Andean Geology ISSN: 0718-7092 [email protected] Servicio Nacional de Geología y Minería

Kobayashi, Chiaki; Orihashi, Yuji; Hiarata, Daiji; Naranjo, José A; Kobayashi, Makoto; Anma, Ryo Compositional variations revealed by ASTER image analysis of the Viedma Volcano, southern Volcanic Zone Andean Geology, vol. 37, núm. 2, julio, 2010, pp. 433-441 Servicio Nacional de Geología y Minería Santiago, Chile

Available in: http://www.redalyc.org/articulo.oa?id=173916371009

How to cite Complete issue Scientific Information System More information about this article Network of Scientific Journals from Latin America, the Caribbean, Spain and Portugal Journal's homepage in redalyc.org Non-profit academic project, developed under the open access initiative Andean Geology 37 (2): 433-441. July, 2010 Andean Geology formerly Revista Geológica de Chile www.scielo.cl/andgeol.htm

Compositional variations revealed by ASTER image analysis of the Viedma Volcano, southern Andes Volcanic Zone

Chiaki Kobayashi1, Yuji Orihashi2, Daiji Hiarata3, José A. Naranjo4, Makoto Kobayashi5, Ryo Anma6

1 Earth Remote Sensing Data Analysis Center, Chuo, Tokyo 104-0054, Japan. [email protected] 2 Earthquake Research Institute, the University of Tokyo, Bunkyo, Tokyo 113-0032, Japan. [email protected] 3 Kanagawa Prefectural Museum of Natural History, Odawara, Kanagawa 250-0031, Japan. [email protected] 4 Servicio Nacional de Geología y Minería, Av. Santa María 0104, Providencia, Santiago, Chile. [email protected] 5 Dia Consultants Company Limited, Saitama, Saitama 331-8638, Japan. [email protected] 6 Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan. [email protected]

Abstract. We conducted a lithological mapping of the Viedma volcano, one of five volcanoes in the Andean Austral Volcanic Zone (AVZ), using remote sensing techniques. We used data of the Advanced Spaceborne Thermal Emission and Reflection radiometer (ASTER) sensor which is highly effective in geological research, to understand build-up processes and to deduce compositional variation of the Viedma volcano emerging from the South Patagonian . The volcanic edifice was divided into bright parts that mainly compose the eastern flank of the volcano and dark parts

at the central crater area based on the observation in visible and near infrared ranges. The SiO2 concentration was cal- culated using the bands in the visible and thermal infrared regions. The dark part and the bright part have approximately

51 wt% and 63 wt% average SiO2 content respectively, indicating that the exposures of the Viedma volcano have a

wide variation in SiO2 concentration. Although, according to other authors, ejecta from the Viedma volcano have 64-66

wt% SiO2 and other AVZ volcanoes are essentially monolithologic / volcanoes, the edifice of the Viedma volcano appears to be composed mostly of basalts or older rocks/basement with low silica contents.

Keywords: ASTER, Viedma volcano, SiO2 wt%, Andean Austral Volcanic Zone, Chile.

Resumen. Variaciones composicionales reveladas mediante análisis de imágenes ASTER del volcán Viedma, Zona Volcánica Andina Austral. Mediante el uso de técnica de sensoría remota se ha desarrollado un mapeo litoló- gico del volcán Viedma, uno de los cinco volcanes de la Zona Volcánica Andina Austral (ZVA). Para este efecto, se ha utilizado el radiómetro ‘Advanced Spaceborne Thermal Emission and Reflection’ (ASTER) que es muy efectivo en investigación geológica, para entender los procesos que han controlado la estructura y deducir las variaciones composi- cionales del volcán Viedma, que sobresale levemente de la superficie del campo de hielo Patagónico Sur. Sobre la base de la observación en el intervalo del espectro visible e infrarrojo cercano, en el edificio se distinguen partes brillantes

que corresponden al flanco oriental del volcán y partes oscuras en el área del cráter central. La concentración de SiO2 se calculó a través del uso de las bandas en las regiones visibles e infrarrojo termal. Las partes oscura y brillante tienen

un promedio aproximado de 51% y 63% de SiO2 en peso, respectivamente, lo que indica que el volcán Viedma expone

una amplia diferenciación de las concentraciones de SiO2. Aunque, según otros autores, las muestras analizadas de

material piroclástico del volcán Viedma tienen contenidos de 64-66% SiO2 y que otros aparatos de la ZVA son volcanes esencialmente monolitológicos dacíticos/andesíticos, el edificio del volcán Viedma estaría compuesto principalmente de basaltos o las rocas más antiguas de su basamento tienen un bajo contenido de sílice.

