This article was downloaded by: [University of Tasmania]On: 30 November 2014, At: 04:05Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Geocarto InternationalPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tgei20

Monitoring disasters with aconstellation of satellites – typeexamples from the InternationalCharter ‘Space and Major Disasters’Ahmed Mahmood aa Canadian Space Agency, Space Utilization , 6767 route del'aéroport, Saint-Hubert , Quebec , J3Y 8Y9 , CanadaAccepted author version posted online: 23 Sep 2011.Publishedonline: 26 Oct 2011.

To cite this article: Ahmed Mahmood (2012) Monitoring disasters with a constellation of satellites –type examples from the International Charter ‘Space and Major Disasters’, Geocarto International,27:2, 91-101, DOI: 10.1080/10106049.2011.622051

To link to this article: http://dx.doi.org/10.1080/10106049.2011.622051

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Monitoring disasters with a constellation of satellites – type examples

from the International Charter ‘Space and Major Disasters’

Ahmed Mahmood*

Canadian Space Agency, Space Utilization, 6767 route de l’aeroport, Saint-Hubert,Quebec J3Y 8Y9, Canada

(Received 7 July 2011; final version received 6 September 2011)

In the case of a major disaster, information derived from satellite observation isnot only highly useful, it may at times be indispensable because of the damagecaused by the disaster to ground infrastructure. The International Charter ‘Spaceand Major Disasters’ (‘the Charter’) has been one of the primary sources ofsatellite data for the past 11 years to cover events like floods, fires, tsunamis,ocean storms, earthquakes, volcanic eruptions and oil spills. With the growingmembership of the Charter, an increasingly large number of sensors are nowavailable, which can be planned with the required temporal frequency andspectral range to cover a disaster event. Some of the type Charter activation casesare reported in this article to demonstrate the innovative use of multi-satelliteimagery for disaster response.

Keywords: Space Charter; satellite disaster response; change detection

1. Introduction

The idea of the Charter as a mechanism to supply satellite data-derived informationin situations of crisis was introduced by the European (ESA) and the French (CNES)space agencies during the UNISPACE III conference held in 1999 (Bessis et al. 2003,2004). Soon after the signing of the Charter by these two, the Canadian SpaceAgency (CSA) came on board in October 2000 as its third member. The threefounding members then established the necessary infrastructure to implementthe Charter. Since its inception, the Charter membership has grown to 14 with theIndian Space Research Organization (ISRO), the US Geological Survey (USGS) andthe National Oceanic and Atmospheric Administration (NOAA), the Argentineanspace agency (CONAE), UK Space Agency together with the Disaster MonitoringConstellation (DMCii), Japan Aerospace Exploration Agency (JAXA), ChinaNational Space Administration (CNSA), the German Aerospace Centre (DLR),Korea Aerospace Research Institute (KARI), Brazilian National Institute forSpace Research (INPE), and the Russian Federal Space Agency ROSCOSMOS. TheCharter is therefore benefiting from a growing number of satellites that increasethe revisit frequency and the choice of sensor for spectral and spatial resolution.

*Email: ahmed.mahmood@asc-csa.gc.ca

Published with the permission of Her Majesty the Queen in Right of Canada, � Governmentof Canada 2011.

Geocarto International

Vol. 27, No. 2, April 2012, 91–101

ISSN 1010-6049 print/ISSN 1752-0762 online

� Canadian Space Agency, 2011

� Agence spatiale canadienne, 2011

http://dx.doi.org/10.1080/10106049.2011.622051

http://www.tandfonline.com

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

Data from these sensors are processed, merged and interpreted in a variety of waysto extract the best possible information on the effects of a given disaster, as brieflydescribed in this article.

The Charter is a multi-satellite operational system to provide space-bornedata on disasters that may have the potential of causing significant loss of life orproperty.

