Sensores remotos y Riesgo ambiental Serie

Conference Proceedings

REMOTE SENSING, NATURAL HAZARDS AND ENVIRONMENTAL CHANGE

28 – 29 July 2011

Centre for Remote Imaging, Sensing and Processing, National University of Singapore, Singapore

Laboratoire Magmas et Volcans, CNRS UMR6524,and CLERVOLC, Université Blaise-Pascal, France

An international conference on remote sensing, natural haz-ards and environmental change took place in Singapore on 28-29 July 2011. It was organized by Centre for Remote Imaging,Sensing and Processing, National University of Singapore;Laboratoire Magmas et Volcans (LMV CNRS UMR 6524),Université Blaise Pascal, Clermont-Ferrand, France; and theWorking Group for Large Rivers and Climate Change, Inter-national Association of Geomorphologists (IAG).

Scientists use remote sensing to observe, analyze andrecord changes on the surface of the Earth. The recent arrivalof a new generation of satellites with extreme high resolu-tions has increased the ability to study hazardous phenomenaof nature such as volcanoes, earthquakes, tsunamis and largefloods. Remote sensing is also an efficient tool for investigat-ing the effects of anthropogenic modifications of the envi-ronment. Such modifications include climate change. Bothnatural hazards and environmental modifications are com-mon in Southeast Asia, and studies from the region formedthe core of the conference.

Forty-eight scientists from 11 countries examined remotesensing, natural hazards and environmental change in this two-day meeting held at the National University of Singapore. Inthe framework of the ICT-STIC Asia (Information and Com-munication Technologies) research and exchange programsponsored by the French Foreign Office in Southeast Asia, ourproject termed ‘ImagerleRisk’ focused on the application of re-mote sensing on geological hazard studies in institutions anduniversities based in Indonesia, the Philippines, Singapore and

IntroductionJ.-C. Thouret*, S. C. Liew**, A. Gupta**,***

Une conférence internationale « Télédétection, aléas, risquesnaturels et changement environnemental » a eu lieu à Singa-pour les 28 et 29 juillet 2011. Cette conférence a été organisépar le Centre for Remote Imaging, Sensing and Processing(CRISP), Université Nationalede Singapour (NUS), Singapour,le Laboratoire Magmas et Volcans (LMV CNRS UMR 6524),Université Blaise Pascal, Clermont (France) et par le groupede travail « Large Rivers and Climate Change » de l’Associa-tion Internationale des Géomorphologues (AIG).

Les chercheurs utilisent la télédétection dans le but d’obser-ver, d’analyser et d’enregistrer les changements qui survien-nent à la surface de la Terre. Le lancement récent d’une nou-velle génération de satellites capables d’acquérir une imageriede très haute résolution spatiale (submétrique) a considérable-ment accru les moyens d’analyse des phénomènes domma-geables, tels les effets des éruptions volcaniques, des séismes,des tsunamis et des grandes inondations. La télédétection estaussi un outil qui permet d’étudier de façon synoptique et ré-currente les effets des modifications d’origine anthropiquesur l’environnement à grande et petite échelle. Ces modifica-tions actuelles incluent le changement climatique dont els ef-fets son redoutés. Or, les effets des aléas naturels et les modifi-cations environnementales engendrées sont communs en Asiedu SE ; c’est pourquoi les études conduites dans cette régionforment le cœur de cette conférence internationale.

Quarante-huit scientifiques de onze pays ont examiné la té-lédétection, les aléas naturels, leurs effets et les changementsenvironnementaux durant une conférence de deux jours tenue à

l’Université Nationale de Singapour (NUS). L’organi-sation de cette réunion était l’un des objectifs du pro-gramme d’échanges et de recherche ICT STIC Asie(Technologies de l’information et de la communication)promu par le Ministère des Affaires étrangères en Asiedu SE. Au sein de ce programme, notre projet « Ima-gerleRisk » est consacré à l’application de la télédé-tection à l’évaluation des risques volcaniques et hy-drologiques en collaboration avec quatre institutionset universités basées en Indonésie, aux Philippines, àSingapour et en France. Seize étudiants deMaster etdoctorants issus de ces pays partenaires ainsi que del’Australie et de la Nouvelle Zélande ont été encoura-gés à présenter leurs résultats lors de la conférenceinternationale, grâce à l’aide des Ambassades deFrance à Jakarta, Manille et Singapour. Trois agences

Remote sensing, natural hazards and environmental change, p. 3-4

* PRES Clermont, Université Blaise Pascal, Laboratoire Magmas et Volcans, CNRS-UMR 6524, IRD-UR163, 5 rue Kessler, 63038 Clermont-FerrandCedex, France.** CRISP, Centre for Remote Imaging, Sensing and Processing, 10 Lower Kent Ridge Road, National University of Singapore, Singapore.*** School of Earth and Enrivonmental Sciences, University of Wollongon, Australia.

Fig. 1 – 2000 flood in the vicinity of Phnom Penh (Courtesy of CNES,CRISP, A. Gupta).

4 Remote sensing, natural hazards and environmental change

J.-C. Thouret, Soo Chin Liew, Avijit Gupta

France. Sixteen graduate students from the partner countriesand from Australia and New Zealand were encouraged to pre-sent papers at this meeting. Three space agencies were repre-sented: Centre National d’Etudes Spatiales (CNES), JapanAerospace Exploration Agency (JAXA), National Aeronauticsand Space administration (NASA). The four keynotes were on(1) International charter for earthquakes in Japan (M. Nagai,JAXA), (2) Quantifying volcanic hazard and risk (C. McGill,Macquarie University), (3) Remote sensing of volcanic emis-sions in the Asia-Pacific region (S.A. Carn, Michigan TechnicalUniversity) and (4) Spatial information for analyzing changinghydro-meteorological risk (C. Van Westen, University of Twente).Twenty-seven oral papers and six posters formed the rest of thepresentations.

After the conference, the participants were requested tosubmit an extended version of their conference abstract witha limited number of supporting diagrams to make this publi-cation possible. The publication thus includes the conferenceprogramme, the keynote presentations, extended abstractsfor a substantive number of papers presented at the confer-ence, and brief abstracts of the other papers. We would liketo acknowledge the support from CNES which made thiscollection possible. We also thank Kenix Koh Poh Loo andFrédérique Van Celst for organizing and preparing confer-ence-related publications.

The meeting explored possible technical collaboration be-tween institutions and the strengthening of such programs. Itis hoped that similar meetings will take place in the future.

This is Laboratory of Excellence Clervolc contribution n°5.

spatiales étaient représentées : le Centre National d’EtudesSpatiales (CNES), l’agence d’exploration aérospatiale duJapon (JAXA) et la National Aeronautics and Space Adminis-tration (NASA). Quatre exposés centraux ont été livrés sur lessujets suivants : (1) « Charte internationale pour l’étude desséismes au Japon » (M. Nagai, Jaxa, Japon et Thailande),(2)« Quantifier l’aléa et le risque volcanique » (C. Magill, Mac-quarie University, Sydney, Australie), (3) « La télédétectiondes émissions volcaniques dans la région du Pacifique et del’Asie du SE » (S. Carn, Michigan Technical University, USA),et (4) « L’information spatiale pour l’analyse des modifica-tions des risques hydro-météorologiques » (C. Van Westen,University of Twente, Pays Bas). En outre, vingt-sept exposésoraux et six posters ont été présentés durant la conférence.

À la suite de la conférence, les participants ont été priésde soumettre une version étendue de leur résumé donné lorsde la conférence avec un nombre limité d’illustrations, afinde rendre la publication d’un ouvrage possible. La publica-tion de ces Actes de la Conférence rassemble ainsi le pro-gramme de la conférence, les présentations centrales, les ré-sumés étendus d’un certain ombre d’articles présentés lors dela conférence et un bref résumé des autres articles. Nous vou-drions remercier le CNES pour son soutien en faveur de lapublication de ces Actes. Nous remercions également KenixKoh Poh Loh (Singapour) et Frédérique Van Celst (Cler-mont) qui ont bien voulu organiser puis préparer les publica-tions issues de cette conférence.

La conférence internationale a enfin exploré les possibili-tés de coopération technique entre les institutions et à cher-ché à renforcer de tels programmes. Les participants ont émisle vœu que des conférences similaires prennent place en Asiedu SE dans l’avenir proche.

CRISP is honoured to have this opportunity to co-host theConference with Laboratoire Magmas et Volcans, UniversiteBlaise Pascal, and the Working Group for Large Rivers andClimate Change, International Association of Geomorpholo-gists. The two key areas of discussions of this conference areNatural Hazards and Environmental Change. The Asia-Paci-fic is, as we all know, is a disaster-prone region. In view ofthe recent devastating earthquakes and tsunamis in Japan,and the earthquakes in Christchurch, New Zealand, the rele-vance of these topics is beyond question. Singapore, fortuna-tely, is outside the major natural disaster zones. Nevertheless,we still occasionally suffer from flash floods due to extremerainfalls and Singapore is not immune from the potential ef-fects of global climate change.

Satellite remote sensing plays a useful role in disaster mo-nitoring and management. At CRISP, we operate a satelliteground receiving station that receives satellite data from va-rious satellites, ranging from the low and medium resolutionenvironmental satellites such as MODIS to high resolution sa-tellites which include SPOT, Ikonos, GeoEye, and World-View-1 and 2. We have a research team that conducts researchin various aspects of earth observation.

We collaborate with the National Environment Agency inmonitoring the regional fires using a combination of low andhigh resolution satellites. During the December 2004 IndianOcean Tsunami, satellite images of the affected areas receivedby CRISP were sent to the Singapore humanitarian forces forplanning and execution of relief and rescue efforts. In fact,whenever a disaster event occurred in the region, we wouldtry our best to acquire satellite images of the affected areas.These images were promptly dispatched to our counterpart inthe disaster area to aid in the relief and rescue operations. Theimages are also posted on CRISP web site for free access.

CRISP’s scientists actively participate in internationalcollaborative projects. For example, CRISP is a partner inthe Imagerisk project of STIC Asia. Together with scientistsfrom France, Indonesia and Philippines, satellite images areused in risk mapping of regions around volcanoes. CRISPalso participates as a data analysis node of Sentinel Asia, aninitiative led by the APRSAF (Asia-Pacific Regional SpaceAgency Forum), of which JAXA plays a major leading role,to support disaster management activity in the Asia-Pacificregion using satellites data. The first keynote speaker of thismorning session will elaborate on this topic.

Despite the small group we have today, I am delighted thatwe have representations from 19 organizations and 11 coun-tries, including three space agencies, JAXA, CNES andNASA, as well as participants from universities and agen-cies in France, United States, Indonesia, the Philippines,Vietnam, Australia, New Zealand, Netherland, and of course,Singapore. I believe we are all here with the same goal, andthat is, to exchange experiences and to discuss the results ofour study in remote sensing, natural hazards and environ-mental change. Looking ahead, I hope this conference willgenerate increased interest in environmental change andbring about improved strategies for the mitigation of natu-ral disasters.

I sincerely thank our conveners Prof Jean-Claude Thouret,Dr Avijit Gupta and Dr Liew in spearheading this Conferen-ce and the organizing committee who have worked hard tomake this event possible. The kind supports from STICAsia, CNES and the French Embassies in Singapore, Indo-nesia and the Philippines are gratefully acknowledged. Lastbut not least, I thank all participants for your contribution.Many among you have travelled from afar to share yourexperiences and knowledge.

Welcome address

L. K. Kwoh*

* Director, CRISP, National University of Singapore.


Introduction

On the 11th March 2011, the huge earthquake of M9.0 oc-curred offshore of Tohoku area of which epicenter was 500 kmlong and 200 km wide in the Pacific Ocean. Accordingly thedamaged areas were also 500 km long including a part ofHokkaido to Tokyo. The huge earthquake triggered tsunamiwaves that propagated westward toward the Japan coastlineand eastward across the Pacific Ocean. Coastal areas withinIwate, Miyagi, and Fukushima prefectures experienced wavesof over 5 m, with some areas above 10 m, and in localizedareas more than 15 m. There are over 15,000 confirmeddeaths, with approximately 4,000 people missing in Septem-ber 2011. Over 440,000 people have been displaced from theirhomes by the earthquake, tsunami, and radiation alert.

Asian Institute of Technology (AIT) had been playing acritical role in helping recovery efforts in Japan for the In-ternational Charter and Sentinel Asia. The Cabinet Office ofJapan had activated the International Charter and the Char-ter has appointed Geoinformatics Center of AIT as projectmanager of this disaster for the International Charter on ‘Space

and Major disasters’. International space organizations wereworking along with AIT to provide maps and satellite imagesfor rescue and relief operations. The International Charter wasactivated following a disaster and partner agencies immediate-ly start sharing satellite information and data on the disaster.Apart from JAXA, ESA, USGS, CNES, DLR, CSA, ISRO,CNSA, KARI, GISTDA, and NSPO are among the agenciesactively involved in helping for disaster relief operations.AIT is also acting as a Principal Data Analysis Node (P-DAN) of Sentinel Asia that is a coordinator of data analysisin Sentinel Asia activity. Table 1 shows the list of satellites,which was kindly provided for the disaster. The blue colorwas provided from Sentinel Asia, the bright yellow was pro-vided from the International Charter, and the right green isprovided by bilateral agreement.

International Charter for Earthquakein Japan

Emergency satellite observation and rapid analysis isbeing coordinated by AIT along with major international

space agencies, United Nations,research institutions, universities,and commercial sectors to providecritical support in the recovery op-erations. Satellite images are help-ing detect changes in the land-forms, particularly in the coastlinein North West Japan. Using changedetection techniques, where thesatellite image before the disaster iscompared with the satellite imageafter the disaster, the difference inlandforms becomes very apparent.New threats are identified and newmapping helps guide rescue and re-

International charter for earthquake in Japan

M. Nagai*

Short abstract: On the 11th March 2011, the huge earthquake of M9.0 occurred offshore of Tohoku area of which epicenter was 500kmlong and 200km wide in the Pacific Ocean. Accordingly the damaged areas were also 500km long including a part of Hokkaido to Tokyo.The International Charter is activated just after the earthquake. The International Charter aims at providing a unified system of spacedata acquisition and delivery to those affected by natural or man-made disasters through Authorized Users. The International Charter isactivated following a disaster and partner agencies immediately start sharing satellite information and data on the disaster. Apart fromJAXA, ESA, USGS, CNES, DLR, CSA, ISRO, CNSA, KARI, GISTDA, and NSPO are among the agencies actively involved in helpingfor disaster relief operations.

Keywords: earthquake, tsunami, disaster management, international charter, sentinel Asia.

Organization Satellite Organization Satellite Organization Satellite

JAXA ALOS

USGS

LANDSAT-5,7 CNES SPOT-4,5

NSPO/NARL FORMOSAT-2 EO-1 CSA Radarsat-1,2

GISTDA THEOS IKONOS-2 ESA ENVISAT

ISRO Cartosat-2 GeoEye-1 ASICOSMO-SkyMed

KARI KOMPSAT-2 QuickBird-2 DEIMOS DEIMOS-1

DLR RapidEye WorldView-1,2 RSOCOSMOS RESURS-DK

TerraSAR-X CNSA HJ EIAST DubaiSat-1

Tab. 1 – List of satellites.


* Satellite Application and Promotion Center (SAPC), Japan Aerospace Exploration Agency (JAXA), Tsukuba Space Center, 2-1-1 Sengen, Tsukuba-shi, Ibaraki-ken, 305-8505, Japan.


M. Nagai

lief teams to reach the destination. Since the devastationcaused by the earthquake and tsunami is huge, new mapshave to create in the affected areas. This is where satellite im-agery assumes critical importance. Table 2 shows the list ofemergency observations and provided satellite data for thefirst 10 days after the disaster. Figure 1 and Figure 2 wasemergency mapping products created by GeoinformaticsCenter of AIT. Figure 1 shows the inundated areas due to the

Tsunami by RADARSAT-2 (RADARSAT-2 Data and Prod-ucts © MacDONALD, DETTWILER AND ASSOCIATESLTD. (2011) – All Rights Reserved and ‘RADARSAT is anofficial mark of the Canadian Space Agency’) and WorldView multi spectral image (© USGS), which was acquiredon 12 March, 2011. The inundated area can be seen in redcolor patches. It was equal to approximately 71 square kmarea of extent. Figure 2 shows Tsunami affected area near the

Date Satellites

12 Mar.ALOS(AV2, PSM), FORMOSAT-2, THEOS, IKONOS, WorldView-2, GeoEye-1, RapidEye, LANDSAT-7, SPOT-5, TerraSAR-X, Radarsat_

13 Mar. ALOS(PSR), FORMOSAT-2, THEOS, WorldView-2, RapidEye, EO-1, LANDSAT-5, SPOT-5, TerraSAR-X_

14 Mar. ALOS(AV2, PSR), FORMOSAT-2, Cartosat-2, KOMPSAT-2, GeoEye-1, RapidEye, SPOT-5, HJ_

15 Mar. FORMOSAT-2, SPOT-4, DubaiSat-1_

16 Mar. ALOS(AV2,PSR), FORMOSAT-2, QuickBird-2_

17 Mar. ALOS(AV2), FORMOSAT-2, SPOT-5_

18 Mar. ALOS(PSR), FORMOSAT-2, EO-1, SPOT-5, DEIMOS-1_

19 Mar. ALOS(AV2), FORMOSAT-2, KOMPSAT-2, WorldView-2, GeoEye-1, RESURS-DK

20 Mar. ALOS(AV2, PSR), FORMOSAT-2, IKONOS, WorldView-2, LANDSAT-5, RESURS-DK

Tab. 2 – List of emergency observations.

Fig. 1 – Tsunami inundated area.

9Remote sensing, natural hazards and environmental change

International Charter for Earthquake in Japan

Sendai Airport from IKONOS Panchromatic image (© USGS)acquired on 12 March 2011.

Sentinel Asia

Sentinel Asia is conducting emergency observation by earthobservation satellites in case of major disasters. Currently par-ticipating satellites are expected to be ALOS (JAXA), IRS

(ISRO), THEOS (GISTDA), KOMPSAT (KARI) and FOR-MOSAT (NARL), which are called DPN (Data ProviderNode). Those agencies accept observation requests for majordisasters in the Asia-Pacific region from ADRC member orga-nizations and representative organizations of JPT (Joint Pro-ject Team) members. On the other hand, DAN (Data AnalysisNode) analyzes the satellite data provided by DPN, makesvalue added product and discloses the result through the Sen-

Fig. 2 – Sendai airport area.


M. Nagai

tinel Asia System within the domestic legislation of each DANpermits. AIT (Asian Institute of Technology) coordinates dataanalysis as a P-DAN (Principal Data Analysis Node).

DAN (Data Analysis Node) is the Node for the data anal-ysis. DPN provide its own satellite data to the DAN and datapolicy of each DPN is decided by each DPN. DAN memberimplements the following tasks; to analyze the satellite dataprovided by DPN, to make value added product, and to dis-close the result through the Sentinel Asia System within thedomestic legislation of each DAN permits. For creatingvalue added products, data should be aware of various DPNsatellite data. It is very important to have analysis experi-ence in handling DPN satellite data and DAN needs toprepare resources to analyze satellite data. Information inte-gration is key for making value-added products to havevarious other data helpful for the given disaster. Validationof products is needed to carryout ground verification as soonas possible before product distribution. Reliability has to beproduced within limited time.

Conclusion

On March 11, 2011, the massive earthquake was occurrednear the east cost of Honshu, Japan. It caused a massiveTsunami with widespread destruction of human live andproperties. Figure 3 shows the coverage of satellite imagesacquired by different satellites after 5 days of the disaster.The emergency response of space agencies of the Interna-tional Charter and Sentinel Asia was really quick and veryeffective to understand damages in the recovery operations.More than 6,000 images had been provided including pro-grammed and archived images in the International Charterand Sentinel Asia. In addition, more than 150 analyzed mapshad been created.

Finally, I would like to express my condolence to those vic-tims and their family lost by Japan’s Earthquake on 11 March2011. Also, I thank many friends from foreign countries andregions to have sent me kind words to encourage me as wellas Japanese people.

Fig. 3 – Coverage of emergency observation.

Introduction

Over the past twenty years with the significant develop-ment of remote sensing, many actors have addressed thetheme of geohazards. The space data used have ranged fromradar and optical remote sensing imagery to more recentgravity field or ionospheric perturbation data. Over this peri-od, the CNES has accompanied the French scientific commu-nity in the development of missions, but also in space dataanalysis with the ultimate aim of better understanding geo-hazard processes. This paper will review some of this workand propose areas where space agencies can, working hand inhand with their land based counterparts, help to move forwardresearch on this important theme.

While geological hazards have rarely been the primary ap-plication of space missions, many satellites provide usefuldata for the study of earthquakes, volcanoes and landslides.High resolution optical imagery (e.g. SPOT 5) is currentlyused to provide source geometry after major earthquakesand work is currently underway on integrating systematicinterferometric SAR displacement measurements in the mon-itoring of high risk faults. In these cases, the generally highspatial coverage of InSAR data provides a unique view oflocal displacement. This, coupled with in situ data from seis-mic networks and GPS, gives an unprecedented insight intothe way in which many earthquakes take place. Today, mea-suring displacements for all types of geohazard is a potential-ly major application of remote sensing imagery – unfortu-nately the supply of such imagery is unable to match the widerange of demand. While the solid Earth research community in-ternationally remains dispersed, various opportunities (GMESin Europe or the GEO geohazard ‘Supersites’ initiative in-ternationally) could potentially improve this situation in thecoming years.

Current missions and methods

With the development of remote sensing over the past twen-ty five years, scientists have developed many interesting waysto extract information from satellite data useful in the study ofgeohazard processes. These range from deriving ground dis-placement data in different ways, to following the effects in theionosphere of the displacement of large masses of water byearthquakes resulting in Tsunami. In this paper a review ofsome of these techniques will be presented.

Ground displacement

Different methods that can be used to derive quantitativeground displacement maps from satellite data are describedbelow:

Historic events. In the case of earthquakes, these often leavegeomorphological markers (stream channels, terraces) that can,in certain instances, be used to derive quantitative estimationsof displacements induced by these events. This type of inter-pretation (Klinger et al., 2011), of optical satellite imagery mostfrequently, is particularly effective in dry areas where thesemarkers have remained unaffected by erosion processes and arenot hidden by vegetation. Under these ideal circumstances andwith a high resolution image (>5 m) it is possible to estimatedisplacements of less than a metre. While these ideal conditionsallow more accurate quantitative estimations, geomorphologi-cal markers are used routinely in this way to obtain key infor-mation in seismotectonic studies across the world.

Optical image correlation. When the event is recent andarchive imagery of the zone affected is available for the pe-riod before the event, an image acquired subsequently to theevent can be used, under certain circumstances, to measurethe ground displacement. Depending on the spatial resolu-

The contribution of space based observations to understanding and addressing geohazards:

a CNES perspective

S. Hosford*

Short abstract: Earth observation data acquired by diverse satellites has been used over many years to address the important study areaof geohazards. Many methods have been developed that provide key information in the estimation and monitoring of hazard and thestudy of geohazard processes. Over the next ten to fifteen years new opportunities will arise in this domain with access to huge newdatasets provided on a ‘free and open’ basis. A huge challenge lies ahead for the international geohazards science community to makethe most of this data : improved international coordination is undoubtedly required to face this challenge.

Keywords: geohazards, satellites, future missions.


* French Space Agency CNES, 18 avenue Edouard Belin, 31000 Toulouse, France.


S. Hosford

tion of the images used, displacements up to several 10’s ofcentimeters can be measured (Delacourt et al., 2007). Thisis true largely regardless of the type of event that has in-duced the ground displacement, whether it be an earthquake,volcanic eruption, or landslide, for example. The COSI-Corr[1] package is currently the most widely used and freelyavailable package to apply these methods.

Differential Interferometric SAR. The DInSAR techniqueand its various close relatives (Permanent Scatterer, SBAS etc.)emerged in the mid 90’s (Massonet et al., 1993) as a revolution-ary new technique based on processing of at least two SAR im-ages spanning a period within which some ground displacementhas occurred. Centimetre or smaller displacements can be mea-sured (a fraction of the wavelength of the SAR imager used – C,L or X band) and large area coverage products can be generatedin a near-automated. These can provide critical information nowused in earthquake source modeling as well as to derive dis-placement over time for landslide and volcanic hazard assess-ment. Due to their capacity to measure even smaller movements,PS InSAR and other related techniques exploiting large stacks ofSAR images are now being used to investigate previously uniden-tified seismic phenomena such as ‘silent earthquakes’. Thesetechniques are providing new insight into the seismic cycle. TheROI_PAC [2] processing package is freely available for aca-demic use to apply some of these methods.

Given the sensitivity of the optical and SAR image process-ing techniques described, these methods complement eachother. When studying earthquakes, for example, large near-fielddisplacements (too large for InSAR which loses coherence) canbe mapped using optical data and smaller displacement furtherfrom the surface rupture is captured by InSAR.

High resolution DEMs

Various types of image data can be used to derive high res-olution DEM products by classic photogrammetry or SAR in-terferometric techniques. Those providing the highest spatial

resolution coupled with the best height accuracy are pro-duced, in the optical domain by data from the HRS instrumenton board the SPOT 5 satellite and, in the radar domain, by therecent TanDEM-X and Cosmo-Skymed missions.

Geological mapping

High spatial and spectral resolution data from the Hyperionand ASTER missions are most commonly used for geologicalmapping. As was mentioned previously geomorphological map-ping can be carried out with standard multispectral imagery suchas that provided by the SPOT series, Landsat and the high res-olution commercial data providers.

Monitoring volcanic eruptions and ash clouds

During the recent volcanic eruptions in Iceland and Chile in-volving disruption to commercial air transport, new data sourceswere tested for measuring volcanic plumes including thermaldata (Meteo satellites, Aster) and cloud and SO2 data (IASI)

‘Exotic’ datasets

Among new research areas, two types of data stand out forgeohazard studies: time varying gravity data and ionosphericparameters. Data from the GRACE gravity mission have beenshown to provide unique information on mass redistributionfollowing large (> magnitude 8) earthquakes (Panet et al.,2007). While this limits their use somewhat in the short term,future time-varying gravity missions should provide better res-olution allowing the study of smaller events. Monitoring of thestate of the ionosphere has proven to be a new domain of inter-est for some physicists working on solid Earth processes suchas earthquakes and tsunami. While the Demeter mission aimedat finding precursor signals of seismic and volcanic activity,monitoring of the Total Electron Content of the lower iono-

Fig. 1 – Continuous wavelet analysis coefficients at 500 km scale of the geoid difference between 2005 and 2004, stacked over9 months (right panel), and of the geoid difference between 2004 and 2003, stacked over 9 months (left panel). Image adapted fromPanet et al. (2007) with the author's consent.


The contribution of space based observations to understanding and addressing geohazards

sphere using GPS tomography and airglow imaging (Makela etal., 2011) has been shown to provide a potentially interestingmirror of seismic waves generated by earthquakes. This in-cludes possible very long period events as reported recently fol-lowing the Tohoku Great Japanese earthquake and tsunami.

Future missions

Over the next decade many new earth observing satelliteswith significant interest for the solid Earth community willbe launched. These include :

Pleiades. The French component of the italo-French ORFEOconstellation (comprising of Pleiades and Cosmo-Skymedsatellites), the first Pleiades satellite is currently scheduledfor launch from Kourou in December 2011. This pair ofsatellites (Pleiades-2 scheduled for spring 2013) will ulti-mately provide a 24 hrs revisit at any point on the globe withhigh resolution panchromatic (70 cm) and multi-spectral(2.8 m) imagery. The Pleiades satellites are highly agile andwill be able to acquire up to three images of the same scenein one pass. This will open the way to the development ofhigh resolution 3-D products potentially very useful in nat-ural hazard applications [3].

Sentinel-2. The second satellite developed in the contextof the space component of Europe’s Global Monitoring forEnvironment and Security programme (GMES), Sentinel-2will provide 10 m resolution multi-spectral imagery of allland-surfaces with a temporal revisit of around 5 days at theequator. All land surfaces will be systematically acquiredand data will be made available under a ‘Free and Open’da-ta policy. Sentinel-2 will be launched in 2013 [4].

Sentinel-1. Sentinel-1 will provide continuity of C-bandSAR imagery following on from the ERS and ENVISATseries of missions. With a revist of several days once the satel-lite pair are launched, this mission will be of particular benefitfor ground displacement studies based on the SAR Interfer-ometry technique. Sentinel-1 will be launched in 2013 [5].

In addition to these missions many other private or state-owned satellite systems imaging our planet in the visible, in-frared and microwave part of the spectrum are planned forlaunch in the next 5 to 7 years. Other missions which cangreatly contribute to our understanding of geohazard process-es can also be cited: gravity field missions such as Grace Fol-low-on currently under development by USA/Germany; orthe Chinese Seismo-Electromagnetic Satellite which is afollow on to the Demeter mission developed by the ChineseEarthquake Administration.

Conclusion

This paper has presented a very brief review of data andmethods currently used in the study of geohazards. It has high-

lighted the utility of the information it is possible to extractfrom satellite data whether it be satellite imagery or more ‘ex-otic’ data types such as gravity data or ionospheric parame-ters. The information described must be used in conjunctionwith ground based monitoring and measurement systems inorder to derive the most complete and useful monitoring sys-tem possible for a given hazard area.

In the next ten to fifteen years many new satellite systemsare to be launched by diverse actors. In order to fully harnessthe potential that this data could bring to improving our col-lective understanding of various geohazard processes, it maybe useful to improve coordination between science teams inorder to define high priority areas where data should be acquiredand avoid duplicating effort in producing information products.Currently, several potential frameworks exist to achieve this in-cluding GEO geohazard ‘Supersites’ initiative internationally orthe GMES programme within Europe.

References

Klinger Y., Etchebes M., Tapponnier P., Narteau C. (2011) –Characteristic slip for five great earthquakes along the Fuyunfault in China. Nature Geoscience 4, 389-392, DOI:10.1038/ngeo 1158, 2011.

Delacourt C., Allemand P., Berthier E., Raucoules D., Casson B.,Grandjean P., Pambrun C. and Varel E. (2007) – Remote-sensing techniques for analysing landslide kinematics: a review.Bulletin de la Societe Geologique de France 178, 2, p. 89-100,DOI: 10.2113/gssgfbull.178.2.89.

Makela J.J., Lognonné P., Hébert H., Gehrels T., Rolland L., All-geyer S., Kherani A., Occhipinti O., Astafyeva E., Coisson P.,Loevenbruck A., Clévédé E., Kelley M.C., Lamouroux J.(2011) – Imaging and modeling the ionospheric airglow responseover Hawaii to the tsunami generated by the Tohoku Earthquakeof 11 March 2011. Geophysical Research Letters 38, L00G02,DOI: 10.1029/2011GL047860.

Massonnet, D., M. Rossi, C. Carmona, F. Adragna, G. Peltzer,K. Feigl and T. Rabaute (1993) – The displacement field of theLanders earthquake mapped by radar interferometry. Nature364, 138-142.

Panet I., Mikhailov V., Diament M., Pollitz F., King G., deViron O., Holschneider M., Biancale R., Lemoine J.M.(2007) – Co-seismic and post-seismic signatures of the SumatraDecember 2004 and March 2005 earthquakes in GRACE satel-lite gravity, Geophysical Journal International 171, 1, 177-190,DOI:10.1111/ j.1365- 246X.2007. 03525 .x.

[1] http://www.tectonics.caltech.edu/slip_history/spot_coseis/down-load_software.html

[2] http://www.cnes.fr/web/CNES-en/3236-pleiades.php [3] http://www.esa.int/esaLP/SEMM4T4KXMF_LPgmes_0.html [4] http://www.esa.int/esaLP/SEMBRS4KXMF_LPgmes_0.html [5] http://www.roipac.org/

Abstract: The International Charter (Space and Major Disasters)has been designed to provide a unified system of space data ac-quisition and delivery of products to those affected by natural andman-made disasters. The Charter operates on a volunteer, best-ef-forts basis by the member organizations. There are currentlymany Space Agencies and national or international space systemoperators, that are members of the Charter. The total is 40+ userorganizations from 36 countries. Each member agency has pled-ged resources to support the articles of the Charter. The goal is tomitigate the effects of disasters on human life and property. Theonly agencies authorized to activate the services of the Charter arethe authorized users. An authorized user is a civil protection, re-scue, defense or security body from the country of a Charter mem-ber. A single phone call activates the Charter and mobilizes thespace and associated ground resources of the member agencies toobtain data and information on a disaster event. Data are proces-

sed into images and maps by a Value Added Reseller, and the dataare delivered to the end user.In the United States, The U.S. Geological Survey and the Natio-nal Oceanic and Atmospheric Administration are AuthorizedUsers. NASA responds to internal requests from these two agen-cies to acquire data from its resources. ASTER is one of the ins-truments frequently asked to provide data on behalf of the USAuthorized Users.For the first 11 months of 2010, there have been 48 activations ofthe Charter. Over half of these have been for flood events, such asthe August flooding in Pakistan. Other events that have resulted inactivating the Charter include earthquakes, snow, cyclones, volca-nic eruptions, landslides and tsunamis. This presentation will des-cribe the details of how the Charter operates, will provide numerousillustrations of the types of products provided to end users, and willdiscuss ASTER’s participation in the Charter’s activations.

