malang (35 km timur), yang membanjit rumah dan menghancurkan Iahan pertanian dan beberapa jembatan. Tujuh korban jiwa dilaporkan di Malang karena runtuhnya atap ga untuk Kelud diturunkan menjadi level 2, dan zona eksklusif dikurangi menjadi dari akumulasi abu jatuhan, Pada tanggal 18 dan 19 Februari 2014, tingkat radius 5 km. Erupsi semakin menurun hingga 20 Februari, dengan abu vulkanik hanya mencapai ketinggian 1 km (Jeffery, 2013),BABII METODOLOGI PENELITIAN Pengumpulan sampel dilakukan pada bulan April 2014, Akses ke kawah puncak dilarang pada saat itu, sehingga sampel dikumpulkan di dekat, dan di dalam, lembah sungai di barat, selatan, dan barat daya puncak. Sampel dari berbagai tahap letusan 2014, termasuk jatuhan dan endapan PDC dikumpulkan dari sebelas daerah yang berbeda (Gambar. 3.1) pep a te ils Gambar 3.1. Titik lokasi pengambilan sampel (Goode, Handley, Cronin, & Abdurrachman, 2019). Data sedimentologi di setiap lokasi dikumpulkan mulai dari ketebalan lapisan, ukuran butir, struktur sedimen, penyortiran, hubungan kontak batuan, dan adanya aktivitas fumarol. Bukti pengerjaan ulang deposit dan kerusakan vegetasi dan struktur bangunan juga diperhatikan, Kronologi erupsi. secara langsung dibandingkan dengan. stratigrafi Endapan piroklastik 2014 yang diamati di lapangan, Gambar di bawah ini merupakan gambar-gambar dari endapan piroklastik yang ditemukan di lapangan.Gambar 3.2. Foto endapan vulkanik di titik-titik lokasi pengambilan sampel. a) endapan PDC di dekat lubang di K. Putih, mengisi lembah, dan mungkin tergores oleh curah hujan pasca-erupsi . b) Overbanking PDC dan endapan abu jatuh di lembah K. Badak, topograli terpal, dan menunjukkan perusakan dan perataan pohon. c) Dua PDC (1-PDC dan 2-PDC) di FI di K.Badak. d) Fumarol di atas endapan aliran piroklastik (2-PDC) di lokasi Fl. e) Breadcrust bomb di atas cendapan aliran piroklastik (1-PDC) di lokasi F1. f) Terfragmentasi blok batu apung di atas endapan aliran piroklastik kaya batu apung (I-PDC) di lokasi Fl. g) Kerusakan akibat jatuhan tephra pada atap gubuk (Goode, Handley, Cronin, & Abdurrachman, 2019).BABIV HASIL DAN PEMBAHASAN 4.1. Stratigrafi dan Dinamika Erupsi Gunung Kelud Tahun 2014. Dari hasil penelitian ini didapatkan data sampel yang diambil dari 11 titik lokasi yang berbeda dengan kode lokasi F5, F4, F2, F1, F3, F6, T1, $1, $2, $3, dan S4. Pada FS terdapat 3 sampel yaitu sampel K127, K128, dan K129 dengan tipe endapan overbank PDC-surge. Pada F4 terdapat 5 sampel yaitu K120, K121, K122, K123 dengan tipe PDC-surge dan K125 bertipe PD- flow. Pada F2 terdapat sampel K44 dengan tipe overbank PDC-surge (kaya litik). Pada Fl terdapat sampel K40 dan K37 bertipe PDC-marrix, KA0 dan K38_bertipe PDC-pumice, K42 dan K39 bertipe PDC-litik. Pada F3 terdapat sampel K113, K114 bertipe overbank PDC- surge, K115 bertipe overbank PDC - ash, dan sampel K116 bertipe tefra jatuhan. Pada F6 terdapat sampel K117, K118, dan K119 dengan tipe PDC- surge. Pada T1 terdapat sampel K110, K111, dan K112 dengan tipe endapan tefra jatuhan, Pada $1 terdapat sampel K35 dengan tipe endapan tefra jatuhan. Pada S2 terdapat sampel K62 dengan tipe endapan tefira jatuhan, Pada S3 terdapat sampel K75 dengan tipe endapan tefra jatuhan, Pada S4 terdapat sampel K105 dengan tipe endapan tefra jatuhan, Gambar di bawah ini merupakan kenampakan secara visual dari endapan- endapan tersebut Gambar 4.1, Gambar endapan piroklastik Gunung Kelud 2014 (Goode, Handley, Cronin, & Abdurrachman, 2019).Adapun urutan stratigrafi dari endapan piroklastik Gunung Kelud 2014 dapat dilihat dari gambar di bawah ini. ow Stage 3 cimactplnian phase Fina tephra fallout from the main, sstaned plinan eruption column \,__Foundsout- west primal, west mid south-west distal and north west distal ofthe \ vent ‘Access oad 2.3.3.8, 7 hm from vent; KB ddak-62, 68m from vent 1 1 1) Stocations, 7 samples - 51 (K35); 52 (RE2); 53 (e754 cos) dK 2.90 the partial overloading of basal portion of the (easel) pinion eruption column du to vent instbi- thes, and ater the pena af dome destruction Poce Found south prenimal, southwest proximal = ard E ofthe vent ? \ Putin -2.1 km from vent \ Badak- 2.6 km, 44,47 km fom vent 190¢ \ locations, 1 samples -F 13, 125; F1 (upper) \ (xa? - 9, F323 -K128) 6 (KA Stage 1, Onset phase - nial destruction of the 2007-2008 dome, and rita conduit and vent widening associated OCS Found south proximal and south-west prox malof the vent Putin 1.9 02.1 km from vent KC Badak- 2.6m from vent ‘locations, 1 samples FS (127-229); Fa (120 22); 1 (kD =A); Fk, KAS) 675 100 (Companentry (Wt. Gambar 4.2, Urutan stratigrafi endapan piroklastik untuk letusan Kelud 2014, (Goode, Handley, Cronin, & Abdurrachman, 2019) 10Berdasarkan stratigrafi batuan (Gambar 4.2) secara umum dinamika erupsi Gunung kelud dibagi menjadi 3 tahap. Tahap 1 merupakan fase awal, tahap 2 mengenai endapan dan pembentukan kolom plinian, dan tahap 3 merupakan fase klimaks dari peristiwa plinian dan Jatuhan tefra. 4.1.1, Tahap 1: Fase Awal Fase awal terjadi pada tanggal 13 Februari 2014 pada pukul 22:32 WIB, berlangsung sekitar 15 - 30 menit, Endapan piroklastik yang terkait dengan fase awal meliputi urutan endapan PDC (1-PS1, 1-PSL dan 1-PDC) dan overbanking.). Selama erupsi awal (tahap 1), PDC energi rendah dihasilkan hingga < 2,6 km. Lapisan basal menunjukkan karakteristik endapan yang mirip dengan lonjakan awan abu yang membawa sebagian besar abu halus dan fragmen Kristal. . Ini dibatasi oleh simpanan yang khas dari aliran piroklastik terkonsentrasi partikel tinggi. Semua endapan tahap | memiliki kandungan fragmen litik padat yang tinggi (44%), bersumber dari kubah Java 2007-2008 dan dinding saluran, menunjukkan bahwa letusan onset didorong oleh pelepasan eksplosif dari tekanan berlebih gas di bawah kubah tersebut. 4.1.2. Tahap 2: Pembentukan kolom plinian Tahap 2 adalah dimulai pada pukul 22:58 WIB. Erupsi tahap 2 ini menyebabkan pembentukan kolom plinian dan runtuhnya kolom marginal. Pada tahap 2 ini, endapan PDC yang kaya akan pumice (70%) mengalir hingga 4,7 km dari ventilasi. Endapan tersebut mencerminkan peningkatan keluaran magma segar yang terfragmentasi, dan beberapa pelebaran saluran yang dibuktikan dengan padatnya fragmen litik. Ketidakstabilan ventilasi dan penggabungan material padat ke dalam margin basal plume inilah yang menyebabkan keruntuhan marginal dan pembentukan PDC ini. 4.13. Tahap 3: Fase Klimaks Fase klimaks plinian akhir terjadi pada jam terakhir di fase puncak erupsi plinian, sekitar pukul 01:00 - 02:00 WIB. menghasilkan kolom plinian hingga ketinggian 26 km yang membentuk seperti payung. Produk dari Tahap ini mengandung, endapan jatuhan lapili dengan struktur invers, debu sedang hingga lapili kasar, batu pumice (290%) dan tefra jatuhan, Endapan ini menggambarkan bahwa adanya raletusan yang semakin lebih besar seiring berjalanya waktu, dengan peningkatan fluks massa dan ekspansi gas segar magma, dan fragmentasi lanjutan yang ‘mendukung terjadinya letusan plinian. Di bawah ini disajikan gambar mengenai dinamika erupsi dari tahap 1 sampai tahap 3 Gambar 4.3. Diagram skema dinamika yang terjadi selama Tahap 1 hingga 3 letusan dengan waktu mulai dan durasi letusan (dalam UTC dan waktu setempat) untuk setiap fase (Goode, Handley, Cronin, & Abdurrachman, 2019) 4.2. Pola Erupsi Gunung Kelud dan Implikasi Bahaya Erupsi Kedepannya Salah satu cara untuk memprediksi pola erupsi gunung berapi adalah dengan melihat pola erupsinya, walaupun cara tersebut belum tentu akurat, Jika melihat dari sejarah dan pola erupsi (abel 2.1), Kelud memiliki_ sifat gaya erupsi yang umumnya siklik antara eksplosif hingga efusif sejak tahun 1900-an, menunjukkan bahwa letusan berikutnya Kelud bisa menjadi periode pertumbuhan kubah atau istirahat yang panjang, berlangsung antara 7 hingga 31 tahun. Ini akan tergantung pada kondisi degassing magma dan beberapa faktor lainnya. Meskipun pola di Kelud tampak sangat teratur baru-baru ini. Namun, ada juga periode semi-kontinu dari letusan kecil dengan peristiwa yang lebih besar secara sporadis (VEI 3 ke 5). Umumnya letusan paling besar (1586) diikuti oleh periode istirahat terpanjang (dalam hal ini 75 tahun). Bahaya dari erupsi Kelud ini yang berpotensi terjadi di masa yang akan datang berupa bencana aliran lahar dan awan panas. Pada tahun 1586 M dan 1919 lahar pasca-erupsi menyebabkan kematian hingga 10.000 korban jiwa. Bahaya dari 12aliran labar ini diprediksi akan terus berpotensi jika terjadi erupsi kembali, apalagi dibeberapa bagian gunung kelud banyak terdapat danau kawah yang tentunya dapat ‘memperparah bencana aliran lahar tersebut jika material piroklastiknya tercampur dengan air danau kawah, Bencana awan panas juga harus diwaspadai, Awan panas ‘mengandung gas fluor, belerang, H2S, magnesium, dan kalium yang sangat berbahaya untuk dihirup. Awan panas ini juga memiliki suhu yang sangat tinggi bisa mencapai 100° C dan dapat mencapai kecepatan sebesar 200 m/ detik, tentunya hal tersebut sangat sulit dihidari jika terdapat manusia dalam radius awan panas tersebut. Untuk kedepannya, diharapkan adanya peningkatan sistem peringatan di aliran lahar atau awan panas agar dapat meminimal ir jumtah korban jika terjadi erupsi di masa yang akan datang, 13BABY KESIMPULAN Erupsi Gunung Kelud 2014 dibagi menjadi 3 tahap. Tahap 1 merupakan fase awal, tahap 2 mengenai pembentukan kolom plinian dan runtuhnya kolom marginal , dan tahap 3 merupakan fase klimaks dari peristiwa plinian dan Jatuhan tefra. Pada endapan piroklastik Tahap pertama memiliki kandungan fragmen litik padat yang tinggi (44%), endapan_piroklastik tahap kedua kaya batu apung (pumice) (70% ) dan tahap ketiga merupakan endapan jatuhan lapili bergradasi terbalik (invers) dengan batu apung (90%) dari gumpalan abu vulkanik setinggi 26 km. Jika melihat Sifat gaya erupsi yang umummnya siklik antara eksplosif hingga varian efusif sejak tahun 1900-an, menunjukkan bahwa letusan berikutnya Kelud masuk periode pertumbuhan kubah atau istirahat yang panjang, berlangsung antara 7 hingga 31 tahun. Sementara itu, Bahaya dari erupsi gunung seperti aliran lahar dan ‘awan panas juga harus diperhatikan karena tetap akan berpotensi terjadi lagi di masa yang akan datang. 14DAFTAR PUSTAKA Bourdier, J.-L. P.