Palabras claves: ASTER, Volcán Viedma, Contenido SiO2, Zona Volcánica Andina Austral, Chile.

Ms.249 Kobayashi et al.indd 433 08-07-2010 9:08:56 434 COMPOsmONAL VAJlL(TIONS REYEALED BY ASTER IMÁGE AJ¿4LYSIS OF 11IE VlEDMA VOLCANO, ...

1. Introduction TABLE 1. BAND-WAVELENGTH..sPATIALRESOLUTION.

The Viedma volcano, located at 49'22'S, 73'19'W, sensor ASTER wavelength spalial is a Holocene active volcano lhat belongs to lhe range band no. (11m) resolution Austral Volcanic Zone (AVZ) (e.g., González-Ferrán, 1 0.520 - 0.600 1995; Stern, 2004). The Viedma volcano is siluated 2 0.630 - 0.690 VNIR 15 m northeast oflhe , and surroonded by !he 3n 0.760 - 0.860 io !he Soulhern Patagonian Ice Field. 3b 0.760 - 0.860 The volcano is 1,500 m bigh aboye sea level and 4 1.600 - 1.700 exposes only a par! of its edifice aboye lhe surface 5 2.145 - 2.1 85 oflhe glacier. Killian (1990) reported !hat lhe latest 6 2.1 85 - 2.225 eruption occurred from a subglacial veot io 1988 SWIR 30 m deposited ash and on lhe Viedma glacier, 7 2.235 - 2.285 and lhe deposits were re-mobilized and flowed ioto 8 2.295 - 2.365 lhe Viedma Lake as alabar. Subsequeotly, Stern and 9 2.360 - 2.430 Killian (1996) reported geochemical data haviog an 10 8.125 - 8.4 75 8.4 75 - 8.825 adakitic signature and 64-66 wt"1o Si02 contents forlhe 11 ash and pumice from lhe Viedma volcano. However, TIR 12 8.925 - 9.275 90m lilhological vatiations !hat compose lhe edifice of 13 10.25 - 10.95 lhe Viedma volcano have not beeo well-documeoted 14 10.95 - 11.65 ye!, due lo ioaccessibility and lhe prevailiog heavy wealher conditions. Aimiog lo onderstand lhe lilhological vatiation over lhe volcano, we ernployed lhe remote seosiog lhe TIR. The absorption ofTIR by silicate mioerals techoiques. The Advanced Spacebome Thermal is correlated lo lhe molecular structore and silica Emission and Reflection radiometer (ASTER) conteot. For example, mafic minerals have large sensor, recently developed by Japanese Ministry of absorption atband 13 (11.5 ¡lm) and felsic miner­ Economy, Trade and Industry (METI), is a powerfol als atband 12 (9 ¡lm)(e.g., Conel, 1969; Hont and tool to map lhe composition of surface exposures. Vmcent, 1968; Hon!, 1980). Since igneous rocks ASTER has 3 banda (band 1-3) io visible andnear are classified by lhe assemblages of silicate mio­ iofrared range (VNIR) wilh spatial resolution ofl5 m eraIs and lheir quantity, lhe type of igneous rocks x 15 ID, 6 banda (band 4-9) io short wave iofrared aod approximate silica content cao be indirectly range (SWIR) wilh spatial resolution 000 m x 30 ID, inferred from lhe TIR absorption position. Figure and 5 banda (band 10-14) io lherma! iofrared range I shows lhe ernission spectrum of vatious igneous (TIR) wilh spatial resolution of90 m x 90 m (Table rocks io the TIR. 1). The ASTER sensor, having 11 bands in SWIR Based on tbis theory, the rock types aod ap­ and TIR waveleogth range, enables bighly effee!ive proximate Si02 contents of igneous rocks can be geological mapping, because rocks and mioerals estimated using lhe emissivity ratio (bandI3/bandI2). have its own inhereot spectral pattern io SWIR However, lhis melhod yielda relatively large er­ and TIR range. Usiog tbis characteristic, we can rors when a larget area has high Si02 content. To classify and detee! lhe type of exposed rocks and minimize the error, an approach to estimate Si02 mineralization, and estimate the mineral contents wt% using ASTER VNIR, SWIR aod TIR data quantitatively in sorne case. In lhis paper, we used was developed by means of multivatiate analysis ASTER images coveriog lhe overall body of lhe (ERSDAC, 2005). Stepwise regression analysis