The International charter for disaster mitigation:Participation by ASTER project

M. Abrams*, K. Duda*

Abstract: Spatial data are central to understanding, monitoring,and responding to natural hazards events. Remote sensing playsa pivotal role in obtaining spatial data over large areas at mul-tiple time intervals, and is widely used in earthquake researchand disaster management. Satellite imagery provides an excel-lent baseline for mapping geological structures and identifyingactive faults. High-resolution multispectral imagery, LiDAR(light-detection and ranging) and InSAR (interferometric syn-thetic aperture radar) enable the precise measurement of earthsurface deformation following seismic events, as well as asso-ciated effects such as landslides, liquefaction and structural da-mage. Developing better understanding of the earth’s structure

and dynamics places us in stronger position to understand futu-re seismic risk.We utilized a broad range of passive and active remote sensingtechniques to better understand the nature of the 2010 Mw 7.1and 2011 Mw 6.3 seismic events in Canterbury, New Zealand,and monitor their associated effects. Both events were characte-rised by substantial soil liquefaction, and the 2011 Mw 6.3 eventcaused major structural damage to the city of Christchurch andtriggered landslips in the adjacent hills. Here we present our re-sults on mapping and quantifying these effects through the ob-ject-based fusion and analysis of high resolution aerial photo-graphy, satellite imagery, LiDAR and X-band SAR.

Remote sensing of earthquake effects following the 2010 Mw 7.1 and 2011 Mw 6.3 events

in Canterbury, New Zealand

S. Levick**


*NASA/Jet Propulsion Laboratory, United States.**GNS Science, New Zealand.

Introduction

Supertyphoon Durian, packing maximum sustained windsof 190 km/hr and gusts as high as 230 km/hr, made landfall inthe eastern coast of Luzon island, Philippines on 29 Novem-ber 2006. It brought 495.8 mm of rain over a period of 36 hoursto Albay province, where Mayon volcano is located (PAGASA,2006). The amount of rainfall was more than enough toremobilize volcanic debris into lahar flows as it exceededthe initiation threshold set for Mayon lahars (Rodolfo andArguden, 1991) and for debris flows worldwide (Balducci,2007). This resulted in severe loss of life and property, with1,266 people killed, including 740 that were missing and pre-sumed dead (Rabonza, 2006), and incurring PhP 1.13 billionworth of damages in agriculture, aquaculture and infrastruc-ture (PHIVOLCS, 2008).

The amount of lahar generated by this event is by far thelargest in terms of their volumes, runout lengths and devas-tative effects (Paguican et al., 2009). Only the 1825 laharevent which buried the town of Cagsawa, 11 km southeastof Mayon’s summit (Ramos-Villarta, 1985), which left1,500 people dead (Task Group for the International Deca-de of Natural Disaster Reduction, 1990), comes close to themagnitude of fatalities and destruction caused by the 2006disaster.

This study examines the lahar deposits left in the wake ofsupertyphoon Durian using high-resolution SPOT 5 image-ries. This work builds on the paper published by Paguican etal. (2009) where they described in detail and analyzed thelahar deposits left by Durian.

Methods

Pre- and post-disaster SPOT 5 imageries dated 17 Februa-ry 2003 and 12 December 2006, were provided by the Mani-la Observatory and the Center for Remote Imaging, Sensingand Processing (CRISP), through the IMAGERLERISK /STIC ASIA project, respectively. These imageries were in-terpreted to examine the extent of lahar flows that affectedthe downstream areas of Mayon volcano.

Dikes, drainage and lahar deposits were outlined manual-ly from the satellite imageries. Unsupervised classificationof the lahars was not possible because of similarities of theirappearance with flooded crop lands. Manual delineation ofthe lahars was based on their association with channels andplatform distribution.

Results

Flood prone areas characterize the lower and gentler slopesof Mayon volcano. These areas, located in between drainagechannels appear to be rice and coconut plantations. Imageriestaken in the aftermath of typhoon Durian show heavily inun-dated croplands. Because the spectral reflectance of water inflooded areas is nearly the same as those of the lahar deposits,they can easily be mistaken as the same material. The distinc-tion between the lahar fields and croplands was assessedthrough examination of the 2003 SPOT 5 imagery and groundvalidation performed by Paguican et al (2009).

The interpreted post-disaster SPOT 5 imagery shows the ra-dial distribution of channels typically terminating into fan-sha-

Understanding the fatal 2006 dike breaching of Mayon Volcano using high-resolution imageries

R. Eco*, A.M.F. Lagmay*, E. Paguican**

Short abstract: Heavy rains delivered by supertyphoon Durian from 29-30 November 2006 remobilized volcanic debris on the sou-thern and eastern slopes of Mayon Volcano, generating devastating lahars that caused severe loss of life and property in downstreamcommunities. Floods and lahars from the intense rainfall overtopped river bends, breaching six dikes through which they created newpaths, buried downstream communities in thick, widespread deposits, and caused most of the 1,266 fatalities. Using high-resolutionSPOT imageries provided by CRISP through the IMAGERLERISK/STIC ASIA project, we determine places of dike breaching and thelahar outbreak deposits. Barangays (villages) Maipon and Tandarora in Guinobatan municipality, Sua in Camalig municipality, Budiaoand Busay in Daraga municipality, Pawa and Padang in Legaspi City and San Antonio in Sto. Domingo municipality experienced heavycasualties. These are the same places where dike breaching occurred. Lessons learned from this study could be used to mitigate futurelahar hazards at Mayon Volcano.

Keywords: Mayon volcano, lahar, dike breaching, supertyphoon Durian, SPOT 5.


* National Institute of Geological Sciences, College of Science, University of the Philippines, Diliman, Quezon City 1101, Philippines.**Clermont Université, Université Blaise-Pascal, Laboratoire Magmas et Volcans, BP 10448, F-63000 Clermont-Ferrand.


R. Eco, A.M.F. Lagnay, E. Paguican

ped structures at the lower slopes of Mayon’s edifice fromabout 245 to 100 m elevation to sea level at gradients of 5.5˚-2.0˚. New channels can be observed to have formed splayingfrom existing channels. These freshly developed channels actedas alternative pathways of lahars generated on 30 November2006. Drainage along the Mabinit, Padang, Lidong and Basudchannels reached the coastline of Albay Gulf.

Large clusters of human settlements are clearly visiblearound the lower slopes of Mayon volcano, approximately8-14 km from the summit. Due to their proximity to the vol-cano, most of these villages and towns were inundated by la-hars, as can be seen in the post-disaster imagery. In addition,the manner by which the dikes were constructed suggeststhat the behavior of lahars may have been underestimated,as seen from their locations vis-a-vis the path generated bythe lahar flows.

Discussion

Field investigation by Paguican et al (2009) revealed brea-ched sections in several of the dikes around Mayon. In Sua,for example, the breach was about 20 m wide, forming achannel 50 m wide and 7 m deep. Meanwhile at Padang, theyfound that an 80 m wide and 15 m deep channel was cutthrough rice paddies and coconut groves. Based on the ti-ming of the arrival of debris flows at the communities, they

surmised that the dikes at the southern and eastern sectors ofMayon were breached at about 1400H, several hours afterthe first lahars were initiated (Paguican et al, 2009).

Based on these observations, we believe that the lahars for-med during typhoon Durian were initially confined in existing‘diked’ river channels and later breached the dikes upon swel-ling of the river. Perhaps like the erosive lahars observed inPinatubo, debris flows that overtopped the banks of channelsalong river bends, may have eroded the dikes from its outerside then back towards the river channel until that section ofthe dike was breached. The opening thus created a pathwayfor debris flows to cut through fields and overrun settle-ments. Since the timing of the lahar inundation of all the vil-lages occurred mostly at the same time in the afternoon of30 November, the extreme swelling of the rivers channelsand eventual breaching of dikes must have also occurredjust before this period.

Conclusion

With the intensity and duration of rainfall brought bysupertyphoon Durian to Albay province on 30 November2006 exceeding the initiation threshold for Mayon laharsand debris flows worldwide, volcanic sediments aroundMayon volcano were remobilized into lahars, causing seve-re damage and fatalities to communities around the volcano.

Fig. 1 – Left to right clockwise: Municipalities of Basud, Guinobatan, Camalig, Daraga, Padang, and Bongga. Stream flows prior tothe November 2006 lahar flows.

pre - 30 November 2006 stream flow post - 30 November 2006 lahar flow dyke


Understanding the fatal 2006 dike breaching of Mayou Volcano using high-resolution imageries

Clearly visible from the satellite imageries are the new path-ways generated by these lahar flows, resulting in overtoppedriver bends, breached dikes and inundated communities andplantations. It is apparent that the infrastructures constructedto protect the populace around Mayon were unable to withs-tand an extreme event such as this. In addition, this eventserves to illustrate the urgency by which concerned authori-ties must formulate better and more effective solutions tomitigate these kinds of disasters. But given the sheer num-ber of other potential hazards that threaten the peoplearound Mayon, more comprehensive actions may be nee-ded, such as the gradual movement of development andsettlement away from the slopes of Mayon volcano.

References

Arboleda R, Martinez M (1996) – 1992 lahars in the Pasig-Potre-ro River system. In: Newhall C., Punongbayan R. (eds), Fire andmud: eruptions and lahars of Mount Pinatubo. Philippine Institu-te of Volcanology and Seismology, Quezon City, p. 1045–1052.

Arguden A.T., Rodolfo K.S. (1990) – Sedimentologic and dyna-mic differences between hot and cold laharic debris flows ofMayon Volcano, Philippines. Geological Society of AmericaBulletin 102, 865-876.

Balducci V. (2007) – Rainfall thresholds for the initiation of land-slides. http://rainfallthresholds.irpi.cnr.it/credit.htm. Accessed26 September 2011.

PAGASA (2006) – Philippine Atmospheric, Geophysical and As-tronomical Services Rainfall archive. Data record.

Paguican E.M.R., Lagmay A.M.F., Rodolfo K.S., Rodolfo R.S.,Tengonciang A.M.P., Lapus M.R., Baliatan E.G., Obille E.C.Jr. (2009) – Extreme Rainfall-induced lahars and dike-brea-ching, 30 November 2006, Mayon Volcano, Philippines. Bulletinof Volcanology 71, 8, 845-857.

PHIVOLCS (2008) – The 30 November 2006 Supertyphoon Re-ming Lahars of Mayon Volcano, Philippines. http://ypws.tao-pi-lipinas.org/downloads/2008%20YP/lectures/scientific%20as-sessment.pdf. Accessed 26 September 2011.

Rabonza G (2006) – Philippines: NDCC media update-TyphoonReming (Durian). Technical Report. Office of Civil Defense.

Ramos-Villarta S., Corpuz E., Newhall C. (1985) – Eruptive histo-ry of Mayon volcano, Philippines. Philippine Journal of Volcano-logy 2, 1-35.

Rodolfo K. (1989) – Origin and early evolution of lahar channel atMabinit, Mayon Volcano, Philippines. Philippine Journal of Volca-nology 2, 1-35.

Rodolfo K., Arguden A. (1991) – Rain-lahar generation and sedi-ment-delivery systems at Mayon Volcano, Philippines. In: FisherR., Smith G. (eds), Sedimentation in volcanic settings. Society ofEconomic Paleontologists and Mineralogists Special Publication,45, p. 71-87.

Rodolfo K., Umbal J., Alonso R., Remotigue C., Melosantos M.,Salvador J., Evangelista D., Miller Y. (1996) – Two years of la-hars on the western flank of Mount Pinatubo: initiation, flow pro-cesses, deposits, and attendant geomorphic and hydraulic changes.In: Newhall C., Punongbayan R. (eds), Fire and mud: eruptionsand lahars of Mount Pinatubo. Philippine Inst Volcanol Seismol,Quezon City, p. 989-1013.

Smith G.A., Lowe D.R. (1991) – Lahars: Volcano-hydrologic eventsand deposition in the debris flow-hyperconcentrated flow conti-nuum. In: Fisher R.V., Smith G. (eds), Sedimentation in VolcanicSettings. Society of Economic Paleontologists and MineralogistsSpecial Publication, 45, p. 59-70.

Task Group for the International Decade of Natural DisasterReduction (1990) – Report. Bulletin of Volconalogical Societyof Japan Ser. 2, 35, 80-95.

Tungol N., Regalado M. (1996) – Rainfall, acoustic flow monitorrecords, and observed lahars of the Sacobia River in 1992. In:Newhall C., Punongbayan R. (eds), Fire and mud: eruptions andlahars of Mount Pinatubo. University of Washington Press,Seattle, p. 1045-1052.

Umbal J.V., Rodolfo K.S. (1996) – The 1991 lahars of southwesternMount Pinatubo. Philippines and the evolution of a lahar-dammedlake: p. 951–970. In: Newhall C., Punongbayan R. (eds), Fire andmud: eruptions and lahars of Mount Pinatubo. Philippine Instituteof Volcanology and Seismology, Quezon City, p. 1045-1052.

Van Westen C.J., Daag A.S. (2005) – Analyzing the relationshipbetween rainfall characteristics and lahar activity at MountPinatubo, Philippines. Earth Surface Processes and Landform30, 1663-1674.

IntroductionActive and dangerous volcanoes can be studied in safety

using remote sensing. This is particularly true for Semeru, anextremely active volcano in Indonesia, where the access to thesummit cone is difficult and dangerous (Solikhin et al., accept-ed). The launch of IKONOS (1 m panchromatic resolution) in1999, Quick-Bird2 in 2001 (60 cm) and SPOT5 in 2002 (2.5 m),made three very high-spatial-resolution satellite data sets avail-able for the analysis of volcanic terrains and eruptive phenome-na. However, only a limited number of images with such high-spatial-resolution have been used to examine the products ofvolcanic eruptions (Thouret et al., 2007, 2010).

The 3676 m high Semeru volcano (8°06’05”S, 112°55’E) isthe southernmost edifice of the Semeru-Tengger volcanic massif.Its ring plain of about 1790 km2, located below 400 m, supportsmore than one million people including 85,000 people in thetown of Lumajang to the east. Another 600,000 people live in ornear the city of Malang 45 km WNW of the summit. Semeru’seruptive activity has been recorded since 1818 and eruptionshave been persistent since at least 1967.

Methodology

This study is based on remote sensing of optical and thermalimagery, field observations and measurements, geologic map-ping based on air photos, computation of DEMs, and interpre-tation of landforms and structures. High-spatial resolution im-

ages (IKONOS and SPOT5) and aerial photos have beenused to analyse the structure of the Semeru volcano and mapits deposits. Geological and tectonic mapping is based ontwo DEMs and on the interpretation of air photos plus fourSPOT and IKONOS optical satellite images acquired between1996 and 2002. The satellite images used in this study are: (1)five high-spatial resolution (1 m) IKONOS satellite imagesdated 14-11-2002, 25-09-2004, 16-06-2006, 10-04-2008 and20-08-2009; (2) Two SPOT5 images, including one panchro-matic one at 2.5 m resolution dated 24-10-2003 and one mul-tispectral one at 5 m dated 26-07-2008; (3) One SPOT2 imagedated 11-05-1996 and one SPOT1 scene dated 08-08-1997,both at 10 m; (4) Three AST08 or ASTER TIR Surface Kinet-ic Temperature products at 90 m, dated 16-08-2002, 12-03-2003 and 25-09-2005. The ASTER images have been used toestimate the volume of the 2002-2003 block-and-ash flow de-posits. We have also compared two 10 m pixel images ac-quired before and after the event to describe the extent and im-pact of the 2002 block-and-ash flows.

Geology, tectonics, and the 2002-2003eruption of the Semeru volcano

The geologic map (Fig. 1) depicts the historic and present-day deposits of Semeru’s composite cone and ring plain basedon previous work (references in Thouret et al., 2007), and ourinterpretation of satellite images, aerial photos, and field ob-


Geology, tectonics, and the 2002-2003 eruption of the Semeru volcano, Indonesia: Interpreted from high-spatial resolution satellite imagery

A. Solikhin*, J.-C. Thouret**, A.J.L. Harris**, A. Gupta***, S.C. Liew****

Short abstract: We used high-spatial resolution images of Semeru volcano in Java, Indonesia in order to analyse its structures, map thedeposits, and record the effects of the 2002-2003 eruption. A structural map, based on two DEMs and four optical satellite images,encompasses four groups of faults. The Semeru composite cone is located on and buttressed against the Mahameru edifice at the headof a large scar on the SE flank that may reflect a failure plane at shallow depth. The deformation pattern of Semeru and its large scarmay be induced by flank spreading over the weak basal layer of Tertiary sediment. The last eruption took place in December 2002 -January 2003, and involved emplacement of block-and-ash flows and wet pyroclastic surges. We estimated the volume of the 2002-2003block-and-ash flow deposits to be c.5.45 x 106 m3, using the 2003 ASTER Surface Kinetic Temperature image. The paper illustrates theapplication of high-resolution satellite images in interpreting volcanic structures and eruption impacts.

Keywords: Semeru volcano, remote sensing, 2003 eruption, geology, tectonics.

*Center for Volcanology and Geological Hazard Mitigation, Bandung, Java, Indonesia.**PRES Clermont,Université Blaise Pascal, Laboratoire Magmas et Volcans, UMR6524 CNRS and IRD-UR163, 5 rue Kessler, 63038, Clermont-Ferrand cedex, France.***School of Earth and Environmental Sciences, University of Wollongong, Australia. ****CRISP, National University of Singapore, Singapore.


A. Solikhin, J.-C. Thouret, A. Gupta, A.J.L. Harris et S.C. Liew

servations. The prehistoricactivity of Semeru is poorlyknown, the oldest reportederuption dating back to 1818.Since 1967, cycles of domegrowth and collapse have oc-curred every 5 to 7 years in theJonggring-Seloko crater, feed-ing pyroclastic flows that hadtravelled up to 12 km: firstalong the SE-trending scar andthen following the drainagenetwork formed by the Kembar-Bang and Koboan-Lengkongrivers and their tributaries to-wards the southeast.

The structural map (Fig. 2)encompasses four groups offaults. Three of these extendN40, N160 and N75. Thefourth is a group of faultstrending N105 to N140, withprevailing N120 faults. Thefirst and second fault groupscan be related to the com-pressive stress regime in-duced by the motion of theIndia-Australian plate beingsubducted towards the northbeneath the Sunda plate andthe third fault group to the ex-tension of the Madura Basinnorth of the Bromo-Tenggercaldera. The last fault group ispossibly related to crustalspreading caused by mag-matic intrusions from earli-er volcanoes such as Jam-bangan, whose growth pre-ceded the formation of theTengger caldera and Se-meru composite cone. Se-meru’s principal structuralfeatures may be due to the tec-tonic setting of the volcano.Conspicuous structures suchas the SE-trending horseshoe-shaped scar on Semeru’s sum-mit cone coincide with theN160-trending faults. The Se-meru composite cone whichhosts the currently activeJonggring-Seloko vent is lo-cated on and buttressed against the Mahameru edifice at thehead of a large scar that may reflect a failure plane at shal-low depth. Dipping 35° towards the southeast, the failureplane may correspond to a weak basal layer of weathered

volcaniclastic rocks of Tertiary age. We suggest that the de-formation pattern of the Semeru and its large scar may be in-duced by flank spreading over the weak basal layer of thevolcano. It is therefore necessary to consider the potential

Fig. 1 – Geological map of the Semeru’s composite cone and ring plain based on previous basedon previous works (references in Thouret et al., 2007) and on our interpretation of satellite images,aerial photos, and on field observations.

Fig. 2 – A. Structural map of the Semeru-Tengger volcanic massif, inferred from SRTM-DEM, TOPO-DEM and optical satellite images. B. Rose diagram of faults showing four groups of faults F1 to F4.C. The sketch diagram depicts the regional tectonic setting around Semeru.


Geology, tectonics and the 2002-2003 eruption of the Semeru Volcano, Indonesia

for flank and / or summit collapse inthe future.

The last eruption took place inDecember 2002-January 2003, andinvolved emplacement of block-and-ash flows. On 30 December 2002,pyroclastic flows travelled 12 kmdown the Kembar and Bang valleys.Based on the 2003 ASTER SurfaceKinetic Temperature image (Fig. 3),the volume of the 2002-2003 block-and-ash flow deposits is aboutc.5.45 x 106 m3. A wet, fine ash-richpyroclastic surge escaped from oneof the valley-confined block-and ashflows at a distance of 5 to 8 km fromthe crater and swept across the forestand tilled land on the southwesternside of the Bang river valley. Thetemperature of the pyroclastic surgedecreased along the valley, and amud-rich deposit coated the banksof the Bang River.

Conclusion

We make the following summary statements. (1) The high-spatial resolution imagery enabled us to safely study a per-sistently active and dangerous composite volcano, the Se-meru. (2) Four groups of faults trending N40, N160, N75and N120 occur on the Semeru-Tengger volcanic massif. (3)Structures visible on the Semeru’s summit cone may be re-lated to the regional tectonic setting. (4) The Semeru com-posite cone has been built on and is buttressed against theMahameru edifice. Some structures such as summit normalfaults, and thrust faults at the base of the southw0est to eastflank of the Semeru indicate an asymmetric deformationpattern possibly induced by flank spreading of the weak-cored volcano. (5) An example of the hazards posed the Se-meru is the 2002 block-and-ash flows with a volume of5.45x106 m3 that caused the evacuation of 500 people anddamaged the forest and tilled land on the west side of theBang valley. (6) Hazard mitigation at Semeru should includecontinuous monitoring of the eruptive activity through an

early-warning system and continuous remote sensing of themorphological changes in the drainage system due to theimpact of frequent pyroclastic flows and lahars.

References

Solikhin A., Thouret J.-C., Gupta A., Harris A.J.L., Liew S.C.,(2012) – Geology, Tectonics, and the 2002-2003 Eruption of theSemeru Volcano, Indonesia: Interpreted from High-Spatial Resolu-tion Satellite Imagery. Geomorphology 138, 364-379.

Thouret J.-C., Lavigne F., Suwa H., Sukatja B., Surono (2007)– Volcanic hazards at Mount Semeru, East Java (Indonesia), withemphasis on lahars. Bulletin of Volcanology 70, 221-244.

Thouret J.-C., Gupta A., Lube G., Cronin S.J., Surono (2010) –Analysis of the 2006 eruption deposits of Merapi Volcano, Java, In-donesia, using high-resolution IKONOS images and complemen-tary ground based observations. Remote Sensing of Environment114, 1949-1967, DOI:10.1016/j.rse.2010.03.016.

Fig. 3 – A. The ASTER TIR surface kinetic temperature image taken on 12 March 2003 witha blue-red color code shows thermal anomalies induced by the 2002 block-and-ash flowson Semeru along the SE-trending scar and into the Bang river valley as far as Supit, a sub-urb of Pronojiwo. B. and C. Three-D surface view of 16 August 2002 and 12 March 2003ASTER TIR surface kinetic temperature images on TOPO-DEM, showing thermal anomalieson Semeru before and after 29 December 2002.

Abstract: Located in Central Java, Indonesia, Merapi Volcano is one of the most active volcanoes in the world. More than eighty erup-tions have been identified since the 16th century, of which a dozen caused casualties. In October and November 2010, ashfall, pyro-clastic surges and a series of pyroclastic flows swept the flanks of the entire volcano, causing 353 fatalities, significant environmentaland infrastructural damage. During the year preceding this eruption, a field census was conducted on all slopes of Merapi volcano in theframe of the MIA VITA FP7 European project. Data were acquired at the level of the hamlet (dusun) where no official statistics exist.Therefore a series of maps, including the spatial distribution of population, have been drawn and further integrated into a GIS. In a sec-ond step of the study, the areas affected by the 2010 volcanic flows and surges were mapped in the field and on high-resolution geospa-tial images (SPOT5, WorldView 2 from Digital Globe acquired on the aftermath of the main explosion. The number of people who couldhave been killed in case of a failure of the evacuation has been estimated for each type of volcanic hazards (i.e. pyroclastic flows, py-roclastic surges, and ashfall) through the cross-analysis of satellite remote-sensing data and field data of population distribution. Thedamaged buildings were automatically extracted from the images using the new Feature Extraction Module of ENVI 4.8., and also in-tegrated into the GIS. An attribute table displays a mean number of people in each house before the eruption, based on field data ex-trapolation acquired at the level of the hamlet. Hence it is possible to compare the level of houses damage with the number of casual-ties and assess the success of crisis mitigation with the phased implementation of a safety evacuation zone that concerned an unprece-dented large population of about 1.5 million.


Satellite remote-sensing analysis of casualties anddamage from the 2010 eruption of Merapi volcano

F. Lavigne*

Risk microzonation of Yogyakarta city following the 2010 eruption of Merapi volcano

D.S. Hadmoko**, L.W. Santosa**, M.A. Marfai**, F. Lavigne**

Abstract: The 2010 Eruption of Merapi volcano expelled 150 million cubic meters of pyroclastic materials. The remobilization of thesematerials in Code River following the 2010 eruption would constitute a significant hazard to critical infrastructure within the urban areaof Yogyakarta located 25 km south of the Merapi crater. A series of lahars of hyper concentrated-flows type have already causedsignificant sediment aggradation in the Code river channel that decreases the river capacity to transport future lahars. The risk of laharis enhanced due to the high population density along the river, the expansion of settlements and infrastructures, as well as the highvulnerability of the people. Therefore, risk mapping is fundamental at the local scale, in minimizing both loss of life and damage toproperty. In this on-going research, we attempt to: (1) assess the lahar and flood hazard through several scenarios of lahar discharge; (2)to identify, map and quantify the elements at risk (e.g., settlements and vital infrastructures) through remote sensing; (3) to assess thevulnerability of people and properties along the river; and (4) to draw a risk map of the threatened area. Collected data have proven tobe extremely useful for identifying the element at risk along the Code River. The expected result of this research will be useful for laharrisk mitigation, providing a basis for emergency plans.

* Université Paris 1 Panthéon-Sorbonne, Laboratoire de Géographie Physique, France.** Faculty of Geography, Gadjah Mada University, Indonesia.

Introduction

The Smithsonian ‘Volcanoes of the World’Catalogue listsover 400 terrestrial volcanoes in the Asia-Pacific Region with200 having erupted historically (Siebert and Simkin 2002).Many of these volcanoes are situated in areasof high-population density leading to increasedlevels of risk. Jenkins et al. (in press-b) estimat-ed 2 billion people to be living within 1000 kmof 190 potentially active volcanoes in the Re-gion (Fig. 1).

Of more recent volcanic events affecting theRegion, the eruption of Merapi in Indonesia thatbegan in late October 2010 killed more than 350people, most as the result of pyroclastic densitycurrents. In January 2011, ash from Shinmoe-dake severely impacted agricultural communi-ties in Miyazaki prefecture, Japan (Magill andOkada 2011). Disruptions from the May 2010Eyjafjallajökull eruption in Iceland were felt asfar away as Asia and Australasia when flightswere cancelled due to closure of Europeanairspace. The total impact on global GDP esti-mated by Oxford Economics (2010) was US$5billion with $517 million within Asia. As a fur-ther blow to the airline industry, the Puyehue-Cordón Caulle eruption, Chile, disrupted flightswithin South America, South Africa, Australiaand New Zealand throughout June 2011.

Risk is a function of hazard, exposure and vulnerability;where hazard ideally considers the magnitude, footprint, prob-ability, timing and duration of potential events. To improveour understanding of hazard and risk, statistical analysis andmodelling can build upon geological, engineering and soci-


Quantifying volcanic hazard and risk

C. Magill*

Short abstract: Volcanic risk can be thought of as a function of hazard, exposure and vulnerability; where hazard ideally considers themagnitude, footprint, probability, timing and duration of potential events. Modelling volcanic hazard and risk poses many unique chal-lenges – eruption duration may be extended over months or even years, eruptive volumes for a particular volcano may vary by manyorders of magnitude, style may range from effusive to highly explosive, multiple and secondary hazards may occur, and short eruptionhistories mean that identifying patterns in activity can be difficult. Statistical studies and recent advances is stochastic modelling haveallowed many of these challenges to be addressed. Probabilistic assessments of risk incorporate stochastic simulations of volcanic pro-cesses, spatial exposure information and vulnerability functions relating hazard intensity to likely impacts. High resolution modelling,integration with virtual globes, time-dependent eruption probability calculations and multidisciplinary studies are all contributing toimprove assessment of volcanic hazard and risk.

Keywords: loss modelling, risk assessmet, stochastic modelling, volcanic hazard, volcanic risk.

Fig. 1 – Volcano locations in the Asia-Pacific region,overlain with a histogram of population countswithin 1000 km of each volcano and within 5 de-gree latitude bands. Figure from Jenkins et al. (inpress-b).

*Risk Frontiers, Macquarie University, NSW 2109, Australia.


C. Magill

ological investigations. Quantifying risk by including prob-ability estimates for events of various magnitudes can assistdisaster planning and allow the benefits of various prepara-tion measures and strategies to be analysed.

Challenges

Modelling volcanic hazard poses many challenges. In par-ticular, and in contrast to other natural hazards, volcaniceruptions may continue for many months or even years(Jenkins et al., 2007; Siebert et al., 2010). Soufrière Hills inMontserrat has now been erupting for 15 years, with contin-ued dome growth and collapse, pyroclastic flow and laharevents (Sparks and Young, 2002; Wadge et al., 2010). The2000 eruption of Miyake-jima Island, Japan, forced theevacuation of 3845 people, the entire population of the is-land. Not until 2005 when gas levels returned to safer lev-els were residents allowed to return permanently and the re-covery could begin (Miyake Village 2008). Hazards can con-tinue even after the end of eruptive activity. For example,151 train passengers were killed at Tangiwai, New Zealand,in December 1953 after the destruction of a bridge by a laharfrom Mount Ruapehu (Graettinger et al., 2010) even thoughRuapehu had not erupted since 1952.

In addition to event duration, it must be borne in mind thateruptive volumes for a particular volcano may vary by manyorders of magnitude and the style of activity may range fromeffusive to highly explosive. Eruptions from the Taupo Vol-canic Centre, for example, have varied in bulk volume be-tween 0.1 and >45 km3 since the c. 26.5 ka Oruanui eruptionwhich erupted c. 400 km3 of magma (Sutton et al., 2000).

Multiple hazards are likely to occur throughout an event,each with varying magnitudes, extents, durations and conse-quences. Hazards may range from small ash and gasemissions through to destructive pyroclastic currents. Disas-ter planning for volcanoes such as Mount Fuji, Japan, mustconsider hazards including tephra-falls, lava flows, debrisavalanches, lahars and pyroclastic flows (Cabinet Office2004). Secondary hazards such as tsunami (Choi et al.,2003; Unzen Restoration Office 2002), or climatic coolingfrom sulphuric acid aerosols (Self et al. 1981; Zielinski etal., 1994), may also occur during some eruptions.

Determining probabilities for future volcanic events, aswell as assessing likely eruption styles, hazards and magni-tudes, relies largely on evidence from past eruptions fromthe volcano in question. For volcanic systems where an ex-tended eruption history is known, there is often considerableevidence for cyclicity and/or temporal clustering on varyingtimescales, which needs to be considered. However, reposeperiods for some volcanoes may be in the 100s or 1000s ofyears, exceeding human history and with little evidenceeven in the geological record, meaning that these patternsare difficult to identify.

All these possibilities complicate the calculation of vol-canic hazard and therefore of risk. We cannot assume thatfuture events will be the same as in the past and therefore,although useful for detailed planning exercises, it is not ad-equate to rely on deterministic hazard simulations. Studies

must consider the many combinations of hazards, phases andduration rather than resting upon deterministic analysis ofdiscrete event scenarios. We must consider the evolution ofeach volcanic system and potential time-dependent changesin magnitude, style and eruption probability.

Modelling approaches

Statistical studies and recent advances is stochasticmodelling have allowed many of these challenges to beaddressed. An event tree methodology was developed byNewhall and Hoblitt (2002) where each branch was as-signed a probability and led from a general to more specificoutcome, e.g. from magmatic intrusion to magmatic erup-tion to VEI level and then the probability of pyroclasticflow. Neri et al. (2008) built upon this for Vesuvius byadding probability distributions to each branch based on theresults of expert elicitation. Marzocchi et al. (2008) usedevent tree principles to develop the software package BET-EF, which combines all available volcanological data to es-timate probabilities for risk assessment and planning pur-poses.