-C. (1997). Observations Stratigraphy and Eruptive Proces of 1990's Eruption of kelud volcano. Volcano Geothermal, 181-203. Goode, L. R., Handley, H. K., Cronin, S. J., & Abdurrahman, M. (2019). Insight Into Eruption Dynamics From the 2014 Pyroclastie Deposits of Kelut Vulcano, Java, Indonesia, and Implications for Future Hazards. Journal of Volcano and Geothermal Research, 6-23. GVP, G. V. (2008). Report on Kelut (Indonesia). In: McClelland, L. , Bulletin of the Global Volcanism Network. Jeffery, A. G. (2013). volcano. The pre-eruptive magma plumbing system of the 2007-2008 dome-forming eruption of Kelut volcano, East Java, Indonesia, 275-308. Zaenuddin. (2007). Penyelidikan Endapan Piroklastika Gunung Kelud, Jawa Timur. Laporan Pengamatan dan Penyelidikan Gunungapi. Bandung: Pusat Volkanologi dan Mitigasi Bencana Geologi. 15Journal of Vleanology and Geothermal Research 382 (2019) 6-23 Contents lists available at ScienceDirect Journal of Volcanology and Geothermal Research ELSEVIER journal homepage: www.elsevier.com/loc: elivolgeores Insights into eruption dynamics from the 2014 pyroclastic deposits of Kelut volcano, Java, Indonesia, and implications for future hazards Louise R. Goode *“, Heather K. Handley *, Shane J. Cronin ®, Mirzam Abdurrachman ‘ * Deparment of Ear and Planetary Sciences Mecquarie University Sey. NSW 2108, Ausra ® Schoo of Enronmnt University of clan, Private Bog 92019, Acland 112, New Zelend “ Depatent of Geolpcal Ener Bonding stu ofTechcegy endure 4012. donesio at ARTICLE INFO ABSTRACT ‘ae ay Received 13 November 2016 Received in eied fr 19 December 2017 Accepted 6 February 2018 ‘Ava online 21 Febuary 2018 (0n13t0 14 February2014a~A/n long, VEI4 eruption occured at Kelut volcano (va, Indonesia, Pyroclastic der- sity currents (PDCs) and extensive ash fllled to 7 fatalities and disruption to flghts across the Acis-Pacic region. "New sedimentological descriptions ofthe pyroclastic deposits from the 2014 eruption were compared with eye- Witness and satelite reports to elucidate temporal variations in eruptive dynamics. The stratigraphy ofthe de Posts is presented in 3 stages, associated with two eruptions that occurted approximately ~15-30 min apart. Faywors Stage 1 PDC deposits originate ftom the smaller onset eruption. The PDC deposits fom Stage 2, and tephra fall pel ‘deposits from Stage 3 originate from the second, plinian eruption. During the onset eruption (Stage 1) lw en- ergy POCS were produced tht ran outro <2.6 km, Basal layers show characteristics of deposits similar to ash cloud surges that carried dominantly fine ash and crystal fragments. These ae capped by deposits typical of high-partile concentrated pyroclastic lows. All Stage 1 deposits have high contents of dense lithic fragments (up to 44% by vol), sourced from the 2007-2008 lava dome and conduit walls, indicating that the eruption ‘onset was driven by an explosive release ofgas-overpressue below the vent-capping dome. Increases in the ‘magma flux and transition to a more constant eruption ed to a growing eruption column during the -2-hour long Stage 2plinian eruption. Pumice rich (70% by vol.) PDC deposits ran out to 7 km fom the vent. The de Posts refect an increased output of fresh fragmented magma, and some conduit widening evidenced by dense lithic fragments. Vent instabilities and incorporation of dense material into basal margins ofthe plume led to the marginal collapse and formation of these PDCS. tage 3 occurred inthe final hour at the peak ofthe plinian etuption, around 01:00 to 02:00, and produced reversely graded lap all deposits with <90 volx pumice fiom a 26 kin-high plume. This indicates that there was a sustained flux of juvenile magi tothe now open ‘vent system, and expansion and fragmentation ofthe gas-ich magma was a its most efficient. Our study of the eruptive sequence of Kelut provides some constraints on predicted pattems for future explosive activity, crit- ical for further hazard assessment ofthe voleano. Since 1901 Kelut has erupted on intervals f | to 23 years, ad the 2014 event characterises atypical “explosive” style of eruption that alternates regularly wit effusive dome- formation and collapse events. Ths pattern depends on the dynamics of magma renewal othe system, and degassing conditionsin the shallow magia reservoir and upper conduit. ths pattern hols, and the next erup- {on occurs within the next two decades, a return of prolonged dome growth could be anticipate. ‘©2018 The Authors Published by Elsevier BV. This isan open access article under the CCBY-NC-ND license (ht) aeativecommonsorgliensesfby-ne-nd/0). Uva dome Precast Bow yoda surge yoda denstycurest Pinan “Toparafa 1. Introduction Kelut volcano in East Java is renowned for its unpredictable, often destructive and hazardous explosive eruptions, and the production of deadly syn- and post-eruptive lahars (Bourdier etal, 1997; Bourdier et al, 1997a: De Bélizal et al, 2012; Jeffery et al, 2013: GVP, 2014). Due tothe frequent presence ofa crater lake and densely populated sur= rounds (Fig. 1), Kelutis considered one of Indonesia's most deadly vol- canoes (Bourdier et al, 1997; Siebert etal, 2011; Jeffery etal, 2013). To * Ceresponding author. sal adress use goode@maeduau (LR. Goode) tps: feior/101016Avlgrores2018.02005 address the lahar hazard a sophisticated tunnel drainage system for the lake in started construction in 1919 and was rebuilt following its de- struction by subsequent eruptions (Zen and Hadikusumo, 1963) Historic eruptions have varied in magnitude with Volcanic Explosivty Index (EI) ranging from 2to 5 (Table 1). ince 1919, erup- tions of Kelut have shown a cyclic pattern of alternating VEI-4 explosive eruptions (destroying lava domes) and periods of lava dome effusion (Table 1; Zen and Hadikusumo, 1965; Pratomo, 1992; Bourdier et al, 1997; De Bélizal et al. 2012: GVP, 2012, 2014: Jeffery etal, 2013; VSI 2014) Similar alternating patterns are common regionally and globally at dome-forming atc volcanoes such as Merapi, Indonesia (Newhall et al, 2000; Andreastuti et al, 2000; Gertisser et al, 2012; Cronin (037-4273/0 208 Te Authors Puli by Ekever BV. This an ope access are under the CCBY-NC.ND license (hp/ceavecommonsngicensesy-n- 40,{LR God ura of Volcanology and Geothermal Rescrch 32 (2018) 6-23 ae ee ee @ ste 120: sanone A. 2014 en 2007/8 dome aon ig. 1 Topographic map of Kea olan show Merap! sove 110 tak = im Semers Pervaieterd we ao" t90" 200" aa" (A ckentsumann(a7aim — — Mines 7-7 catesin A. von summasloser somes the lcton of sampling sites (Modified from 8D, 2013) The main rater rim opens to the south-west of the ede Contours rin 200 minterals. Grey arow shows the prevailing wind ection reported in Krstansen ta (2015). Sampling ites 51 and T are al samples, 51-54 denote sites where POC {epost ate absent and the l depos est on thn layer asl above the 1950 deposits Sites FI*=FE are srples fund in 22014 stratigraphic Sequence of aPDC depesit (not I ‘nots fw PDC sequences observed atone lean), with some tephra fal manting the tops ofthe PDC Inset shows a DEM of Java with the locaton of Kelutvokane marked. Tecerson inserted map show the locations of K = Kei, B~ Bitar and M ~ Malang around Kea volar, et al, 2013; Preece et al, 2014), Colima, Mexico (Robin et al, 1991; ‘Arambula-Mendoza et al, 2011; Lavallée et al, 2012), Montserrat, Lesser Antilles (Druitt and Kokelaar, 2002; Edmonds and Herd, 2007) and Shiveluch, Kamchatka (Gorshkov and Dubik, 1970; Dirksen etal, 2006; Zharinov and Demyanchuk, 2013). Previous research on volcanic activity and deposits at Kelut has concentrated on the explosive 1990 and effusive 2007-2008 eruptions (Bourdier et al, 1997; Bourdier et al, 1997a; Thouret et al, 1998; Hidayati et al, 2009; De Bélizal cet al, 2012; Jeffery et al, 2013) After the 1990 (VEI 4) plinian eruption (described by Bourdier et al, 1997) there was a 17-year quiescence, followed by a dome-excrusion episode in 2007-2008. The growing ‘dome gradually displaced most ofthe crater lake, without producing la- hrs (Hidayatiet a, 2009; Jeffery et al,2013).On 13 February 2014, an- ‘other VEI 4 plinian eruption occurred, destroying the 2007-2008 dome (GvP, 2014, 2014, 20140), with two eruptive pulses occurring over ap- proximately 4.5 h (GVP, 2014; Kristiansen et al, 2015). In order to understand the eruption dynamics and mechanisms of ‘emplacement of pyroclastic deposits during large explosive eruptions at volcanoes like those of Kelutin 1990 and 2014, it isnecessary to com- bine detailed descriptions ofthe volcanic deposits (stratigraphy, lithol- ogy, size and distribution) with geophysical and eyewitness accounts (eg, Bourdier et al, 1997; Carey et al, 2007; Pardo et al, 2008, 2012; Tuer et al, 2011; Cronin et al, 2013). Here we present a summary of the observational accounts, field sedimentological data, and ‘componentry and grain size analysis of the 2014 pyroclastic deposits to reconstruct the eruption chronology, and provide insight into the ‘eruption dynamics. The significant features of the 2014 eruption are de- scribed inthe context ofthe overall pattem of apparent cyclic behaviour atthe volcano, which will asist in forecasting future eruption styles and hazards, 1.1, Geological seting ‘Volcanism on Java is attributed to the Sunda arc subduction system, formed by the northward subduction of the Indo-Australian plate be- neath the Eurasian plate at ~6-7 cm/yr~' (Van Bemmelen, 1949; Hamilton, 1979; Hall, 002,201). Although it appears to have the gen- eral form ofa stratovolcano, Kelutis best described as adome-complex. Ithas a complex morphology and summit profile, with several peaks ‘composed of large lava domes, hosting numerous craters, and steep val- leys (Fig. 1; Wirakusumah, 1991; GVP, 2014b). The main edifice was constructed by the growth of three lava domes and multiple viscous and large andesitic lava flows. Explosive eruptions have produced era- ters Co the east, south, and west of the summit area, The eastern crater area is dormant as there has been a clockwise migration ofthe active ‘vent location (Wirakusumah, 1991; Humaida et al, 2011; De Bélizal cet al, 2012; Jeffery et al, 2013). Current activity occurs within a hhorseshoe-shaped scar that opens to the west (Fig 2). This crater con- tains a 1 km-wide lake that has often affected the style of eruptions and produced voluminous lahars (eg. the 1586 eruption) (Pratomeo,8 {LR Goode ett Jour of Vero nd Geothermal Research 382 (2018) 6-23, ‘able own erupie history of Kut velcao, with deta of indvidual eruptions, ruption Year Erion VEI__Volume (kn) Depo Compson 1099 Pe 7? wT 7 mm Poe, 3} ry 2 04 Pe xD 3? unt 3 176 Poexp exp 3} iz 2 185 Pex ex? 3? uy 3 185, Poesy exp 3 ? 