Viedma volcano to address lhe vatiation of Si02 was used to detennine lhe parameters usiog Si02 contents of lhe edifice. wt% as a dependent vatiable, and lhe emissivity ratio, bands 1-9, aod their ratios such as band2/ 2, Estimation of SiO, contents bandl and band4/baod5 as iodependent vatiables. The significant iodependent vatiables identified The differeoce of emission spectrum character­ by lhe aoalysis were baod13/baodI2 aod band2/ istics caused by lhe Si02 conteots is remarkable io bandl. The resnlting Si02 wt% estimation equa- Kobayashi et al.! Andean Geology 37 (2): 433-441,2010 435

Stereoscopic image is acquired using nadir­ looking and backward-Iooking telescopes of - GRANITE High band 3 (0.76 to 0.86 ¡.un) in the same orbit. We ~ generated Digital Elevation Model (DEM) from _ ~ RHYOLlTE the stereoscopic images. We then created a three ~~ONZONITE dimensional image ofViedma volcano by super­ .. -~ ~ ~ i5 --..'----.J '--/~ ANDESITE SiO, imposing the DEM data on the false color image (Fig. 3). The 3D image shows that the north side of ~ - W¡¡;~ BASALT the edifice is, compared to the south side, rougher due to erosion by the glacier. Several irregular concave structures are preserved in the southem ~ ~ P::::: 'o. part ofthe edifice, sorne are considered to be recent craters. This interpretation agrees with the report of recent occurrence of eruption on the south side 6.25 7 8 10 15 2025 ofthe edifice by González-Ferrán (1995). A N-S WAVELENGTH ( ¡lm) trending 5 km-long lineament was observed in the middle ofthe volcanic body in this image. Several FIG. l. Spectral characteristics of igneous rocks in thermal infrared region. Vertical gray bands indicate the ranges irregular craters were observed to align on the of five ASTER TIR bands. lineament. This alignment of craters corresponds to the observation ofGonzález-Ferrán (1995) who showed that four large craters/, between tion is given below. The estimation accuracy of 1.5 and 4 km in diameter, are located along a the equation is RZ=0.74. N-S lineo However 3D observation in this study shows that diameter ofthe craters is 1.5 km at the SiOz (wt«>/o)= 138.42x(band13/band12)+34.61x (band2/ maximum. Sorne ofthe craters/calderas illustrated band1) - 129.02 (eq.1) in González-Ferrán (1995) were not recognized in this study. where band2/band1 is a proxy for the state of oxida­ tion of iron in rock surface and hence a proxy for 3.2. Si02 contents the weathering intensity. Preprocessing of ASTER data is illustrated in 3. Results figure 4. First, we applied atmospheric correction to VNIR data to acquire reflectance data, and 3.1. False color image and Digital Elevation atmospheric correction and temperature emissi­ Model(DEM) vity separation to TIR data to acquire emissivity data. Second, to apply the equation 1 to the rock We used VNIR data and assigned band 1 for blue, exposures, we removed areas affected by cloud, band 2 for red and (3 x band1 +band3)/4 for green to vegetation, ice and side moraine from the data. produce false color image (Fig. 2). Objects larger than For noise elimination, a conventional method used 15 m x 15 mis shown as a dot with speci:fic color. Oreen in the remote sensing analysis was employed as color in the figure 1 indicates vegetation, white is snow follows. We excluded the vegetation pixels from and glacier, black is water and gray is rock exposures. calculation using the threshold derived from Because this ASTER data was acquired in summer NDVI (vegetation index), i.e., (band3-band2)/ (January 03, 2005), outcrops of the Viedma volcano (band3+band2) (Tucker, 1979). Pixels consistently emerge more widely. Based on the VNIR observation, having a high value for bands 1-3 in the VNIR and the volcanic edifice was divided into bright parts and band 4 in the SWIR indicate cloud. Similarly, ice dark parts (Fig. 2). The boundary of two colors did and snow constantly shows high value in the VNIR not correlate with the topographical elements such three bands. By excluding the pixels having high as elevation, steepness of slopes and obliquity of sun values in these three bands of the VNIR range, radiation (shadows). The color difference most likely cloud, snow and ice pixels were removed. Moraine originated from the color of the outcrops. was eliminated by visual examination based on the t; 0'1 Legend 111 :mountain body ~ 111 :morain c=J : ice and snow 111 : lake ( inferred vent) ~ : crater wa 11 (broken 1ine is inference) , " : 1i neament (b roken 1i ne i s i nference)

g ~ ~ ~ ~

:;¡~ ~ Ell ~ 1» ::..."< VJ ¡;;j ::ó ~ &l ~ ¡,:j "i ~

t>i~ ~ ~ 1, 9 2 ,km ~ ___~ ____2 , km ~

t"~ ~ FIG. 2. Left: Color composite irnage ofthe Viedrna volcano generated using data ofASTER VNIR bands acquired on January 03, 2005. Right: Landforrn analysis rnap ofViedrna volcano. ~ Kobayashi et al.! Andean Geology 37 (2): 433-441,2010 437