Hazard assessments involving monogenetic fields consid-er both temporal and spatial eruption probability. Proba-bilistic assessments of vent locations for the Auckland Vol-canic Field, centred on Auckland City, New Zealand, havebeen carried out using cluster analysis (Magill et al., 2005)and by incorporating improved tephra dating information,correlation to source volcanoes and a spatial density kernel(Bebbington and Cronin 2011). A number of similar assess-ments involving kernel smoothing have been carried out inrespect to the siting of nuclear facilities (e.g. Connor andHill 1995; Connor et al., 2000; Martin et al., 2004; Weller etal., 2006), were annual probability calculations are criticaland values of up to 10-8 can be considered high.

Stochastic modelling of volcanic processes takes into ac-count a volcano’s eruptive history, trends in recent activity,data from analogous volcanoes and aleatory variability. Inthe case of modelling volcanic ash dispersal (e.g. Bonadon-na 2006; Hurst and Smith 2004; Magill et al., 2006), distri-butions are created describing eruption probability, eruptivevolume, eruption column height, particle size and windspeed and direction with height. Each event may be mod-elled as a series of phases in order that a probabilistic as-sessment of event duration can also be made (Bonadonna etal., 2005; Jenkins et al., 2008). Tens of thousands of simu-lations are performed, also taking into account environmen-tal conditions, such as weather and topography, allowing theprobability of exceeding various ash fall thicknesses or loadsto be calculated for any location. Similar probabilistic simu-lations have been carried out for other volcanic hazards, in-cluding lava flows (e.g. Favalli et al., 2009; Felpeto et al.,2001; Wadge et al., 1994) and, although computation de-mands are intensive, pyroclastic density currents (Rossano etal., 2004).

Probabilistic simulation techniques have been applied tomultiple volcanoes to obtain regional hazard assessments(Magill et al., 2006; Volentik et al., 2009). On a larger scale,


Quantifying Volcanic Hazard and Risk

1000 tephra-dispersal simulations were carried out for eachpotentially dangerous volcano in the Asia-Pacific region;where tephra thickness exceedance estimates were com-bined with population data to obtain population-weightedhazard maps for the region (Jenkins et al. in press-a; Jenkinset al. in press-b).

Probabilistic methods can also be used to calculate volcanicrisk and to estimate the likely damage and disruption arisingfrom future volcanic events. Probability density distributionsare created that relate mean damage, cost, resources, etc. tohazard magnitude, with uncertainty given by the spread ofvalues about the mean. These distributions are combined withhazard layers and population, building and land-use informa-tion to provide a spatial distribution of risk.

Volcanic risk models have been developed for the NorthIsland, New Zealand, and the Greater Tokyo Region, Japan (seehttp://www.riskfrontiers.com/volcNZ.html and www.riskfron-tiers.com/kazanrisk.htm). These models combine probabilistictephra dispersal simulations with exposure information andvulnerability functions to calculate loss exceedance statis-tics. In the case of the KazanRisk model developed forGreater Tokyo (Fig. 2), for each of 60,000 simulations, weare able to calculate the total amount of tephra falling onvarious land-use types including residential, non-residential,

agriculture, forestry and roads. Vulnerability functions havebeen developed for different building types allowing us toestimate damage and loss for each simulated event. Func-tions have also been created that relate the volume and massloading of tephra to the costs and resources needed forclean-up activities and seasonally-dependent losses to agri-cultural/horticultural production. Hazard and risk resultsmay be displayed by plotting against Average RecurrenceIntervals (ARIs) or, alternatively, mapped at a 1 km meshlevel. KML files are created for given impact thresholds andARIs, which can then be displayed on a virtual globe.

Discussion and future trends

Probabilistic hazard and risk (or loss) results can benefitplanning by governments, business owners, transport organ-isations, farmers and emergency managers. Results such astephra load, building loss or clean-up time for a given loca-tion can be plotted against conditional or annual probability,or alternatively, ARI. Probabilities of exceeding varioushazard or risk thresholds can also be mapped. Future mod-elling will incorporate a layered approached that allowsexpansion and improvements to be made as more informa-tion becomes available and requirements expand.

Fig. 2 – Probabilistic hazard results for the Greater Tokyo Region generated by KazanRisk and displayed in a Google Earth frame-work.


C. Magill

As observational data, computing resources and simula-tion techniques improve, hazard simulations will become in-creasing more complex. High resolution modelling is al-ready being carried out for tephra dispersal (e.g. Folch et al.,in press) and pyroclastic density currents (e.g. Esposti On-garo et al., 2008). With time, probabilistic techniques willbe able to be applied to simulations such as these, allowingthe full range of eruption possibilities to be considered ingreater detail. Visualisation techniques are also becomingmore sophisticated with simulations increasingly displayedon virtual globes (e.g. Webley 2011). Both 2- and 3-dimen-sional simulations displayed in this way allow underlyingexposure, including critical infrastructure and transportroutes, to be easily identified and have the potential to playan important role in education and disaster planning.

The timing of volcanic events cannot be thought of as asimple Poisson or random process and probability calcula-tions, where possible, need to consider clustering and cyclic-ity within volcanic systems. Estimates of eruption probabili-ty are best considered to be time-dependent, with the lengthof time since the last eruption being important (Bebbington2010; Turner et al. 2008). Geological, geochemical and dat-ing studies all add to the reliability of these estimates.

A multidisciplinary approach will increasingly be appliedto risk assessment. Engineering (e.g. Baxter et al., 2005;Spence et al., 2004; Wardman et al., in press) and sociologi-cal (Gaillard 2008; Jenkins and Haynes 2011) studies will beincorporated to better understand the physical and social im-pacts to communities from future events. Improving the com-munication of risk assessment studies to communities and de-cision makers is a fundamental step in the successful imple-mentation of mitigation measures and increased resilience(Barclay et al., 2008; Cronin et al., 2004; Haynes et al., 2007).However, an important contribution is increasingly providedby those scientists investigating the underlying social vulner-abilities that increase disaster risks (Wilson 2009; Wisner etal. 2003). This includes the incorporation of social develop-ment approaches which actively reduce people’s exposure tohazards, and increase their capacity to anticipate, cope withand recover from volcanic impacts.

The potential extended nature of volcanic events, large rangeof eruptive volumes and styles, multiple hazards and limitedevidence due to short eruption histories, all contribute to theuncertainties of any volcanic hazard and risk estimation. Manyof these uncertainties can be addressed though statistical anal-ysis and probabilistic modelling, fields of research that aregrowing rapidly and benefiting from contributions frommany fields of science.

References

Barclay J., Haynes K., Mitchell T., Darnell A. (2008) – Framingvolcanic risk communication within disaster risk reduction:Finding ways for the social and physical sciences to work to-gether. In: Liverman D.G.E., Pereira C.P.G., Marker B. (eds),Communicating Environmental Geoscience. Geological SocietySpecial Publication, 305, Geological Society, London, 214 p.

Baxter P.J., Boyle R., Cole P., Neri A., Spence R., Zuccaro G.(2005) – The impacts of pyroclastic surges on buildings at theeruption of the Soufrière Hills volcano, Montserrat. Bulletin ofVolcanology 67(4), 292-313.

Bebbington M., Cronin S.J. (2011) – Spatio-temporal hazard esti-mation in the Auckland Volcanic Field, New Zealand, with anew event-order model. Bulletin of Volcanology 73(1), 55-72.

Bebbington M.S. (2010) – Trends and clustering in the onsets of vol-canic eruptions. Journal of Geophysical Research 115 (B01203),DOI: 10.1029/2009JB006581.

Bonadonna C. (2006) – Probabilistic modelling of tephra disper-sion. In: Mader H.M., Coles S.G., Connor C.B., Connor L.J.(eds), Statistics in Volcanology. IAVCEI Publications, Geologi-cal Society of London, p. 243-259.

Bonadonna C., Connor C.B., Houghton B.F., Connor L.J.,Byrne M., Laing A., Hinks T. (2005) – Probabilistic modelingof tephra dispersion: hazard assessment for a multi-phase erup-tion at Tarawera, New Zealand. Journal of Geophysical Re-search 11(B3), B0320310.0321029/ 0322003JB0002896.

Cabinet Office (2004) – Report of the committee for hazard mapsof Fuji Volcano. Cabinet Office, Tokyo, 240 p.

Choi B.H., Pelinovsky E., Kim K.O., Lee J.S. (2003) – Simula-tion of the trans-oceanic tsunami propagation due to the 1883Krakatau volcanic eruption. Natural Hazards and Earth SystemSciences 3(5), 321-332.

Connor C.B., Hill B.E. (1995) – Three nonhomogeneous Poissonmodels for the probability of basaltic volcanism: Application tothe Yucca Mountain region, Nevada. Journal of GeophysicalResearch,100(B6), 10107-10125.

Connor C.B., Stamatakos J.A., Ferrill D.A., Hill B.E., OfoegbuG.I., Conway F.M., Sagar B., Trapp J. (2000) – Geologic fac-tors controlling patterns of small-volume basaltic volcanism:Application to a volcanic hazards assessment at Yucca Mountain,Nevada. Journal of Geophysical Research 105(1), 417-432.

Cronin S.J., Gaylord D.R., Charley P., Wallez S., AllowayB.V., Esau J.W. (2004) – Participatory methods of incorporat-ing scientific with traditional knowledge for volcanic hazard as-sessment on Ambae Island, Vanuatu. Bulletin of Volcanology 66,652-668.

Esposti Ongaro T., Neri A., Menconi G., de’Michieli Vitturi P.,Marianelle P., Cavazzoni C., Erbacci G., Baxter P.J. (2008) –Transient 3D numerical simulations of column collapse andpyroclastic density current scenarios at Vesuvius. Journal of Vol-canology and Geothermal Research 178(3), 378-396.

Favalli M., Chirico G.D., Papale P., Pareschi M.T., Boschi E.(2009) – Lava flow hazard at Nyiragongo volcano, D.R.C., 1.Model calibration and hazard mapping. Bulletin of Volcanology71(4), 363-374.

Felpeto A., Araña V., Ortiz M., Astiz M., García A. (2001) –Assessment and modelling of lava flow hazard on Lanzarote(Canary Islands). Natural Hazards 23(2-3), 247-257.

Folch A., Costa A., Basart S. (in press) – Validation of theFALL3D ash dispersion model using observations of the 2010Eyjafjallajökull volcanic ash clouds. Atmospheric Environment.

Gaillard J.-C. (2008) – Alternative paragrams of volcanic riskperception: The case of Mt. Pinatubo in the Philippines. Journalof Volcanology and Geothermal Research 172, 315-328.


Quantifying Volcanic Hazard and Risk

Graettinger A.H., Manville V., Briggs R.M. (2010) – Deposi-tional record of historic lahars in the upper Whangaehu Valley,Mt. Ruapehu, New Zealand: implications for trigger mecha-nisms, flow dynamics and lahar hazards. Bulletin of Volcanology72, 279-296.

Haynes K., Barclay J., Pidgeon N.F. (2007) – An evaluation ofvolcanic hazard maps as a communication tool on the CaribbeanIsland of Montserrat. Bulletin of Volcanology 70(2), 123-138.

Hurst A.W., Smith W. (2004) – A Monte Carlo methodology formodelling ashfall hazards. Journal of Volcanology and Geother-mal Research 138(3-4), 393-403.

Jenkins S.F., Haynes K. (2011) – Volcanic eruption. In: Wisner B.,Gaillard J.-C., Kelman I. (eds), Handbook of hazards and disas-ter risk reduction. Routledge, 872 p.

Jenkins S.F., Magill C.R., McAneney K.J. (2007) – Multi-stagevolcanic events: A statistical investigation. Journal of Volcanol-ogy and Geothermal Research 161, 275-288.

Jenkins S.F., Magill C.R., McAneney K.J., Blong R.J. (in press-a) – Regional ash fall hazard I: A probabilistic assessmentmethodology. Bulletin of Volcanology.

Jenkins S.F., Magill C.R., McAneney K.J., Hurst A.W. (2008) –Multistage volcanic events: Tephra hazard simulations for theOkataina Volcanic Center, New Zealand. Journal of GeophysicalResearch 113(F04012), DOI: 10.1029/2007JF000787.

Jenkins S.F., McAneney K.J., Magill C.R., Blong R.J. (in press-b) – Regional ash fall hazard II. Asia-Pacific modelling resultsand implications. Bulletin of Volcanology.

Magill C.R., Hurst A.W., Hunter L.J., Blong R.J. (2006) – Prob-abilistic tephra fall simulation for that Auckland Region, NewZealand. Journal of Volcanology and Geothermal Research 153,370-386.

Magill C.R., McAneney K.J., Smith I.E.M. (2005) – Probabilis-tic assessment of vent locations for the next Auckland VolcanicField event. Mathematical Geology 37(3), 227-242.

Magill C.R., Okada T. (2011) – Preliminary investigation ofthe impacts of the 2011 Shinmoedake eruption: January29th to February 3rd. Risk Frontiers, Macquarie Universi-ty, 31 p.

Martin A.J., Umeda K., Connor C.B., Weller J.N., Zhao D.,Takahashi M. (2004) – Modeling long-term volcanic hazardsthrough Bayesian inference: An example from the Tohoku vol-canic arc, Japan. Journal of Geophysical Research 109 (B10208),DOI: 10.1029/2004JB003201

Marzocchi W., Sandri L., Selva J. (2008) – BET_EF: a proba-bilistic tool for long- and short-term eruption forecasting. Bul-letin of Volcanology 70, 623-632.

Miyake Village (2008) – Volcanic eruptions on Miyakejima Islandin 2000: the year 2000 record of disastrous eruption MiyakejimaIsland/overview, 29 p.

Neri A., Aspinall W.P., Cioni R., Bertagnini A., Baxter P.J.,Zuccaro G., Andronico D., Barsotti S., Cole P.D., EspostiOngaro T., Hinks T.K., Macedonio G., Papale P., Rosi M.,Santacroce R., Woo G. (2008) – Developing an event tree forprobabilistic hazard and risk assessment at Vesuvius. Journal ofVolcanology and Geothermal Research 178(3), 397-415.

Newhall C.G., Hoblitt R.P. (2002) – Constructing event trees forvolcanic crises. Bulletin of Volcanology 64(1), 3-30.

Oxford Economics (2010) –The economic impacts of air travelrestrictions due to volcanic ash: A report prepared for Airbus.Oxford Economics, 12 p.

Rossano S., Mastrolorenzo G., De Natale G. (2004) – Numericalsimulation of pyroclastic density currents on Campi Flegreitopography: a tool for statistical hazard estimation. Journal ofVolcanology and Geothermal Research 131(1), 1-14.

Self S., Rampino M.R., Barbera J.J. (1981) – The possibleeffects of large 19th and 20th Century volcanic eruptions onzonal and hemispheric surface temperatures. Journal of Vol-canology and Geothermal Research 11(1), 41-60.

Siebert L., Simkin T. (2002) – Volcanoes of the World: an Illustrat-ed Catalogue of Holocene Volcanoes and their Eruptions. In:Smithsonian Institution, Global Volcanism Program Digital Infor-mation Series, GVP-3 (http://www.volcano.si.edu /world/)

Siebert L., Simkin T., Kimberly P. (2010) – Volcanoes of theworld: third edition. Smithsonian Institution, 551 p.

Sparks R.S.J., Young S.R. (2002) –The eruption of SoufrièreHills Volcano, Montserrat (1995-1999): overview of scientificresults. Geological Society London, Memoirs, 21, 45-69.

Spence R.J.S., Zuccaro G., Petrazzuoli S., Baxter P.J. (2004) –Resistance of buildings to pyroclastic flows: Analytical andexperimental studies and their application to Vesuvius. NaturalHazards Review 5(1), 48-59.

Sutton A.N., Blake S., Wilson C.J.N., Charlier B.L.A. (2000) –Late Quaternary evolution of a hyperactive rhyolite magmaticsystem: Taupo volcanic centre, New Zealand. Journal of theGeological Society, London, 157(3), 537-552.

Turner M.B., Cronin S.J., Bebbington M.S., Platz T. (2008) –Developing probabilistic eruption forecasts for dormant volca-noes: a case study from Mt Taranaki, New Zealand. Bulletin ofVolcanology 70(4), 507-515.

Unzen Restoration Office (2002) – The catastrophe in Shimabara:the 1791-92 eruption of Unzen-Fugendake and the sector col-lapse of Mayu-Yama. Ministry of Land, Infrastructure and Trans-port in Japan, 20 p.

Volentik A.C.M., Connor C.B., Connor L.J., Bonadonna C. (2009)– Aspects of volcanic hazard assessment for the Bataan nuclear powerplant, Luzon Peninsula, Philippines. In: Connor C.B., Chapman N.A.,Connor L.J. (eds), Volcanic and tectonic hazard assessment for nucle-ar facilities. Cambridge University Press, p. 326-345.

Wadge G., Herd R., Ryan G., Calder E.S., Komorowski J.C.(2010) – Lava production at Soufrière Hills Volcano, Montser-rat: 1995-2009. Geophysical Research Letters 37(L00E03), doi,10.1029/ 2009GL041466.

Wadge G., Young P.A.V., McKendrick I.J. (1994) – Mappinglava flow hazards using computer simulation. Journal of Geo-physical Research 99(B1):489-504.

Wardman J.B., Wilson T.M., Bodger P.S., Cole J.W., JohnstonD.M. (in press) – Investigating the electrical conductivity of vol-canic ash and its effect on HV power systems. Physics and Chem-istry of the Earth.

Webley P.W. (2011) – Virtual Globe visualization of ash-aviationencounters, with the special case of the 1989 Redoubt-KML in-cident. Computers and Geosciences 37(1), 25-37.

Weller J.N., Martin A.J., Connor C.B., Connor L.J., Karakha-nian A. (2006) – Modelling the spatial distribution of volcanoes:


C. Magill

an example from Armenia. In: Mader H.M., Coles S.G., Con-nor C.B., Connor L.J. (eds), Statistics in Volcanology. IAVCEIspecial publication, p. 77-87.

Wilson T.M. (2009) – Vulnerability of pastoral farming systemsto volcanic ashfall hazards. PhD thesis submitted to the Natu-ral Hazard Research Centre, University of Canterbury, NewZealand, 261 p.

Wisner B., Blaikie P.M., Davis I., Cannon T. (2003) – At Risk: Natu-ral Hazards, People’s Vulnerability and Disasters. Routledge, 464 p.

Zielinski G.A., Mayewski P.A., Meeker L.D., Whitlow S.,Twickler M.S., Morrison M., Meese D.A., Gow A.J., AlleyR.B. (1994) – Record of volcanism since 7000 B.C. from theGISP2 Greenland ice core and implications for the volcano-cli-mate system. Science 264(5161), 948-952.

Absttract: Monogenetic volcanic fields are common on the Earth’s surface and typically consist of basaltic lava flows which can range in lengthfrom tens of metres to up to ~160 km (e.g. Undara in Australia). Lava flows are most commonly derived from lava spatter cones and scoria cones.Tuff rings, tuff cones and maars can also produce sizeable lava flows if the lava volume is large enough to spill out of the basin. Lava flow sizeand behaviour depend primarily on (i) effusion rate, (ii) geochemistry, (iii) magma volume and (iv) topography. Lava flows can have a signifi-cant and long lasting impact on the anthropogenic and natural environment, but rarely cause loss of human life.The Auckland Volcanic Field (AVF) is comprised of at least 50 monogenetic volcanoes in the form of scoria cones, phreatomagmatic tuffrings and maars. The AVF’s ca. 360 km2 area coincides with Auckland City, hosting a population of 1.4 million. The volcanism appears tohave peaked in frequency around 30 ka before present, forming most of the central part of the AVF. The most recent eruption (600–550 yearsago) took place in the northern part of the field and formed a basaltic shield volcano, Rangitoto. While the eruption of Rangitoto produceda distinctive shield cone in which lava is distributed radially from the vents with no preferred flow orientation or direction, elsewhere in theAVF lava flows seem to follow paleo-topography, forming single longitudinal flow lobes.The majority of the volcanic hazard scenarios for the AVF have focused on the potential of a sustained phase of explosive eruption (e.g.considering an already formed scoria cone), while the style of the onset of the eruption (e.g. phreatomagmatic maar-forming) or theeffect of potential (long-lived) lava flow effusion have thus far not been the focus of detailed study. While there is no doubt that initialphreatomagmatic explosive phases and related phenomena (e.g. base surge) represent the major hazard to life and infrastructure, effu-sive lava flows may actually represent a more enduring hazard and impact, impeding recovery from eruption-induced disasters. Fromthe known 50 monogenetic volcanoes, 15 individual lava flows have been recognized by geological mapping.In this study, we examine the main morphometric characteristics of the lava flows of the AVF and calculate the maximum and mean length,maximum and mean thickness, areal extent and volume in order to estimate the size of an average lava flow. The length and area have beencalculated from geological maps on a vector basis. Based on the DEVORA Borehole Database (link: http://pet.gns.cri.nz/), we reconstructthe bottom surface of the lava flows in order to estimate more reliably the lava flow volumes. An average AVF lava flow is characterized by~1,500 m length, ~ 19 m thickness and occupies an area ~5.2 km2. The average volume is ~0.1 km3.Remote sensing data (e.g. LiDAR) are utilised to provide a new database that allows systematic characterization of the present surface in orderto locate topographically the areas where any future lava flows could pond, i.e., topographic depressions, obstacles that can change flow direc-tions, and unconfined areas where flows could spread in unpredictable ways. The high resolution physical characteristics of the present topogra-phy, along with the characterisation of past AVF lava flows allow us to compile a new lava flow hazard map for the area.


Topographic characterization of the Auckland VolcanicField (New Zealand) – Implications for lava flow hazard

mapping

G. Kereszturi*,**, J. Procter*, K. Németh*, J. Lindsay***, J. Kenny****, S. J. Cronin*, M. Bebbington*,*****, G. Jordán**

*Volcanic Risk Solutions, Institute of Natural Resources, Massey University, Private Bag 11 222, Palmerston North, New Zealand.**Geological Institute of Hungary, Stefánia út 14, H-1143, Budapest, Hungary. ***School of Environment, The University of Auckland, PB92019, Auckland Mail Center 1142, Auckland, New Zealand.****Geomarine Research, 49 Swainston Rd, St Johns, Auckland, New Zealand.*****Institute of Fundamental Sciences–Statistics, Massey University, Palmerston North, New Zealand.

Introduction

Volcanic flows include lava flows, debris flows and avalanch-es, pyroclastic density currents and surges, all of which mayhave devastating consequences for local communities, theireconomy, transportation, and the natural environment. Theapplication of geophysical mass flow models (GMFMs) tosimulate volcanic flows is fundamental to better understandthe key conditions that control flow behavior and to improveassessment of their potential hazards. Recent advances havebeen made in creating computational models of these flowsfor the purpose of hazard mitigation (e.g., McDougall andHungr, 2004; Kelfoun and Druitt, 2005; Patra et al., 2005).Applications of these models have been performed at sever-al volcanoes over topographies obtained from stereo coverageof satellite imagery (e.g., ASTER— Advanced SpaceborneThermal Emission and Reflection Radiometer), radar data(e.g., SRTM—Shuttle Radar Topography mapping Mission),laser altimetry (e.g., LIDAR— Light Detection and Ranging),and generic vector data (e.g., contour lines from topographicmaps). Despite the variability of the input parameters requiredby each routine, in order to perform a numeric simulation, adigital elevation model (DEM) is a common input for simu-lation algorithms.

Previous authors have shown the importance of the choiceof the DEM on computational routines for reconstructing thedifferent paths, velocities and extents of various flows, andfor correctly estimating the areas and levels of hazards asso-ciated with future volcanic activity (e.g., Stevens et al., 2002;Capra et al., 2011). The age and resolution of the DEM is animportant issue when testing the accuracy of any terrain-de-pendent model and developing an appropriate DEM pro-vides the basis for any realistic flow modeling. As with allsimulation studies that attempt to use existing depositional

records to evaluate model outputs, the topography represen-tation or DEM used is normally that of the present day, ratherthan the ideal of a pre-event terrain model. Therefore, a prop-er understanding and respect for model uncertainty arisingfrom poor parameter estimation, topographic description ormechanical understanding is critical when determiningwhether the use of GMFMs is appropriate for volcanic haz-ards assessment.

Methods

In this work, we study in a systematic way how differentDEM resolutions influence the output of simulations repro-ducing past volcanic flows, especially where topography ischaracterized by sudden changes in slope or close curves indeep ravines. The 2006 block-and-ash flows (BAFs) of Mer-api Volcano, located in Central Java, Indonesia, presented arare opportunity to test the validity of some of these DEMsagainst a well-constrained field example (Charbonnier andGertisser, 2009). The May-June 2006 eruption of MerapiVolcano consisted of three eruption phases that producedtwo main types of BAFs (short- to medium-runout BAFsthat show similar behavior as granular-free surface flows onunconfined planes and long-runout BAFs interpreted as un-steady, modified grainflows) that have been recognizedbased on various parameters such as their generation mech-anisms, flow volume, travel distance, deposit morphology,distribution, lithology and grain size distributions (Charbon-nier, 2009; Charbonnier and Gertisser, 2011). The influenceof various types of topographic settings on transport and de-position mechanisms of these two types of BAFs was ex-amined through the development of two conceptual models(Charbonnier and Gertisser, 2011). Based on these models,a new classification scheme for the different types of BAFs


Application of field observations and remote sensing tonumerical modeling and hazard assessment of volcanic

flows: an example from Merapi volcano, Indonesia

S. J. Charbonnier*, C. B. Connor*, L. Connor*, T. Dixon*, R. Gertisser**

Short abstract: The performance of geophysical mass flow models in simulating actual events is critically dependent on: (1) the cali-bration of the model by using extensive field-based data such as deposit distribution, processes of flow generation, transport anddeposition; (2) the incorporation of a suitable numerical topographic dataset through multiple acquisition of remote sensing data (i.e.,high-resolution digital elevation models); and (3) the choice of model input parameters and source characteristics.

Keywords: volcanic flows, numerical modeling, remote sensing, Merapi volcano, hazard assessment.

* Department of Geology, University of South Florida, Tampa, FL, 33620, USA.** School of Physical and Geographical Sciences, Earth Sciences and Geography, Keele University, Keele, Staffordshire, ST5 5BG, United Kingdom.


S.J. Charbonnier et al.

observed at Merapi has been proposed that can be directlyintegrated into numerical simulations using three differentcomputational routines: Titan2D (Patra et al., 2005), Vol-cFlow (Kelfoun and Druitt, 2005) and DAN3D (McDougalland Hungr, 2004). The DEMs we used for the simulationsare: (1) a local DEM (LDEM) with a spatial resolution of 15m, a vertical accuracy of ± 9 m; (2) an ASTER Global DEMof ~30 m spatial resolution and ± 11 m vertical accuracy;and (3) a SRTM WRS-2 DEM with a spatial resolution of 90m, a vertical accuracy of ± 10 m. Sensitivity tests have beencarried out by running numerical simulations using input pa-rameters defined in Charbonnier and Gertisser (2009).

Results

Figure 1 shows the results of sensitivity tests carried outusing the Titan2D routine for reproducing the inundationarea of the 14 June 2006 BAF at Merapi Volcano over threeDEMs with different spatial resolutions. Low-resolutionDEMs, such as the 90-m DEM, are inappropriate to repro-duce the BAF inundation area over irregular topographicpath where obstacles and abrupt turns can suddenly changethe simulated flow path. Inaccuracy in both, altitude of topo-graphic obstacles and depth of ravines, made impossible toreproduce past flows for DEM with coarse spatial resolu-tions (> 15 m). Table 1 shows the effects of the DEM spatial

resolution on two of the model output variables obtainedduring the Titan2D simulations, maximum flow velocitiesand maximum flow depths. Results show increasing maxi-mum flow velocities (from 45.8 to 60.7 m/s) and decreasingflow depths (from 25.1 to 8.5 m) when using DEMs withcoarser spatial resolutions (from 15 to 90 m). Similar resultswere obtained with VolcFlow and DAN3D simulation codeswhen using similar input parameters over the same threeDEMs. These results represent a significant contribution,evidence that the source data, used to obtain the DEM overwhich the flow is simulated, are a key input parameter thatcontrols the simulation results. In the particular case ofrugged topography, such as at subduction-zone stratovolca-noes like Merapi, a DEM with high spatial resolution (i.e.,5–15 m) should be acquired in order to obtain confidentsimulation results.

Discussion and conclusions

Results show the importance of the DEM accuracy oncomputational routines for correctly reproducing the flowpaths and areas covered from a well constrained eruptiveevent at Merapi Volcano. The results suggest that the per-formance of numerical models in simulating actual events iscritically dependent on: (1) the calibration of the model byusing extensive field-based data such as deposit distribution,

Fig. 1 – Sensitivity tests carried out using the Titan2D routine for reproducing the inundation area of the 14 June 2006 BAF at Mer-api volcano over DEMs with different spatial resolutions. The red outline is the mapped extent of the 14 June 2006 BAF deposits. The‘flow coverage match’ is obtained by dividing the intersection area of the mapped and simulated inundation areas by their union area.


Application of field observations and remote sensing to numerical modeling and hazard assessment of volcanic flows

processes of flow generation, transport and deposition; (2)the incorporation of a suitable numerical topographic dataset(i.e., high-resolution DEM); and (3) the choice of modelinput parameters, such as location and volume of the initialpile of material and source characteristics.

The newly available combination of TerraSAR-X satellitedata with its add-on TanDEM-X data (from the GermanAerospace Center DLR) can generate accurate numerical to-pography (DEMs with up to 6 m spatial resolution and 0.8 mvertical accuracy) and/or capture rapid topographic changesassociated with the emplacement of volcanic deposits overshort (< two weeks) periods (Wadge et al., 2011). This givesvolcanologists the tremendous potential to better under-stand the dynamics of hazardous volcanic flows. Our maingoal is now to demonstrate the utility of such enhanced qual-ity DEMs for volcano hazards and crisis management. Wewill test and validate the high resolution DEM products fromthe current TanDEM-X mission, by investigating the sensitiv-ity of standard volcanic flow models to various DEMs. Bycombining numerical simulations using freely available mod-els, probability modeling and statistical methods for definingbest-fit input parameters, we can develop a systematic ap-proach for correctly estimating model uncertainties arisingfrom poor parameter estimation, pre-event topography andmechanical understanding of volcanic flows. This methodcan also be used for detecting and measuring rapid topo-graphic changes occurring during short eruptive periods anddefining hazard zonations for key areas at risk from futurevolcanic activity.

Consequently, the work proposed here will be of immediatebenefit to all groups involved in assessing volcano hazards ei-ther directly (at observatories on some of the most active vol-canoes around the world) or through remote sensing tech-niques. Ultimately, the dataset obtained in this study is consid-

ered not only instrumental for characterizingvolcanic flows and related hazards at Merapi,but will allow comparisons with similar vol-canic phenomena at other volcanoes aroundthe globe.

References

Capra L., Manea V.C., Manea M., Norini G.(2011) – The importance of digital elevation modelresolution on granular flow simulations: a test casefor Colima volcano using TITAN2D computational

routine. Natural Hazards Review, DOI: 10.1007/s11069-011-9788-6.Charbonnier S. (2009) – The dynamics and hazards of small-vol-

ume pyroclastic flows: a case study of the 2006 eruption of Merapivolcano, Java, Indonesia. PhD Dissertation, Keele University,United Kingdom, 347 p.

Charbonnier S., Gertisser R. (2009) – Numerical simulations ofblock-and-ash flows using the Titan2D flow model: examplesfrom the 2006 eruption of Merapi Volcano, Java, Indonesia. Bul-letin of Volcanology 71, 953-959.

Charbonnier, S., Gertisser, R. (2011) – Deposit architecture anddynamics of the 2006 block-and-ash flows of Merapi Volcano,Java, Indonesia. Sedimentology, DOI: 10.1111/j.1365-3091.2011.01226.

Kelfoun K., Druitt T.H. (2005) – Numerical modeling of theemplacement of Socompa rock avalanche, Chile. Journal ofGeophysical Research 112, B12202.

McDougall S., Hungr O. (2004) – A model for the analysis ofrapid landslide motion across three-dimensional terrain. Canadi-an Geotechnics J. 41, 1084-1097.

Patra A.K., Bauer A.C., Nichita C.C., Pitman E.B., SheridanM.F., Bursik M.I., Rupp B., Webber A., Stinton A.J.,Namikawa L.M., Renschler C.S. (2005) – Parallel adaptivesimulation of dry avalanches over natural terrain. Journal of Vol-canology and Geothermal Research 139, 1-22.

Stevens N.F., Manville V., Heron, D.W. (2002) – The sensitivityof a volcanic flow model to digital elevation model accuracy:experiments with digitized map contours and interferometricSAR at Ruapehu and Taranaki volcanoes, New Zealand. Journalof Volcanology and Geothermal Research 119, 89-105.

Wadge G., Cole P., Stinton A., Komorowski J.-C., Stewart R.,Toombs A.C., Legendre Y. (2011) – Rapid topographic changemeasured by high-resolution satellite radar at Soufriere HillsVolcano, Montserrat, 2008-2010. Journal of Volcanology andGeothermal Research 199, 142-152.

Tab. 1 – Titan2D model output variables for the 14 June 2006 BAFs at Merapi usingdifferent DEM resolutions.