2 ep 3? > 2 ep 30 ? 7 ' ep 3? ? 2 ' ep 30? > 2 7 ep 3? ry 2 t ep 5 os us 3 1 ep «on ? 2 X op 2 2 a 7 1 op 2 2 > ? t ep 2 2 w 2 1 ep 2 2 ? ? t ep 2 2 > 7 ' ep 2 2 1 2 1 op 2 2 ny 2 3 ep 4 on Pen > 1238 185 ep 2 7 its > ns 88 ep a? us > 3 1831 op 2 2 ing > t 1804 ep 2 2 ry > and references within) 1801 =p 3 02 mertun 2 5 (and references itn), 8.9 1919 ep 4 oo Teun 2 5 (and references within), 8.9 10 ep: 2 7 Nosubstantal deposits 7 1951 exp Da ao merih > 13,5 (and references within) 8 1996 ep 4 ao mouket > 35.89 1967 1? Nosubstantial deposits? 3 1967 7 1? Nosubstanialdeposis 7 8 10 Pexpexp (pinion a = 4 a3 TEDL PS Lh asicandeste —2.3.5.7.8 2007-2008 =D 2 03s Some dome remnants Basic andesite 3,45 (and references within, 6,7.8, ao Pexpexp (pinion): Da 4 Ot TEDL PS Un asicandeste 69,10, ‘oie for eusions oop nal pretonagraticrupton exp: expsve eruption acetal cater, D: dome son; Do: destruction of ome Code for deposits: tephra 2 pyrecase Now: Laka s: prose surge Volunesreerted at ul deposit vole. Al Vaan Explsiny nes (VETS fem GVP (2014), ferences: Zen and Hacikusuo, 19652: Prtome, 1092; 3: Source ta, 1997; 4: GVP 20085: De BAizalet al. 2012; 6: GVP, 2012; 7: ery etal 20188 GP, 2014;0: VSL. 2014; To: Kestnsen eal, 2015, 1992; Bourdier etal. 1997; De Bélizal etal, 2012). The volcanoes flanks comprise bedded pyroclastic density current (PDC) deposits tephra fall ‘deposits, with lahar sequences forming several large radial laharc fans (Bourdier etal, 1997; Thouret eta, 1998), 1.2. The pre-2014 eruptive history of Kelut Kelut has been periodically active since 1000 CE with 34 recorded eruptions (Table 1). Most recorded episodes were short-lived, violent, ‘ypically subplinian to plinian eruptions (VEI 3 to 4). ix eruptions of this style have occurred since 1901 also forming destructive syn- and, post-eruption lahars related to the summit crater lake (Bourdier eta, 1997; Siebert etal, 2011; GVP, 2014a). The largest and most deadly re- corded eruption of 1586 (VEI 5), and its associated lahars, caused 10,000 casualties (Zen and Hadikusumo, 1965; Bourdier etal, 199; Van Bergen et al, 2000), Pyroclastic deposits were mainly andesitic in composition, with the more recent eruptions producing basalt andesites (Bourdier et al, 1997; Jeffery et al, 2013). The 1919, 1951, 1966 and 1990 eruptions produced similar extensive tephra fall and PDC deposits and syn- and post-eruptive lahars (Bourdier etal, 1997; Siebert et a, 2011). 12.1. The 1990 eruption ‘The VEI 4 1990 eruption occurred after 24 years of dormancy at 11:41 local time, 10 February 1990, and although it was short lived it caused 32 fatalities. The eruption completely removed a growing em- bryonic lava dome and the crater lake, excavating a new 400 m x 500 m-wide crater (GVP, 1990; Bourdier et al, 1997). The eruption produced a total of 0.13 km’ of basaltic andesite volcanic rocks, includ- ing extensive pumice- and scoria-rch PDC deposits reaching kmsouth and west of the crater. Widespread fallout comprised 0.10 km? of the total deposit volume (GVP, 1990a; Bourdier et al, 1997). The first phase of the eruption was a phreatomagmatic explosion, followed by seven discrete eruptions and associated “wet ash” fallout. The eruptive products became progressively darker in colour throughout the eruption, although no compositional variation was noted. At 12:35 a major eruption produced a sustained, plinian eruption col- ‘umn that was dispersed by wind to the south-west. A further pulse of activity occurred around 15:10 to 19:30, and was accompanied, by ash and accretionary lapilli fallout from a 12 km-high plinian plume. The eruption decreased in intensity throughout the night, ceasing on 12 Februaty (Bourdier et al, 1997). Bourdier et al. (1997) noted that on the 7 April 1990, a small, embryonic lava dome was growing in the summit crater in muddy water, measuring 2 few tens of metres in height. The new dome was compositionally and petrographically similar to the material erupted in February and was subsequently covered as the lake re-formed (Pratomo, 1992; Bourdier et al, 1897) 1.22, The 2007-2008 eruption ‘Starting in November 2007 and continuing over a seven-month pe- riod to 2008, a lava dome was extruded in the crater, which almost completely displaced the crater lake (Fig. 2a) (GVP, 2008; Hidayati et al, 2009; jeffery etal, 2013). Plumes of steam rose 2 km from phre- atic activity, although there was no explosive magmatic activity assoct- ated with this eruption (GVP, 2008). The dome grew rapidly until[LR Goode Journal of Vanco and Geothermal Resarch 32 (2019) Fig.2 Satelite imagery showing Keli vokano and peripheral areas surrounding the ‘ter bale (a ater (0) the February 2014 etupton 3 Caprted on the 2378 2012 showing the 2007-2008 dome sl ocupying the sum cater ake.) Captured ‘on the 19}052014 after the 2014 eruption showing the desuucin ofthe 2007-2008, {dome and creation of te new eater andthe 2014 vet lcation. Re line how he ‘ater ates of Kelut in both a) and b). White dashed line ina) shows the ote ofthe 2007-2008 avs dome. Pn line nb) shows the ection of the 2014 eruption vent ‘range ie nb shows the extent of verank and aly confined PDCs inthe K Path and K Baga rom tis sty (incloding ash lanketing the peripheral proximal rater ue) Yellow lin sows the extent of pyroclastic deposits in he proximal ata nth “nd north-east ofthe went, not sampled inthis study, White dashed ines and arrows Stow the deposition of futher extensive valey-coained PDC, andar post-eruptive Secondary ars dawn the surrounding ever valleys. Acquired wig Case Earth CopyigheDigalcabe Ine. 2015), December 2007, when seismic activity and growth rate decreased and growth ceased around May 2008 (Hidayat a zal et al., 2012; Je 201 3). A this stage it was 260 m concluded from petrologica studies thatthe eruption was preceded by a complex interplay of magma mixing, crustal assimilation and decompression-driven resorption of the crystal cargo. The domi- iantly basaltic-andesite dome rocks contained zoned, and texturally varied plagioclase, orthopyroxene, clinopyroxene, Ti-magnetite and minor apatite (Jeffery et al, 2013). The dome was closely monitored in fear of renewed sudden explosive activity (GVP, 2012: De 12). However, only small white plumes rose from the dom¢ after 2008, up until it was completely destroyed inthe Febru- ary 2014 eruption (Fig. 2b) 2 9 2. The February 2014 eruption ‘The most recent explosive eruption of Kelut occurred on 13 February 2014, which ejected the remaining water of the crater lake and completely excavated the 2007-2008 dame (GVP, 20 2). The ‘general nature and extent of the deposits are similar to those ofthe Feb ruary 1990 eruption 2.1. Pre-eruption observations ‘The Indonesian Centre for Volcanology and Geological Hazard Mit gation (CVGHM: known locally as Pusat Vulkanologi dan Mitigasi Bencana Geologi, PVMBG) described seismic events at Kelut from mid- January onwards (GVP, 2014). Prior to the eruption, volcanic earth- quakes occurred between 2 and 8 km deep, with 14 seismic events re corded in the first week of January, rising up to 693 the week before the eruption. Titmeter data showed inflation ofthe dome atone station. The remaining crater-lake water warmed by 5.5 °C from September to 2 February 2014. At 21:15 local time on 13 February 2014, the Indonesian National Board for Disaster Management (Badan Nasional Penanggulangan Bencana, BNPB) raised the alert level to the highest level (4) meaning, “Awas” or danger/warning, with eruption expected, within 24 h (see De Beli (2012) and Andreastuti etal, (2019 for a description of alert levels). A 10 km-radius exclusion zone was set up, leading to the evacuation of -166,000 people (Andreastuti 22. Syn- and postenuption observations ‘The BNPB reported that the first eruption occurred at -22:50 local time on 13 February 2014, followed by the second, main plinian phase starting -23:30, lasting 4-45 h in total (GVP, 2014), The 2007-2008 dome was destroyed during this time, and a new 400 m-diameter crater excavated (Fig. 2b). Satelite imagery (CALIPSO) taken at -01:12 local time showed that the eruption plume reached ~18-19 km in altitude Imaged hourly by MTSAT-2 IR after 23:32 local time, the plume rapidly expanded, with parts of the plume rising to a maximum altitude of 26 km. A series of photographs from a local eyewitness (Putra, 2014) in Bitar taken from around 00:38 to 01:00 local time on 14 February 2014, show the eruption gradually increasing in intensity with an initial dark-grey plume that became progressively paler in colour. The phume dispersed primarily west of the volcano, out over the Indian Ocean, with pale-brown coloured ash fallin as far as 240 km from the volcano to the north-east, north-west, south-west and west (GVP, 201 Kristiansen et al, 2015; Fig. 1), The ash disrupted 40 flights from air- ports across indonesia, including Yogyakarta (-215 km west), Surabaya (-91 km north) and around the Asia-Pacific area 201 ristiansen et al, 2015). Satellite imagery from The International Charter Space and Major Disasters (2014) shows extensive PDC de- posits, which destroyed forested areas proximal to the crater, and were directed down the Kali (meaning river, abbreviated to K.) Sumberagung, K. Badak and K. Puti valleys, reaching approximately ‘4 km from the crater (GVP, 2014; Fig, 2). Grey-to-black plumes accom= pany successive eruptions, reaching 400-600 m in height above the cra- ter on 14 February, with further 3 km-high, grey-white steam-rich plumes observed on 15 February 2014, On 18 February, heavy rainfall ‘generated lahars that reached Kediri (35 km west north-west), Blitar (20 km south-west), and Malang (35 km east; Fig. 1 insert), which flooded houses and destroyed arable land and several bridges (Dibyosaputro et al, 2015). Seven fatalities were reported in Malang, due to the collapse of