FIG. 3. Bird's eye views. Upper image: view from fue SE. Lower image: view from fue NW. These 3D images are produced by overlaying VNIR image on fue DEM (Digital EIevation Model). The vertical scaIe is exaggerated by a factor of2 to emphasize relief contrasts. 438 COMPOSI110NAL VARIA110NS REVEALED BY ASTER IMAGE ANALYSIS OF TllE VIEDMA VOLCANO, ...

(ERSDAC, 2005). However, partems of estimated Flowchart of Si02 estimation Si02 concentration seem to have no strong control / ASTER VNIR / AS TE R TIR of altitude or slope angle, by comparing the pat­ / / tem in figure 5 to the topographic image in figure 1 Atmospheric correction 1 Atmospheric correction 6. Therefore, we infer that the pattems reflect the Temperature Emissivity average Si02 concentration ofthe exposure. 1 Separotion There is a possibility that the Viedma volcano / Refl ectance data/ / Emissivity data / erupted basaltic ejecta or . The Viedma vol­ cano, however, is a smaIl volcano and unlikely to

Ortho correction & have basaltic or variable compositions that tend to Bond registrotion I I form large volcanic edifice. Furthermore, most other volcanoes in theAVZ are monolithic dacitic/andesite I Removol processing (Cloud, vegetotion. ice, moroine) I in composition (Stem and Killian, 1996). Another explanation for the low Si02 exposures

11 Estimotion equotion 11 is distribution ofbasement rocks having low Si02 content. In fact, exposures of Paleozoic metamor­ phic rocks (metasediments and metabasites) were Contents map of Si0 2 reported around the Viedma volcano (e.g., Vivallo et al., 1999). Carbonate rock showing no absorp­ tion feature around 9 ~m was considered to be the FIG. 4. Flowchart ofSi0 wt% estimation. 2 potential host rock with low Si02 contents, but its diagnostic emissivity absorption near 11.2 ~m due

to C03 vibration was not observed at the Vieldma image overlaid on DEM because emissivity oftalus volcano. Therefore, we conclude that the expected slopes such as moraine is known to be different from basement exposure is likely metapelites or rocks the original rock as described in ERSDAC (2005). with basaltic composition (old basaltic edifice or

The result of the Si02 estimation is shown in basement ofmetabasalt). figure 5. Noises were marked in whitish color and There are other factors that may lead to under­ water in dark green. Although rock types (igneous, estimation of apparent Si02 concentration. At least, metamorphic or sedimentary) cannot be distinguis­ two possibilities were assessed before. The frrst is hed, the Viedma volcano has much wider variety soil cover on the exposure surface. The second is ofSi02 wt% (51-63 wt%), compared to previously the influence of hydrothermal alteration, such as ; published data (64-66 wt% in Si02 n=2) (Stem sulfide mineralization on the surface. To eliminate and KiIlian, 1996). This variation is almost similar such possibilities, validation study that compares to that of the Hudson volcano (50.5-66.6 wt% in the results estimated by ASTER and data sampled ; Si02 Orihashi et al., 2004, 2008), which is one of at the Viedma volcano wiIl be required in the future. the largest stratovolcanoes in the southernmost SVZ. We demonstrated that the remote sensing techni­ que based on ASTER data is useful tool for recon­ 4. Discussion and concluding remarks naissance and lithological mapping of a volcano as

weIl as semi-quantitative estimation of Si02 wt%.

The rocks with high Si02 contents estimated In this study, we estimated the Si02 wt% ofViedma in this study were distributed at the eastem flank volcano to vary from 51 % to 63%. Although there of the edifice and the maximum Si02 content was are sorne possibilities for lower Si02 value (e.g., consistent with the reported values (64-66 wt'llo: e.g., metamorphic rocks). Stem and Killian, 1996). In contrast, we obtained exposures in the central crater area with much lower Acknowledgements ¡.e., Si02 contents, basaltic equivalent composition. We are grateful to Dr. Y. Arakawa ofKangean Energy In general, talus, regardless if they are basaltic Indonesia Ltd. and eolleagues of the CHRISTMASSY or rhyolitic, tend to exhibit spectral characteristic (Chile Ridge Subduetion to Supply System) Pro­ closer to the rocks with intermediate Si02 contents jeet Group for diseussions and supports. We should like to Kobayashi et al.! Andean Geology 37 (2): 433-441,2010 439

' km L-~~~~-L ______L- ______-" !