TITAN 2D model

Output variable Max. velocity (m/s) Max. flow depth (m)

LDEM 15 m 45.8 25.1

ATSER GDEM 30m 50.8 11.7

SRTM WRS-2 90 m 60.7 8.5

Introduction

Merapi volcano, located in densely populated of Central Javaand Yogyakarta Provinces, is one of Indonesia’s most activevolcanoes. The summit at 2968 m above sea level has beenbuilt up by andesitic lava flows and domes. The most dange-rous and high risk area is southeast to southwest slopes, sincethe last 100 years pyroclastic flows due to dome collapse di-rected mainly towards these slopes area. During historicaltime, the volcano has frequently erupted and the eruption styleis characterized by dome growth and collapse, subsequentlyproducing pyroclastic flows. Since 1768, there have beenmore than 80 recorded eruptions. Some are categorized aslarge eruptions with VEI ≥ 3 such as in 1768, 1822, 1849,1872 and 1930-1931 (Andreastuti, et al., 2000; Voight et al.,2000). At least seventeen of Merapi’s past eruptions, includ-ing the latest eruption in 2010, have caused fatalities due todensely population (about 400,000 people) in its hazard zones.

2010 Merapi eruption: monitoring,chronology, warning and impacts

Center for Volcanology and Geological Hazard Mitigation(CVGHM) has operated a seismic network since 1982 onMerapi to monitor volcanic activity. Beside seismicity, defor-mation, geochemistry, geological and visual observation meth-ods are applied. Electronic Distance Measurement (EDM) is

conducted regularly with the reflectors installed on thesouthern and western flanks. Meanwhile gas and temperaturemonitoring is carried out by Ground SO2 – UV DOAS dis-crete measurements, satellite SO2 (IASI, AIRS, and OMI),temperature measurements of fumaroles are acquired at thesummit. Based on these data, CVGHM issues early warningto the people living around the volcano.

Early sign of unrest volcanic activity is shown by a signifi-cant number of volcano-tectonic earthquakes (VT) whichwere observed in October 31, 2009, December 9, 2009, June10 and September 9, 2010. On September 20, 2010, Merapiactivity increased from the alert Level I (Normal) to Level II(‘waspada’ stage) where the number of volcano-tectonic ear-thquake (VT) reached 11 events, multiphase (MP) attained 38events and rockfall 3 events in a day. On 21 October 2010, thestate of activity was raised to Level III (‘siaga’ stage) since wi-thin a day the number of VT earthquake reached 17 events,MP attained 150 events, 29 rockfall events were recorded andthe inflation rate of deformation was 17 cm. The high activityof the volcano still continued as shown by an increasing num-ber of VT earthquake (80 events), MP (588), rockfall (194)and the rate of deformation inflated 42 cm in a day causing arise of alert level to its highest level, Level IV (‘Awas’ stage)on October 25, 2010. CVGHM recommended that people beevacuated to a safe zone beyond 10 km in radius from thesummit.

During the 2010 crisis, the first eruption on October 26,2010 generated pyroclastic flows, which travelled down to

The role of remote sensing data during the 2010 crisis at Merapi volcano

Surono*, A. Solikhin*, A. B. Santoso*, P. Jousset**, J. S. Pallister***, M. Boichu****, S. Carn*****

Short abstract: Merapi volcano (Indonesia) is one of the most active and hazardous volcanoes in the world, known for its frequent rel-atively small eruptions, characteristic pyroclastic flows and large population at risk. Center for Volcanology and Geological HazardMitigation (CVGHM) declares alert level of Merapi volcano based on ground observations (seismic, deformation, SO2 emission) andsatellite remote sensing data provided from international collaboration in a worldwide scale. Between 26 October and 8 November 2010,Merapi Volcano produced its largest eruption in more than a century, requiring the evacuation of an area of about 1300 km2 and dis-placing 410,388 people. Our data show that the eruption had a rapid onset and relatively short duration as consequences of rapid ascentof unusually gas-rich magma for Merapi. The integration of ground observations and satellite remote sensing data for real-time and near-real time monitoring of the eruption played a vital role it in decision support, especially regarding ranges of exclusion zones.

Keywords: Merapi volcano, remote sensing, 2010 crisis.


* Center of Volcanology and Geological Hazard Mitigation, Jalan Diponegoro 57, 40122 Bandung, Indonesia. ** BRGM, RNSC, 3 Avenue Claude Guillemin, BP36009, 45060 Orléans Cedex 2, France. *** U.S. Geological Survey, Cascades Volcano Observatory, USA.**** Institut Pierre Simon Laplace, Laboratoire de Météorologie Dynamique, Ecole Polytechnique, France.***** Michigan Technological University, Department of Geological/Mining Engineering & Sciences, USA.

8 km into Kali Gendol, on southern flank of the volcano.Since November 3, 2010, continuous and “over scale” tremoroccurred and thought to be associated with high-rate lavadome growth. This activity reached its peak on November 5,2010 when lava dome collapse generated pyroclastic flowswhich traveled distances up to 15-16 km from the summit intoKali Gendol. Between 26 October and 8 November 2010, theseries of eruption caused 367 victims, 277 injured, 410,388evacuated as well as heavy damaged to infrastructures, housesand agriculture areas. After November 8, Merapi activity fluctu-ated and tended to decrease. On December 3, 2010, the alertlevel was downgraded from Level IV to Level III. The total vol-ume of the 2010 deposits, estimated fromfield mapping is about 0.13 km3(Aisyah etal., 2010). The deposits covering the slopesof Merapi have and will generate laharsduring the rainy seasons and will threatenpeople living along the riverbanks.

Remote sensing duringthe 2010 crisis

Satellite remote sensing and other tech-nical support were provided by interna-tional collaboration to CVGHM duringthe October to November 2010 erup-tion. The high-resolution SAR systemsof the COSMO-SkyMed constellation,RADARSAT and TerraSAR-X satellites,supplied very detailed images of the vol-cano summit crater, rapidly growinglava domes, vent features, and pyroclas-tic-flow deposits.

The explosive eruptions on 26 and 30 Oc-tober removed the 2006 lava dome, deep-ened the summit crater, and deeply inci-

sed the headwall of the Kali Gendol drainage (Figs. 1a, 1b).These data confirmed that the 2010 eruption did not beginwith extrusion of lava (as a characteristic for other recenteruptions of Merapi) but instead with an explosive, crateringevent. This fact, along with subsequent evidence for veryrapid rates of dome growth (peaking at > 25 m3 s-1), reinfor-ced CVGHM concerns that the 2010 eruption would be muchlarger and more hazardous than those of the past century.

Between 26 October and 4 November, the lava dome grewto ~5 × 106 m3 in volume (Fig. 1c). It was destroyed, howev-er, during the explosive eruption on the night of 4–5 Novem-ber, which greatly enlarged the new summit crater (Fig. 1c).


Surono et al.

Fig. 1 – Synthetic Aperture Radar (SAR) im-ages of the summit of Mount Merapi brack-eting the time of the 26 October explosiveeruption and the 4–5 November explosiveeruptions. G (Kali Gendol), K (Kali Kuning). a,RADARSAT image, 11 October, 2010. Arrow in-dicates the 2006 lava dome. b, TerraSAR-Ximage, 26 October, showing new summit crater(arrow) produced by explosive eruption of26 October. c, TerraSAR-X image, 4 Nov 2010,showing large (~5 × 106 m3) lava dome (D) thatgrew rapidly after the 26 October eruption andwas destroyed by the explosive eruption of4–5 November. Pyroclastic flow deposits (PF)from the 26 October eruption appear dark inthe radar images. d, RADARSAT image of 5November, 2010, showing pyroclastic flow de-posits (PF, dark gray) and surge deposits (S,light gray). These deposits formed earlier dur-ing the main phase of the 4–5 November ex-plosive eruption. An enlarged, elongate crater,produced by the November 4–5 eruption isalso evident at the summit.


The role of remote sensing data during the 2010 crisis at Merapi Volcano

It also produced a pyroclastic flow and surge that travelled15-16 km from the summit in the direction of Yogyakarta,shortly after the evacuation zone was extended to 20 km.The surge affected an area of ~13 km2 (Fig. 1c) and was res-ponsible for many of the fatalities. Post-eruption SAR ima-gery shows a new, roughly circular crater with a diameter of~400 m, breeched on the southeast by a sloping trough thatextends 400 m down slope (Fig. 1c). The eruptions removedlarge parts (~10–20 × 106 m3) of the previously emplacedsummit dome complex. RADARSAT images collected on 6November show that rapid extrusion resumed and produced anew ~1.5 × 106 m3ava dome in <12 h at a minimum effusionrate of ~35 m3 s-1. Dome growth ceased by 8 November butwas followed a period of dome subsidence and gas and ashemissions from several vents adjacent to or penetrating thenew lava dome.

Combination of gas and ash plume remote sensing fromthe ground (ultraviolet differential optical absorption spec-

troscopy, DOAS) and satellites (using IASI, AIRS and OMIinfrared and ultraviolet sensors), provided crucial informa-tion on degassing during the four stages of Merapi’s activi-ty (Fig. 2). The late October plumes were fuelled by exter-nal water vaporized by ascending magma, and consistent withthe phreatomagmatic character of this phase of the eruption.The SO2 flux then decreased to a relatively ‘low’ level for thiseruption, but still at elevated levels compared to past Merapieruptions. It then increased significantly on 3 November, lessthan two days before the climax on 4–5 November. The lar-gest SO2 output, of ~ 0.28 Tg, was recorded on 5 Novemberby AIRS. Intriguingly, Fig. 2 suggests similar trends of de-gassing and RSAM during all phases of the eruption exceptfor one episode during the waning stage on 6–7 Novemberwhen the SO2 flux continued decreasing despite renewal ofseismicity. This could be explained by associated subsi-dence of the final lava dome, by hindered gas release, moreenergetic harmonic tremor and lesser Vulcanian explosions.

Fig. 2 – Comparison between SO2 fluxes and RSAM data. a, Overview of Merapi volcano 2010 eruption degassing, described throughSO2 flux observations and upper-bound values of the plume altitude, which give an indirect estimation of the total flux of gases. b, RSAMcomputed for the Plawangan station (6 km from the summit). A clear correspondence between RSAM and SO2 flux is demonstrated, sup-porting our identification of four distinct phases to the eruption (indicated by PHASE I to IV). E stands for explosion; L for Lahar.


Surono et al.

Conclusion

Between 26 October and 8 November 2010, Merapi Vol-cano produced its largest eruption in more than a century,caused 367 fatalities, requiring the evacuation of an area ofabout 1300 km2 and displacing 410,388 people. The eruptionhad a rapid onset and relatively short duration as consequencesof rapid ascent of unusually gas-rich magma for Merapi. Dur-ing the 2010 Merapi eruption, integration of ground observa-tions and satellite remote sensing data for real-time and near-real time monitoring has an important role in decision support,especially regarding the ranges of exclusion zones. The re-peated acquisition of spaceborne Synthetic Aperture Radar(SAR) image data enabled monitoring of changes at the vol-cano’s summit, despite the cloud cover during much of theeruptive episode, and also mapping of the extent of pyroclas-tic density currents. Combination of gas and ash plume remote

sensing from the ground and satellites provided crucial infor-mation on degassing during the four stages of its activity.

References

Aisyah N., Sumarti S., Sayudi D. S., Budisantoso A., Muzani M.,Dwiyono S., Sunarto, K. Aktivitas (2010) – G. Merapi PeriodeSeptember – Desember 2010 (Erupsi G. Merapi 26 oktober –7 November 2010). Bulletin Berkala Merapi 07/03, December.

Andreastuti S.D., Alloway B.V., Smith I.E.M. (2000) – A detailedtephrostratigraphic framework at Merapi Volcano, Central Java,Indonesia: implications for eruption predictions and hazard assess-ment. Journal of Volcanology and Geothermal Research 100, 51-67.

Voight B., Constantine E. K., Sismowidjoyo S., Torley R. (2000)– Historical eruptions of Merapi Volcano, Central Java, Indone-sia, 1768-1998. Journal of Volcanology and Geothermal Re-search 100, 69-138.

Introduction

One of the hazards associated with a steeply-sided volcanois large-scale failure of its edifice. Known occurences of sec-tor failure in volcanoes worldwide occur with or without anassociated eruption. Mayon, with its greater than 35° slopes,has inherent instability (Siebert, 1984) that can be aggravat-ed by hydrothermal processes and tectonic movement (Lag-may et al., 2000). To investigate the role of tectonics on thestability of Mayon, Permanent Scatterer Interferometric Syn-thetic Aperture Radar (PSInSAR) was used to measure groundmotion in the region where its edifice is erected. In particu-lar, the fault blocks comprising the northwest-trending OasGraben is examined in terms of sense and rate of movement.The importance of determining activity of the upper faultboundary of the Oas Graben is highlighted since its trace pro-jects into the base of Mayon’s edifice, possibly traversing thecone and contributing further to its instability.

Methods

PSInSAR is a new method of interferometric processingthat overcomes the limitations of conventional SyntheticAperture Radar differential interferometry (DInSAR) and iscapable of detecting millimeter scale ground displacements.It measures the change in the distance from the satellite to theground by determining the phase difference in the radarwavelength between different satellite passes. PSInSAR elim-inate anomalies due to atmospheric delays and temporal andgeometric decorrelation (Burgmann et al., 2000) eminent intropical regions by exploiting the temporal and spatial char-

acteristics of radar interferometric signatures derived fromtime-coherent point-wise targets. These point-wise targetsare radar bright and radar-phase stable such as built-ups,rock-outcrops and corner reflectors (Ferretti et al., 2000; Fer-retti et al., 2001; Ferretti et al., 2004). Persistent point-wisetargets that exist in multiple SAR acquisitions are called‘persistent scatterers’ and are used to improve the signal-to-noise ratio by separating a modeled deformation rate from at-mospheric and elevation error components in the measuredrange change (Colesanti et al., 2003). Changes in the phasesignal of persistent scatterers is required not to exceed 2π (ra-diation wavelength), to prevent ambiguities in the measuredphase difference. The principle behind PSInSAR makes itsuitable to perform time-series analysis of ground deforma-tion in areas where conventional InSAR or dInSAR may fail.

In this study, a total of 47 combined descending ERS1-2and ascending ENVISAT raw 0 data were processed usingStaMPS (Hooper et al., 2004; Hooper 2006), now named asMAINSAR. The imageries were archived from 1993-2000and 2003-2006, for ERS1-2 and ENVISAT, respectively.Moreover, fieldwork on the flanks of Mayon as well as onthe surrounding areas was conducted to relate the PSInSARresults from the regional stress regime in the volcano.

Results

Interferograms

The interferograms generated from both the ascendingand descending data show coherence in the urban areas ofthe cities in Legaspi, Tabaco and Ligao as well as along the

Detecting slip movement of Mayon volcano from persistent scatterer interferometry

A.M.F. Lagmay*, M.G. Bato*,**, E.M.R Paguican*,**, H. Zebker***


Short abstract: Permanent Scatterer Interferometry conducted in Albay Province, Bicol, Philippines, reveal tectonic deformation of theOas graben, a northwest-trending structural depression on which Mayon Volcano is built. Differential movement between the northernhorst and graben, measured in terms of line-of-sight (LOS) change in the radar signal, is as much 2.5 cm/year. The northern horst movesnorthwest while the graben moves mostly downward. Coupled with morphological interpretation, the results of the PSInSAR study,suggests left-lateral oblique-slip movement of the northern bounding fault of the Oas graben. PSInSAR is a relatively new method ofinterferometric processing that enables the detection of mm-scale deformation. The technique applies even in tropical regions whereconventional InSAR normally doesn’t work due to temporal and spatial decorrelation. This paper demonstrates the functionality ofPSInSAR in a humid tropical region and highlights the probable landslide hazards associated with an oversteepened volcano that mayhave been further destabilized by tectonic activity.

Key words: Permanent scatterer interferometry, PSInSAR, Mayon volcano, Oas graben, Legaspi lineament.

* National Institute of Geological Sciences, College of Science, University of the Philippines, Diliman, Quezon City 1101, Philippines.** Laboratoire Magmas et Volcans, Université Blaise Pascal, 63000 Clermont-Ferrand, France.*** Department of Geophysics, Stanford University, Panama Mall, Stanford, 94045, CA, USA.


A.M.F. Lagmay et al.

roads that interconnect these cities where houses have beenbuilt. Other areas which show coherence are the 1984, 1993and 1999 lava flows that have been deposited in the south-west flank of the volcano.

Ascending data pairs

Permanent scatterers in the interferograms from 2003-2006 show distinct movement as suggested by the change inthe line-of-sight of the radar sensor, which looks down at anincidence angle of 23.5° from the west-southwest. The north-ern horst of the Oas Graben (red dots) moves toward the radarsensor with a line-of-sight change as much as -4.7 radians(Fig. 1a). The graben (downthrown block), on the other hand,is generally dotted with blue, indicating movement awayfrom the radar sensor by as much as 3.35 radians change,respectively. Yellow- and red-colored permanent scatterersare also found on Mayon’s cone suggesting inflation to-wards the satellite.

Descending data pairs

The descending data show a distinct distribution of per-manent scatterers represented by blue- to red-colored dots(Fig. 1b). The northern horst of the Oas Graben is dominat-ed by permanent scatterers (blue dots) that show an increase

in line-of-sight of a radar satellite, which looks down from theeast-southeast with an incidence angle of 24.8°. This changein range is equivalent to 3 radians. The graben (downthrownblock) is also characterized by an increase in line-of-sight ofthe radar sensor with the exception of orange to yellow-col-ored permanent scatterers on Mayon volcano which depictmovement towards the imaging satellite.

Fieldwork results

Fieldwork conducted within the vicinity of Mayon volcanoshows the presence of fractures both on the west and eastflanks of the volcano. Basud river, which is located at the east-ern flank of Mayon, exposed fractures hosted in indurated py-roclastic flow and lahar deposits of up to 900 m in length.Relay and right stepping fractures in lahar deposit, Riedelshears and structural wedges are found near the bend of theBasud river. Fractures on the western flank exposed alongBaligang and Masarawag rivers are steeply dipping and mea-sured an aperture of about 1 to 2 mm. The deposits observedare 3 to 4 m thick and lies in an indurated lahar and terrace de-posits. The mean orientations of the fractures in the west flankare N57°W and N66°E, whereas the east flank recordedN44°W and N48°E. When plotted in a rose diagram (Fig. 2), adominant northwest trending measurement and a minor north-east trending can be observed. Geophysical surveys conducted

Fig. 1a and 1b – Velocity of permanent scatterers for the ascending and descending data. The letter A in the arrow symbol is theazimuth of the radar sensor and L is the look direction. Opposing colors in the ascending and descending data indicate a horizontal compo-nent.


Detecting Slip Movement of Mayon Volcano from Persistent Scatterer Interferometry

around the base of Mayon revealgravity and magnetic anomalieson both the east and west flanksof Mayon (Fig. 3).

Discussion

The differential change inLOS of the unwrapped phasesignal from permanent scatter-ers in urban areas of Tabaco,Legaspi, Ligao and Oas indicatestrain along the northern faultboundary of the Oas Graben.The northern horst, with a decrease in LOS from the east-look-ing ascending imageries and increase in LOS from the west-looking descending imageries suggest horizontal movement andpossible uplift. The differential rate in range change observed inthe ascending and descending imageries further suggest left-lateral motion. This interpretation is valid on the premisethat the motion vector of the northern horst from the period

1993-2000 and 2003-2006 has been constant in direction.The graben, with the exception of areas on Mayon’s cone,is interpreted to be moving downward. This is suggested bythe increasing LOS from the ground to the satellite sensorin both ascending and descending imageries. Subsidence insome areas of the graben is as much as 10.1 mm/yr. Infla-tion observed on Mayon’s cone may be due to the rise of

Fig. 2 – Rose diagrams of frac-tures measured on the westand east flanks of Mayon Vol-cano.

Fig.3 – Results of gravity and magnetic surveys on the west and east flanks of Mayon Volcano. Lower profiles shows gravity and mag-netic anomalies as the survey line transects lineaments corresponding to the northern fault of the Oas Graben.


A.M.F. Lagmay et al.

magma associated with eruptive periods in 1993, 1999,2000 and 2006. The inflation on Mayon cone masks detec-tion of possible horizontal movement of the graben. Theoccurence of differential movement between the northernhorst and the downthrown block of the Oas Graben suggestthe presence of a fault underlying Mayon volcano. This isconsistent with the left-lateral shear of the northern faultboundary of the Oas Graben detected from dGPS data col-lected around Mayon Volcano with a maximum rate ofmovement of about 1 cm per year (Bacolcol, 2011, person-al communication).

Field structures hosted in the volcanic products of Mayonare interpreted as shear fractures. The consistency of its ori-entation associated to the regional stress indicates a tecton-ic origin of its existence. The consequence of tectonic move-ment on Mayon volcano is further destabilization of an in-herently unstable cone. Deformation of an edifice by eithernormal or strike-slip faulting (Francis and Self, 1987; Tibal-di, 1995; Lagmay et al., 2000; Lagmay et al., 2005) canweaken the structure of the volcanic edifice and create struc-tural discontinuities that may provide sliding planes in theevent of gravitational collapse.

Conclusion

Permanent Scatterer Interferometry on a time series ofERS-1, ERS-2 and ENVISAT imageries was applied forvolcano tectonic investigation of Mayon volcano. The re-sults indicate differential motion between the northern horstand the downthrown block of the Oas Graben. Coupled withmorphological analysis and the field data in the area ofstudy, there is evidence to believe the presence of an activefault, hereby called as the Oas Fault. This structure strikesnorthwest and can be seen to project towards the base of thewestern flank of Mayon, where it disappears beneath thecone. PSInSAR analysis, however, reveal the continuationof the Oas Fault beneath Mayon volcano as suggested bydifferential movement of permanent scatterers north andsouth of its edifice. The presence of an active fault has im-plications in the hazards assessment of Mayon volcano. Thisvolcano with its steep-sided flanks is inherently unstable,and is further being destablized by tectonic movement. Suchinstability may eventually lead to catastrophic failure. Assuch, the results of this study can be used as basis for themitigation of a hazard at Mayon, which may or may not beaccompanied by an eruptive event.

References

Bürgmann R., Rosen P., Fielding E. (2000) – Synthetic apertureradar interferometry to measure Earth’s surface topography andits deformation. Annual Review of Earth and Planetary Sciences28, 169–209.

Colesanti C., Ferretti A., Novali F., Prati C., and Rocca F. (2003)– SAR monitoring of progressive and seasonal ground deforma-tion using the permanent scatterers technique. IEEE Transanc-tions Geoscience Remote Sensing 41, 7, 1685–1701.

Ferretti F., Novali F., Burgmann R., Hilley G., Prati C. (2004)– InSAR Permanent Scatterer analysis reveals ups and downs inSan Francisco Bay area. EOS, 85, 34, 1–3.

Ferretti F., Prati C., Rocca F. (2000) – Non-linear subsidence rateestimation using permanent scatterers in differential SAR Inter-ferometry. IEEE Transactions Geoscience Remote Sensing 38, 5,2202–2212.

Ferretti F., Prati C., Rocca F. (2001) – Permanent scatterers in SAR In-terferometry. IEEE IEEE Transactions Geoscience Remote Sens-ing 39, 1, 8–20.

Francis P., Self S. (1987) – Collapsing Volcanoes. Scientific Amer-ican 256, 72–89.

Hooper A. (2006) – Persistent Scatterer Radar Interferometry forCrustal Deformation Studies and Modeling of Volcanic Defor-mation. Ph.D. thesis, Stanford University.

Hooper A., Zebker, Segall P., Kampes B. (2004) – A NewMethod for Measuring Deformation on Volcanoes and OtherNatural Terrains Using InSAR Persistent Scatterers. Geophysi-cal Research Letters, L23611.

Kampes B., Hanssen R., Perski Z. (2003) – Radar Interferometrywith Public Domain Tools. In: Proceeding of FRINGE 2003.December 1-5, Frascati, Italy, 2003.

Lagmay A., Tengonciang A., Uy H. (2005) – Structural setting ofthe Bicol Basin and kinematic analysis of fractures on Mayon vol-cano, Philippines. Journal of Volcanology and Geothermal Re-search 144, 23–36.

Lagmay A., van Wyk de Vries B., Kerle N., Pyle D. (2000) – Vol-cano instability induced by strike-slip faulting. Bulletin ofVolcanology 62, 331–346.

Siebert L. (1984) – Large volcanic debris avalanches: characteris-tic of source areas, deposits and associated eruptions. Journal ofVolcanology and Geothermal Research 22, 163–197.

Tibaldi A. (1995) – Morphology of pyroclastic cones and tecton-ics. Journal of Geophysical Research 100, 24521–24535.

Zebker H. (2000), Studying the Earth with interferometric radar.Computing in Science and Enginee0ring 2, 3, 52–60.

Introduction

Big Earthquake generating tsunami occurred in the south-ern coastal of Java on 17 July 2006, there was a Tsunamiwaves occurred in the South Java Coastal Area and destruct-ed the coastal area of West Java, Central Java, and Yo-gyakarta Special Province. Tsunami also occured in Pangan-daran, Cilacap and Kebumen. Many coastal areas in WestJava (Pangandaran), Central Java (Cilacap and Kebumen),and Yogyakarta were affected. More than 600 people died

and hundreds of houses were completely damaged. Run-upgeneration was recorded in various heights, from 3 – 5 me-ters and in some places more than 5 meters. The impacts ofthe Tsunami depend on how far Tsunami can move directlyto inland. It can be more than 300 meters inland and makesmore destruction in the coastal area. Tempo (19th July 2006)reported that the most effected area in Tasikmalaya, wereCipatujah and CikalongSub-districts[1].

Coastal area is very dynamic and constantly changing (Birdet al., 1980). The change in the coast of Indonesia consists


Sand dune conservation zone based on tsunamiinundation hazard in Parangtritis coastal area,

Bantul regency, Yogyakarta special province:Remote sensing and Geographic Information

System application

R. F. Putri*, D. Mardiatno*, J. Sartohadi**, J. T. Sri Sumantyo***

* Faculty of Geography, Gadjah Mada University, Indonesia.** Research Center For Disaster, Gadjah Mada University, Indonesia.*** Center for Environmental Remote Sensing, Chiba University, Japan.

Short abstract: Tsunami waves occurred in the South Java Coastal Area and destructed the coastal area of West Java, Central Java, andYogyakarta Special Province. Sand dune in Parangtritis coastal area has a function such as a barrier to threat tsunami hazard. Remotesensing technology and Geographic Information System become one of data source, in particular like Disaster Information System. Theobjectives of this research are: (1) To make spatial distribution of tsunami inundation hazard using Remote Sensing & GIS application(Direction wave), (2) To make Tsunami Inundation Scenario Hazard based on the Tsunami events in the Parangtritis Coastal Area usingdistance and friction map function. (Elevation inundation scenario wave), (3) To make Sand dune zone mapping using topographic datamap analysis, and (4) To make sand dune conservation zone based on distribution tsunami inundation hazard zone and coastal regulationzone analysis. Sand dune conservation zone mapping will be correlated with coastal regulation zone in Parangtritis coastal area,Determination of the sand dune conservation zone has purpose to optimize sand dune function as a barrier tsunami inundation hazard.CRZ-I are ecologically sensitive areas where activities are largely prohibited (explicitly mandates the protection of sand dunes). CRZ-II are developed areas for agriculture land cultivation. CRZ-III comprise all rural areas as well as undeveloped areas in urban limits.The result of this research are: (1) Tsunami inundation zone with southeast wave direction scenario, it can to determination area for sanddunes area will be conserved, such as CRZ I approximately 362.100 hectares (87.12%), CRZ II approximately 53.217 hectares (12.80%)and CRZ III approximately 0.307 hectares (0.07%); (2) The tsunami inundation scenario (southwest wave direction) has 87.12% CRZI with an area of about 362.058 ha, CRZ II (12.80%) with an area around 53.213 ha, and CRZ III (0.073%) with an area of 0.307 ha;(3) Coastal Regulation Zone based on tsunami inundation of west wave direction scenario such as CRZ I approximately362.193 hectares (78.48%) are ecologically sensitive areas where activities are largely prohibited (explicitly mandates the protection ofsand dunes), CRZ II approximately 77.923 Ha (18, 74%) are developed acres agriculture land areas for cultivation and CRZ IIIapproximately 11.503 hectares (2.76%); (4) Coastal Regulation Zone based on tsunami inundation of south wave direction scenario suchas CRZ I approximately 360,107 hectares (86,56 %) are ecologically sensitive areas where activities are largely prohibited (explicitlymandates the protection of sand dunes), CRZ II approximately 55,606 Ha (13,36 %) are developed acres agriculture land areas forcultivation and CRZ III approximately 0,307 hectares (0,07 %) comprise all rural areas as well as undeveloped areas in the urban limits.

Keywords: sand dune conservation, tsunami inundation hazard, remote sensing, Geographic Information System.


R.T. Putri et al.

of short term, medium term and long term changes that can beseen as various kind of natural hazards, gradual or intermit-tent advance or retreat shoreline and those such as land upliftof subsidence or sea level rise and fall. Southern part coastalzone of Parangtritis has an unique sand dune phenomena.Sand dune in this area is considered as the only one sand dunein South East Asia. Sand deposits in Parangtritis consists ofbeach sand and sand dune. That sand deposit in Parangtritiswas composed of volcanic sand. Sand dune in Parangtritis areahas a function such as a barrier to threat tsunami hazard. Thestudy area of research can be shown in Figure 1.

Methods

The inundation zone due to tsunami would be determinedusing the predicted water depth scenario. This study intends toidentify the inundation zone of the hypothetical water depthscenario and sand dune actual mapping using SAR data anal-ysis. Unfortunately, we exclude the physical mechanisms orhydrodynamic characteristics of tsunami during generation,propagation, or inundation. Moreover, we do not consider fac-tor such as tsunami source region and coastal configurationduring inundation. Determination of the conservation zonehas purpose to optimize sand dune function as a barriertsunami inundation hazard. This research area will be clas-sified into 3 Coastal Regulation Zone (CRZ) (Namboothri et

al., 2008). CRZ-I are ecologically sensitive areas where ac-tivities are largely prohibited (explicitly mandates the protec-tion of sand dunes). CRZ-II are developed areas for agricul-ture land cultivation. CRZ-III comprise all rural areas as wellas undeveloped areas in urban limits. However, this wouldmake the methodology attractive for local authorities andcoastal manager to use.

Results

Total percentage of CRZ area is obtained based on data cal-culation sand dunes area is flooded by the tsunami inundationscenario. Based on Table 3.2. indicates that the tsunami inun-dation scenario (southwest wave direction) has 87.12% CRZI with an area of about 362.058 ha, CRZ II (12.80%) with anarea arround 53.213 ha, and CRZ III (0.073%) with an area of0.307 ha. Sand dune conservation is needed in Coastal Reg-ulation Zone I, this areas are ecologically sensitive areaswhere activities are largely prohibited (explicitly mandatesthe protection of sand dunes). In Parangtritis Coastal Area,though the dune systems are protected by CRZ regulationsthey are still facing many anthropogenic stresses. There hasbeen considerable amount of citizen action over the protec-tion of sand dunes. Coastal Regulation Zone based on tsuna-mi inundation of west wave direction scenario such as CRZI approximately 362.193 hectares (78.48%) are ecologically

Fig. 1 – Parangtritis Study Area.


Sand dune conservation zone based on tsunami inundation hazard

sensitive areas where activities arelargely prohibited (explicitly man-dates the protection of sand dunes),CRZ II approximately 77.923 Ha(18, 74%) are developed acres agri-culture land areas for cultivationand CRZ III approximately 11.503hectares (2.76%) comprise all ruralareas as well as undeveloped areas inurban limits. Table 1 and figure 2shows coastal regulation zone ofParangtritis Coastal Area.

Discussion

Coastal Regulation Zone based on tsunami inundation ofsouth wave direction scenario such as CRZ I approximately360,107 hectares (86,56 %) are ecologically sensitive areaswhere activities are largely prohibited (explicitly mandates theprotection of sand dunes), CRZ II approximately 55,606 Ha(13,36 %) are developed acres agriculture land areas for culti-vation and CRZ III approximately 0,307 hectares (0,07 %)comprise all rural areas as well as undeveloped areas in theurban limits. Based on the analysis of tsunami inudation zonewith southeast wave direction scenario, it can to determination

area for sand dunes area will be conserved, such as CRZ I ap-proximately 362.100 hectares (87.12%), CRZ II approximate-ly 53.217 hectares (12.80%) and CRZ III approximately 0.307hectares (0.07%). This result shows that 87.12% of the totalcoastal area used as a san dune conservation zone. Determina-tion of the conservation zone has purpose to optimize sand dunefunction as a barrier tsunami inundation hazard. The salient rec-ommendations and suggested action of this research area are(a) Develop and promote planning policies and procedureswhich will aim to prevent or minimize further losses of sanddune habitat because of development, (b) Develop and promotecoastal zone management policies which allow the maximumpossible free movement of coastal sediment and pay full re-

TsunamiInundation Area(Wave Direction)

CRZ 1 CRZ II CRZ III

Hectare (Ha) % Hectare (Ha) % Hectare (Ha) %

West 326.19 78.48 77.92 18.74 11.50 2.76

Southwest 361.05 87.12 53.21 12.80 0.31 0.07

South 360.10 86.56 55.60 13.36 0.317 0.07

Southeast 362.10 87.12 53.21 12.80 0.31 0.07

Tab. 1– Coastal regulation zone based on tsunami inundation area (Source : Data Calcula-tion, 2010).