roofs from accurmulated ash fall deposits (C 2014). On 18 and 19 February 2014, the alert level for Kelut was lowered to 2, and the exclusion zone was reduced to a 5 km radius (GVP, 2014). Eruptions further declined until 20 February, with steam plumes reaching only 1 km altitude.0 {LR Goode ett Jour of Verlag nd Geothermal Research 382 (2018) 6-23, 4. Strategy and methodology Fieldwork and sample collection took place in April 2014. Access to the summit crater was prohibited atthe time, thus samples were col- lected near to, and within, river valleys to the west, south and south- west of the summit (GVP, 2014; Fig. 1). Samples from different stages of the 2014 eruption, including fall and PDC deposits were collected from eleven different localities (Fig 1). At each locality sedimentological data was collected: individual bed thickness, deposit grain size (using the classification scheme of White and Houghton, 2006), bedding and sedimentary structures, sorting and contact relationships, and the pres ence of fumarolic activity, charcoal and breadcrust bombs was noted (Fig. 3). Evidence of deposit reworking and damage to vegetation and ‘building structures was also noted, Observational accounts ofthe 2014 eruption chronology were directly compared tothe stratigraphy of the 2014 pyroclastic deposits observed in the field Grain siz distributions (GSDs) were obtained of 20 samples from the 2014 pyroclastic deposits by sieving at 16 intervals from <63 mm. (—66) to <32 ym (54). The componentry of 18 samples were deter- ‘mined by counting (where possible) up to 500 grains in each grain size fraction selected, where for most samples the <8 mm (34) to '500 um (Oi) fractions were used, However, forthe finer ash fall deposits the 1 mm (08) to 125 jum (36) fractions were used for componentry analysis. The lithological classes used in componentry analysis are listed Fig 3. Photos ofthe 2014 vlcani depots. Steaming PDC deposit near the vent inthe K. Pt nling the valley, and possibly ince by post-rupon rfl rar ewerking ‘material b] Overanking DC and ash ll deposts down the Hada val shetng topographic highs and showing ree destruction and faeningc) Pwo PDS (1-PDCand 2-P0) at FL incheK Bada 6) Furmarcles ontop ofthe pyrocasic ow depos (2-PDC at acaion Fi. readers bam on top ofthe pyro Now deposit (1-FDC) asi FI.) Fragmented mie beckon op ofthe purice-eh preci w depos (1-PO) at st Fs) BalisicRopya fall damage onthe oa faut lng the main access ead othe VERE Se xt ox futher discussion Phitograps taken between the O204/2014 an O90472014{LR God {ura of ocanolgy and Geothermal Research 382 (2018) 6-23 ‘able tology cssfcaons ued in componenuy ana induing desaipcon of sub-clasifcatons ‘Gein oe eras ‘ee ah Daraae poe Acieatirave ceaney ck) Acido (county sc) = aT, i” (Colour Lighgeyorkghttnont ———-Dalkyellowsh ro Dar greylack gh ey ask greyback ‘Textae Viti prow vecuke Vin pertfonny vesicat High veser Perec Parnoeite Shape Elongded~sbspencl——loptet—sabsehea Sibspieea bee —Nocky Elongated chy Rounding Sohmgar-soboumied ——_Angulr—satranod Sihnded Song Sitges ese shape Aner oct cnn and mmm ee flog ond ey ore, aniaiamiiél ‘sub-sphesical to spherical ‘with smaller angular vesicles. Comte Vi Ving ten neied Vii Mower afplaeciac md Merry of placa nd iin eed! Mitel atitnt MAME totes El reser i tempos Semel ain ans ment - ns eine nerve nrioprnee, STIS Yeisen dy ein to us viata» Me maul vest Steciartmistons Skesin lmweiniskin vie epee. Saha wel pty Oc Deernn acronis tite saaeetiet say wth nication map Spence Gea shaped clasts Very vn: with ‘CHS porous. Often cenfimetric | plagiockaac thon pumice clasts G10 (enn. Less ‘commen, Less vesicular than pumice bm ae ‘erysal ‘clots’ appear in large vesicle Vescikes ate also — = ‘yea ony tune 53 mm) were counted fr each section, Note that tic ad pumice have been sampled separately, The lower POC Section atF(1-PDC) has pumice casts (sample Kal) and dense lithic ess (sample 14} are represented by red brs. The second PDC 2 Fl (2-PDC) has pumice casts (18 and danse ities (sample X30) ae vpresentd by ble bars (se key, 422, Plinian-phase PDC units “The plinian-phase PDC units and associated overbank sub-units (see Table 3 starting with 2- unit code) are observed up to 2.1 km from the vent inthe K. Puih valley (section at F), and 26 ta 47 km down the K. Badak valley (upper section at FI, and sections at F3 and F6; Figs. 1,20 and 4). There is one main vertical PDC deposit succession associated ‘with this stage of the eruption, and its defined by vertical facies varia- tions: 2-PDC is a bipartite deposit (similar to 1-PDC), with a basal low angle laminated medium ash, and a coarse, massive and blocky, ppumice-rich upper deposit with litle matrix material (Figs. 4, 5 and 6a, b).Proximally, the basal facies of 2-PDC have discrete low-angle ‘ross-bedding, but laterally continuous overbank deposits at greater distance from the vent have clearer low-angle t horizontal laminations in sub-facies (at F3,F6) (Fig. 4). Dissimilar to the onset-phase units, the ‘deposits become thinner with distance from source, from 0.8 m thick in the K.Putih (Fé) and proximal K. Badak locations (F1), to 0.3m thick at FB and F6 (Figs. 1 and 4). ‘The 2-PDC deposits occur above the onset-phase PDC deposits at lo- ‘cation Fé and Ft (Figs. 1 4, 5, 6b). The plinian phase PDC at locality F1 6c) lies directly an top of the onset phase PDC deposit sequence (Fig. 30), and is distinguishable from 1-PDC due to the larger volume of the deposit (Fig 6a, c), that is also more pumice-rich and contains less block-sized clasts (Fig. 7). The lower-most climactic stage PDC de- posit at Fis 40 cm thick, with low-angle cross bedding of the fine- ‘medium ash matrix (125-500 um), and represents the basal deposits of the climactic PDC sequence. The deposits consist of mostly juvenile, pale to dark grey-brown pumice clasts, with dense lithics and crystals, (Appendix Table 1; Fig 5). The fine lapili (4-8 mm) inthis basal facies ‘comprise 71% pumice and 25% litics. In both sections at Fé and Fi, this finer-grained facies coarsens upwards into a very pootly sorted, massive ‘deposit with a matrix of medium ash, with randomly dispersed fine lapili6 {LR Goode ett Jour of Verlag nd Geothermal Research 382 (2018) 6-23, acedenaouneyrock Bugresense ies risied dese cs ig cena 8) pees ae 2 6 vesyane—ersums BOCAS nism Dean EL Fig.8. Pot oftox popotons of varying ithe cast (for 1-2 mm sie action} pe unit, ‘with sample name labelled. Amounts ofthe follwing ltholoie re shown (weit Detcentage): Dark, dense lithic Tava class: light, dense lithic lava dats: brown, ‘cient this curry rock and cud danse older, weathered uc Cs tw block-sized clasts (Figs. 4,5, 6c). ln general, the climactic PC deposits contain greater pumice contents chan the onset-phase deposits (Fig. 5). ‘The coarser grained clasts within the upper faces of 2-PDC are very poorly-sorted lapill and blocks from 3.5-13.5 cm in diameter of dense light grey, vitric juvenile pumice, and darker brownish-grey oxidised [pumice (sample K38), and dense grey to black lava (sample K39) ithics| 14-20 em in diameter (Figs. Ge and 7) This deposit contains less oxidised pumice (by ~5~G%) than the onset phase PDC deposits. ithic lpili and blocks within this deposit are dominantly lightly coloured, dense lava clasts. There is also a higher percentage of oxidised dense lithics and brown accidental lthcs than that recorded in the upper 1-PDC deposit (Fig. 8). The flow front of 2-PDC ison top of the flow front from 1-PDC at the same location (both 3 km from the vent), and comprises a ‘matrix-poor levee of blocks of pale grey or greyish-brown pumice, ‘The laterally continuous associated overbank sub-facies from 2-PDC include 2-oPS1, 2-oPS2, 2-PF3 and 2-oPDCa are observed at localities 3 and F6 (Figs. 1 4, and 5) and show similar facies and componentry to the 2-PDC unit at Fi, but are better sorted. At F3, the beds are 10-15 em-thick, consisting of grey to greenish-grey, medium ash show- ing low-angle laminations (Fig. Gd). The lowermost deposits (2-oPS1) hhave a matrix of medium to coarse ash, with 20-40 mm pumice fine a- pill (74%), and minor lithic lapilli (26%). The 2-oPS2 deposits directly above 2-oP I, and is very similar in componentry, but has a finer ash ‘matrix (125-250 jm) (Appendix Table 1), Above 2-oPS2 at F3, there is a thin layer of crystal-rich fine ash (2-PDCa), which shows a gradation "upwards to pale grey, pumice- and crystal-rch, fine ash (63-125 mm) with low-angle laminations (Fig. 6d). The section at location F6 (eleva- tion 741 m, Fig. 4) is 200 m downstream of F3 (782 m) and shows fur- ther lateral continuation of the overbank flow sequence and sub-facis. ‘The 2-oPS1 deposit at F3 thins to -2 cm at F6, and is dominated by coarse ash. The 2-oPS2 deposi is 5 cm thicker at F6 than at F3. Pumice blocks up to -20 cm in diameter and minor dense fine lithic lapilli cap the sequence at FG (2-oPF3). All the climactic-phase PDC units are characterised by higher pumice to lithic contents (Fig. 5). Pumice con- tents within the deposits of the overall eruption increase from 62 to 70% at Stage 1, to 85% at Stage 2 (Fig. 5). 