These data are viewed as an important complement to ground-based informa-tion, which is often difficult to generate because the infrastructures installed togenerate the information are among the first victims of a disaster event.Furthermore, satellites are the most effective means for synoptic viewing for disasterresponse on an emergency and priority basis. The Charter data are delivered to userson the ground with fast turnaround and at no cost. As can be understood from theforegoing account, the Charter covers only the response phase of a disaster, the laterrecovery and rehabilitation efforts are excluded. Only basic data (georeferencedimagery) are provided, although any further value-adding and informationextraction is generously sponsored by the individual Charter members. The satellitedata are provided with due regard to the policies and procedures of the Charterand its member space agencies. A complete review of these policies, procedures andprogrammes are given in a comprehensive account on the Charter (Mahmood andShokr 2008).

2. Charter implementation

This section describes the administrative and operational architecture of the Charter,in terms of its governing bodies and implementation functions.

2.1. Administration

The Charter implementation is carried out primarily by an Executive Secretariat(Mahmood et al. 2002, 2005) comprising the representatives of the member spaceagencies that have assumed full Charter functions. The Executive Secretariat (ES)works under the supervision of a Board where all the members are represented.The Charter business is conducted by regular conference calls of the ExecutiveSecretariat and biannual Board meetings hosted in turn by the member agencies.The agency hosting a Charter Board meeting assumes the lead role and becomesthe Charter point of contact until the following meeting. A well structured ftpsite serves as a virtual Charter hub and its repository of all the current andarchived information and documentation. The Charter website (http://www.disasterscharter.org) is kept up-to-date with the latest Charter activities andactivations.

2.2. Operation

The Charter can be activated by a group of predefined users for obtaining datacovering the crisis that follows a major disaster event. These users are called‘authorized users’ (AUs). They are primarily institutions or services responsible forrescue and civil protection, defense and security under the authority of a State whosejurisdiction covers an agency or a space operator that is a member of the Charter.A series of operational entities come into play as soon as a Charter activation request

92 A. Mahmood

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

is received from an AU, and this process has been described (Mahmood et al. 2004,2008). In summary, the activation request is sent by the AU to a centralised 24 h/daycall receiving unit staffed with an ‘on-duty operator’ (ODO) and, after initialvalidation of the request, is passed on to an ‘emergency on-call officer’ (ECO), who isalso accessible 24 h/day. The ECO carries out in-depth verification of the request byinteracting with the AU and by using the ECO’s own resources. The ECO prepares averbose record of the request and establishes a multi-satellite data acquisition planfor spacecraft tasking and archive retrievals by the individual member agency staff,the whole of which is referred to as the ‘Dossier’. As soon as an activation requesthas been accepted, the ES nominates a Project Manager (PM). The ECO thentransfers the Dossier to the PM, who takes over the responsibility of managing theentire activation project, including supervision of data processing, any value-adding,data delivery, end user assistance and feedback. The PM finally submits a report onthe activation to the ES.

3. Disaster coverage

The Charter AUs have placed activation requests not only for their own countriesbut also for others with which they have arrangements of cooperation in the event ofmajor disasters. In addition, the Charter Board has assigned to United NationsOffice for Outer Space Affairs (UNOOSA) and UNITAR’s Operational SatelliteApplications Program (UNOSAT)/United Nations Office for Training and Research(UNITAR) the privilege to channel UN humanitarian assistance requests in the faceof similar disasters. As a result, the Charter’s reach has been truly worldwide, andthe multi-satellite data have been acquired for a variety of disaster events: floods,fires, earthquakes, ocean storms, ocean waves, volcanic eruptions and oil spills.

Typically, satellite data are analyzed for change detection by merging the newlyacquired, post-disaster imagery with reference imagery retrieved from the archives.This signifies the wealth of satellite data archives that are available to the Charter.The resulting products comprise maps and geocoded image overlays showingflooded surfaces and derived water depths, lava flows, linear elements (affected roadsand bridges), hot spots, burnt areas, landslide scars and eruptive edifice.