Fig. 2 – Coastal regulation zone map based on tsunami Inundation scenario (A) West wave direction, (B) Southwest wave direction,(C) South wave direction, and (D) Southeast wave direction.


R.T. Putri et al.

gard to the conservation of sand dunes. Include in ShorelineManagement Plans where they have a role to play in flood de-fence, and (c) Raise public awareness about the importance ofsand dunes, and the essential mobility of coasts and the valueof maintaining unrestricted coastal processes. Promote aware-ness of the implications of the policies outlined in this planamong decision-makers.

Conclusion

Remote Sensing and Geographic Information System appli-cation can be used to generate tsunami hazard map includedwith mitigating the natural disaster. The study of sand duneareas is challenging, as not only does the land cover vary inspace and time, the topographic features of the landscape maychange over short time spans of several years with interfero-metric SAR technique can investigate the presence of subtlesurface changes. The availability of a recent DEM, site surveyand control data are very useful for maximizing the accuracy ofthe horizontal positioning InSAR data to make sand dune zonemapping and these data help to provide a means to interpret theInSAR-derived movement information to determine the over-all impact of any significant movement. Tsunami inundationzone with southeast wave direction scenario, it can to determi-nation area for sand dunes area will be conserved, such as CRZ

I are ecologically sensitive areas where activities are largelyprohibited (explicitly mandates the protection of sand dunes),CRZ II are developed areas for agriculture land cultivation andCRZ III are comprise all rural areas as well as undevelopedareas in urban limits.

Acknowledgment

The authors would like to thank Beasiswa Unggulan BPKLN– DIKNAS – Indonesia (Bureau of Planning and Internation-al Cooperation of Ministry of National Education), Center forEnvironmental Remote Sensing Chiba University and Facul-ty of Geography Gadjah Mada University (Double DegreeMPPDAS UGM) for the support of this paper.

References

Bird E.C.F., O.S.R. Ongkosongo (1980) – Environmental Changeson The Coast of Indonesia. United Nation University: United Na-tion University Press.

Namboothri N., Subramanian D., Muthuraman B. (2008) –Beyond the Tsunami: Coastal Sand Dunes of Tamil Nadu, India(An Overview). United Nations India, United Nations Develop-ment Programme and Ashoka Trust for Research in Ecology andthe Environment. India.

[1] http://www.ngdc.noaa.gov; spatial data source of tsunami eventsdatabase, 27 July 2006.

Introduction

South Java is a part of the Java Island, Indonesia, which isexposed directly to the subduction zone between the India-Australian Plate and Eurasian Plate. This region is veryvulnerable to earthquakes and tsunamis triggered by earth-quakes. Since 1920 twelve earthquakes have been recordedwhile two of them causing deadly tsunamis, i.e. in 1994 and2006. The 3 June 1994 tsunami in Banyuwangi (East Java)triggered by an earthquake with moment magnitude (Mw)7.6 killed more than 200 people. The second tsunami affecteda wider area, from eastern part of West Java Province (Garutdistrict) through Yogyakarta Special Region (Gunungkiduldistrict). It was triggered by an earthquake with moment mag-nitude (Mw) 7.7, located in 34 km of depth in the IndianOcean.

As South Java has significant experience of tsunamis, i.e.in 1994 and 2006, it is necessary to establish such mitigationprograms as an inseparable part of tsunami risk reduction.The lowland area at the southern coast of Java is at very highrisk to tsunamis (Mardiatno, 2008), as shown by the occur-rence of tsunami events in 1994 and 2006 which causedmany fatalities. Since 2008, German-Indonesia TsunamiEarly Warning Systems (GITEWS) and Ina TEWS have pre-pared the setting-up of TEWS which is derived from the as-sumption that most tsunamis are preceded by a clear signalfrom a major earthquake.

Parangtritis is one of the most popular tourist destinationslocated at South Java region, i.e. in Yogyakarta Special Re-gion Province. It is bordered by coastal alluvial plain in thenorth, escarpment in the east, Opak river in the west and In-dian Ocean in the south (fig. 1). This area is utilized by localinhabitants mainly for tourism supporting activities andagriculture.

This coastal area was selected as the research area sincethat location experienced to Java tsunami 2006. Some in-frastructures were damaged and several people had injuries.Based on that experience, it is necessary to strengthen moreefforts at mitigating tsunami, such as by utilizing more evac-uation way. This research aims to generate tsunami hazardmodel by applying different run up scenarios and to deter-mine the evacuation route for tsunami mitigation.

Method

This research used some data such as high resolutionimage, SRTM image, topographical map, and statistical data.Image processing software and GIS software were used fordata analysis. Elevation data was derived from SRTM data.Visual and digital image interpretation was conducted forenhancing several objects related to the evacuation effort iftsunami occurs. Both elevation and image interpretationresults were validated by conducting a couple days of fieldwork.


Evacuation route determination for tsunami mitigation using remote sensing data and

Geographic Information Systems at Parangtritis coastal area, Yogyakarta-Indonesia

D. Mardiatno*, R.F. Putri*, M. Susmayadi**, D.S. Sayudi***

Short abstract: This research aims to generate tsunami hazard model by applying different run up scenarios and to determine theevacuation route for tsunami mitigation. Parangtritis coastal area was selected as the research area since that location experienced toJava tsunami 2006. Some data such as high resolution image, SRTM image, topographical map, and statistical data are used in thisresearch. The 15 m value was estimated as the possible maximum run up with regard to historical tsunami data in south Java. The resultsshow that some locations will be at the high and very high hazard to tsunami for 5 m, 10 m and 15 m run up scenario. There are10 vulnerable points for evacuation, i.e. nearby shoreline, at the middle of the coastal area, and at the side of Opak River. Both horizontaland vertical evacuations are possibly applied in this area. Pathways should also be utilized more effectively for evacuation routes.

Keywords: evacuation route, horizontal evacuation, tsunami mitigation, run up, Parangtritis.

* Geography Faculty, Universitas Gadjah Mada Yogyakarta, Indonesia.** Research Center for Disaster, Universitas Gadjah Mada Yogyakarta, Indonesia.*** National Coordination Agency for Surveying and Mapping (BAKOSURTANAL), Cibinong, Indonesia.


D. Mardiatno et al.

For tsunami hazard assessment, run up scenarios weredetermined by selecting 1 m, 2 m, 5 m, 10 m and 15 m ofheight. The 15 m value was estimated as the possible maxi-mum run up with regard to historical tsunami data in southJava. Evacuation time was calculated by referring to Sugi-moto et al. (2003), estimating the walking speed for groupof elderly people is 0,75 m/sec

Results and Discussion

Based on the tsunami hazard analysis result, some loca-tions will be at the high and very high hazard to tsunami for5m, 10m and 15m run up scenario. For 1m and 2m run upscenarios, the lower hazard levels become more dominant.Giyanto et al. (2008) mentioned that 75% of Parangtritisarea is very vulnerable to tsunami disaster with regard tohistorical tsunami database, distance from the shoreline, andmorphological condition of this coastal area.Evacuation effort is one of several alternativesto reduce the tsunami risk.

Time for evacuation was planned for 40 min-utes maximum after the earthquake (GITEWS,in Dewi, 2010). Cahyono (2009) found thatthere are 10 vulnerable points for evacuation incase of tsunami. They are located nearby shore-line, at the middle of the coastal area, and at theside of Opak River. Based on the evacuationroutes proposed by Cahyono (2009), more de-tail routes were generated by utilizing a highresolution image (Fig. 2a). The routes were di-vided into 14 segments, which are explained inTable 1. The detail features of both two sampleplaces are shown in Figure 2b.

As shown in Figure 2 (b), it is necessary toidentify more narrow pathways for evacuationroutes. Although several pathways are very dif-ficult interpreted from the image, they can beidentified from field work. They are always as-sociated to the settlements or main road. These

pathways are located either on the flat or hilly topographyand they have a relative good condition.

Furthermore, according to Table 1, it can be seen that seg-ment 2 has the longest distance to the shelter while segment 3has the shortest. If evacuation will be carried out on the hillytopography, it is assumed that the time will be more 10% fromthe initial estimation. The longest time will be about 53 min-utes and the shortest time is about 4 minutes. This is a roughestimation. If hill slope is steeper, time for people movementon the roads or pathways can be up to 50% from the initialestimation. With regard to GITEWS - Ina TEWS role, thetravel time to the shelter or safe area is about twenty minutes.According to that role, four routes (segment 1, segment 2,segment 8, and segment 13) should be improved by addingsome shelters to reduce the evacuation time.

Local community in this area has a good capacity to responseto the threat such as tsunami (Sunarto et al., 2010). People have

Fig. 1 – Location and landscape of Parangtritis area.

SegmentDistance to shelter

(m)Time to nearest

shelter (min)Topography

1 1315,6 29,2 Flat

2 2199,8 48,9 Flat and hilly



5 330,9 7,4 Flat



8 951,8 21,2 Flat


10 585,4 13,0 Hilly

11 665,5 14,8 Hilly

12 540,9 12,0 Hilly

13 878,5 19,5 Hilly

14 312,3 6,9 Hilly

Tab. 1 – The characteristic of evacuation route segment.


Evacuation Route Determination of Tsunami Mitigation

improved their capacity to reduce the tsunami risk by strength-ening the institutional system as well as the supporting facilities.Thus, improvements on evacuation systems will be more easilyapplied since the people awareness are quite high.

Conclusions

High resolution image can provide useful information forevacuation planning. Some relevant objects can be easilyidentified, such as main roads and wide pathways, althoughnarrow pathways are difficult interpreted from the image.Narrow pathways could be identified during field verifica-tion and then added at re-interpretation phase. Both hori-zontal and vertical evacuations are possibly applied in thisarea. It is necessary to utilize more pathways for verticalevacuation because of their availability at the hillsides.

Acknowledgements

The authors would like to thank to the Head of Parangtri-tis Geospatial Laboratory (Mr Ari Dartoyo) for the partialsupport in conducting field work.

References

Cahyono A. (2009) – Penentuan Jalur Evakuasi Tsunami dalamBerbagai Variasi Ketinggian Gelombang Tsunami di Wilayah Pe-

sisir Bantul Yogyakarta. B.Sc. Thesis, Faculty of Geography, Uni-versitas Gadjah Mada, Yogyakarta.

Dewi R.S. (2010) – A GIS-based Approach to the Selection ofEvacuation Shelter Buildings and Routes for Tsunami Risk Re-duction, a case study of Cilacap Coastal Area, Indonesia. M.Sc.Thesis, Double Degree M.Sc. Program of Universitas GadjahMada and International Institute for Geo-Information Scienceand Earth Observation (ITC), Yogyakarta-Enschede.

Giyanto R.C.S., Santosa L.W., Sartohadi J., Suratman (2008) –Identification of Coastal Area Damage Using Remote Sensingand Geographic Information System in Parangtritis, Yogyakarta,in Umitsu M. and Takahashi M. (eds.), Geomorphological Com-parative Research on Natural Disaster Mitigation in the CoastalRegions of Tropical Asia. Proceedings of Phuket, Ho Chi Minh,and Pattaya Conferences, JSPS Asia and Africa Science Plat-form Program, Nagoya University, Japan.

Mardiatno D. (2008) – Tsunami Risk Assessment Using Scenario-based Approach, Geomorphological Analysis, and GeographicInformation System – A case study in the South Coastal Area ofJava Island, Indonesia. PhD Thesis, Faculty of Geo- and Atmo-spheric Science, University of Innsbruck, Austria.

Sugimoto T., Murakami H., Kozuki Y., Nishikawa K. (2003) – AHuman Damage Prediction Method for Tsunami Disasters Incor-porating Evacuation Activities. Natural Hazards 29, 585–600.

Sunarto, Marfai M.A., Mardiatno D. (eds.) (2010) – MultiriskAssessment of Disasters in Parangtritis Coastal Area. GadjahMada University Press, Yogyakarta.

Fig. 2 – A. Modified evacuation route plan adopted from Cahyono (2009). B. Detail images of two samples in the west part (above)and east part (below).

Introduction

The SE Asia-Pacific region contains the world’s highestconcentration of active or potentially active volcanoes. In-donesia alone contains 76 historically active volcanoes; thehighest concentration of any country, and has more dated erup-tions than any nation except Japan (Simkin and Siebert, 1994).These eruptions include rare, cataclysmic paroxysms such asthe 1883 Krakatau eruption, but also the frequent smaller ex-plosive eruptions that can affect the region’s airspace on aweekly basis. Coupled with the growth in air traffic within andacross the region (e.g., between Japan and Australia), and thepresence of major hub airports such as Singapore, the result isa high vulnerability to the hazards associated with driftingvolcanic eruption clouds. These include direct hazards to avi-ation from drifting volcanic ash, which can melt upon inges-tion into the hot sections of jet engines and cause engine fail-ure. In June 1982, a British Airways Boeing 747 en route fromKuala Lumpur (Malaysia) to Perth (Australia) encountered anash cloud from the erupting Galunggung volcano (West Java),causing a quadruple engine shut-down and a near catastrophe,before the engines were restarted and a safe landing in Jakartawas achieved (Johnson and Casadevall, 1994). This event andother volcanic ash encounters triggered the development of theInternational Airways Volcano Watch (IAVW), which set upthe infrastructure required to provide timely warnings of vol-canic ash hazards to aviation (e.g., Tupper et al., 2007). Satel-lite remote sensing plays a crucial role in tracking volcanicash clouds and here we review recent developments in thefield, providing some specific examples from SoutheastAsia, and also show how satellite data can contribute to rou-tine volcano monitoring in the region.

Satellite remote sensing of volcanicemissions

The synoptic perspective of satellite remote sensing pro-vides the most effective means of detecting and trackinghazardous volcanic clouds, and there are currently numer-ous space-borne instruments capable of measuring volcanicemissions of ash and sulfur dioxide (SO2), some in near real-time (NRT) (Carn et al., 2009). Although SO2 is typicallythe third most abundant gas species in volcanic emissions,after water vapor (H2O) and carbon dioxide (CO2), the lat-ter species are abundant in the ambient atmosphere andhence the volcanic signal is very difficult to isolate via re-mote sensing. However, SO2, which has very few majorsources in most regions other than degassing volcanoes, andwhich also has strong absorption bands at ultraviolet (UV)and infrared (IR) wavelengths, can be easily measured usingremote sensing from a range of platforms (ground-based,airborne or spaceborne). Measurements of SO2 are also ad-vantageous since volcanic SO2 emissions tend to increaseprior to a magmatic eruption as rising magma nears the sur-face, providing some potential for eruption warnings (e.g.,Daag et al., 1996). A significant development over the lastdecade has been the emergence of satellite instruments ca-pable of detecting volcanic SO2 emissions in the lower tro-posphere on a daily basis, such as the UV Ozone Monitor-ing Instrument (OMI) on NASA’s Aura satellite (launched inJuly 2004). This capability permits detection of passive vol-canic degassing (e.g., Fig. 1), which broadens the applica-tions of satellite measurements to include routine volcanicsurveillance (e.g., Carn et al., 2008). Satellite data have alsobeen used to quantify hydrogen chloride (HCl) and brominemonoxide (BrO) in volcanic emissions (e.g., Theys et al.,


Remote sensing of volcanic emissions in the Asia-Pacific region

S.A. Carn*

Short abstract: Numerous satellite sensors now provide SO2 measurements, some with sensitivity to passive degassing. Satellite con-stellations (e.g., the A-Train) permit sensor synergy and 3D analysis of volcanic emissions. Tropical volcanic SO2 emissions areimpacted by sub-surface scrubbing and atmospheric cloud-processing. CO2 and H2S are better candidates for geochemical monitoring,but are difficult to measure by remote sensing. Satellite SO2 data provided critical observations during the 2010 Merapi and 2005 Anata-han eruptions.

Keyword: remote sensing, sulfur dioxide, volcanic emissions, Merapi, Anatahan.

*Department of Geological and Mining Engineering and Sciences, Michigan Technological University, Houghton, MI 49931, USA.


S. A. Carn

2009), but these measurements are relatively rare and notuseful for routine monitoring. Prata (2009) provides anoverview of the techniques and sensors used to retrieve vol-canic ash abundance from satellite measurements.

As mentioned above there are numerous UV and IR in-struments on polar-orbiting and geostationary satellites capa-ble of quantifying volcanic SO2 emissions (including OMI,GOME-2, SCIAMACHY, AIRS, IASI, MODIS, ASTER andSEVIRI), which in concert provide several overpasses perday and improve temporal resolution. A further recent devel-opment has been the deployment of a satellite constellationcalled the A-Train (the ‘A’ refers to the local afternoon over-pass time of the satellites), consisting (currently) of 4 space-craft spaced a few minutes apart in polar orbit at ~700 km al-titude. The satellites (Aura, CALIPSO, CloudSat and Aqua)carry several sensors capable of mapping volcanic SO2 andash emissions (including OMI, AIRS and MODIS), in addi-tion to a space-borne lidar (CALIOP on CALIPSO) and radar(CloudSat) that can provide vertical profiles through volcanicclouds. The proximity of the A-Train satellite orbits providesnear-coincident measurements from each instrument. Syner-gy between UV measurements of SO2, which are sensitive tothe total atmospheric column, and IR soundings at wave-lengths of 7.3 μm, which are sensitive to the upper tropo-sphere and lower stratosphere, provides information on SO2altitude and hence potential hazards and climate impacts.The result is a new capability for 3D visualization of vol-canic clouds and multi-spectral data synergy, given optimaltiming of the A-Train measurements relative to the eruptiontime. The A-Train observations of volcanic cloud altitude areparticularly valuable given the importance of this parameterfor aviation hazards. Synergy between A-Train sensors hasalso been valuable for validation of retrievalalgorithms.

There are substantial challenges involvedwith remote sensing of volcanic emissionsin the moist tropics. These include abundantatmospheric water vapor, which can com-pletely obscure the volcanic emissions (e.g.,through deposition of ice onto volcanic ashparticles; Rose et al., 1995), and near-per-manent meteorological cloud cover, whichobscures active volcanic vents (orographicclouds) and may prevent detection of pre-cursory eruptive activity. Furthermore, con-vective instability in tropical atmospheres

can amplify the column height and resulting impact of smallvolcanic eruptions through release of latent heat (Tupper etal., 2009). Solutions to these problems include using SO2measurements as a proxy for volcanic ash, since SO2 is typ-ically less affected by water vapor interference, or exploit-ing the effects of volcanic aerosol on cloud microphysics.Volcanic SO2 emissions indicate shallow magmatic de-gassing and hence also provide information on pre-eruptiveunrest, and therefore may be used for eruption forecasting.UV satellite measurements have sufficient sensitivity to de-tect such quiescent SO2 degassing, and thus contribute tovolcano monitoring (Fig. 1). However, many volcanoes inthe moist tropics have large hydrothermal systems, whichcan sequester SO2 (known as ‘scrubbing’) and complicatethe use of this gas as an indicator of volcanic unrest (e.g.,Symonds et al., 2001). Other less soluble volcanic gases,such as hydrogen sulfide (H2S) and carbon dioxide (CO2),may be more appropriate indicators of magmatic degassingin the tropics but are not easily measured via remote sensing(e.g., O’Dwyer et al., 2003), and must be quantified by di-rect sampling of volcanic gases. After emission, SO2 even-tually converts to sulfate aerosol, which can impact air qual-ity and visibility downwind of degassing volcanoes and is ahealth hazard. Monitoring of tropospheric SO2 emissionsfrom volcanoes and the derived sulfate aerosol is required tomitigate these hazards.

An analysis of OMI SO2 measurements for Indonesia col-lected over the duration of the Aura satellite mission (sinceSeptember 2004) reveals that detected SO2 emissions fromthe Indonesian archipelago are relatively low compared tosome other volcanic regions, despite the high levels of vol-canic activity. This is likely due in part to the persistent

Fig. 1 – Ozone Monitoring Instrument (OMI)SO2 data for the Southwest Pacific region onApril 23, 2006. SO2 column amounts are reportedin Dobson Units (DU) where 1 DU = 2.69×1016

molecules cm-2 = 0.0285 g m-2 SO2. In this singleimage, tropospheric SO2 plumes are visible ema-nating from (north to south) Anatahan (Mariana Is),Manam, Ulawun and Bagana (Papua New Guinea),Tinakula (Solomon Is), Gaua and Ambrym (Vanu-atu). OMI has a spatial resolution of 13×24 km,which is sufficient to resolve plumes from individu-al volcanic sources.



cloud cover on many Indonesian volcanoes, which masksemissions at lower altitudes and also removes emitted SO2from the atmosphere rapidly via aqueous phase reactions incloud water (e.g., Carn et al., 2011). Significant scrubbingof SO2 in the hydrothermal systems of many volcanoes priorto emission is also probable, and it is likely that a consider-able fraction of the sulfur budget of Indonesian volcanoes isemitted in the form of H2S or other reduced sulfur species.However, the OMI SO2 measurements do permit identifica-tion of the strongest volcanic SO2 sources in Indonesia in2004-2011, which include Soputan (Sulawesi), Kawah Ijen(Java) and Bromo (Java). Soputan in particular is character-ized by frequent small-to-moderate explosive eruptions(with no precursory activity apparent in the OMI SO2 data)that represent a significant aviation hazard in the region. Themajor eruption of Merapi (Java) in November 2010 alsoresulted in large SO2 emissions and is discussed furtherbelow. When comparing time-averaged SO2 emissions inIndonesia measured by OMI with model simulations ofSO2burdens initialized with existing emissions inventories(e.g., Pfeffer et al., 2006), we find notable discrepancieswhich suggest that the satellite measurements could con-tribute to more accurate emissions inventories.

The November 2010 eruption of Merapi (Java)

The explosive eruption of Merapi volcano (Java) in Octo-ber-November 2010 was Merapi’s largest since 1872, cate-gorized with a Volcanic Explosivity Index (VEI; Newhalland Self, 1982) of 4 (~0.1 km3). The explosive phases of theeruption began in late October and culminated with theparoxysmal event on November 4-5. Ash emissions fromthe eruption resulted in major disruption to domestic and in-ternational flights, but were difficult to track using conven-tional satellite remote sensing techniques due to abundantmeteorological clouds and ice-rich volcanic clouds. Satellitemeasurements of SO2 from OMI, the Infrared AtmosphericSounding Interferometer (IASI on MetOp) and the Atmo-spheric Infrared Sounder (AIRS) on Aqua proved valuableduring the Merapi eruption since, unlike ash particles, SO2does not nucleate ice. Hence even ice-rich volcanic cloudstypically maintain a strong SO2 signature that can be used foreruption detection and cloud tracking. The SO2 data weretherefore used by proxy to monitor eruption intensity, and dis-played a good correlation with coincident seismic data. An in-crease in SO2 production detected from space prior to the

Fig. 2 – Total attenuated backscatter at 532 nm measured by the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP)aboard the CALIPSO satellite on November 5, 2010 at ~19:30 UTC. Latitudes and longitudes of this lidar curtain are shown along the x-axis. Aerosol in the Merapi volcanic eruption cloud (yellow / orange colors) is detected near the tropopause (white solid line) at altitudes of~13-17 km. Depolarization measurements indicate that the volcanic aerosol particles are solid, and could be ice, solid sulfate, ash, or a mix-ture of these. Attenuating features at lower altitudes are meteorological clouds.


S. A. Carn

paroxysmal event on November 4-5 contributed to a fore-cast of the eruption, which permitted timely evacuations ofthe volcano’s flanks that saved many lives. The Merapieruption produced a total of ~0.2-0.3 Tg of SO2; the largestSO2 release measured from an Indonesian volcano since the1988 eruption of Banda Api, detected by the Total OzoneMapping Spectrometer (TOMS). The Microwave LimbSounder (MLS) instrument (also on the Aura satellite withOMI) was also able to detect HCl in the Merapi volcaniccloud, which is consistent with the Cl-rich composition ofMerapi volcanic gases measured by direct sampling. Furtherdetails are provided in Surono et al. (2011). A-Train measurements of the Merapi volcanic cloud provided

some unique constraints on the altitude of the cloud, whichreached the tropopause at ~17 km (Fig. 2). Solid aerosol particleswere detected in the volcanic cloud by the Cloud-Aerosol Lidarwith Orthogonal Polarization (CALIOP) aboard the CALIPSOsatellite, but the composition of these particles remains unclear.Ash was not explicitly detected in the volcanic cloud using otherremote sensing techniques, but ice nucleated on ash is a possibleexplanation. Sulfate aerosol in the solid phase could perhaps ex-plain the observations. To attempt to resolve this issue we (in col-laboration with J. Wang, University of Nebraska

– Lincoln) are conducting simulations of the Merapi erup-tion using the GEOS-Chem chemistry-transport model (CTM;http://www-as.harvard.edu/chemistry/trop/geos/), which in-cludes advanced treatment of sulfate aerosol compositionand phase (Wang et al., 2008). The CTM is initialized usingaccurate estimates of volcanic SO2 emission magnitude andaltitude based on A-Train satellite observations, and themodel then simulates the sulfate aerosol evolution with time.By comparing the modeled sulfate aerosol optical depth(AOD) with the CALIOP and other aerosol observations, wehope to gain insight into the composition of volcanic cloudsand better interpret the satellite measurements. This will im-prove our understanding of volcanic cloud hazards to avia-tion and also permit calculation of the radiative forcing as-sociated with such eruptions.

The 2005 eruption of Anatahan(Mariana Islands)

The April-May 2010 eruption of Eyjafjallajökull (Iceland)highlighted the vulnerability of countries that depend on airtransport to volcanic cloud hazards. Given that SoutheastAsia contains some of the world’s most volcanically active

Fig. 3 – SO2 emissions from Anatahan (Mariana Islands) measured by OMI from January-September 2005. Also shown are Real-timeSeismic Amplitude (RSAM) data from a seismic station on the volcano (courtesy of the U.S. Geological Survey), estimated plume altitudes(courtesy of M. Guffanti, USGS), and periods impacted by Pacific ocean typhoons.



regions, one may ask the question: could an Eyjafjalla-jökull-type event could occur in the region? Although thedensity of air routes in Southeast Asia is considerably lowerthan in mainland Europe, there is still significant potentialfor disruption at busy hubs such as Singapore, with the vol-canoes of Sumatra situated to the west. However, some ofthe factors that exacerbated the impact of the Eyjafjalla-jökull eruption, such as a persistent northwesterly wind as-sociated with a high-pressure system that transported vol-canic ash towards Europe, are less probable in the tropicswhere meteorology (at low altitudes) is dominated by con-vection and vertical transport and there are no synoptic pres-sure systems. Small eruptions in the tropics result in greatercolumn heights relative to drier environments due to the im-pact of latent heat release from condensing water vapor(Tupper et al., 2009), so even a modest eruption (of a simi-lar intensity to Eyjafjallajökull) would produce a relativelyhigh column and could promote rapid long-range transportaway from the region.

One recent period of volcanic unrest that resembles theEyjafjallajökull event in some respects is the activity ofAnatahan volcano (Mariana Islands) in April-August 2005.Anatahan, which erupted for the first time in recorded his-tory in May 2003, is remote and uninhabited and most ofthe observations of this activity are derived from satelliteremote sensing. An explosive eruption on April 5-6, 2005presaged an extended period of ash and gas emissions thatcontinued for ~5 months (Fig. 3; the Eyjafjallajökull erup-tion lasted ~6 weeks). Emissions of SO2 from Anatahan dur-ing this period were tracked by OMI on a daily basis andshowed long-range transport of volcanic plumes to Japan, thePhilippines, and beyond. Volcanic ash was likely removedfrom the plumes relatively rapidly, so the degree of distaltransport of volcanic ash is unclear, but it is clear that anyimpact of the eruption was mitigated by its remote location,far from major population centers. However, the volcanicplumes caused significant reductions in surface visibilitythroughout the western Pacific, impacting shipping, andsulfate aerosol derived from Anatahan’s SO2 emissionswas apparently detected by sun photometers in Taiwan,some 3000 km downwind (part of the AERONET network;http://aeronet.gsfc.nasa.gov). This event demonstrates thatlong-lived eruptions characterized by continuous ash andgas emissions have the potential to impact air quality atgreat distances from the source, and can also impact ship-ping operations in addition to the recognized hazards toaviation.

Conclusions

The capabilities of satellite remote sensing techniques haveadvanced significantly in the past decade, and satellite obser-vations now play a key role in monitoring volcanic SO2 and ashemissions. Tropical environments provide particular challengesfor remote sensing, but the examples shown here demonstratethat critical observations can still be made, resulting in valuablecontributions to hazard mitigation efforts and new insights intothe composition and fate of volcanic emissions.

References

Carn S.A., Krueger A.J., Krotkov N.A., Arellano S., Yang K.(2008) – Daily monitoring of Ecuadorian volcanic degassingfrom space. Journal of Volcanology and Geothermal Research176(1), 141-150, DOI:10.1016/j.jvolgeores.2008.01.029.

Carn S.A., Krueger A.J., Krotkov N.A., Yang K., Evans K.(2009) – Tracking volcanic sulfur dioxide clouds for aviationhazard mitigation. Natural Hazards 51(2), 325-343, DOI:10.1007/s11069-008-9228-4.

Daag A. S. et al. (1996) – Monitoring sulfur dioxide emission atMount Pinatubo. In: Newhall C.G., and Punongbayan R.S., (eds.),1996, Fire and Mud – Eruptions and Lahars of Mount Pinatubo,Philippines. Philippine Institute of Volcanology and Seismologyand the University of Washington Press, 1126 p.

Hair J.W., Diskin G., Sachse G., Vay S. (2011) – In-situ measure-ments of tropospheric volcanic plumes in Ecuador and Colombiaduring TC4. Journal of Geophysical Research 116, D00J24, DOI:10.1029/2010JD014718.

Johnson R. W.,Casadevall T. J. (1994) – Aviation safety and vol-canic ash clouds in the Indonesia-Australia region. Proceedingsof the First International Symposium on Volcanic Ash and Avia-tion Safety, Seattle, WA, Office of the Federal Coordinator forMeteorological Services and Supporting Research, 191–197.

O’Dwyer M., Padgett M. J., McGonigle A. J. S., OppenheimerC., Inguaggiato S. (2003) – Real-time measurement of volcanicH2S and SO2 concentrations by UV spectroscopy, GeophysicalResearch Letters 30(12), 1652, DOI: 10.1029/ 2003GL017246.

Pfeffer M. A., Langmann B., Graf H.-F. (2006) – Atmospherictransport and deposition of Indonesian volcanic emissions. At-mospheric Chemistry and Physics 6, 2525-2537.

Prata A. J. (2009) – Satellite detection of hazardous volcanicclouds and the risk to global air traffic. Natural Hazards 51(2),303-324.

Rose W. I. et al. (1995) – Ice in the 1994 Rabaul eruption cloud:Implications for volcano hazard and atmospheric effects. Nature375, 477–479.

Simkin T., Siebert L. (1994) – Volcanoes of the World, 2nd Edition.Geoscience Press, Tucson, AZ.

Surono, Jousset P., Pallister J., Boichu M., Buongiorno M.F.,Budisantoso A., Costa F., Andreastuti S., Prata F., SchneiderD., Clarisse L., Humaida H., Sumarti S., BignamiC., Gris-wold J., Carn S., Oppenheimer C. (2011) – The 2010 explosiveeruption of Java’s Merapi volcano – a ‘100-year’ event, Journalof Volcanology and Geothermal Research (in review).

Symonds R. B., Gerlach T. M., Reed M. H. (2001) – Magmaticgas scrubbing: implications for volcano monitoring. Journal ofVolcanology and Geothermal Research 108, 303-341.

Theys N.,Van Roozendael M., Dils B., Hendrick F., Hao N., DeMazière M. (2009) – First satellite detection of volcanicbromine monoxide emission after the Kasatochi eruption, Geo-physical Research Letters 36, L03809, DOI:10.1029/2008GL036552.

Tupper A., Itikarai I., Richards M, Prata A.J., Carn S.A.,Rosenfeld D. (2007) – Facing the challenges of the Interna-tional Airways Volcano Watch: the 2004/05 eruptions ofManam, Papua New Guinea. Weather and Forecasting 22(1),175-191.


S. A. Carn

Tupper A., Textor C., Herzog M., Graf H., and Richards M. S.(2009) – Tall clouds from small eruptions: the sensitivity oferuption height and fine ash content to tropospheric instability.Natural Hazards 51, 375-401.