423. Climacticplnian tephra fall ‘There is one main deposit associated from this Stage ofthe eruption; the final plinian tephra fall unit, 3-TF, is found at sites S1-S4,F2, F3, FS andT1 (Figs land 4; Table 3; Appendix Table 1) Allthe deposits are sam- pled under the broad dispersal axis ofthe plume (Fig. 1). Where PDC de- posits are absent (at S1-S4), the fll deposits rest ona thin layer of sol, forming on top of the 1990 pyroclastic deposits, and found along the -main access road parallel to K. Sumberagung (Figs. 1 and 4). The 3-TF ‘nit varies in thickness from ~0.5-5 cm thick. At 2.3 km south-west (S1),and 3.8 km west ofthe vent (S2) the 3-TF units 5 cm thick, thinning t03.cmat6 2-68 km south-west (T! and $4), and 7 km north-westofthe vent ($3), Unit 3-TF is ~3 cm thick and found directly on top of the ‘overbank PDC sequences at F2,F3 and F5 (Fig. 4; 6d). Modal pumice la- pill are 8-16 mm in size in proximal sites, wth the mode decreasing t0 2-4 mm aver a distance of 43 km from the vent. ‘allocations where sampled (S154, F2,F3, F5 and T1) unit 3-TFis the uppermost 2014 pyroclastic deposit, and is comprised predomi- nantly of reversely graded, moderately-sorted, grey or pale greyish brown, vesicular (and tube-vesicular), angular to sub-rounded, pumice lapilli (4-16 mm in diameter). The clasts are highly vtric and crystal~ rich, with 1-5 mm euhedral-subhedral pyroxene and plagioclase phe- ngcrysts (Table 2). These angular, grey-light brown, viric pumice clasts represent the only definitive juvenile clast identified, similar to the blocky pumice clasts in 1-PDC and 2-PDC, ‘The tephra fall deposit at site (Figs. 6e and 9) comprises 3 layers sub layers 3-TFa and 3-TPb, and the main fall deposit 3-TF (Fg. 4). The lowermost 3-TFa (sample K110), and 3-TFb (sample K111) sub-units are only observed at this location, appraximately 6.2 km south-west of ‘the vent, Deposit 3-TFa isa pate grey, 0.5 em-thick, well-sorted medium ash (250-500 jun) deposit chat is weakly-stratfied and reversely graded. Angular crystal aggregates dominates (of mostly clinapyraxene and orthopyroxene and lesser plagioclase feldspar) the ash-sized parti- cles, with some pumice and minor lithics (Fig. 2a). This deposit grades up to deposit 3-TFb above (Fig. 9b). 3-TFb is a reversely-graded, 1-2 m-thick, coarse ash ta fine lapill (0.5->2 mm) deposit rich in dense and poorly-vesicular lithic clasts of mostly light and dark dense lithics (Fig. 8), with lesser pale greyish-brown, juvenile pumice lapilli. This srades upwards into the uppermost fall deposit, 3-TF (sample K112). This deposit is similar to the tephra fall deposits observed at all other lo- cations examined (and mentioned above at S1-S4) (Figs 1,4 and 6e). It consists of <90% fine to medium lapili-sized (4-8 mm) grey and oxidised pumice, with minor lithic lpili. The lithic lapili clasts include grey to black fresh lava (93%), with the remainder ofthe lithic clasts including ‘weathered and oxidised brown, accidental country rock fragments (7%) (Fig. 8). Similar to the PDC sequences this sequence of fall units from 3- Tra to 3-TF atT1 becomes more pumice-tich upward (from 66 to 87% forthe same grain siz clas), witha corresponding drop in the content ‘of dense ities. Stratigraphically, the tephra fall deposits from tis stage also show some of the highest pumice contents compared to deposits from the fist and second stages ofthe eruption (Fig. 5) 5. Discussion 5.1. Correlating the stratigraphy to the eruption chronology and dynamics “The 2014 deposit characteristics detailed above are compared with eyewitness accounts and satellite report ofthe eruption, in orderto es timate the timing of eruptive events and place constraints on the mech- anisms, processes and dynamics leading to the origin, formation and ‘emplacement of the PDC and tephra fall deposits 5.1.1, Stage 1: eruption initiation and dome destruction ‘The earliest reported start time ofthe onset phase ofthe 13 February 2014 eruption is at 22:32 local time (Kristiansen et al, 2015), lasting ap- proximately 15 to 30 min (Fig. 10). This period included five shor, im= plsive infrasound signals atthe International Monitoring System (IMS) at the International Data Centre (1DC) (Hidayat, 2014; Caudron etal, 2015). Nakashima etal. (2016) detected acoustic waves from the erup- tion in the ionosphere, observed as changes in the Total Electron Con= tent (TEC) using data from the Global Navigation Satellite System (GNSS) receivers. The first wave was recorded at ~22:46 local time and was associated with an explosion (Nakashima et al, 2076). This seismic signal was taken as the eruption start time by Maeno et al{LR God {Jura of Volcanology and Geothermal Reszrch 32 (2018) 6-23 ” ‘BPumice (ight grey) Pumice oxidised) ‘Grystals (aggregates) ‘Buthics (dense ight grey) 1B Pumice darker grey-brown) scorn ‘Withies (dense dark grey) Lithes(orown/accdental), ‘Uthics (dense oxidised) 'BNo componentry dat crystals (singe free) 4 a z i : | Bi » | aer=pRn10 2 5-0 5.0 gg gs ee fo ee ae ein to "ig. 9. Gran size dso and componcny af the tephra il sequence a1 (se Fi 4), showing 2) 3-Tay sample [110 b) S-1F, sample KITT c) 3-TF, sample KII2 The ‘componeamy forthe 5 t9 1mm gain sz lass dsplayed in ple chats fo each unit ‘See key for ily ad See Table fora description ofeach component dscns, (2017), however earlier infiasound signals suggest this eruption may hhave started -14 min earlier (~22:32). Further seismic signals were re- ‘corded at 22:51 local time (Ms of -23; Caudron et al, 2015) (Fig. 10). ‘The pyroclastic deposits associated with the onset phase includes se- ‘quences of PDC deposits (1-PS1, 1-PSL and 1-PDC) and the averbanking ‘sub-facies ofthese units (Fig. 4). The first deposited 1-PS1 and 1-PSL ‘surge facies have characteristics of deposits of low-partice density sim- ilar to ash-cloud surge units, carrying dominantly fine ash and crystal fragments (up to 47%) in the upper depositing part of the cloud. This is similar tothe laminated, fine-grained and poorly.sorted surge de- Dosis observed by Lube et al. (2014) at Mt Tongariro, and Cro etal: (2013) at Merapi. Te poor sorting and gente low-angle lamina tions ofthese deposits ae consistent with low-level turbulence in the parental surge (eg. Branney and Kokelaar, 2002; Sulpiio etal, 2016) Deposits 1-PS1 and 1-PSL stop at 2.1 km from the vent, although thin deposits may have been eroded farther away. Similar pre-plinian ‘urge deposits, that are fines-rch, 10-20 cm thick in valleys, and 1-5.mthickon adjacent ridges, were observed from the 1990 eruption, but traveled upto 5 km from the vent (Bourdier et a, 1997). Maeno «etal. (2017) present four major depositional units from the 2014 erup- tion attributed to three eruptive stages. They identity dilute high-energy PDCS associated with the pre-plinian, partial destruction of the existing lava dome (Unit A) in the fist stage ofthe eruption, beginning around 22:45 local time, similar to this study. These deposits and subunits ‘were found tothe north-east the vent in proximal ridge-top sections, and are described as originating from a small lateral blast, similar to that ‘of the 2010 Merapi dome-destroying eruption. It could be possible that units 1-51 and 1-PSL are related to an upper and more distal facies of this lateral blast, due to the similar occurrence of juvenile and non- juvenile clasts However, since the deposits described in this study ‘were found south-west of the vent, it may be likely that Maeno et al. (2017) are recording an earlier blast based on near-vent stratigraphy ‘that was only exposed on rdge-tops. ‘The 1-PDC valley infil (at F1) and valley margin (at F4) faces show typical characteristics for deposits from highly-particle concentrated PDC or pyrocastc lows, with a fine-grained basal portion grading up into a more poorly sorted massive deposit and overall carse-tailre- verse gracing There i a higher proportion of dense, thc clasts and ve- sicular pumice (including blocks) inthe upper 1-PDC faces than inthe lower 1-PDC facie (Fig. 5). The lower, basal facies of 1-PDC is likely «emplaced by traction fom the base of a curent, whilst the upper, mas sive partsikely emplaced en masse asthe overriding low came to rest similar to the mechanism explained by Sparks etal. (1873) responsible for the inverse grading of larger dat in ignimbrites. The presence of ‘oxidised pumice casts in upper most layer ofthe PDC deposit could sug- gest temperature variation within the flow (Fink, 1987). The onset phase unit (apart from the lower facies from 1-PDC and 1-PS1) include high contents of dens light grey and dark grey lithic fragments lopli to block sized), similar to the 2007-2008 lava dome, which was likely destroyed during this phase (Table 2: Figs. 5, 7, and 8). Calder et al (2000) interpret similar lithicrich deposits from Lascar, Chile, as being produced by a highly concentrated granular fow. Other dense lava and brown country rock ities are commonly altered and weath- cred (Fig 8), and were probably derived from fragmentation of older vent wall rock. Vent clearing and blockage along with low mass- «jection rates, meant that the onset phase PDCs were of low energy and runout (=2.6 km from the vent) (Fig. 10; eg. Carey etal, 2007; ‘Wallace et al, 2013). The higher abundance of dense lava and also brown country rock lithcs in the Stage 1 deposits may suggest that ini- tial partial vent and conduit widening occurred during this time, which initially lowered explosive fragmentation, and maybe responsible for the greater pumice clatsize observed when compared to later deposits (Smith and Kokelaar, 2013) “The overbank PDC deposits are similar in componentry, but finer- ‘rained than the initial 1-PSL. and 1-PS2 surge units, and derive from the sill out and detachment of higher (up t0 20 and 85 m in elevation from the valley floor, more turbulent and pumice-rich, and les con- ‘entrated parts ofthe density currents to the adjacent valley tops and banks, similar to the Merapi overbank deposit as described by Lube etal. (2011) and Cronin etal. (2013). [eis kely that due to transport time such overbanking units may have been deposited ata later stage. however, the parent current and plume was generated during this stage. The absence ofco-PDC fine ash at the top of the FDC succession Iikely reflects there being only a brief interval (~<30 min) before further DCS were formed in Stage 2{LR Goode ett Jour of Verlag nd Geothermal Research 382 (2018) 6-23, Eruption Stage: stage : ‘Stage 2 a SS SAE | ian Sb mtee ' sow \ \ \ \ \ ' sim] 5 1 \ ' is | ' ow? : \ owe O56 \ pais LTR a agem amt sem ates \ Srcme i erect copa BETES Sas Honea te tee \ BERT eon Semen | \ cen ae 0 ! son 2x vee Ge oF by 1 . om = i 4 castonestioome | I 1 \Nakashirs etal TEC oxcitaton: | HA aA_h_—S st crnmeemee Ly Fig 10 Schomatc lagram of dyna ecuring dung Sages 1 woof the exption with start anes an eruption dation in UTC and local ine fo each phase as documented in sepliyseal and ser atinal eprts (CYP, 2014 Caudron eta, 2015; Kestisen ea, 2015; Natasha eal, 2016). Filed polyhedron indicate dese hc content (percentage, ‘minus accdentalivown lithic), white speckled plyhedronsindiate pumice content (percentage) after normalisation (et Table 2) This component datas taken frm the 2am gain size action of epesentative samples fr each tage Stage 1 taken om ample KI20; Sage 