FIG. 5. Map of Si02 contents. Viedma volcano has wide variety of Si02 contents frorn 45 wt% to 67 wt%. Typical average Si02 content is 51 wt%-63 wt% in rnost area. 440 COMPOsmONAL VARIATIONS REVEALED BY ASTER lMAGE ANALYSIS OF THE VIEDMA VOLCANO, ••.

'km L-~~~~~L-______~ ______~I

FIG. 6. Contour map. This map is produced by relative ASTER DEM. Red and violet contours are at 50 m and 10m intervals, respectively. Kobayashi el al.! Andean Geology 37 (2): 433-441, 2010 441

express OUT appreciation to Dr. C. Stern, Dr. A. Demant Orihashi, Y.; Nakai, S.; Shinjoe, H.; Naranjo, l.A.; Mo­ and the anonymous referee for their helpful comments. toki, A.; CHRISTMASSY Group 2008. Magmatic Tbis research was partly supported by the Earthquake evolution afthe Quaternary volcamcs from Hudson Research Institulo ofthe University ofTokyo cooperative and volcanoes, Austral Andean Cordillera. research program (to D.H.). In Goldschmidt Conference, No.18, Gecchimica el CosmochimicaActa Special Supplement 72 (12): p. References A709. Vancouver, Canada. Orihashi, Y.; Naranjo, l.A.; Motoki, A.; Sumino, H.; Cone~ lE. 1969. Infrared emissivities of silicates: Experi­ Hirata, D.; Anma, R.; Nagao, K. 2004. Quatemary mental resu1ts and cloudy _osphere model of speclra volcanic activity ofHudson and Lautaro volcanoes, emission from condensed particulate mediums. Journa1 Chilean Patagonia: ncw constraints from K-Ar ages. ofGeophysical Researeh 74 (B6): 1614-1634. Revista Geológica de Chile 31 (2): 207-224. ERSDAC.2005.Remoresensingapplicationfordevelopment Stem, C. 2004. Active Adean volcanism: its geologic ofextractingmethodofhighpotentialresourceareafrom and lectonic setting. Revista Geclógica de Chile 31 ASTER TIR data, Rcport onReseareh andDevelopment (2): 161-206. of Remote Sensing Technology for Non-renewable Stern, C.; Killian, R. 1996. Role ofthe subduclod slab, Resources (3/4): 171-187. Tokyo, Japan. mantle wedge and continental in the genera­ Goozález-Ferrán, O. 1995. Volcanoes de Chile. Instituto tion of adakitc from the Andcan Austral Volcanic Gecgráfico Militar: 640 p. Santiago. Zanc. Contributions to Mineralogy and Petrology Hunt, G.R 1980. Electromagnctic Radiation: The Cotnmu­ 123: 263-281. nication Link in Remoto Scnsing. In Remoto Scnsing Tucker, C.J. 1979. Red and photographic infrnrcd linear in Geclogy (Sie~ B.S.; Gil1espie, A.R; editors). John combinations for monitoring vegetation. Remote Wiley: 5-45. Ncw York. Sensing ofEuviromnent 8: 127-150. Hun~ G.R; Vmcen~ RK. 1968. The behavior of spectra1 Vivallo, W.; Gardcweg, M.; Tassara, A.; Zanettini, J.; fca1ures in !he infrared emission from particulate SUIface Márquez, M.; González, R. 1999. Mapa de recursos ofvarious grain size. Journa1 of Geophysical Researeh minerales del área fronteriza Argentino-Chilena 73 (B18): 6039-6046. entre los 34' y 56'S. Servicio Nacional de Geologia Killian, R 1990. ThcAustralAndcan Volcanic Zooe (South y Mineria y Servicio Geológico Minero Argentiuo. Patagonia).InlntemationalSymposiumonAndcanGecI­ PublicaciÓll Geológica Multiuacional No. 1, escala ogy(ISAG), No. I,AbsIract: 301-305. Grenoble, Fnmce. 1:1.000.000.

Manuscript received: July 29, 2009; rcviscd/accepted: Aprill9, 2010.