Wang J., Hoffmann A. A., Park R., Jacob D. J., Martin S. T.(2008) – Global distribution of solid and aqueous sulfate aerosols:effect of the hysteresis of particle phase transitions. Journal of Geo-physical Research 113, D11206, DOI: 10.1029/2007JD009367.

Abstract: Volcanoes are highly variable but continuous emitters of gaseous compounds, during both episodic eruptions and quiescent contin-uous activity, from their vents and craters as well as their flanks. To monitor their emissions, spectroscopic methods have long been used fromthe ground, from aircraft, and from space, mainly measuring sulfur dioxide (SO2). However, fewer studies have focused on the more abundantand stable carbon dioxide (CO2). The application of such observations is both in improving the currently very poorly understood global CO2

source strength estimates for volcanoes, as well as in volcano monitoring, because volcanic CO2 emissions before eruptions are the potential-ly earliest indicators of unrest available. Early detection of anomalies permits ground-based monitoring at the best place and time because intimes of crisis, the time is extremely short for decisions, validation and response.The rationale for using CO2 as an early unrest indicator is complex: Among the first potential signals of ascending magma is the exsolu-tion of volatiles contained in magma induced by dynamic depressurization, crystallization, and temperature variations. The three most abun-dant gas species in these emissions are usually water (H2O), CO2, and SO2. SO2 monitoring methods are widespread, using COSPEC, mini-DOAS, SO2 cameras, and space-borne SO2 data. However, since H2O and SO2 are frequently scrubbed out by near-surface processes, theymay be obscured unless the magma is already near the surface. SO2 is most useful for volcanoes that erupt frequently and have a drychimney for easy gas escape. CO2 is more difficult to measure remotely than SO2 because the atmospheric background concentration ofCO2 is so much higher than for SO2. Nevertheless, CO2 is important because it is the first gas to exsolve from magma (together withhelium), and it is minimally affected by scrubbing and other near-surface processes. CO2 monitoring has been attempted by ground-based CO2 flux monitoring and by crater plume CO2 measurements using ground-based open-path FTIR and airborne closed-path IRmeasurements.Recent advances in satellite-borne measurements of atmospheric CO2 have resulted in smaller field of views, approaching the domainsize necessary to detect and quantify emissions from volcanoes. The Japanese GOSAT instrument aboard the IBUKI satellite has beenoperational since January 2009, producing CO2 total column measurements at a repeat cycle of 3 days and a field of view of 10km.GOSAT has thus the potential to provide spatially integrated data for entire volcanic edifices. In target mode, repeat observation requestshave a great potential to detect volcanic anomalies. We present data of target mode observation requests on a number of selected vol-canoes world-wide, using GOSAT FTS SWIR data. We discuss a new comprehensive approach to instrumental gas monitoring and datatreatment, and our progress on implementation in the Philippines and Indonesia.

Remote sensing, natural hazards and environmental changee, p. 61-62

Monitoring carbon dioxide emissions from volcanoesfrom space and from ground based networks

F.M. Schwandner*, C.G. Newhal*, S.S. Marcial*

Detection of volcanic dust by AERONET sunphotometers

S. Salinas** and S.C. Liew**

Abstract: Aerosol optical depth and its first and second spectral derivatives respect to wavelength, are often used to describe the interaction ofaerosol particles present on a given size distribution. During volcanic eruptions, large amount of dust particles are emitted into the atmosphe-re. At source eruption sites, it might be difficult if not impossible to apply photometric measurements due to the high optical depth of volcanicdust and the difficulty of tracking the Sun. However, at receptor sites, located at places far enough to allow photometric measurements, it ispossible to quantify the particle size distribution of volcanic dust from direct Sun photometry.

* Earth Observatory of Singapore, Nanyang Technological University, Singapore.**CRISP, National University of Singapore, Singapore.

IntroductionIn 1994 several industrial companies in Muntinlupa and

vicinities (Fig. 1) complained about fissures on the walls andfloors of their establishments prompting government authori-ties to map out the area. By the end of 1999, mapping in theMuntinlupa revealed ground rupture with a total length ofabout 2.6 km with additional ground fissures with lengthsof about 50 m to 1.4 km trending north-northeast(Oquendo, 2004). Movement in Muntinlupa wasidentified to be along the trace of the West MarikinaValley Fault with vertical fractures offsets rangingfrom 12 to 116 cm with an average of about 50 cm.Horizontal displacements were up to 0.25 m. InBiñan and San Pedro, Laguna, further south ofMuntinlupa and west of Laguna de Bay, the sameground rupture features were identified cuttingthrough houses in villages. Field work in the area re-vealed up to 50 cm vertical displacement with later-al offset up to 46 cm over a 21 year period.

With measured displacement rates faster than creepmovement in other known fault areas, authoritieswere prompted to suspect that the observed groundrupture is aided by excessive ground water extraction(Ramos, 1998). It is still unknown today whetherground rupture in the southern Metro Manila is pri-marily due to tectonics or excessive groundwaterextraction. To contribute in the discussion on thenature of the observed fractures, this work providesadditional information in terms of measuring move-ment in the area of Muntinlupa down to San Pedrousing Permanent Scatterer Interferometry. The studycomplements leveling measurements, differentialGlobal Positioning System (dGPS) and conventional

interferometric measurements conducted in the region (Deguchiet al., 2008).

Methodology

A total of 24 descending ENVISAT level 0 data, spanningthe period from 2003-2006 were processed using MAIN-


PSInSAR detection of ground subsidence and faultmovement in Muntinlupa City, MM and Biñan, Laguna

A.M.F. Lagmay*, R. Eco*, J. Adeppa*

Abstract: Permanent Scatterers Interferometric Synthetic Aperture Radar (PSInSAR) is a radar processing method for quantitative, highprecision, multi-scale monitoring of land deformation. In the PSInSAR approach multiple acquisitions gathered repeatedly over time inthe same target area are processed. For this work, we used 24 ENVISAT descending imageries from 2003 to 2006 processed with MultiAcquisition InSAR (MAINSAR). Using this technique, we were able to resolve ground surface motions in the area of West Laguna alongthe trace of the West Marikina Valley Fault, including motions of individual targets, not previously recognised by traditional InSAR. Theline-of-sight (LOS) movement measured in Muntinlupa City, MM and Biñan, and San Pedro, Laguna using PSInSAR is measured tohave rates of about 15-25 mm/year and is used to complement the existing knowledge on ground deformation of the Marikina ValleyFault System in this area.

Keywords: PSInSAR, ground subsidence, Marikina Valley Fault, fault movement, permanent scatterer interferometry.

* National Institute of Geological Sciences, College of Science, University of the Philippines, Diliman, Quezon City 1101, Philippines.

Fig. 1 – Index map of the study area. Muntinlupa is the southernmost cityof Metro Manila. The white hashed trace is the Marikina Valley Fault.


A.M.F. Lagmay, R. Eco and J. Agdeppa

SAR to extract ground displacements for persis-tent scatterer pixels from multiple syntheticaperture radar acquisitions (Hooper et al., 2004;Hooper, 2006). Prior to the persistent scattererprocessing step, the raw data were focused intosingle-look complex (SLC) images using Delftprecision orbits with the Repeat Orbit Interfer-ometry package (ROI PAC; Rosen et al., 2004).Interferograms were then generated from theSLC dataset with DORIS (Kampes et al., 2003)using a master image selected on the basis ofminimizing perpendicular, Doppler, and tempo-ral baselines (Hooper, 2006). The digital eleva-tion model (DEM) used for removing the topo-graphic signal in the interferogram is a 3 arc sec-ond Shuttle Radar Topographic Mission (SRTM)image interpolated to 30 m postings. The DEMswere downloaded from the USGS seamless datadistribution system (USGS, 2004-2007).

Results

The interferograms for the descending im-ageries show coherence in the urban areas of the cities ofMuntinlupa, San Pedro, Biñan, Paranaque, and Cavite (Fig.1). Permanent scatterers in the time-series interferogramsfrom 2003-2006 show distinct movement as suggested by thechange in the line-of-sight of the radar sensor, which looksdown at an incidence angle of 23.5º from the east (Fig. 2).

Noticeable in the velocity map of permanent scatterersimaged over the period 2003-2006, is a red bowl-shapedfeature in Paranaque (Figure 3). This red-colored area showsan increase in the line-of-sight (LOS) of the radar satellite tothe permanent scatterers by as much as 27.8 mm/year. Otherareas that depict a strong change in the line-of-sight of theradar satellite (red dots) are near the trace of the WestMarikina Valley Fault (WMVF) in Muntinlupa, San Pedroand Biñan. These places show an increase in LOS, whichappears to be associated with the en-echelon trace of theWMVFS (Fig. 3) as mapped out by Phivolcs. Other areasalong the WMVFS trace that show an increase in LOS butdo not have a corresponding fault trace may mean that theseplaces were not mapped and need to be checked in the field.Near the trace of the WMVFS, increase in LOS rates is asmuch as 15-25 mm/year.

If movement seen in the PSInSAR is purely vertical, theserates of movement amount to nearly 30-50 cm of groundsubsidence over a period of 20 years along the fault trace. Insuch a case, the measured rates of movement as seen by theradar satellite would be comparable to the field-measuredvertical displacements in villages that straddle the WMVFS.

However, it is still not possible to determine whether thesechanges in LOS are the same as subsidence since more radarimages need to be processed. In particular, PSInSAR-pro-cessed ascending radar satellite images are needed, to assesslateral movement and its contribution to changes in LOS.The most that can be surmised at this point is that the sig-

nificant change in LOS (red areas in Fig. 3) in Muntinlupa,Biñan, and San Pedro seem to correspond to fractures andfissures of the WMVFS.

Discussion and Conclusions

The PSInSAR processing demonstrates that it is possibleto determine ground deformation in the area of Muntinlupa,Biñan and San Pedro. The reported measurements werederived from processing descending ENVISAT imageriesfrom 2003-2006 using an SRTM DEM.

The results can be further improved with the addition ofascending imageries. The use of higher resolution topogra-phy with horizontal postings of less than 20 m and verticalaccuracy not exceeding ± 10 m is also desired. A DEM gen-erated from 1:10,000 scale topographic maps is intended to beused in future processing of the ENVISAT imageries. Further-more, it is necessary to process ENVISAT ascending imageriesto assess the contribution of lateral movement to the increasein LOS of the radar signal. Also, ERS1 and ERS2, as well asALOS PALSAR can be used in future PSInSAR processingto extend the time frame of deformation analysis in the studyarea. ERS1 and ERS2 radar imageries are available as earlyas 1992.

References

Deguchi T., Kinugasa Y., Omura M. (2008) – Monitoring of Val-ley Fault system and land subsidence in Metro Manila, Republicof the Philippines by InSAR and leveling surve. International Ge-ological Congress 2008, Oslo.

Colesanti C., Ferretti A., Novali F., Prati C., Rocca F. (2003) –SAR Monitoring of Progressive and Seasonal Ground Deforma-tion Using the Permanet Scatterers Technique, IEEE Transactionson Geoscience and Remote Sensing 41, 7, 1685-1701.

Fig. 2 – Velocity of permanent scatterers for the ascending data. Increase inline-of-sight of the radar satellite is as much as 27.8 mm/year and in areas corre-sponding to the enchelon trace of the WMFVS as much as 15-25 mm/year. The redtrace is the official trace of the Marikina Valley Fault according to Phivolcs. Theradar satellite is looking downward to the west.


PSInSAR detection of ground subsidence and fault movement in Muntinlupa City, MM and Biñan, Laguna

Ferretti A., Prati C., Rocca F. (2001) – Permanent Scatterers inSAR Interferometry, IEEE Transactions on Geoscience andRemote Sensing 39, 1, 8-20,.

Ferretti A., Prati C., Rocca F. (2000) – Non-linear SubsidenceRate Estimation Using Permanent Scatterers in Differential SARInterferometry, IEEE Transactions on Geoscience and RemoteSensing 38, 5, 2202-2212.

Ferretti A., Novali F., Bürgmann R., Hilley G., Prati C. (2004)– InSAR Permanent Scatterer Analysis Reveals Ups and Downsin San Francisco Bay Area. EOS 85, 34.

Massonnet D., Rossi M., Carmona C., Adragna F., Peltzer G.,Feigl K., Rabaute T. (1993) – The displacement field of the

Landers earthquake mapped by radar interferometry. Nature364, 138-142.

Oquendo O. (2004) – Earthquake disaster preparedness plan.Muntinlupa City Disaster Coordinating Council.

Ramos E.G. (1998) – Subsidence, a serious hazard caused bygroundwater extraction in Metro-Manila. Manuscripts Volume,11th Annual Convention, Geological. Society of the Philippines,18–30.

Zebker H. A., Rosen P. A., Hensley S. (1997) – Atmospheric ef-fects in interferometric synthetic aperture radar surface defor-mation and topographic maps. Journal of Geophysical Research102(B4), 7547-7563.

Abstract: Deforestation in the world’s tropics is an urgent international issue. One response has been the development of satellite based mo-nitoring initiatives largely focused on the carbon rich forests of western Indonesia. In contrast this study focuses on one eastern Indonesiandistrict, Kabupaten Kupang, which has some of the largest and least studied tracts of remaining forest in West Timor. This study used a com-bination of remote sensing, GIS and social science methods to describe the state of forests in Kupang district, how and why they are chan-ging. Through integrating satellite imagery, case studies and on-ground interviews, this study explores the proposition that local social, cul-tural and biophysical knowledge is important for effectively using remotely sensed data as a tool to inform management. When compared tosome other parts of Indonesia, the rate and extent of deforestation in West Timor was found to be relatively small and a satellite based as-sessment alone could conclude that it is not a critical issue. However when coupled with on-ground social data a more complex picture emer-ged related to key livelihood issues. The causes of forest cover change were found to be multivariate and location specific, requiring mana-gement approaches tailored to local social issues. This finding has implications in view of the substantial resources being devoted to produ-cing satellite derived forest cover change metrics which generally provide little insight into underlying dynamics. This study suggests that in-tegrative research incorporating local knowledge is key to maximizing the utility of satellite data for understanding causation and effective-ly informing management strategies.


Tropical forest monitoring, socializing the pixel to inform management and livelihood implications:

A case study from Indonesian West Timor

R. Fisher*

* Research Institute for Environment and Livelihoods, Charles Darwin University, Australia.** CRISP, National University of Singapore, Singapore.

El Niño and rainfall influence on the temporal and spatial patterns of vegetation fires

in insular Southeast Asia

S.C. Liew** and J. Miettinen**

Abstract: Vegetation fires are regular occurrences in the Southeast Asian region. Fire is traditionally used as a tool in land clearing by far-mers and shifting cultivators. However, the small scale clearing of land is increasingly being replaced by modern large-scale conversion offorests into plantations/agricultural land, usually also by fires. The fires get out of control in periods of extreme drought, especially duringthe El Nino periods, resulting in severe episodes of transboundary air pollution in the form of smoke haze. Vegetation fires in SoutheastAsia is also a significant source of carbon emission. In this paper, we use the MODIS thermal anomaly product (hotspots data) to study thetemporal and spatial patterns of vegetation fires in the western part of Insular Southeast Asia for a decade from 2001 to 2010. Fire occur-rence exhibits a negative correlation with rainfall, and is more severe overall during the El-Nino periods. However, not all regions areequally affected by El-Nino. In Southern Sumatra and Southern Borneo the correlation with El-Nino is high. However, fires in some regionssuch as Riau, Jambi and Sarawak do not appear to be influenced by El-Nino. These regions are also experiencing rapid conversion of forest,especially those in peat areas, to large scale plantations.

Introduction

Wetlands have been impacted globally as humans need formore land and resources increased to meet the growing de-mand for food, housing and recreation. This global trend ofwetland deterioration is mainly driven by agricultural andurban area expansion. In south Florida, the conversions ofthe wetlands began over the last century mainly for floodcontrol and expansion of the agricultural and urban areas.

The conversion of wetlands to agricultural and urban areasin the early 19th century, the channelization of the KissimmeeRiver between 1962 and 1970 for flood control, and the cur-rent restoration activities have caused ecologicaldisturbances. Successful restoration of the ecosys-tem requires achieving ecohydrological integrity,the ecosystem’s capability to support biodiversity,structural and functional organization comparableto that of the natural habitat of the region throughhydrologic and vegetation restoration.

The Kissimmee River Restoration Project start-ed with the goals of reversing the environmentaldamage that was brought by the channelization inorder to restore the once thriving headwaters of theEverglades ecosystem (Stover, 1992). The projectaimed at restoring over 102 km2 of river/floodplainecosystem including 69 km of meandering riverchannel and 10, 900 ha of wetlands.

The main objective of this analysis was to usethe MODIS based products (surface temperature,albedo, emissivity and normalized differencevegetation index (NDVI)) to compute latent heatflux using surface energy flux modeling ap-proach and characterize the change in latent heatchanges over the period of the Kissimmee Riverrestoration.

Study Area

The study was conducted in the Kissimmee River basin lo-cated north of Lake Okeechobee in South Florida (Figure 1).

Latent Heat Mapping

Remote sensing-based evapotranspiration (ET) estima-tions using the surface energy budget equation are proven tobe one of the most recently accepted techniques for areal ETestimation for large areas. Surface Energy Balance Algo-rithms for Land (SEBAL) is one of such models utilizing


Assessing the hydrologic response of wetlands torestoration: A remote sensing perspective

A.M. Melesse* and F. Miralles-Wilhelm**

* Department of Earth and Environment, Florida International University, Miami, FL, USA.

Abstract: The Kissimmee-everglades natural ecosystems of south Florida have experienced various historical changes that led to theshrinkage of the wetland systems and the current ongoing restoration efforts. The first and main goal of a wetland restoration is to restorethe hydrology and vegetation back to their original condition. In this study, satellite based latent heat flux was used to evaluate the successof restoration of the hydrology and wetland vegetation. Results have shown that increase in latent heat flux was observed for areas whererestoration is taking place.

Key words: Restoration, energy flux, NDVI, Everglades, wetland, remote sensing.

Fig. 1 – The Kissimmee River basin.


A.M. Melesse and F. Miralles-Wilhelm

Landsat, MODIS and ASTER imagesand images from others sensors with athermal infrared band to solve equation(1) and hence generate areal maps of ET(Bastiaanssen et al., 1998a; Bastiaanssenet al., 1998b; Melesse et al., 2006; 2007).

SEBAL requires weather data such assolar radiation, wind speed, precipita-tion, air temperature, and relative humid-ity in addition to satellite imagery withvisible, near infrared and thermal bands.SEBAL uses the model routine ofERDAS Imagine in order to solve thedifferent components of the energy bud-get equations.

In the absence of horizontally advec-tive energy, the surface energy budget ofland surface satisfying the law of con-servation of energy can be expressed as,

Rn – LE – H – G = 0 (1)where Rn is net radiation at the surface,LE is latent heat or moisture flux (ET inenergy units), H is sensible heat flux tothe air, and G is soil heat flux. MODISbased surface temperature, NDVI, albe-do, emissivity were acquired and usedto generate the latent heat for the monthof April for the period 2000-2004.

Results

Latent heat grids were generated from MODIS imageryfor the month of April (2000-2004). Figures 2 show maps oflatent heat in watts per square meter for the month of April.As it is depicted in Figures 2, latent heat values were higherin 2002 and 2004 than 2000 on areas along the rivers. Theremoval of flood control structures and rechannalization ofthe river to its natural course will increase the floodplainarea and in turn higher latent heat flux. It is shown that high-er latent heat flux along the river can be attributed to theincreased flood plain areas and vegetation cover. The rain-fall volume for the month of April (2000, 2002 and 2004)was 40, 10 and 35 mm, respectively.

Conclusion

Response of the Kissimmee basin’s hydrology and vege-tation to the recent restoration was evaluated using datafrom MODIS spatial latent heat flux. The spatial latent heatflux, which is evapotranspiration in energy units, has alsoshown an increase in 2002, 20003 and 2004 compared to2000, which can be attributed to large areas of vegetatedsurface. This change was mainly seen along the river wheremost of the restoration work is going and changes in the

hydrology are expected. Understanding the complete ecohy-drological response of the basin due to the restoration workwill require collection and analysis of vegetation cover atfiner scales than reported in this study.

References

Bastiaanssen W.G.M., Menenti M., Feddes R.A., Holtslag A. A.M. (1998a) – The Surface Energy Balance Algorithm for Land(SEBAL): Part 1 formulation, Journal of Hydrology 212-213,198-212.

Bastiaanssen W.G.M., Pelgrum H., Wang J., Moreno J., Ma Y.,Roerink G.J., van der Wal. T. (1998b) – The Surface EnergyBalance Algorithm for Land (SEBAL): Part 2 validation. Jour-nal of Hydrology 212-213, 213-229.

Melesse A., Nangia V. Wang X., McClain M. (2007) – WetlandRestoration Response Analysis using MODIS and GroundwaterData. SENSORS 7, 1916-1933

Melesse A.M., Oberg J., Beeri O., Nangia V., Baumgartner D.(2006) – Spatiotemporal Dynamics of Evapotranspiration andVegetation at the Glacial Ridge Prairie Restoration. Hydrologi-cal Processes 20(7), 1451-1464

Stover D. (1992) – Engineering the Everglades. Popular Science.46-49 & 94-95.

Fig. 2 – Latent heat flux map of the Kissimmee River for the month of April.

Abstract: Hundreds and thousands of rreservoirs were constructed for flood control, power generation, and water supply etc. However,information such as the location and water storage capacity of these reservoirs is often not available. This talk will introduce our recentwork on remote sensing of reservoirs conducted in large Chinese river basins such as Huanghe, Changjiang and Zhujiang wherereservoirs are extensive. In view of the shortcomings of conventional approaches in locating reservoirs’ spatial location and quantifyingtheir storage volume, remote sensing technique has several advantages, either for a single reservoir or for a group of reservoirs withina basin. The satellite-based remote sensing data, encompassing spatial, spectral and temporal attributes, can provide high-resolutionsynoptic and repetitive information with short time intervals in a large scale. By use of remote sensing images in conjunction withGoogle Earth and field check, the spatial distribution of constructed reservoirs in these large river basins was delineated. Their storagevolume and the residence time of the stored water were estimated. The ultimate goal is to evaluate their impact on water flow andsediment flux changes.


Remote sensing of reservoirs in large Asian river basins

X.X. Lu*, L. Ran*, X. Yang*, S. Liu*

* Department of Geography, National University of Singapore, Singapore.** Charles Darwin University, Australia.

Large floods and average climate: Is there a relationshipin the Asia-Australia monsoon region?

R. Wasson**

Abstract: Are large floods randomly distributed in time or do they follow temporal patterns of average climate? The answer to thisquestion is given particular significance by projections of future climate.Proxy records (speleothems, river sediments, and documents)of past precipitation in the Asia-Australia Monsoon Region (Australia, India and China: AAMR) show an average trend that was wetbetween 1000 and 1200 AD during the Medieval Climate Anomaly (MCA), a transition to a dry climate between 1400 and 1600 AD,and a transition to wet conditions between 1800 AD and the present. In Peninsular India there is a record of large floods inradiometrically dated river deposits (also known as slackwater deposits), from before 1000 to 1240AD, then a complete absence of largefloods until ~1875AD (from the work of Vishwas Kale and co-workers).Thereafter the largest floods on record have occurred. InAustralia slackwater deposits from 7 catchments in the area affected by the monsoon show a different pattern from that in India. Thedry period in the average record of precipitation has a low number of large floods,but there is no equivalent of the MCA pattern of floodsin India.The number of large floods increases from 1500AD and is at its highest between 1900 and 2000AD.The most recent increasein floods is consistent with the record from India.It is important to note that both sets of flood records are non-random but show differentpatterns in relation to the trend of average precipitation. Reasons for this difference will be explored including the completeness of theslackwater record, the limitations of radiocarbon dating, and spatial variability in the climatic reorganization that occurred during theMCA.The recent increase in large floods is consistent with climate projections for a warmer Earth but a more complete understandingof the relationship between large floods and average climate on long time scales is needed to complement atmospheric/oceanicmodelling, and will rely upon better dating and hydraulic modelling of slackwater deposits at more sites in the AAMR.

Changing hazards

It is evident that Europe undergoes rapid changes in terms offast population growth, urbanization, economic developmentand socio- political structures. On top of that, there isconvincing evidence that the emission of greenhousegasses (GHSs) causes changes in the earth’s climatethat are expected to lead to an increase in hazardousevents with a hydro-meteorological trigger. It is the dif-ficulty of the prediction of the magnitude of thesechanges and the frequency of the occurrence of extremeevents that urges a thorough change in our adaptationmanagement of hydro-meteorological risks (EEA,2004). According to Intergovernmental Panel on Cli-mate Change (IPCC) reports climate change is expect-ed to cause a rise in temperature ranging from 2.5 –5.4° C in Europe by 2080 depending on the uncertaintyassociated with the driving forces of global emissionsand the sensitivity of climate models to GHG concen-tration (Christensen et al., 2007). Several studies areavailable that evaluate the impacts of climate change inEurope (e.g. Beniston et al., 2007; Alcamo et al., 2007).Several EU project have studied the possible impacts ofclimate change in Europe (e.g. PRUDENCE, DINAS-

COAST, NewExt and CASH); according to recent studies at aEuropean level the projected impact of flooding in Europewould increase dramatically in the coming decades. By 2080


Spatial information for analyzing changing hydro-meteorological risk

C.J. van Westen*

* Director UNU-ITC School for Disaster Geo-Information Management, Faculty of Geo-Information Science and Earth Observation (ITC), Universityof Twente, Enschede, The Netherlands.

Short abstract: Environmental changes due to global change and resulting reactions in ecosystems, combined with expected changesin socio-economic development will lead to adjustments in land use in areas that are exposed to hydro-meteorological hazards such asflooding, mass movements, severe erosion, snow avalanches and wind storms. These hazards will also have domino effects (e.g. theeffect of land-use change on runoff severe erosion and consequent landslides and river damming leading to flooding) that are still notproperly understood. The effects of these changes need to be analyzed and modeled with probabilistic hazard and risk methods that canbe used by stakeholders from different sectors. The probabilistic models should incorporate the uncertainties in temporal probability,spatial extend and magnitude of the hazards, as well as the uncertainties of the vulnerability of the exposed elements at risk. Themodeled changes in hazard and risk patterns need to be incorporated into disaster risks management strategies and will form animportant factor in land use planning activities at stakeholder relevant levels. They also have a large impact on risk governance policiesthat need to be adapted.This paper gives an overview of work related to the analysis of changes in hydro-meteorological risk in Europe, focusing on analysisof changing hazards, changing exposure of elements at risk, and their vulnerabilities. The structure and workplan of a recent EU FP7Marie Curie Initial Training Network called ‘CHANGES’ is presented. This project (Changing Hydro-meteorological Risks – asAnalyzed by a New Generation of European Scientists) intends to develop an advanced understanding of how global changes (relatedto both environmental and climate change as well as socio-economical developments) will affect the temporal and spatial patterns ofhydro-meteorological hazards and associated risks in Europe, how these changes can be assessed, modeled and how these can beincorporated in sustainable risk management strategies, focusing on spatial planning, emergency preparedness and risk communication.

Key words: Risk, climate change, hazard, landuse, vulnerability.

Fig. 1 – Conceptual framework for the analysis of changing hydro-mete-orological risk.


C. J. van Westen

it is estimated that between 250,000 and 400,000 people areaffected each year by flooding in Europe, with highest con-centration in the British Isles and Central Europe. The total an-nual expected flood damage in Europe is estimated to rangebetween 7.7 and 15 billion Euros. These values are more thandouble of those in the period 1961-1990 (Ciscar, 2009). Thelocal effects of flooding have not been taken into account yet,due to the coarse resolution of the model, and there is a needto downscale such models to make them applicable for localrisk assessment and management.

For mass movements, such global impact studies have notbeen carried out so far. Mass movements are one of themajor soil threats that are considered within the EU The-matic Strategy for Soil Protection (EC, 2006). The directiveindicates that Member States should carry out the identifi-cation of risk areas based on empirical evidence or on mod-eling. However, the focus is more on evaluation of the cur-rent situation than on assessing the changes in risk that arelikely to occur in the coming decades. Landslide suscepti-bility studies for some individual countries have been car-ried out (Malet et al., 2009). Nadim et al. (2006) carried outa general evaluation of landslide susceptibility for the wholeof Europe as part of the Global Hotspots study using a scaleof 1 x 1 km by combining the triggering factors (precipita-tion, human activity, seismicity) and susceptibility factors(slope, lithology, soil moisture, vegetation cover) in a qual-itative manner. General studies (Beniston & Douglas, 2006)indicate that the number, frequency and intensity of massmovements are likely to change, but the variation in the pat-tern is not well defined. The impact of climate change onmass movements is being studied in a number of EU re-search projects (e.g. ClimChAlp in the Alps, SIGMA andGACH2C in France, ESPRC in the UK). So far, the ap-proach to assess the impact of environmental change onlandslide risk has been relatively narrow focused on changesin landslide hazard (e.g., van Beek & van Asch, 2004, Dixon& Brook 2007). This distinct weakness can be addressed byincluding socio-economic change and interactions betweenclimate and land use through scenarios. Very limited workhas been carried out up to now to include the cascading orconjoint (also called domino) effects into account in the anal-ysis of future impacts of environmental changes to hydro-meteorological hazards. For instance through changes invegetation patterns the probability of wild fires may increaseleading to more severe run-off, erosion and mass movementproblems; also the analysis of landslide dams and consequentdam break flooding is an important topic to be considered.

Changing elements at risk

The exposure of elements at risk also increase and there-fore the risk of natural hazards is constantly growing. Landuse changes are predicted for Europe as a result of technolog-ical, socio-economic and political developments as well asglobal environmental change. The type and effects of thesechanges will strongly depend on policy decisions which aregoverned. The recent EU research project ACCELERATEScompared the impact of several scenarios on the prediction of

land use changes in Europe in 2050 (Audsley et al., 2006).EEA (2004) concludes that many environmental problems inEurope are caused by rapidly expanding urban areas. By 2020,approximately 80 % of all Europeans will be living in urbanareas, while in seven countries the proportion will be 90 % ormore. The global economy, cross border transport networks,large scale societal, economic and demographic changes anddifferences in national planning laws are some of the majordrivers of change to the urban environment. Land use changesmay have various detrimental effects on the quality of land-scape and environment. Studies aiming to predict land usechanges are of great use to European policy-makers to antic-ipate such possible prejudicial effects and to engage adaptedactions for their prevention.

Changing vulnerability

The vulnerability to hydro-meteorological hazards of the ex-posed elements has different components (Birkmann, 2006),including the systems or the community’s physical (struc-tural), economic, social and environmental susceptibility todamage. Studies on vulnerability related to environmentalchange indicate that these have a very high level of uncertain-ty. Whereas flood vulnerability has been defined in a rather de-tailed manner (Moel et al., 2009) there are still many uncer-tainties involved. For mass movement there is much lesswork done on defining vulnerability (Glade, 2003), partly dueto the large variation in physical mass movement processes,the difficulty in expressing landslide intensity versus the de-gree of damage, and also related to the purely non existenceof data. Some approaches exist for single elements (e.g Fuchset al. 2007), but an integrated methodology is still lacking.

Uncertainty and risk assessment

As the level of uncertainty of the components used in therisk equation (hazard, vulnerability, quantification of the ex-

Fig. 2 – Overview of partners in the CHANGES project.


Spatial information for analyzing changing hydro-meteorological risk

posed elements at risk) is very high, the analysis of thechanges in future risk should incorporate these uncertaintiesin a probabilistic manner. Methods for probabilistic risk as-sessment for flooding have been developed (e.g. Moel et al.,2009). The European Parliament adopted a new Flood Di-rective (2007/60/EC) with the objective to establish a frame-work for the assessment and management of flood risk inEurope, emphasizing both the frequency and magnitude of aflood as well as its consequences (Lavalle et al., 2005). How-ever, such methodologies should be downscaled for practicaluse in risk management at local levels. Impacts of naturalhazards on the Environment and on the Society are still tack-led by mono-disciplinary approaches. The focus is reflectedin the domains of scientific research (single approach andtools for each type of threat), in the existing managementtools and in the legislative basis of these activities. Manage-ment tools, models, and local-to-regional technical solutionshave been proposed by numerous projects for single hazards.However only a few of them have tackled the issue of risk as-sessment and management in a multi-hazard perspective, in-cluding possible combined and domino effects. Probabilistictools for multi-hazard risk assessment are not available tostakeholders at the local level. Insurance companies and spe-cialized risk assessment consultants have developed modelsbut these are not open for public use. Internationally, severalinitiatives for multi-hazard risk assessment platforms exist,such as HAZUS-MH (FEMA, 2009) and CAPRA (WorldBank, 2009). HAZUS-MH is a powerful risk assessmentmethodology for analyzing potential losses from floods, hurri-cane winds and earthquakes. CAPRA is a system which uti-lizes state-of the-art technology in Geographic InformationSystems, Web-GIS and catastrophe models, used to generatean open platform for disaster risk assessment. The CHANGESnetwork will further built on these experiences and adaptsuch an open system for probabilistic risk assessment forhydro-meteorological hazards that can be used by stakehold-ers at a regional and local level.