2 aten fom sample KIZ3; Stage 3 sen rom ample KU. In Stages | and 2 moda ran sizes forthe matrix ash api and blocks for FOC unt re shown. tage 3. moda lp grain ize shown forte ai fal unison Red arrows indicate ‘sprsl of pyochstic mate Back rows indate condi and vent widening during the eruption Manu height ofthe eruption clu (25m) taken om GVP (204) and Istansen ea (2015 01:10 00:13 focal tne. ee the Discusion section fr a eta desernon ofthe eruption chronology. References Isted ae (CUP, 2014 Caron a Jeffery etal. (2013) suggested that pre-eruptive processes in the 2007-2008 lava dome effusion involved fractional crystallisation, along with magma mixing and mafic magma recharge at shallow- levels. This model for magmatic processes was based on binary bulk mixing models, and mineral textures and zoning patterns in the 2007-2008 dome lavas. Therefore, Caudron etal, (2015) hypothesised that the first 2014 eruption episode was similar to that in 2007-2008, and could have been induced by mixing of replenishing magma into fan existing shallow reservoir, causing gas-overpressure, and destabilising the conduit capping dome (eg. Girona etal, 2015). Simi Jar eruptions have occurred in 1919 and 1990 at Kelut, after a small ex- plosion displaced the crater-lake, and unroofed a capped shallow ‘magma in the conduit (Bourdier etal, 1997; Caudron et al, 2015). A weak seismic signal (Ms = -23) likely indicated the vent-opening ex- plosion (Caudron et al, 2015). There was no evidence for large-scale ‘magma-water interaction inthe onset-phase deposits, despite the com- ‘mon white clouds observed, Accretionary apil, blocky shards or wet fine ash emplacement structures were not observed, unlike that of the phreatomagmatic phase of the 1990 eruption (Bourdier etal, 199: Porreca et al, 2006). 5.12. Stage 2: plinian column formation and marginal column collapses ‘The earliest reported start time for Stage 2 is at 22:58 local time (Caudron et al, 2015), with the latest start time reported at 23:30 local time (GVP, 2014) (Fig. 10). Variations in reported start and end times for this eruption, and Stage 1, are shown in Fig, 10. The IDC de- tected 14 global infrasound signals associated with a separate, larger event, ~15 min after the onset phase (Hidayat, 2014; Caudron et 2015). Furthermore, a sustained seismic signal was recorded from 22:58, overlapping with the arrival times of the infrasound signals (Caudron et al, 2015). The CNSS also detected a large acoustic wave in the ionosphere at -23:01 (Nakashima et al, 2016) and a further brief seismic event was recorded at 23:11 (Ms of ~4.7; Caudron et a., 2015). Satellite imagery showed the plinian ash cloud clearly from (00:30 (GVP, 2014). The Stage 2 pyroclastic density current generation and subsequent deposition likely began around 22:58 local time and continued until around 01:00 local time, when ionosphere data sug- gested that the plume reached its maximum height at 26 km (Figs. 5 and 10). ‘The ist Stage 2 PDC deposits are deposited directly on top of tage 1 deposits (where present; Fig. 4). The upper part of the 2-PDC deposit shows a more massive structure, with a poorly sorted concentration of ‘more block-sized clasts than in the cross-stratified, medium-ash facies that define the basal part of the 2-PDC deposit. Ths is similar to the onset phase PDC deposit at Fl with coarse-tail grading. Similar to 1- PDC, the lower, basal faces is likely emplaced by traction from the base ofa current, whilst the upper, massive part is likely emplaced en ‘masse as the overriding flow came to rest (e.g. Sparks et al, 1973). Such PDCs and similar pumice-rich deposits were observed at Lascar, Chile (Calder et al, 2000). large block-tich low front and similar mar= ginal levees occur in 2-PDC at 3 km from the vent, and at 4.7 km from the vent (deposit 2-oPF3). These reflect deflection and concentration of segregated large, often low density clasts towards the flow margins and levees due (0 the internal velocity gradients in dense-granular flows (eg. Calder etal, 2000; Lube etal, 2007; Sulpizio et al, 2007; Cronin etal, 2013). AtFI the Stage 2 deposits (2-PDC) are thicker, but less coarse than the Stage 1 deposits (1-PDC). The differences in the sizes of blocky clasts in the deposits between these Stages could possi bly be linked different fragmentation conditions in both eruptions (McBimney and Murase, 1970). Inthe Stage 2 PDC deposits, pumice contents in ash and laplli-sized clasts are -25% higher than in most earlier units (Fig 5). The continued presence of rare, fine, dark and pale grey dense lava lithic clasts, along, ‘with fragments of country rock and oxidised dense material within these deposits suggests on-going conduit erosion and vent widening (Figs 8 and 10), such as that inferred at Tarawera, (Carey etal, 2007), And at Monte dei Port (Sulpizio et al, 2016). This is confirmed by satel- lite imagery of the vent region after the eruption (Fig. 2b: GVP, 2014),[LR God ura of Volcanology and GethealRescrch 32 (2018) 6-23 6 ‘which shows a deep vent crater, similar to that fllowing other intense ‘eruptions from Kelut (eg. 1951, Hadikusumo, 1974; 1990, Bourdier ‘ec al, 1997), Such instabilities in the vent can cause changes in mass «ejection rates. Due to the increased vent radius the incorporation of 2d- ditional blocky material into the lower portions of the building plume ‘may have generated an increase inthe lateral extent of the jet region at the star of ths second eruption. This can limit the ability of pyroclas- ‘tic material mixing with the atmosphere. tis therefore possible that the incorporation of conduit wall rock, an increase in the proportion of blocky pumice clasts, and vent widening may have contributed to col- lapse or overloading of this eruptive column at the beginning of the plinian eruption (eg. Cioni tal, 2004) (Fig. 10). Similar valley infillin, ‘coarse-grained pumice rich flows reaching 3 km from the crater were «also formed during the beginning ofthe plinian phase of the 1990 erup- tion at Kelut (Bourdier etal, 1997). Photographic accounts of the eruption from 00:38 ta 01:00 local time on 14 February, taken by an eyewitness in Blitar (Putra, 2014), show a tall eruption column, that progressively becomes paler in colour as the eruption progresses. Voluminous and extensive PDCs are visible {rom 00:45 onwards to the west and south-west of Blitar. The longer runout of Stage 2 PDCs likely reflects a more-sustained feeding and boi ing over of material, Higher pumice content might also have played a role, as found by Calder et al. (1999) for a sequence of pumice- or lithic-rich PDCs at Montserrat, where pumice-tich PDCs were more mo- bile ue to the segregation of denser lithic clasts proximal ta the vent. ‘This may account for observing more pumice-rich, less blocky PDCs at ‘greater runout distances (Figs. 4 and 5). These deposits are similar to the pumiceous lobate deposits (Unit C) described by Maeno et al (2017), containing abundant pumice and lava blocks up to 20 cm in ameter. These deposits are interpreted similarly to our 2-PDC deposits, generated by the collapse of the plinian eruption column. However, the deposits described by Maeno et al. (2017) are attributed to the third stage of the eruption, not the second as in this study. The laterally- continuous, fining and better preserved bedding, pumice- rich overbank PDC lobes found 4.7 km from the vent in the K. Badak (Fig. 4), originated from the channelized block-rich PDC deposits of ‘which were found ~2 km further up inthe valley (at F1). The thin, cap- ping 2-oPDCa co-PDC fall unit is rich in crystals (>60%), representing fines that were clutriated from the parent pyroclastic flows (eg. Branney and Kokelaar, 2002; Edmonds et al, 2006; Cronin et al. 2013). Maeno et al. (2017) also describe grey, thinner, poorly-sorted asherich layers on ridge tops that are similar to these thin units. They are interpreted similarly as accompanying dilute portions ofthe current thathave travelled over ridges, or also as co-PDCash (Unit D). Ths loal- ised fll layer (2-oPDCa) records the cessation ofa semi-continuous se- ries of PDCs. It is possible that further PDCs were generated during this time, but travelled farther from the vent (Fig. 2b) and therefore, were ‘deposited later on during the course of the eruption. Such was the ‘ase inthe 1990 eruption, where several series of ash and pumice lapil falls were intercalated within pumice surge and flow deposits, as well as scoria-tich low deposits, with initial evidence of phreatomagmatism (Bourdier e¢ al, 1997). Intercalations of fall and PDC deposits are not a ‘common feature ofthe 2014 deposits although co-PDC falls indicate a brief pause between each ofthe 2014 eruption stages. ‘513, tage 3: climactic phase of he plnian event and tephra fallout ‘The final stage of the 13-14 February 2014 eruption involved the production ofthe tallest plnian eruption and umbrella cloud. This is as- ‘sociated with the injection of tephra into the stratosphere, imaged from (00:30 (GVP, 2014). The main part of the pumice rich lait fall (3-TF unit; Figs. 4 and Ge) was deposited once the plume reached its ma» ‘mumaltitude (26 km above sea level) at -01:12 local time on 14 Febru- _aty (seen by CALIPSO satellite with the CALIPOP lida instrument; GVP, 2014) (Fig. 10) Tephra fall started during the earlier Stage 2, butts not found as an obvious layer within PDC and co-PDC ash sequences de- scribed above. Nakashima et al. (2016) estimate the duration of this ‘climactic phase at 1 h. Dare et al. (2016) estimated that the tephra plume was depositing between 23:00 to 03:00 local time, with Kristiansen etal. (2015) reporting a strong MTSAT ash mass loading. ‘west ofthe coast of Java at02:32.A plane travelling from Perth to Jakarta ‘encountered the dispersed plume off the coast of Java, west of Kelut, at around 04:50 (GVP, 2014; Kristiansen etal, 2015). Further ash detected ‘was reported distally west of Kelut at 09:32 on 14 February 2014 (GVP, 2014). This inal stage of the 13-14 February 2014 eruption therefore likely ended around 02:00 to 03:00 local time, with small phreatic ex- plosive events occurring until 20 February 2014, ‘The 2014 tephra fall deposit, 3-TF.is described in proximal