Changing risk management

The European Commission has identified the need foradaptations in risk management as a consequence to climateand environmental changes in several documents (e.g. EC2009). The implementation of risk management measuressuch as disaster preparedness programmes, land-use plan-ning, regulatory zoning and early warning systems are con-sidered essential. Fleischauer et al. (2006) conclude that spa-tial planning is only one of many aspects in risk managementand that it is, in general, not involved in risk assessment. Fur-ther, multi-risk assessment approaches are not used in plan-ning practice: risk indicators are hardly used and vulnerabil-ity indicators are not at all used. Therefore integrated ap-proaches are needed for integrating spatial planning in disas-ter risk management. Additionally, scientific advances in haz-ard and risk assessment and demands of stakeholders/end-users are still not well connected. In many cases, the scientif-ic outcomes remain rooted solely within the scientific com-munity or new knowledge is not fabricated enough to be im-

plemented by stakeholders and end-users (IRGC, 2005). Akey cause of the gap between the science community andstakeholders/end-users is in the complexity of human-en-nvironment interactions. This has led to the development of adiversity of approaches, often not easy to implement by theend-user community. The CHANGES network recognizes theshared responsibilities of all stakeholders for which sharedknowledge is the key element. Therefore, the network aims ata transparency by putting communication via visualization ofthe whole risk management cycle and scenarios central.There is a need for the development of a harmonised deci-sion-making tool structure for applying hazard and risk mit-igation through spatial planning in risk prone areas and de-velopment of a guideline on natural hazard mitigation in thecontext of the EU Environmental Assessment Directive.

References

Alcamo J., et al. (2007) – Europe. In: M.L. Parry et al. (Eds.), Cli-mate Change 2007: Impacts, Adaptation and Vulnerability. Con-tribution of Working Group II to the Fourth Assessment Reportof the Intergovernmental Panel on Climate Change. CambridgeUniversity Press, Cambridge, UK, 541-580.

Audsley E. et al. (2006) – What can scenario modelling tell usabout future European scale agricultural land use, and what not?Environmental Science and Policy 9(2), 148-162.

Barredo J.I., Petrov L., Sagris V., Lavalle C., Genovese E.(2005) – Towards an integrated scenario approach for spatialplanning and natural hazards mitigation. Joint Research Centreof the European Commission EUR 21900 EN.

Beniston M. et al. (2007) –Future extreme events in European cli-mate: an exploration of regional climate model projections. Cli-matic Change 81, 71-95.

Beniston M., Douglas G.F. (1996) – Impacts of climate change onmountain regions. In: R.T. Watson, M.C. Zinyowera, R.H. Mossand D.J. Dokken (Eds), Climate Change 1995. Impacts, Adapta-tions and Mitigation of Climate Change, Scientific-Technical Anal-ysis, Cambridge Univ. Press, Cambridge 191–213.

Birkmann J. (2006) – Landslides: from Mapping to Loss and RiskEstimation, Crosta G.B. and Frattini P. (eds), LESSLOSS ReportNo. 2007/01, IUSS Press ISBN: 978-88-6198-005-1, 2007.

Ciscar J.C. (Ed.) (2009) – Climate change impacts in Europe.Final report of the research project. European Commission JointResearch Centre. EUR 24093 EN.

Christensen J.H., Carter T., Rummukainen M. (2007) – Evalu-ating the performance and utility of regional climate models: thePRUDENCE project. Climatic Change 81, 1-6.

Dixon N., Brook E. (2007) – Impact of predicted climate changeon landslide reactivation: case study of Mam Tor, UK. Land-slides 4, 137-147

EEA (2004) – Mapping the impacts of recent natural disasters andtechnological accidents in Europe. Environmental Issue Report35, European Environment Agency (EEA), Copenhagen, Den-mark.

European Commission (2006) – Directive of the European Par-liament and of the council: establishing a framework for theprotection of soil and amending. Directive 2004/35/EC.

European Commission (2009a) – White Paper ‘Adapting to cli-mate change: Towards a European framework for action’,


C. J. van Westen

COM(2009b) 147 final and accompanying staff working docu-ment, 70 p.

FEMA (2008) – HAZUS, FEMA’s Methodology for EstimatingPotential Losses from Disasters. http://www.fema.gov/plan/pre-vent/hazus/index.shtm

Fleischhauer M., Greiving S., Wanczura S. (Eds.) (2006) –Natural Hazards and Spatial Planning in Europe. Dortmund206 p.

Fuchs S., Heiss K., Hübl J. (2007) – Towards an empirical vul-nerability function for use in debris flow risk assessment. Natu-ral Hazards and Earth System Sciences 7, 495-506;

Glade T. (2003) – Vulnerability assessment in landslide risk anal-ysis. Die Erde 134, 121-138

Greiving S. (2004) – Risk assessment and management as anImportant Tool for the EU Strategic Environmental Assessment.DISP 157, 11-17.

Greiving S, Fleischhauer M., Wanczura S. (2006) – Managementof Natural Hazards in Europe: The Role of Spatial Planning in Se-lected EU Member States. Journal of Environmental Planningand Management 49, 5, 739-757.

IRGC (2005) – International Risk Governance Council, WhitePaper 1 on Risk Governance: Towards an Integrative Approach,Geneva, IRGC,

ISDR (2009) – Global Assessment Report on Disaster Risk Reduc-tion. United Nations, Geneva, Switzerland.

Lavalle C., Barredo J. I., De Roo A., Niemeyer S., Miguel-AyanzJ. S., Hiederer R., Genovese E., Camia A.(2005) – Towards an

European integrated map of risk from weather driven events,European Commission. Joint Research Centre, EUR 22116 EN.

Malet J.-P., Thiery Y., Puissant A., Hervás J., Günther A.,Grandjean G. (2009) – Landslide susceptibility mapping at1:1M scale over France: exploratory results with a heuristicmodel. In: Malet J.-P., Remaître A., Boogard T. (Eds), Procee-dings International Conference on Landslide Processes: fromGeomorphologic Mapping to Dynamic Modelling, 6 -7 February2009, Strasbourg, France. CERG Editions, Strasbourg, 315-320.

Malet J.-P., Durand Y., Remaître A., Maquaire O., EtcheversP., Guyomarc’h G., Déqué M., van Beek L.P.H. (2007) – As-sessing the influence of climate change on the activity of land-slides in the Ubaye Valley. In: McInnes R., Jakeways J., Fair-bank H., Mathie E. (Eds): Proceedings of the International Con-ference on Landslides and Climate Change – Challenges andSolutions, Taylor & Francis, London, 195-205.

Moel H de, van Alphen J., Aerts J. C. J. H. (2009) – Flood mapsin Europe – methods, availability and use. Natural Hazards andEarth System Sciences 9, 289–301.

Nadim F., Kjekstad O., Peduzzi P., Herold C., Jaedicke C.(2006) – Global landslide and avalanche hotspots. Landslides 3,2, 159-174.

Van Beek L.P.H., Van Asch Th.W.J. (2004) – Regional assessmentof the effects of land-use on landslide hazard assessment by meansof physically-based modelling. Natural Hazards 31, 289-30

World Bank (2009) – CAPRA. Central American ProbabilisticRisk Assessment. http://ecapra.org/en/

Introduction

Large tropical rivers tend to have a common set of prop-erties (Fig.1). A seasonal pattern characterizes the annualdistribution of water and sediment discharge. Effects oflarge floods are superimposed on this seasonal pattern. Thesediment load is derived mostly from the mountains in theupper basin and transferred episodically downstream alongthe channel. The basins are usually polyzonal and riverstend to end in large deltas. Rivers and their basins under-went a series of changes during the Quaternary. Over the lastfew centuries, however, anthropogenic activities have trans-formed a number of these rivers due to changes in basin landuse, closure of dams, and modification of climate.

A selected list of robust changes in climatic characteristicsthat affects the large tropical rivers includes retreat of moun-tain glaciers, melting of snow and ice earlier than expected,changes in annual variations in rainfall, enhanced seasonali-ty, regional increase in droughts, increase in the strength oflarge storms, higher rainfall from extreme events, andsea level rise (IPCC, 2007). The resulting effects vary inintensity and do not always impact all large rivers thesame way. It is crucial to determine the effect of suchchanges on large rivers, at least at a qualitative level.

Methods

The effects of such changes on rivers can be deter-mined by (1) observation, (2) modelling and (3) exa-mining past analogues. Observations are best but weget very few opportunities to see such changes, an ex-ception being the United States Geological Survey ob-servations on streamflow gauges for the 1900s. The

existing models are not in agreement or of required resolu-tion. We can use environmental analogues from the Pleisto-cene for certain cases. We may start determining possiblerobust changes on large rivers from climate change withthree basic premises: (1) rivers are in dynamic equilibrium,(2) a large river is a system, linked from source to sink, (3)analogues exist from geological past, especially Plioceneand Quaternary.

Results

The effect of climate change varies between the upper andthe lower parts of the river but the entire river behaves as anintegrated system. The basin, however, may display polyzo-nal characteristics thus complicating the scenario (Gupta,2010). The impact of climate change along the upper andmiddle rivers may be summarized as:

• less dependence on snowmelt and more on rainfall;• increased duration of low flow;


Global warming and large tropical rivers

A. Gupta*,** and S.C. Liew*

Short abstract: Large tropical rivers carry a common set of properties: a seasonal pattern of discharge, episodic transport of sediment,almost the entire sediment derived from the headwaters, polyzonal basins, large deltas. Anthropogenic activities have transformed anumber of these rivers following changes in basin land use, closure of dams, and climate modification. The Fourth Assessment Reportof the Intergovernmental Panel on Climate Change lists certain consequences of climate change that should impact large rivers. Theseinclude retreat of mountain glaciers, early melting of snow and ice, changes in annual variations in rainfall, enhanced seasonality, region-al increase in droughts, increase in strength of large storms, higher rainfall from extreme events, and sea level rise. It is possible toevaluate the resulting robust changes in morphology and behaviour of river systems using principles of fluvial geomorphology and pastanalogues. We may even be able to construct a new set of rules for understanding, utilizing and managing rivers.

Keywords: Large rivers, tropics, global warming, satellite images, anthropogenic noise.

* Centre for Remote Imaging, Sensing and Processing, National University of Singapore.** School of Earth and Environmental Sciences, University of Wollongong, Australia.

Fig. 1 – Summary description of a large tropical river.


A. Gupta and S.C. Liew

• changes in groundwater discharge (baseflow), at pre-sent not well understood;• possible changes in mean discharge volume;• different types and scales of change among the tribu-taries; • possible increase in flood: small changes in mean cli-mate may give rise to large changes in extreme floods(Knox, 1993);• changes in transfer and storage of sediment.

The present rivers are thus likely to change to wider, shal-lower, floodprone systems with a high rate of periodic sedi-ment transfer separated by long-term storage in between.

We may summarize the expected changes on the lowerriver as:

• a rise in sea level leading to inundation at delta-face andalteration in sediment pattern;• increased impact of cyclonic storms, delta-face ero-sion and saline intrusion;• fine sediment deposited over channel sand in the lowerdelta;• crevasse splays, avulsion and flood basin filling in theupper delta.

This would lead to a dynamic instability over the delta sys-tem which is very likely to affect human occupation of thedelta and utilization of the river channels. For example, amodel for changing avulsion style over time, as proposed byBlum and Aslan (2006), suggests that an early stage of valleyfilling, channel occupation and low aggradation will be substi-tuted by frequent crevassing and high aggradation in channelsand flood basins. This will be replaced in time by reoccupationof old channels and low aggradation in nearly-filled valleys.

A large river, being an integrated system, would need toadjust to all these changes simultaneously.

Discussion

About a quarter of the World currently lives on or near del-taic coastlands or wetlands (Syvitski et al., 2005). It is perti-

nent to enquire about the effect of the suggested changes onthe people of deltas and coastal plains. We also recall that alarge river is an integrated system and the changes that af-fect the upper and middle reaches also need to be amalga-mated. This raises the question whether the inhabitantsconcerned would be able to cope with such changes.

Several complications should be expected to this scenario: • with climate change, river data will lose stationarity (Millyet al., 2008);• the effect of climate change will not be globally uniform; • Morphology and behavior of large river changes alongits length, and these sectors would adjust differently toclimate change. The Mekong is an excellent example.noise from anthropogenic changes may override thesignal from climate change.

Conclusion

Climate change and its effect on large rivers and deltas arelikely to seriously impact the environment and people. Somegeneralizations are possible, but given the morphological andbehavioural differences we need to study large rivers and del-tas individually. We also need to think about consequences atboth the regional and local scales for ecosystems, inhabitantsand economies. We may even need a new set of rules for un-derstanding, managing, and utilizing rivers.

References

Blum M.D., Aslan A. (2006) – Signatures of climate vs. sea-levelchange within incised valley fill successions: Quaternaryexamples from the Texas Gulf Coast. Sedimentary Geology 190,177-211.

Goodbred S.L.,Jr. (2003) – Response of the Ganges dispersal sys-tem to climate change: a source-to-sink view since the lastinterstade. Sedimentary Geology 162, 83-104.

Gupta A. (2010) – The effect of global warming in large rivers anddeltas, In: Developing Countries Facing Global Warming: APost-Kyoto Assessment (M. De Dapper, D. Swinne and P. Ozer,Eds.), Brussels, Royal Academy for Overseas Sciences and Uni-ted Nations, 125-138.

Intergovernmental Panel of Climate Change (2007) – ClimateChange: The Physical Science Basis. Cambridge, CambridgeUniversity Press.

Knox J.C. (1993) – Large increases in flood magnitude in respon-se to modest changes in climate. Nature 361, 430-432.

Milly P.C.D., Betancourt J., Falkenmark M., Hirsch R. M.,Kundzewicz Z. W., Lettenmaier D.P., Stouffer R.J. (2008) –Stationarity is dead: Whither water management? Science 319,573-574.

Syvitski J.P.M., Vörösmarty C.J., Kettner A.J., Green P. (2005)– Impact of humans on the flux of terrestrial sediment to the glo-bal coastal ocean, Science 308, 376-380.

Fig. 2 – Schematic diagram of expected changes in a large riversystems based on Goodbred (2003). Generalized from Gupta (2010).

Abstract: Jakarta is a metropolitan city with the area of 700 km2 and it is the capital of Indonesia housing 9.5 million inhabitants.Despite its importance, it often suffers from heavy flooding. The most devastating flood in the last 3 centuries occurred in Feb 2007.The flood inundated about 70% of the city with some areas under 3-4 m water depth, resulted in at least 57 deaths, and caused about450,000 residents fleeing their premises. Flood events have been occurring regularly in Jakarta. To model the flood in Jakarta, a goodtopographic dataset is needed. The study explores the use of several sources of topographic data. Apart from a quite costly spot heightdata acquired, two free sources, SRTM and ASTER, are considered. The discrepancy between SRTM, ASTER and spot height data forJakarta domain are investigated. The flood depth resulted from 3 sources are also analyzed and compared to the observed data. Theanalysis shows that for a larger domain such as Jakarta, if no surveyed information is available, the SRTM data is a good source of datato achieve adequate estimation of flood depth.

Jakarta flood modeling with different sources of topographic data

C.D. Doan*, S.-Y. Liong*, R. Sanders*


* Tropical Marine Science Institute - National University of Singapore, NUS, Singapore.

** School of Environment and Life Sciences, Charles Darwin University, Australia.

Bank erosion and channel change in the Daly River,Northern Australia

S. Karki**, B. Wasson**, D. Pearson**, S. Maier**, W. Ahmad**

Abstract: Riverbank erosion is an important component of river channel width adjustment, can add significantly to the sediment loadof a river impacting water quality and destroys floodplains widening in on both river banks. The Daly River is a significant river in Aus-tralia’s wet- dry tropics. The Daly River catchment is also of immense interest for having its potential agricultural development. Withthis regard, sediment accumulation in the Daly River is a subject of concern for various community groups, government bodies, resear-chers and land managers. However there is limited information regarding this and other phenomena in this river in a remote part ofAustralia. The top soil is a very small contributor (3%) to the river sediment load and riverbanks and gullies are found to be the majorsources of sediment in this largely undisturbed river. If the riverbanks are eroding rapidly, the channel should be either migrating late-rally or widening or both. We examined channel change and bank erosion for the downstream reach of the Daly River with the help ofhistoric aerial photographs, high resolution Quick Bird imagery and field data. We determined that the channel has been widened for thestudy period of 1963-2010. We also found that the total area of bank erosion and the amount of sediment production has been increasedconsiderably. Slumping caused by both sub-aerial and fluvial mechanism and bank erosion by fluvial mechanism was found to be thedominant process of erosion at this reach. Our results coincided with the increase of precipitation and river discharge for the same per-iod signifying the hydrological influence.

Introduction

Almost every year, Gorontalo city and its vicinity experi-ences flooding from Bone and Tamalate rivers. Before 1996,the accident did not have significant impact to inhabitants.However, flooding tends to be severe, with a notable recordbeing in 2002. The most recent serious inundation was in2006. To avoid such disastrous accidents, mitigation plan-ning is required. The first step to establishing a suitable riskmitigation plan requires evaluation of flood hazard and riskalong the two rivers.

It is widely understood that Gorontalo is a flood-pronearea since it lies in the central basin of North Sulawesi. Thecity has been built on alluvial plains and the Bolango River al-luvial fan. The city is surrounded by hilly to mountainous ter-rain, making it particularly vulnerable to flooding (Tjahjonoet al. 2009). Mapping and study of flood hazard was con-ducted by Bappedalda Gorontalo (Municipal Board for En-vironmental Assessment of Gorontalo) in 2007, using a ge-omorphological approach based on 2006 Landsat data. Thestudy suggested that anthropogenic factors – uncontrolled landuse change, appalling drainage and waste management – con-tributed significantly to increasing flood risk. Hidiya (2011)indicated a significant trend on land use change of LowerBone Watershed. In the site, forests, rice fields and uplandshave been replaced by mixed garden and settlements. Thestudy concluded that urban planning has a major role inavoiding damage and fatalities in the future.

Spatial planning or mitigation planning at municipal levelrequires detailed spatial data or imagery. With the remote-ness of most Indonesian middle sized cities, appropriatescale data is not available. This even occurs in most EasternIndonesian regencies (kabupatens), due to many reasons in-cluding availability of high resolution remote sensing data.

Availability of high resolution data sources such as GoogleEarth leads to diminishing gaps on some locations. This paperdiscusses the use of high-resolution Google Earth image to pro-vide a detailed flood risk map, based on the work of Bappedal-da Gorontalo.

Methods

Gorontalo Municipal is geographically located in 0º 32’ 00.79” Nand 123º 03’ 35.42” E and at the seashore of Tomini Gulf, North-ern Sulawesi. In this research, Google Earth data dated 6th March2010 were used. Digitizing was done online using Google Earthdigitizing facilities. Visual interpretation on land use and land-forms employed common interpretation keys such as tone, tex-ture, shape and association.

Flood hazard map of Bappedalda Gorontalo was used as aninput to estimate risks associated with vulnerability and riskelements of land use. Risk was calculated using the defini-tion of Thouret (1994) as follows: R = H x V x E, where R =risk, H = hazard, V = vulnerability, and E = risk elements (orexposure). Study site was focused along Bone and TamalateRivers which are covered by high resolution imagery.

In this research, simple scoring was implemented to ob-tain risk levels (Table 1, 2, and 3). The hazard levels wereclassified into high, moderate and low.


Flood risk analysis and mapping in Gorontalo city,Indonesia, using high resolution Google Earth’s imagery

B. Tjahjono*, M. Hidiya**, B.H. Trisasongko***

* Soil Science & Land Resources, and PPLH, Bogor Agricultural University, Bogor, Indonesia.** Faculty of Agriculture, Gorontalo State University, Gorontalo, Indonesia.*** Soil Science & Land Resources, and P4W/CRESTPENT, Bogor Agricultural University, Bogor, Indonesia.

Short abstract: We found that Google Earth (2010) images obtained freely from internet are very efficient for risk assessment. About101 hectares of the area studied were considered high risk, while 200 ha were at medium risk. The high risk areas were located in thesettlements. The river embankments (dikes) built along the river for mitigation purposes are built on flood plains. Land utilization shouldbe tightly controlled using strict urban planning legislation.

Keywords: risk, flood, Bone River, Gorontalo, Google Earth.

Flood Hazard Score

High 3

moderate 2

Low 1

Tab. 1 – Scoring for flood hazard.


B. Tjahjono et al.

Risks were classified into three categories based on inter-vals calculated from range divided by designated risk classes,i.e. (27–0)/3 = 9. Risk classes are described in Table 4.

Results and Discussion

It appears that spatial distribution of hazard levels wasclosely related to landforms. Highest level could be found inflood plains of the rivers. Areas with highest level of hazardwere usually flooded every year in the rainy season; inunda-tion frequency is about 7–10 times every decade. Meanwhile,a moderate level of hazard was attributed generally to olderalluvial plains with frequency of 3-6 times/decade (Bappedal-da Gorontalo, 2007). However, we should mention here thatmagnitude of inundation tends to increase in recent yearsand induces substantial cost to local inhabitants who livealong the river.

In this research, vulnerability was solely determined byland use data. Google Earth images provide the possibilityto derive detailed and up-to-date land use information.Using the imagery, 5 major land use classes could be re-trieved, i.e. built-up (dominated by settlements), rice fields,upland (dry) fields, mixed gardens, forest/shrub. Land usedelineation on flood-prone areas concluded that built-up (set-tlement) areas covered 408.9 ha, which was the largest amongland use classes. After settlements, mixed garden (380.5 ha),upland fields (364.5 ha), and rice fields (142.5 ha) were themost abundant. Each land use class was then assigned a des-ignated score and serves as a risk element to obtain overallvulnerability.

Vulnerability Score

Built areas, river channel 3

Agriculture (rice & upland fields) 2

Others 1

Tab. 2 – Scoring for vulnerability.

Risk Elements Score

Settlements (built areas) 3

Rice fields 2

Upland fields 1

Mixed gardens 1

Forest and scrubs, river channel 0

Tab. 3 – Scoring for risk elements.

Risk categories Range

High >18-27

Moderate >9-18

Low 0-9

Tab. 4 – Classification of risk.Fig. 1 – Bone River flood risk map, Gorontalo, Indonesia.


Flood risk analysis and mapping on Gorontalo city, Indonesia

Using simple Boolean overlay in GIS, flood risk of thesite was calculated. The result indicated that highest riskareas were the least abundant, covering around 101.7 ha.These areas were to be found on Tamalate and Bone Riverbanks, especially in the estuary. Most of the study areawas categorized as low risk (1187.6 ha) with followingland uses: upland (dry) fields, mixed garden, rice fieldsand settlements. Moderate risk (about 200.3 ha) was locat-ed in settlements and some in rice fields (Fig. 1). As mostof the medium to high risk areas are to be found in built-up areas, the local government has already built dikes tolimit the impact of most flood events. These are appro-priate for all but exceptional events which may overflowthem. It should be noted here that these dikes have notbeen built on the edge of flood plains, but in the middle. Inthese cases, if urban development takes place behind thedikes, but within the flood plains, the risk is significantlyhigher – here, during exceptional events, these urban areaswill most likely flooded very rapidly after the dikes areoverflowed.

The Google Earth image presented below shows the narrow-ing part of Bone River, near the estuary. This point pays an at-tention and further research is needed to determine its contri-bution to flooding. Restricting the width of the river bodywould dramatically change the river bed, and a slight increaseof water flow could raise water level which, in turn, may createflooding. The situation could be even worse if the tidal influ-ence is also taken into account; however, we cannot currentlyevaluate this impact due to the limited data available.

Conclusion

Newly acquired data (6th March 2010) available throughGoogle Earth was beneficial to assessing flood risk and pro-vided the possibility of deriving detailed land use of a remotearea in Indonesia. In particular, land use information was re-quired as a contributing factor to risk elements (exposure).

This study demonstrates that freely available high resolu-tion imagery were advantageous to assist mitigation andhazard-related planning in remote areas. Availability to suchdata should help reduce damages and casualties in futureflooding events.

References

Bappedalda G. (2007) – Kajian Geomorfologi dan Penutupan /Penggunaan Lahan dalam hubungannya dengan bahaya banjir diDAS Bone Hilir. Badan Penelitian dan Pengendalian DampakLingkungan (Bappedalda), Provinsi Gorontalo (in Indonesian).

Hidiya M. (2011) – Analisis kecocokan penggunaan lahan dantata ruang di DAS Bone Hilir, Provinsi Gorontalo. Thesis. StudyProgram of Regional Planning, Bogor Agricultural University.(in Indonesian).

Thouret J.C. (1994) – Prévision de menace et evaluation desrisques volcaniques. In Bourdier J.-L., Le volcanisme, manuelset méthodes, Edition BRGM, Orléans, France.

Tjahjono B., Hidiya M., Munibah K. (2009) – Identifikasi Ben-tuklahan (Landform) di DAS Bone Hilir, Provinsi Gorontalodengan Data PALSAR, SRTM, dan Landsat. Prosiding SemilokaGeomatika-SAR Nasional. Bogor. (in Indonesian)

Sediment budgets provide a framework for assessing thecontribution of different sediment sources (Walling andCollins, 2008). The use of radionuclide tracers and remotesensing (RS) and GIS have been identified as the best ap-proaches for rapid survey of large catchments (Brown et al.,2009, Brooks et al., 2008). Radionuclide tracers provide abottom-up approach with estimates of the relative contribu-tion of topsoil (rill and sheet) erosion and subsoil sedimentsources upstream from a water storage. Sediment fingerprint-ing can differentiate between geological formations. Howev-er, these approaches cannot differentiate between landslides,gully erosion and river channel change. RS and GIS enable atop down approach to sediment budgets, identifying the typeand location of major erosion features in a catchment and thenexamining likely paths of sediment movement to the waterstorage. However, the sediment delivery analysis is often un-derpinned by large assumptions or broad scale relationships.Recently, rapid development of GIS and RS software (includ-ing free and open source software) and the increasing avail-ability of freely accessible data have created wide ranging op-portunities for analysis of sediment sources. Surprisingly fewstudies have combined GIS techniques with radionuclidetracers, despite the potential benefits of these complimentaryapproaches. This paper will review current methods and ex-plore the potential for application in the often resource anddata poor environments in South East Asia.

Topsoil tracers

Radionuclide tracers Cs-137 and Pb-210 are most com-monly used to determine the relative contribution of topsoiland subsoil erosion to sediment deposited on stream banksand in estuaries (Wasson et al., 2010b, Nawaz, 2010). Cs-

137 and Pb-210 accumulate in the top 20cm of the soil pro-file (Livens and Baxter, 1988) and bind strongly to the soilparticles, including during sediment transport (Motha et al.,2002). Concentrations of radionuclides are measured usingHigh Resolution Gamma Spectrometry (Murray et al., 1987).In the equatorial zone, the lifespan of fallout radionuclideCs-137 as a tracer is limited. Cs-137 has a short half-life of30 years and given fallout peaked in the mid 1960’s andceased in the 1980’s, Cs levels are now less than 35% of orig-inal levels (Tims et al., 2010). Plutonium 239 shows promiseas a replacement for Cs- 137 measured with high sensitivityby Accelerator Mass Spectrometry (Child et al., 2008). Cur-rently the cost of sample analysis may limit the applicationsin SEA in the short term. Lead 210 (excess) is a naturally oc-curring radionuclide with a short half-life of 22 years, it isalso useful as a topsoil tracer. However several recent stud-ies have reported challenges with using Pb-210 includingnegative Pb values (Wasson et al., 2010b, Nawaz, 2010) andthe complication of ongoing fallout increasing Pb-210 con-centrations in sediments that have remained exposed to theatmosphere over a period of time (Douglas et al., 2009).

Sediment ‘fingerprinting’

Radionuclide tracers have been complemented with sedi-ment ‘fingerprinting’ in a number of recent studies (Munks-gaard et al., 2003, Douglas et al., 2009). The concentrations ofrare earth elements along with common elements such as or-ganic matter and nitrogen content of soils can indicate the con-tribution of different geological formations to overall sedimentload. This approach is more successful in geologically hetero-geneous catchments, where there are obvious differences be-tween rock types and compositions (Douglas et al., 2009).


Quantifying sediment budgets in data poor environments of SE Asia:A review of remote sensing,

GIS and isotope based approaches

S. Hobgen*, G. Boggs*, B. Myers*, B. Wasson*

Short abstract: Planning for land and water resource management is becoming increasingly important in the face of a changing climate(IPCC, 2007) . This is particularly relevant in South East Asia where many communities rely on water storages to maintain domesticand irrigation water supplies, and for hydroelectricity generation (Lasco and Boer, 2006). As these storages fill with sediment, commu-nities, governments and international funding agencies debate the most appropriate actions to reduce sedimentation. It is therefore vitalto identify sediment sources and processes of sediment delivery (Brown et al., 2009).

Key words: sediment, budget, erosion, catchment, river.

* Research Institute for Environment and Livelihoods, Charles Darwin University, Australia.


S. Hobgen et al.

Mapping sediment sources and delivery

Mapping of predominantly subsoil sediment sources with-in a catchment requires consideration of the distribution oferosion features, changes in size and shape over time andestimations of volume of sediment production. The suitabil-ity of a method is largely determined by the scale of thestudy, and the availability and accuracy of data.

Mapping the distribution and density of sediment sources

Contemporary distribution of sediment sources can bemapped using a variety of different imagery (Brooks et al.,2008). Medium to high resolution imagery (< 30m pixel eg.Quickbird, Worldview, SPOT 5, Cartosat etc) is required tomap landslide scars, gully erosion features and locations ofchannel migration and can be combined with automated anal-ysis techniques to rapidly map large areas (Joyce et al., 2009).However, this imagery is often prohibitively costly for largescale studies. Quickbird and SPOT 5 imagery are publiclyavailable through Google Earth, providing a valuable sourceof imagery in SEA. However, in the current Google Earthsoftware, analyses are limited to manual methods, hence thisis only practical in smaller catchments.

Analysing changes over time

Comparison of historical and contemporary aerial photog-raphy is the most common method for assessing changesover time, including incidence of landslides (Metternicht etal., 2005) and growth of gully erosion features (McCloskey,2010, Sattar et al., 2010), and movements and bank collapsein river channels (Karki et al., 2011). Acquiring historicaland new aerial photography is often not possible in SouthEast Asia, due to funding and access restrictions.

The Landsat satellite imagery archive provides an invalu-able historical data set for analysis of temporal changes, be-ginning in the 1970’s (USGS, 2011). Recent research on sed-iment sources in Timor Leste (Rouwenhorst et al., 2009) usedLandsat 5/7 imagery to investigate channel change. However,the low resolution of Landsat imagery (30 m) is best suited toidentifying large scale changes (Vrieling, 2006).

SPOT satellite imagery provides opportunities for mediumterm analysis, the high resolution imagery archive beginning in1986 with the Spot 1 satellite (Vrieling, 2006). Some imagesfrom the Spot 5 satellite (5 m resolution) are available for use,through Google Earth. For small scale climate change projectsa limited number of images (5-10) may be provided by the SpotImage ‘Planet Action’ initiative (www.planet-action.org).

Estimating volume of sediment sources

Estimating the volume of sediment produced by individu-al landslides, gullies and channel slumps requires a threedimensional understanding of sediment sources. This can be

obtained through use of three dimensional data, or use oftwo dimensional data supplemented by field measurementsand generalised equations. Three dimensional data is usual-ly obtained from Digital Elevation Models (DEMs). DEMscreated from stereo pairs of aerial photographs are the mostpopular means for this kind of analysis (Martı´nez-Casasno-vas, 2003). LiDAR and Cartosat satellite imagery also havethe capacity to create high resolution DEMs for calculatingchanges in gully depth and sidewall retreat rates, LiDAR alsohas potential to determine long term sedimentation rates(Brown et al., 2009). Sattar (2010) used Cartosat-1 stereosatellite images (2.5 m resolution) and aerial photography for3D mapping of gullies in the Australian Daly River catch-ment. LiDAR and Cartosat imagery are relatively expensiveto purchase, and require resource intensive studies to calcu-late sediment production rates.

Multi-temporal two dimensional mapping of sedimentsources can be supplemented with field measurements and gen-eralised equations for calculating sediment production. Sattar(unpublished data) proposed a relationship between the gullydensity and gully sediment yield based on published studies(cited in Wasson et al., 2010a). Malamud et al. (2004) suggestan empirical equation for estimating sedimentation induced byearthquakes. Karki et al. (2011, this conference) are determin-ing channel change and channel slump contributions from thehorizontal movement of channel banks identified by satelliteimagery and field measurements.