locations from 1.9 to 68 km south-west of the vent, and -7 km to the north-west (Figs. 1 and 4). In these locations, the fall deposits are 5-8 cm thick, fin ing with distance from source from coarse to fine lapii, The broad lat ‘eral uniformity in thickness and clast size of the proximal fall deposit ‘suggests that it sedimented from the umbrella cloud region ofa strong. rapidly expanding plume, driven by a high and sustained mass flux (Walker, 1981; Cioni et al, 2000; Bonadonna and Philips, 2003). A strong south-westerly wind deposited several millimetres of very fine pale brownish-grey ash over the city of Yogakarta, -215 km west of the volcano (GVP, 2014; Kristiansen et al, 2015). Fine ash is missing from the proximal climactic fall deposits, but sedimentation of fine ash 6.2 km south-west ofthe vent (3-TFa) may have resulted from its aggregation due to atmospheric moisture (Bonadonna and Philips, 2003; Brown et al, 2012) during the earlier stages of the plinian phase, similar to that seen at Redoubt volcano, Alaska (Wallace etal, 2013), although accretionary lapilli were not observed. Maeno et al (2017) are also in agreement that most of the tephra fall deposits ‘were dispersed westward of the volcano. They describe a similar umice-rich deposit (Unit B) to 3-TF, with similar grading, a small amount of fresh and altered lithic material, and a clat size that de- ‘creases with distance from the vent. This unit is found overlying the ‘ground surface to the west and south-west, similar to this study. It ‘was also found directly overlying Unit A (pre-plinian PDCs) only in areas north-east of the volcano, not observed in this study. Maeno et al. (2017) attribute these deposits to the second eruptive stage. Stage 2 (reported around 12:31-01:01 local time) and 3 (reported ‘around 01;00-01:30 local time) are reversed in their interpretation of the eruptive events compared to those presented in this study, based ‘on their observation that no pumice-rich PDC deposits were found un- ‘demeath the plinian fll deposit. ‘The main 3-TF climactic deposits have a modal grain size of 4-8 mm (ine lapil), and reversely grade from coarse ash (500 um) to coarse la- pilli (<64 mm). They are dominated by pumice (85-91%), with rare ‘dense lava and country rock lithic lpi (Figs 8 and 9) increasing pum- ice content is common within a deposit from a sustained eruptive blast ‘with great intensity (Walker, 1981), and similar pumice rich fall de~ posits are observed at Nevado de Toluca (Arce et al, 2003), The abun dance of oxidised pumices increases from 3-TFb to 3-TF (Fig. 9), which ‘could reflect temperature changes or increases during the later part of the eruption (Taitet al, 1998).The reverse grading reflects the progres- sive increase in the intensity and mass flux of the plinian eruption (eg, Walker, 1971, 1981; Carey and Sigurdsson, 1989). similar plinian ppumice-rich tephra fall unit was observed from the 1990 eruption, but was strongly normally graded (Bourdier et al, 1997). Bourdier et a (1997) also document 4 separate units of ash fall, probably as a result (of 7 discrete, short-lived explosions when compared to the two mait ‘explosions in the 2074 eruption. The dominance of fine to medium la- pill a varying distances from the vent suggest that the eruption was also sustained fora few houts, allowing large clasts to reach the upper. turbulent portions of the plume (Fig. 10) (eg, Mount St. Helens, 1980; Bursik etal, 1992; Paulick and Franz, 1997: Cioni etal, 2000; Bonadonna and Philips, 2003). Caudron etal. (2015) estimated the in- tensity ofthe eruption, at -10.8 011.0, with magnitude (M) of 43 to 45 (following Pyle, 2000). The high pumice content of the fall deposits ‘shows that during this final stage the conduit was open and stable20 {LR Goode ett Jour of Verlag nd Geothermal Research 382 (2018) 6-23, (Fig. 10; eg. Arce et al, 2003), Elongated vesicles, up to 20 mm-long, show rapid bubble growth, coalescence and shear occurred as magma travelled up through and exited the conduit during continuing magma fragmentation (eg. Klug and Cashman, 1996; Paulick and Franz, 1997; Klug et al, 2002). The greater abundance of juvenile pumice pyroclasts (Fig 5), and decrease in clast size from large blocks to fine a- pill (Fig. 10) over the course of the eruption (from Stage 1 to Stage 3), suggest it is possible that increased fresh magma flux and gas expansion drove further fragmentation, and may have contributed to this later ‘more energetic eruption (Klug etal, 2002; Melnik etal, 2005; Voight et al, 2006; Wallace et al, 2013; Caudron et al, 2015: Girona et al, 2015), ‘The few lithis present in the climactic plan tephra fll unit show a slightly higher percentage of brown country rock or accidental lithics reflecting the earlier loss of the 2007-2008 dome and conduit material during widening and deepening of the vent (Figs. 8 and 10). The freshest lithics of this phase are darker grey and dense, but slightly ‘more porous than those of earlier units, which may represent the exca- vation of shallow degassed magma from the reservoir (e,,Cioni etal, 2000). As the eruption mass flux decreased towards the end of theerup- tion, it is possible that marginal instabilities may have generated subse- quent column collapse pumice-rich PDCs (not observed inthis study). 6. Eruption patterns and hazard implications 6.1. Pre-1900 Little is known about the pre-1901 eruptive activity of Kelut. Historic accounts ofthe eruptions lack significant detail on the style ofthe erup- tions. However, an estimation of VEIs has been made by the Global Vol- canism Program, and based on reports in Indonesian and Dutch, Since 1000 CE, Kelut’s explosive eruptions have often reached VEIs of 3 to 4 (Table 1; Zen and Hadikusumo, 1965; Pratomo, 1992; Bourdier eta, 1997; De Bélizal et al, 2012; GVP, 2014). In this time frame,25 eruptions were recorded (every ~24 years), and at least one also produced adome (1376 CE). The shortest interval between eruptions was 1 year (1825 10 1826, 1919 t0 1920 and 1966 to 1967). Explosive volcanism (VEI3) con- tinued through to 1586, when the largest recorded eruption for Kelut occurred, following a 38-year period of quiescence (VEI 5; Pratomo, 1992: GVP, 2014). The largest repose, 75 years, occurred after this cata- clysmic event, and was followed by a VEI4 event in 1641. After this was 1185-year long period consisting of seven VEI2 eruptions, which could ‘mark longer periods of lava effusion. In 1826, a VEL 4 eruption occurred, followed by 4 eruptions of VEI 3-2 for the next 75 years. The most sig- nificant volcanic hazard duting this period was syn- and post-eruptive lahars (Bourdier et al, 1997; De Bélizal etal, 2012; Jeffery etal, 2013). 62, Recent activity - 1901 onwards More detalis known about the more recent activity of Kelut, with an average 15-year interval between events (and a range of 1-23 years; Table 1) Zen and Hadikusumo, 1965; Pratomo, 1992; Bourdier et al, 1997; De Bélizal etal, 2012; GVP, 2014). There appears to be a repeating pattern of lava dome effusion and short-lived explosive eruptions, For ‘example, in 1966, a VEL 4 eruption occurred, followed by dome effusion in 1967, This dome remained for 23 years, before being completely destroyed in the 1990 plinian eruption (Bourdier et al, 1997; GVP, 2014). A post-1990 dome grew and the latest dome extrusion took place in 2007-2008 (JfTory et al, 2013). 63 Hazard monitoring and implications Monitoring the temperature and chemistry of the crater-lake has been a useful forecasting tool at Kelut (Suryo and Clarke, 1985; Bernard and Mazot, 2004). Degassing from hot springs and subaqueous fumaroles has caused arise in temperature from an ambient 19 °Cto as ‘much as 31-50 °C (Bernard and Mazot, 2004), In 1990, the temperature rose to 41 °C, 3 months before the eruption (GVP, 1990; \Vandemeulebrouck etal, 2000). Between 1993 and 2003 the tempera- ture ofthe crate-lake cycled and occasionally spiked to 50°C, with gas ‘bubbling occurring in 1996 and 2003 (Bernard and Mazet, 2004). Prior to the 2007 dome effusion, the lake temperature rose over four months to 50°C the day before the eruption (GVP, 2008). Also, a threefold in- crease in the carton dioxide flux into che lake was measured by an echo sounding survey ~4 months prior ta the 2007 eruption (Caudron, et al, 2012), Increased COs flux indicates activation of the deep mag- ‘matic system beneath Kelut, similar to that observed at other volcanic crater lakes (eg, El Chichén, Mexico; Jacome Paz et al. 2015). CCaudron et al. (2012) noted discontinuous vertical trains of bubbles or “puffing” from July to August 2007 in normally quiet regions of the crater-lake, whist the main CO, upwelling and temperature anomalies, ‘were located directly above the narrow magmatic conduit, suggesting the renewal of magmatic activity at depth (Caudron et al, 2012). CCaudron et al. (2012) estimated that -30% of the CO: entering the Kelut crater-lake is gaseous, whilst the bulk is dissolved within water. Passive, continuous degassing likely led to the effusion-dominated 2007-2008 eruption episode, Prior tothe 2074 eruption, the number of deep volcanic earthquakes increased and the crater-lake temperature rose by 5.5 °C from Septem- ber 2013 to February 2014 (GVP, 2014). The lake contained 2.1 million m? of water pre-2007 (Bernard and Mazot, 2004). Jeffery et al, (2013) estimated the 2007 dome almost completely displaced the lake, leaving a small crescentic lake forming from rainfall runoff collecting araund the dame margins. The detection af such heightened activity would be difficult if there is not a crater lake present aver the area of emissions to measure the changes in water temperature. 64, Hazard analysis framework ‘The primary hazard to life for proximal and distal populations around Kelut is from large and extensive lahars (Hadikusumo, 1974; Suryo and Clarke, 1985; Bourdier et al, 1997a; De Bélizal et al, 2012). In 1586 CE and 1919 extensive syn- and post-eruptive lahars caused up to ~10,000 casualties. The largest syn-eruptive lahars occurred when the crater lake was a its greatest volume (Hadikusumo, 1974). This led to the construction ofa system of drainage tunnels that started in 1919 with further efforts in 1923 and 1954, to control the lake level and reduce the impact of lahars (Zen and Hadikusumo, 1965; Bernard and Mazot, 2004). ‘Whilst no syn-eruptive, hot primary crater-ake derived lahars were formed during the 13-14 February 2014 eruption, five days after it sec- ondary rain-derived lahars were generated when rain remobilsed py= roclastic material in several river channels (GVP, 2014; Dibyosaputro| et al, 2015). The resulting lahars filled dams and irrigation systems ‘with debris, damaged bridges and roads, and inundated much farmland (Dibyosaputro etal, 2015). Secondary, rain-derived lahars are a persis- tent hazard at Kelut, whilst primary syn-eruptive lahars depend on the status ofthe lake atthe time ofthe eruption, Several generations of haz~ ard maps have been produced for this area (eg, Suryo and Clarke, 1985; ‘Mulyana et al, 2004; De Bélizal et a, 2012), identifying major tisk catchments to the north-west, west, and south-west of the voleano, The official hazard maps by the Centre for Volcanology and Geologic Hazard Mitigation (Mulyana et al., 2004; Primulyana et al, 2014; Andreastuti et al, 2019) are comprehensive, and include hazards in bboth proximal and distal areas. The Primulyana etal. (2014) map was also updated on the basis ofthe effects of the 2014 eruption, ‘The generally cyclic nature of eruption styles between explosive to effusive variants since the 1900's, suggests that the next eruption of Kelut could be a long period of dome growth or repose, lasting any- where from 7 t0 31 years Ths will depend on the degassing conditions of the magma in the upper conduit (Gardner etal, 1996; Klug etal, 2002), and on the occurrence of an interval of magmatic recharge,{LR God ura of Volcanology and Geothermal Rescrch 32 (2018) 6-23 2 ‘Alternations between dome-building and explosive eruptions occur at many other similar volcanoes (eg. Merapi, Indonesia and Colima, Mexico) although the pattern at Kelut appears to be remarkably regular in recent times. There have, however, also been semi-continuous pe- rods of smaller eruptions with sporadically timed larger events (VEI 3 to5). Generally the very largest eruptions (eg, 1586) are followed by the longest repose periods (in this case 75 years). It should also be recognised that the Kelut edifice includes many ‘older domes, centred away from the currently active crater. Although these are not well dated shifts in the eruption site and lateral dome ex- trusions must also be included in hazard analy 7.Conelusions ‘The 2014 Kelut eruption was suelden in onset and relatively brief in ‘comparison to other plinian eruptions, Despite this, t was highly explo- ‘sive and in ts climactic phase produced a stable eruption column with a ‘well-formed and dispersed umbrella region, Three stages ofthe 2014 ‘eruption have been identified. ‘The onset eruption of Stage 1 lasted for only -15-30 min and pre ‘duced PDCs that were rich in dense lava lithics reflecting the progres- sive destruction of the 2007-2008 lava dome, upper conduit and vent walls (Figs. 5 and 10). Gas overpressure drove this eruption, induced ‘bya possible replenishment of magma into a shallow magma reservoir, ‘destabilising the conduit-capping 2007-2008 dome. We hypothesize that this caused a top-down decompression wave, leading to rapid ex: pansion and rise of gas-rich magma (Klug et al, 2002; Caudron et al, 2015), and has also been proposed at ather Javanese volcanoes such {a5 Merapi (Voight et al, 2006; Girona etal, 2015). There was no evi- ‘dence for phreatomagmatism during this onset phase, likely because the crater lake and underlying aquifers were a small fraction oftheir pre-2007 sizes. Increases in the magma flux and transition ta a more ‘constant eruption led to a growing eruption column during the 2- hour-long Stage 2 plinian eruption. Our data indicate that continued ‘magma fragmentation, along with further vent wall and conduit widen- ing led to overloading of unstable, basal column margins. This prevented buoyant rise and atmospheric mixing of blocky pyroclasts, «and thus generated longer runout PCS, with higher pumice contents ‘and decreasing dome and conduit fragments (Fig. 10). The final climac- tic plinian phase produced a plinian column up to 26 km in altitude, ‘with 2 well-developed umbrella region. The deposits from this Stage contain reversely graded, medium-ash to coarse-Iapilli juvenile pumice rich (290%) tephira falls. Widespread proximal and distal fall likely ‘emplaced during the one-hour (to 02:00 local time) climax of the plinian eruption, in the third and final stage (Stage 3) although its pos- sible that fine ash may have fallen proximally from 23:00 local time. ‘These deposits reflect an eruption progressively becoming more ener- ‘etic with time, with increased mass flux and gas expansion of fresh ‘magma, and continued fragmentation sustaining the plinian eruption Although the 2014 eruption was similar in duration, volume and general style to the 1990 event, the 2014 eruption shows a simpler ‘overall stratigraphic succession. Many of the 2014 PDC deposits were ‘emplaced en masse, Therefore, intercalations of contemporaneous fall and PDC deposits common in 1990 deposits were not a feature of the 2014 deposit, although co-PDC fall deposits indicate a brief pause be- tween each of the 2014 eruption stages. The nature of post-1900's Kelut eruptions shows a general cycle be- tween explosive and effusive variants, probably depending on the re rnewal of magma and magma degassing conditions in the shallow ‘magma reservoir and upper conduit. This suggests thatthe next erup- tion, ifit occurs within the next 20 years, may be a long (up to decadal) period of dome growth, In order to understand the dynamics and processes responsible for the cyclic nature of eruptions, and accurately pinpoint triggers for ‘more intense episodes at Kelut, petrological, mineralogical and geo- ‘chemical analyses of pyroclastic deposits from the 1990 and 2014 ‘explosive eruptions are needed, which can then be compared to the 2007-2008 dome. A better insight into the past eruptive activity of Kelut (pre-1990) is also needled to understand the relationship between ‘magma intrusion cycles and eruption patterns. Supplementary data to this article canbe found online at https//dol ‘org/10.10164}jvolgeores.2018.02.005. Acknowledgements Reza Firmansyah and Dermadi Madi from Bandung institute ofTech- nology are acknowledged for their assistance during fieldwork, Lucy McGee is thanked for many supportive conversations. We would also Tike to thank the two anonymous reviewers for their thought provoking ‘comments that improved the manuscript. The Guest Editor Wendy McClausland is also thanked for handling the manuscript. . Goode is funded by an International Macquarie University Research Excellence Scholarship (iMORES) (2012196) associated with an Australian Re- search Council Future Fellowship awarded to H. Handley (F7120100440).5. Cronin is funded by the New Zealand Natural Hazard Research Platform *Quantying exposure to volcanic hazards project. References ‘Andreastut SD, Alloway, BY, Smith LEM, 2000. detailed rephostatgraphic fMamework a Merapl Volcan, Central va Indanest: implications fr erup- tion precicitions and hazard assessment. Volcanl. Geotherm. Res, 100, 51-07 Andiesstt, ,Faipurno, , Gunawan, H, Buin, A, Syn D, Paste, |, 2018. ‘Character afcommmunty response o vlan css Snabung and Keludvlcanoes. [Volenot Getherm Rex 382, 58-310 ‘Admbula Mendoza, R. Lesage, R Valdés-Gnasle, C Varley, NR. Reyes-Dil G. Navara C, 2011. Seismic activi that accompanied te efsive and explsive etop- ‘ons during the 2008-2005 period ar Vole de Calin, Mele, J Volcano, estar es 205, 30-46 ‘Arce JL, Macias JL, Varqez-Selem:L, 2003-The 105 km plinian eruption of Nevado de “Tobia voleano, Meso” suatigapy and hazard impbcatns. Geel Soc A Bull 15, 230-248, Badan Inormasi Geospasial (8), 2013 Bakosutanal (in Indonesian) Topographical Map of Ket Vols Bema Ay Mazo A 2004 Geochemical eolton of the young cater ke of Kel val “ano Indonesia In: Want, RB. Sea l,RB (Eds), Water-roc Ineratin, (WR 11 London Taylor and Frans Gioup Pp. 87-2. Bonadonna C.Piins 2003 Seeimentaton from rong vlan ples. | Geoptys. Res 108-28 oui J-L, tomo |, Thouret,J-C, Bowdon, C, Vice, PM, 1987, Obseiations, statgzaphy and th erupave proceses ofthe 1600 eruption of Klutvelcane, Indonesia Volcan. Gather. Res. 78, 181-205 Sourder J-L. hour JC. Patoma,L, Vincent PM. Boudon, 19572 Volcanic haz- “rds Rei volcano (java sland, Indonesia lessons eared from the 1950 erp branny, Mj, Keel 2022 Pyrocanic dency ccrnt andthe edimentation ft mbit. Geo Sc. iond Mem 27 Brown, RJ. Bonadoana. Durant A 2012 A review of volcanics aggregation. Phys. Chem Fath 45.46 65-78 Buri Mi, Sparks. RS), ber}. Carey, SN, 1952 Sadimentatin of epraby vlea- Tic plumes L theory and is comprson with a tdy ofthe Fogo A plan depos Sm Miguel Azores) ul Noleana 54, 528-348 ‘alder £5. ole FD. Dade, WR, Dri, TH, Hoblt RP. Huppert HE. Richie, Sparks RSJ Yung. SR, 1998 Mobily of pyroclastic flows and sures athe Soultire His Volinn, Montserat Geophys Res. Ltt. 26, 37-540, ‘caléer ES. Sparks, RS], Gardeweg, MLC 200. Erosion, transport and segregation of Pumice and lic cts pyrodaste ws infeed om inimbuse tase Vole ‘ano Cle J. Voleanel Geatherm Res, 104, 201-23. ‘carey-SX Siurdsson 1868 The nensty opin erupons. Bul. Vokcanol 5128-40, (Grey Ry, Houghton BF Sable) Wilton, CJN- 2007, contrasting grain sae and “components in complex proximal deposits of the 1886 Tarawera besa Pian ‘rupon Bull Voleanol 9, 903-026, ‘Caren, C, Matt A, Berard A, 2012, arbor done dynamics in Kei volsnic ie, "}Geopys es. 117, 1-1 ‘Ceuon,€, Tas, B, Garces M, Alexis, LP, Mille. 2015. On the use of remote ntasind and selmi ations to constrain the eruptive sequence and tensity forte 2014 Keto eruption Ceopiys. Res Late 6914-8621 Gini Re Maran P Sartacroce.R Sra, A 200 linia ad sbplinianeropsons Tn: Sgurdsian H Houghton 8. Rymer H, MeNut 5 tx} (E05), Encydopeda a ‘olesnoes Academic Pres pp. 477-404 Gon 8 Cra Lanz 8 Zane E2008 Temperatures of AD. 78 pyrodastc density CUTER aepoits (Vesuus ay). Geophys Res. 10. ‘Comin SI Lube, Dayui DS, Suma. Subrandi. S. Suton, 2073 nits into the October-Navembet 2010 Cunung Merapl eran (Cental av, Indes)