Mapping soil surface erosion risk

Identification of sites of surface erosion and determinationof erosion risk by GIS is well established (eg. Boggs et al.,2001, Witz and Muga, 2009). Many studies use variations ofthe Revised Universal Soil Loss Equation (RUSLE) (Wis-chmeier and Smith, 1978) as all factors are spatially variableand it encompasses both topographic and anthropogenic fac-tors contributing to erosion risk. If the appropriate data areavailable, these can be used to calculate soil erosion rates.More often data are used to map relative erosion risk (Boggset al., 2001) and to evaluate the impacts of land use changeand soil conservation practices (Mati et al., 2000).

GIS based RUSLE modelling is a particularly important toolfor South East Asia, as the data required are mostly publiclyavailable: ASTERV2 30m resolution DEM data (ASTER,2009) for slope steepness and length; Landsat TM imagery forland cover analysis and Quickbird Imagery for identifyingand verifying conservation techniques. The WorldClim (Hi-jmans et al., 2005) global climate model provides a free spa-tial data set for monthly rainfall, which can be used to esti-mate spatially variable R factor values within a catchmentarea (Angulo-Martínez and Beguería, 2009). The exceptionremains access to comprehensive soil mapping which isoften unavailable or incomplete. Some studies have used asubstitute base data layer, such as geological mapping tocreate land units, from which soil properties are determinedby laboratory analysis of representative soil samples (Matiet al., 2000).


Quantifying sediment budgets in data poor environments of SE Asia

Creating an indicative sediment budget

Radionuclide tracer results from the water storage or estu-ary can provide an indication of the relative contribution oftopsoil and subsoil to fine sediments. A nested sampling strat-egy can be developed through a catchment to identify the sitesof major sediment sources: from the water storage, movingupstream, assessing the relative contributions of each majortributary.

Manual investigation of high resolution imagery can quick-ly reveal major gullies, landslide scars, and sites of channelchange, which can then be further investigated using theremote sensing and GIS approaches discussed earlier. Finger-printing the chemical constitution of sediments in the waterstorage can also indicate the relative importance of differentgeological formations where geological diversity exists. Thecombination of approaches can provide maps of high priori-ty areas and erosion types to target appropriate managementactivities to reduce erosion loss and rehabilitate affectedareas. Methods for determining the sediment delivery ratioof the catchment have not been included in this paper, there-fore these methods only estimate the volume of sedimententering the waterway.

Large scale quantification of sediment sources is an inexactscience. There are constant tradeoffs between the accuracy ofthe sediment estimates, and the amount of data and resourcesrequired to produce the estimates. The additional time, effortand resources required to determine rates of change and sedi-ment volumes need to be evaluated against the potential bene-fits, as identifying and mapping of priority areas may be suffi-cient to assist managers in strategic catchment managementand project planning.

This paper provides options for sediment source identificationin data and resource poor environments; however the methodsinvolved still require not insignificant resources, skills and time.In the current context it is highly unlikely that these direct ap-proaches could be applied to every catchment in South EastAsia. However, with further research, increasing data avail-ability and improved technology these approaches may pro-vide a basis from which a more generalised approach to deter-mining sediment budgets for guiding catchment managementpolicies and projects could be developed.

References

Angulo-Martínez M., Beguería S. (2009) – Estimating rainfallerosivity from daily precipitation records: A comparison amongmethods using data from the Ebro Basin (NE Spain). Journal ofHydrology 379, 111-121.

ASTER (2009) – Advanced Spaceborne Thermal Emission andReflection Radiometer. Editor: Ministry of Economy, Trade, andIndustry (METI) of Japan and the United States National Aero-nautics and Space Administration (NASA).

Boggs G., Devonport C., Evans K., Puig P. (2001) – GIS-basedrapid assessment of erosion risk in a small catchment in thewet/dry tropics of Australia. Land Degradation & Development12, 417-434.

Brooks A., Knight J., Spencer J. (2008) – A remote sensingapproach for mapping and classifying riparian gully erosion inTropical Australia. Nathan, Queensland: Australian Rivers Insti-tute, Griffith University.

Brown A.G., Carey C., Erkens G., Fuchs M., Hoffmann T.,Macaire J.J., Moldenhauer K.M., Walling D.E. (2009) – Fromsedimentary records to sediment budgets: Multiple approaches tocatchment sediment flux. Geomorphology 108, 35-47.

Child D.P., Hotchkis M.A.C., Williams M.L. (2008) – High sen-sitivity analysis of plutonium isotopes in environmental samplesusing accelerator mass spectrometry (AMS). Journal of Analyt-ical Atomic Spectrometry 23, 765–768.

Douglas G., Caitcheon G., Palmer M. (2009) – Sediment sourceidentification and residence times in the Maroochy River estu-ary, southeast Queensland, Australia. Environmental Geology57, 629–639.

Hijmans R.J., Cameron S.E., Parra J.L., Jones P.G., Jarvis A.(2005) – WorldClim – Global Climate data [Online]. Museumof Vertebrate Zoology. Available: http://www.worldclim.org[Accessed 25].

IPCC (ed.) (2007) – Climate Change 2007: the physical sciencebasis. Contribution of Working group 1 to the Fourth AssessmentReport of the Intergovernmental Panel on Climate Change Cam-bridge, United Kingdom and New York , NY, USA: CambridgeUniversity Press.

Joyce K.E., Belliss S.R., Samsanov S.V., McNeill S.J., Glassey P.J.(2009) – A review of the status of satellite remote sensing andimage processing techniques for mapping natural hazards and dis-asters. Progress in Physical Geography 33, 183-207.

Karki S., Wasson R. J., Pearson D., Maier S. & Ahmad W. (2011)– Linking pattern to process: The implications of landscapechange for catchment management and water quality. RemoteSensing, Natural Hazards and Environmental Change. CRISPNUS Singapore and LMV-Clervolc, Clermont-Ferand.

Lasco R. D., Boer R. (2006) – An integrated Assessment of Cli-mate Change Impacts, Adaptations and Vulnerability in Water-shed Areas and Communities in Southeast Asia.

Livens F. R., Baxter M. S. (1988) – Particle size and radionuclidelevels in some West Cumbrian Soils. The Science of the TotalEnvironment 70 1-17 1.

Malamud B. D., Turcotte D. L., Guzzetti F., Reichenbach P.(2004) – Landslides, earthquakes, and erosion. Earth and Plan-etary Science Letters 229, 45-59

Martınez-Casasnovas J. A. (2003) – A spatial information tech-nology approach for the mapping and quantification of gullyerosion. Catena 50, 293- 308.

Mati B. M., Morgan R. P. C., Gichuk F. N., Quintor J. N., Brew-er T. R., Liniger H. P. (2000) – Assessment of erosion hazardwith the USLE and GIS: A case study of the Upper Ewaso Ng’iroNorth basin of Kenya. Journal of Applied Earth ObersvationsGeoinformatics 2, 78-86.

McCloskey G. L. (2010) – Riparian Erosion Morphology, Pro-cesses and Causes along the Victoria River, Northern Territory,Australia. PhD, Charles Darwin University.

Metternicht G., Hurni L., Gogu R. (2005) – Remote sensing oflandslides: An analysis of the potential contribution to geo-spa-tial systems for hazard assessment in mountainous environ-


S. Hobgen et al.

ments. Remote Sensing of Environment 98, 284-303.Motha J.A., Wallbrink P.J., Hairsine P.B. & Grayson R.B.

(2002) – Tracer properties of eroded sediment and source mate-rial. Hydrological Processes 16, 1983-2000.

Munksgaard N.C., Lim K. & Parry D.L. (2003) – Rare earth ele-ments as provenance indicators in North Australian estuarineand coastal marine sediments. Estuarine Coastal and Shelf Sci-ence 57, 399-409.

Murray A.S., Marthen R., Johnston A. & Martin P. (1987) –Analysis for naturally occuring radionuclides at environmentalconcentrations by gamma spectronometry. Journal of Radioana-lytical and Nuclear Chemistry 115, 263-288.

Nawaz M. (2010) – Sediment Sources near the Extreme Ends ofthe Catchment Continuum and the Topographic Dependence ofDenudation in a Global Context. PhD, Charles Darwin Uni-versity.

Rouwenhorst J., Boggs G., Wasson R.J., Ahmad W., Mau R., Car-valho N. (2009) – The use of GIS to determine changes in CaraulanRiver Delta, Timor Leste. In: Fisher R., Myers B. et al. (eds.), GISApplications for Sustainable Development and Good Governancein Eastern Indonesia and Timor Leste. Darwin: CDU Press.

Sattar F., Wasson R.J., Pearson D., Boggs G., Ahmad W.,Nawaz M. (2010) – The development oof Geoinformatics basedframewaork to auntify gully erosion. International Multidisci-plinary Scientific Geo-Conference & Expo Surveying Geology &Mining Ecology Management (SGEM). Albena, Bulgaria.

Tims S.G., Everett S.E., Fifield L.K., Hancock G.J., Bartley R.(2010) – Plutonium as a tracer of soil and sediment movement inthe Herbert River, Australia. Nuclear Instruments and Methodsin Physics Research Section B: Beam Interactions with Materi-als and Atoms 268, 1150-1154.

USGS (2011) – Landsat Imagery [Online]. Unites States Geologi-cal Survey. Available: http://glovis.usgs.gov [Accessed].

Vrieling A. (2006) – Satellite remote sensing for water erosionassessment: A review. Catena 65, 2 - 18.

Walling D. E. & Collins A. L. (2008) – The catchment sedimentbudget as a management tool. Environmental Science and Policy11, 136-143.

Wasson R. J., Amaral A., Rouwenhorst J., Fifield K.,Chauhan N., Pietsch T., Singhvi A. K. (2010a) – Timor Leste:A rapidly eroding landscape in the coral traingle.

Wasson R. J., Furlonger L., Parry D., Pietsch T., Valentine E.,Williams D. (2010b) – Sediment sources and channel dynamics,Daly River, Northern Australia. Geomorphology 114, 161-174.

Wischmeier W.H., Smith D.D. (1978) – Predicting rainfall ero-sion losses: a guide to conservation planning, Washington, Dept.of Agriculture, Science and Education Administration.

Witz T., Muga E. (2009) – GIS-based Rapid Risk Assessment of ero-sion potential in Aesesa catchment, Nage Keo - Ngada, Flores,NTT, Indonesia. In: Fisher R., Myers B. et al. (eds.), GIS Applica-tions for Sustainable Development and Good Governance inEastern Indonesia and Timor Leste. Darwin: CDU Press.

Abstract: The Red River is the second largest river in Vietnam with national and international importance. It is considered bothresources, environment and hazard for human survival and development. For sustainable development, the river needs to be fully unders-tood by scientists and managers with the aid of spatial information technologies. This paper describes the use of SPOT satellite imageryto observe the Red River change due to natural and human activities since 1986 when first SPOT satellite was launched. It was foundthat SPOT images are powerful, reliable data sources to derive useful information on the river resource, environment and hazard. Morestudies are required to better understand the river using new data and analysis techniques in the years to come


Exploring the Red River, Vietnam using SPOT data

M. Nguyen Dinh*

* Hanoi University of Science and Technology, C4 Building, No 1 Dai Co Viet Street, Hanoi, Vietnam.** Department of Geological and Mining Engineering and Sciences, Michigan Technological University, Houghton, MI 49931, USA.

Development of a Peace Corps Master’s Internationalprogram focused on volcanic hazard mitigation

in Indonesia

W.I. Rose**, S.A. Carn**, J.J. Wellik Li**

Abstract: The Peace Corps Master’s International (PCMI) program combines a Master’s degree with Peace Corps community service.PCMI graduate students spend one year on a university campus in the U.S. completing academic studies, followed by two years of PeaceCorps service in communities in developing countries. Michigan Technological University (Houghton, MI, USA) currently has morePCMI graduate students actively serving as Peace Corps volunteers than any other university or college in the U.S. Michigan Tech’sPCMI program spans many disciplines, including forestry, civil and environmental engineering, and natural hazards mitigation. Thenatural hazards PCMI program at Michigan Tech is unique and addresses the crosscutting area of disaster preparedness and mitigationby building public awareness and providing technical linkages to schools and communities about geological hazards. The program’sscope includes earthquakes, volcanic hazards, landslides, debris flows, droughts and floods and also indirect linkages such as the impactof these events on infrastructural elements like community development, environmental education, ecotourism, transportation, health,sanitation and water quality. The focus of this program is to improve the effectiveness of geological hazards mitigation in appropriatePeace Corps countries. Michigan Tech has established significant linkage for this work with government agencies in several selectedcountries, particularly in the areas of volcanic hazards. To date, the main regional focus for volcanic hazards projects has been LatinAmerica, but here we describe our efforts to expand Michigan Tech’s PCMI program to Indonesia, which has more historically activevolcanoes (76) than any other country. This, coupled with very high population density (Java is world’s most densely populated island)and limited monitoring resources creates unique and vital challenges for volcanic hazard mitigation. The Peace Corps recently returnedto Indonesia after an absence of several decades, and the first cohort of PCMI volunteers, including one Michigan Tech student, iscurrently in East Java province. Our hope is that the PCMI students can help with technical infrastructure and also with social outreachand hazards communication in rural Indonesian communities. Our partners in this effort are the Indonesian Center for Volcanologicaland Geological Hazard Mitigation (CVGHM) and the U.S. Geological Survey (USGS) Volcano Disaster Assistance Program (VDAP).CVGHM operates 70 volcano observatories across the vast Indonesian archipelago, many on remote islands, but all with significanthuman population at risk. The USGS VDAP is currently engaged in infrastructure development for volcanic hazard mitigation inIndonesia (and East Java in particular), including installation of new monitoring equipment and training of CVGHM scientists. Giventhis tripartite focus on East Java, the expansion of Michigan Tech’s PCMI program into Indonesia is timely and we look forward toincreasing Michigan Tech’s PCMI presence in Indonesia in the future. We also welcome and encourage collaboration between our PCMIstudents and other institutions working on natural hazard mitigation in the SE Asia-Pacific region.

Abstract: Threat from natural disasters is commonly understood as the combination of both population exposure and the probability ofhazardous events occurring. The purpose of this pilot study is to create a consistent, repeatable method for estimating volcanic hazardsexposure that is based on physical models of potential catastrophic events at volcanic edifices. The emergence of GIS technology andthe compilation of consistent population estimates (namely, those provided by the ORNL LandScan project) allow for comparable stu-dies to be made on a global scale.Simple statistical calculations have already been conducted to quantify the overall population living within 30 km of volcanic centers.These figures are most notably reported in the latest edition of the Smithsonian Institution Volcanoes of the World catalog (Simkin andSiebert, 1994). While these data facilitate identification of volcanoes that lie in rural vs. urban settings, the distribution of populationand true hazard exposure—items that are of practical importance to risk managers—are lost due to the lack of detail. This study reachesbeyond the mere 30 km hazard zone and applies the LaharZ, EnergCone, and AshFall models for eruptions between VEI ~1-5 to deter-mine population exposure to volcanic hazards at each volcano. Furthermore, the cumulative populations are represented as curvesplotted against distance from the volcano. This simple improvement shows how populations might be distributed with distance from avolcanic center.The effect of applying physical models to the estimation of population exposure shows the enormous influence of topography and windvectors on the distribution of hazards. The purpose of creating and executing this methodology is two-fold. First, meaningful compari-sons of threat can be made on a global or regional basis. When combined with estimates of probability of occurrence, this has the benefitof being able to more accurately identify which volcanoes may be in need of increased risk management and monitoring resources.Second, the use of physical natural hazard models permits the population distribution to be analyzed in relation to likely hazardous areasand not just proximity to volcanic centers.For the purpose of this poster, examples of our methodology are shown from the island of Java, Indonesia—the world’s most denselypopulated island. This work is a major stepping stone in the relative quantifiable analysis of global volcanic risk. The desired goal is ameaningful tool that risk managers can use to quickly assess the potential threat posed during a developing volcanic crisis.


Quantifying population exposure near volcanoes using physical models of natural hazards

and globally consistent data sets: Case studies from Java, Indonesia

J.J. Wellik Li*, R.E. Wolf*, S.A. Carn*

* Department of Geological and Mining Engineering and Sciences, Michigan Technological University, Houghton, MI 49931, USA.

Introduction: scope and methods

We have traced the evolution of the Lengkong, a tributary ofthe Koboan River, that frequently convey frequent rain-trig-gered lahars to the southeastern ring plain of the active Semeruvolcano, east Java, Indonesia since 2005 (Fig. 1; Thouret et al.,2007). We have studied the geomorphic response of these rivers

to volcanic activity and computed the balance between aggra-dation and degradation at channel scale.

Lavigne (2004) indicated that sediment yield from small-scale eruptions is very difficult to calculate accurately. Long-and short-term post-eruption sediment yield has been studiedMt. St. Helens (Major et al., 2000) and Pinatubo (Pierson etal., 1996), but erosion in basins disturbed by rain-triggered

lahars on persistently active volcanoes has not been prop-erly assessed (Harris et al., 2006; Lavigne and Thouret,2002).

This study on the intermediate and lower reaches ofthe Koboan and Lengkong rivers is based on aerial pho-tographs (1981 and 1990), a SPOT 5 image (2003), ge-ological mapping from low-altitude aerial photos (2005and 2008), and D-GPS and terrestrial LIDAR-basedDEMs acquired between 2005 and 2011. We examine acycle of aggradation and degradation (sensu Pierson etal., 2011), which followed the huge input (3.5 to 6 mil-lion m3) of sediment after the 1994-95 eruptions.

Long-term evolution (1981-2010) of the Koboan-Lengkong Basin

The 1981 and 1990 aerial photos and the 2003 SPOTimage show the following response: (1) separation of theupper basin of the Smut river by the 1941 lava flow; (2)widening of the entrenched channels except along the


Aggradation and degradation from lahars in a catchmenton the active Semeru – mapped and measured from

DEMs, aerial photographs and satellite imagery

J.-C. Thouret*, J.-F. Oehler**, A. Solikhin***, A. Gupta****, S.C. Liew****

* PRES Clermont, Université Blaise Pascal, Laboratoire Magmas et Volcans, 5 rue Kessler, 63038 Clermont-Ferrand, France.** Altran Ouest, Atlantide, Technopôle Brest-Iroise, Brest, France.*** Centre of Volcanology and Geological Hazard Mitigation, Bandung, Indonesia.**** Centre for Remote Imaging, Sensing and Processing, National University of Singapore, Singapor.

Abstract: The geomorphic response of rivers to disturbances from volcanic eruptions has been studied from lahars of the Semeru vol-cano, Indonesia. The study also estimates the sediment balance between aggradation and degradation in stream channels. Degradationincreased in the Lengkong River in 2007-2008, 12 years after the last eruption of the Semeru and the annual downcutting of the riverbed reached 0.5-0.85 m. The annual sediment yield for the surveyed channel and catchment were 3-8x105 and 3.3x104 m3km2, respec-tively. The basin has not yet entirely recovered from disturbances induced by the 1941 lava or the 1994-95 pyroclastic flows. The patternof aggradation and degradation varied between the Lengkong and Koboan rivers. A rapid rate of aggradation was followed by degrada-tion at the rate of 37-80 m3 in the Lengkong, whereas aggradation continued after 16 years in the Koboan. The measured annual sedimentyield (2007-2010) from the catchment area was 1.5-3.3 x104 m3km2, an order less than the published values for similar volcanoes includ-ing the Semeru.

Keywords: lahar, Semeru, aggradation, degradation, DEM.

Fig. 1 – Location of the survey area on the SE flank of Semeru vol-cano.


J.-C. Thouret et al.

lowermost Lengkong, where the nar-row rectangular channel follows aNW-SE trending fault (Fig. 2), and;(3) formation of a set of degradation-al terraces formed along the deposi-tional reach of the lower Koboan andLengkong rivers (Fig. 3).

The observations on the Semerudiffer from examples documentedelsewhere on active volcanoes (Proc-ter et al., 2010; Pierson et al., 2011).This is due to: (1) a regular supplyof tephra from daily eruptions gen-erating a high volume of sedimentin the upper reach of the valley; (2)volcanic gravel and sand transport-ed from the summit area towardsthe lower catchment in weekly la-hars and/or floods during the six-month rainy season; (3) pyroclasticflows, occurring 5 to 7 years on av-erage, that supply 3.5 - 6 106 m3 perbasin leading to a rapid formationof aggradational terraces; (4) rapiddowncutting in the Lengkong chan-nel after a 12-year period of gradu-al aggradation that used up the py-roclastic debris of the upper reach-es; (5) a combination of the dailytrigger rainfall of 50 - 120 mm anddischarges from tributary channelsand aquifers (A in Fig. 2C) leadingto small (<400 m3/s) and large (400- 600 m3/s) lahars.

The channels of the Koboan andLengkong differ morphologically.The Koboan has a wide braided chan-nel, adjusted to the ongoing transferof a huge amount of sediment fol-lowing the emplacement of the1994-95 pyroclastic-flow deposits.This transfer of sediment along themiddle section of Koboan shows upin the 2008 3D-DEM. In contrast,the single channel of the Lengkong is entrenched in sedi-ment deposited in 2007. The terraces at the confluence ofthe two rivers have been eroded but sediment migratingdown the Koboan has buried the volcaniclastic fan betweenthe two rivers in 2005.

Short-term changes (2006-2010) in the Lengkong River

Aerial photos and images of the last thirty-odd years indi-cate that channel aggradation and degradation are governed by

Fig. 2 – Interpreted geological sketchof the basin of the Koboan-Lengkongrivers from 1981 (A) and 1990 (B) aeri-al photos and (C) a SPOT5 (2.5 m pixel)image (23/10/2003). Note: (1) the c.1 kmoverbank area of the 1994-95 pyroclastic-flow deposits; 2) the c.250 m and 100 msouthward shift of Koboan and Lengkongrivers, respectively.


Aggradation and degradation from lahar in a catchment on the active Semeru

the geomorphology of the two rivers (Figs.2 and 3). Figure 4 illustrates such changesin the lower Lengkong between 2006 and2010 as recorded by annual DEMs.

Degradation increased in the Lengkongin 2007-2008, 12 years after the volcaniceruption of 1994-95. The rate of degrada-tion ranged from 37 to 80 m3/m and riverbed downcutting from 0.5 m to 0.85 m peryear. The computed average annual sedi-ment yields for the channel area surveyedand the catchment were) 3-88x105 and 1.5to 3.3x104 m3/km2, respectively (Table 1).

Discussion

The geomorphic response to sedimentloading after the 1994-95 eruption wastwofold: rapid aggradation in the riverchannels followed by degradation that hasincreased in the Lengkong since 2007.Given the larger volume of the input, sed-iment continues to migrate down theKoboan. This study is constrained by (1)lack of information regarding remobiliza-tion of sediment in the upper channels;(2) lack of rainfall data above an eleva-tion of 1000 m, and.; (3) suspected recentchanges in the rainfall pattern. We alsoneed to investigate: (i) possible recentvariations in sediment supply and; (ii) an

Fig. 3 – Mosaic of the 2005 and 2008 low-altitude aerial photographs of the Lengkongand Koboan valleys draped on an 3D-DEM. Processed with SonarScope 3D Viewersoftware (courtesy of IFREMER).

Fig. 4 – A. Shaded relief DEMs of the lowerLenkong reach, calculated from D-GPS (2006to 2010) and LIDAR data (2010). B. Elevationdifference map between 2007, 2008 and 2010,calculated from the DEMs of Fig. 4A.



apparent decrease in lahar frequency and/or magnitude overthe last three years.

Conclusion

Geomorphic responses of rivers to volcanic disturbance ofthe Semeru indicate:

1. River basins have not yet recovered from the effects ofthe 1941 lava flow and the 1994-5 pyroclastic flows,

2. The pattern of aggradation and degradation after a vol-canic disturbance varies. Aggradation has been continuousin the Koboan for 16 years, whereas fast aggradation hasbeen followed by degradation at a rate of 37 -80 m3 / m inthe Lengkong since 2007.

3. The measured annual (2007-2010) catchment sediment yield is 1.5-3.3x104 m3/km2/yr, one order less thanpublished values on similar volcanoesand past values determined from theSemeru itself.

References

Harris A.J.L., Vallance J.W., Kimberly P.G.,Rose W.I., Matias O., Flynn L.P., GarbeilH. (2006) – Downstream aggradation owingto lava dome extrusion and rainfall runoff atVolcano Santiaguito, Guatemala. Geologi-cal Society of America Special Paper 412,85-104.Lavigne F., 2004 – Rate of sediment yield fol-lowing small-scale volcanic eruptions: a quan-titative assessment at the Merapi and Semerustratovolcanoes, Java, Indonesia. Earth Sur-face Processes and Landforms 29, 1045-1058.

Lavigne F., Thouret J.-C. (2002) –Sediment transportation and deposi-tion by rain-triggered lahars at Mer-api volcano, central Java, Indonesia.Geomorphology 49, 45-69. Major J.J., Pierson T., Dine-hart R., Costa J. (2000) – Sedimentyield following severe volcanic dis-turbance – A two decade perspec-tive from Mount St. Helens. Geolo-gy 28, 819-822.Pierson T.C., Pringle P.T., CameronK.A. (2011) – Magnitude and timiongof downstream channel aggradation

and degradation in response to a dome-building eruption at Mt. Hood,Oregon. Geological Society of America Bulletin 123, 3-20.

Pierson T.C., Daag A.S., Delos Reyes P.J., Regalado M.T.M.,Solidum R.U., Tubianosa B.S. (1996) – Flow and deposition ofposteruption hot lahars on the east side of Mt. Pinatubo, July-October 1991. In: Newhall C.G., Punongbayan R.S. (Eds.), Fireand Mud: Eruptions and Lahars of Mt. Pinatubo, Philippines.Phivolcs & Univ. Washington, Seattle, pp. 921-950.

Procter J.N., Cronin S.J., Fuller I.C., Lube G., Manville V.,(2010) – Quantifiying the geomorphic impact of a lake breakoutlahar, Mount Ruapehu, New Zealand. Geology 38, 67-70.

Thouret J.C., Lavigne F., Suwa H., Bambang Sukajat, Surono,(2007) – Volcanic hazards at Mount Semeru, East Java (Indonesia),with emphasis on lahars. Bulletin of Volcanology 70, 221-244.

Fig. 5 – Elevation difference maps between 2010-2008 (A) and 2008-2007 (B) draped on3D-views of the Lengkong valley DEMs.

Tab. 1 – Aggradation - degradation balance between 2007 and 2010 calculated from D-GPSand LIDAR DEMs.

Aggradation Degradation BudgetLinear

aggrad./degrad. rate

Sediment yield

(m3) (m3) (m3) (m3/m)surveyed

area(m3/km2/yr)

catchmentarea

(m3/km2/yr)

2010-2008 31 440 106 980 -75 540 -37,77 -334247,79 -15737,50

2010-2007 16 940 176 280 -159 340 -79,67 -470029,50 -33195,83

2008-2007 31 440 120 590 -89 150 -44,58 -788938,05 -18572,92

Introduction – J.-C. Thouret, S.C. Liew, A. Gupta............................................................................................................

Welcome Address – L.K. Kwoh .......................................................................................................................................

International charter for earthquake in Japan – M. Nagai (Keynote address).............................................................

The contribution of space based observations to understanding and addressing geohazards: a CNESperspective – S. Hosford..................................................................................................................................................

The International Charter for disaster mitigation: Participation by ASTER project – M. Abrams and K. Duda ......

Remote sensing of earthquake effects following the 2010 Mw 7.1 and 2011 Mw 6.3 events in Canterbury, NewZealand – S. Levick...........................................................................................................................................................

Understanding the fatal 2006 dike breaching of Mayon Volcano using high- resolution imageries – R. Eco,A.M.F. Lagmay, E. Paguican..............................................................................................................................................

Geology, tectonics, and the 2002-2003 eruption of the Semeru Volcano, Indonesia: Interpreted from high-spatialresolution satellite imagery – A. Solikhin, J.-C. Thouret, A. Gupta, A.J.L. Harris and S.C. Liew...................................

Satellite remote-sensing analysis of casualties and damage from the 2010 eruption of Merapi volcano –F. Lavigne ..........................................................................................................................................................................

Risk microzonation of Yogyakarta city following to the 2010 eruption of Merapi Volcano – D.S. Hadmoko,L.W. Santosa, M.A. Marfai and F. Lavigne.........................................................................................................................

Quantifying volcanic hazard and risk – C. Magill (Keynote address) ...........................................................................

Topographic characterization of the Auckland Volcanic Field (New Zealand) – Implications for lava flow hazardmapping – G. Kereszturi, J. Procter, K. Németh, J. Lindsay, J. Kenny, S.J. Cronin, M. Bebbington and G. Jordán .........

Application of TanDEM-X data to volcanic hazard assessment and mapping: Example from Merapi Volcano,Indonesia – S.J. Charbonnier, C.B. Connor, L. Connor, T.Dixon and R. Gertisser .........................................................

The role of remote sensing data on the 2010 crisis at Merapi Volcano, Indonesia – Surono, A. Solikhin,A.B. Santoso, P. Jousset, J.S. Pallister, M. Boichu and S. Carn........................................................................................

Detecting fault slip at Mayon Volcano using permanent scatterer interferometry – A.M.F. Lagmay, M.G. Bato,E.M.R. Paguican and H. Zebker ......................................................................................................................................

Sand dune conservation zone based on tsunami inundation hazard in Parangtritis coastal area, Bantulregency, Yogyakarta special province – R.F. Putri, D. Mardiatno, J. Sartohadi and J.T. Sri Sumantyo ........................

Evacuation route determination for tsunami mitigation using remote sensing data and GeographicInformation Systems at Parangtritis coastal area, Yogyakarta-Indonesia – J. Mardiatno, R.F. Putri, M. Susmayadiand D.S. Sayudi..............................................................................................................................................................

Remote sensing of volcanic emissions in the Asia-Pacific region – S.A. Carn (Keynote address)..........................

Monitoring carbon dioxide emissions from volcanoes from space and from ground based networks –F. M. Schwandner, C.G. Newhall and S.S. Marcial............................................................................................................

Detection of volcanic dust by AERONET sunphotometers – S. Salinas and S.C. Liew .............................................

PSInSAR detection of ground subsidence and fault movement in Muntinlupa City, Metro Manila and Biñan,Laguna – A.M.F. Lagmay, R.N. Eco and J. Agdeppa........................................................................................................

Tropical forest monitoring, socializing the pixel to inform management and livelihood implications: A casestudy from Indonesian West Timor – R. Fisher .............................................................................................................

El Niño and rainfall influence on the temporal and spatial patterns of vegetation fires in insular Southeast Asia– S.C. Liew and J. Miettinen ..............................................................................................................................................

Assessing the hydrologic response of wetlands to restoration: A remote sensing perspective – A.M. Melesseand F. Miralles-Wilhelm......................................................................................................................................................

Contents

3

5

7

11

15

15

17

21

25

25

27

33

35

39

43

47

51

55

61

61

63

67

67

69


Remote sensing of reservoirs in large Asian river basins – X.X. Lu, L. Ran, X. Yang and S. Liu.........................................

Large floods and average climate: Is there a relationship in the Asia-Australia Monsoon Region? – R. Wasson ....

Spatial information for analyzing changing hydro-meteorological risk – C.J. Van Westen (Keynote address)........

Global warming on large tropical rivers – A. Gupta and S.C. Liew..............................................................................

Jakarta flood modeling with different sources of topographic data – C.D. Doan, S.-Y. Liong and R. Sanders........

Bank erosion and channel change in the Daly River, Northern Australia – S. Karki, B. Wasson, D. Pearson, S. Maierand W. Ahmad ...................................................................................................................................................................

Flood risk analysis and mapping in Gorontalo city, Indonesia, using high resolution Google Earth’s imagery –B. Tjahjono, M. Hidiya and B.H. Trisasongko .....................................................................................................................

Quantifying sediment budgets in data poor environments of SE Asia and Northern Australia; a review ofRemote Sensing, GIS and isotope based approaches – S. Hobgen, G. Boggs, B. Myers and R. Wasson ...............

Exploring the Red River, Vietnam using SPOT data – M. Nguyen Dinh ......................................................................

Development of a Peace Corps Master’s International program focused on volcanic hazard mitigation inIndonesia – W.I. Rose, S.A. Carn and J.J. Wellik Li .........................................................................................................

Quantifying population exposure near volcanoes using physical models of natural hazards and globallyconsistent data sets: Case studies from Java, Indonesia – J.J. Wellik Li, R.E. Wolf and S.A. Carn ..........................

Aggradation and degradation from lahars in a catchment on the active Semeru: mapped and measured fromDEMs, aerial photographs and satellite imagery – J.-C. Thouret, J.-F. Oehler, A. Solikhin, A. Gupta, S.C. Liew ........

71

71

73

77

79

79

81

85

89

89

91

93

Achevé d’imprimer en décembre 2011Diazo1 Reprographie

©CERAMACClermont-Ferrand, 2011ISBN 978-2-84516-554-0

ISSN 1242-7780

Documents

Sensores remotos y Riesgo ambiental Serie