A limiting factor in change detection is the spatial and spectral resolution, and insome cases, the illumination geometry and the polarisation of the incident andemitted electromagnetic waves. Furthermore, the most recent imagery pre-dating thedisaster event is selected from the archives to ensure scene coherence. Thecharacteristics of the sensors of the current Charter constellation are summarisedin Table 1. These can be broadly divided into electro-optical and microwave sensors.The most common of the latter type are the synthetic aperture radars (SARs). Thedetection of some of the disaster effects can be determined equally well by the twotypes of sensors; for others like burning fires, only thermal and optical devices aredeployed.

Figure 1 shows the distribution of the Charter activations to date with regard tothe various disaster types.

3.1. Flood

Floods have been the most commonly covered disaster, representing roughly 50%of the Charter requests received (Figure 1). Typically, the main feature of the

Geocarto International 93

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

Table 1. Charter constellation sensor characteristics.

Chartermember Satellite sensors

Spatialresolution (range)

Spectral range(nm/mm)/number

of beams

CNES SPOT 1, 2, 3 10–20 m 510–890 nmSPOT 4 10–1165 m 500–1700 nmSPOT 5 2.5–1165 m 500–1700 nmFORMOSAT-2 2–8 m 450–900 nm

CNSA CBERS 20–260 m 500–12500 nmCONAE SAC-C 35–234 m 400–1700 nmCSA RADARSAT-1 8–100 m 33 (C-Band)

RADARSAT-2 3–100 m 4200 (C-Band)DLR TerraSAR-X 1–18 m 229 (X-band)

TanDEM-X

DMCII UK-DMC 32 m 520–900 nmBeijing-1 4–32 m 500–900 nmNigeriaSat-1 32 m 520–900 nmBilsat-1 12–26 m 520–900 nmAlsat-1 32 m 520–900 nm

ESA ERS 25 m N/Aa (C-Band)ENVISAT–ASAR 25–1000 m N/Aa (C-Band)PROBA 5 m 400–1050 nm

ISRO Resourcesat-1 (IRS-P6) 5.8–56 m 520–1770 nmCartosat-1 (IRS-5) 2.5 m 500–850 nm

JAXA ALOS 2.5–100 m 420–890 nmKARI KOMPSAT-2 1m (PAN), 4 m (MS) 0.50–0.90 mm (PAN)

0.45–0.90 mm (MS)NOAA POES 1100 m 580–12500 nm

GOES 1000–8000 m 550–12500 nmUSGS Landsat 7 15–60 m 450–12500 nm

Landsat 5 30–120 m 450–12500 nmQuickBird 60 cm–2.4 m 445–900 nmIKONOS 0.82–3.2 m 445–900 nmGeoEye 1 0.41–1.65 m 450–900 nm

aN/A¼Not available.

Figure 1. Distribution of the total Charter activations to date, by disaster type.

94 A. Mahmood

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

products is based on deriving flooded surfaces, either from the nature of the spectralresponse of these surfaces, and/or by comparison with the reference imagerypredating the disaster. Typically, the flood vectors are presented as final products inGIS layers.

More recently, with the advent of high-resolution SARs, such as RADARSAT-2and TerraSAR with ultra-Fine and Spotlight modes of 3 m and 1 m spatialresolution, respectively, new products showing changing water depths of inundatedareas have been created. One such example is shown in Figure 2. Here, the ultra-FineRADARSAT-2 data were used for the first time by the Charter value-adders for thewater depth estimates in paddy fields that were flooded following the passage ofCyclone Nargis in the Yangon region of Myanmar. The estimates were derived froma content and textural analysis of the RADARSAT-2 image acquired on 7 May2007. Darker blue colouration is deeper water that would have taken longer timeto drain. This information was considered very helpful in the rehabilitation andrestoration period following the response phase of the disaster. SPOT referenceimagery was used for landcover classification and the Shuttle Radar Topography

Figure 2. Qualitative water depth estimates in paddy fields in the Yangon region ofMyanmar. Darker blue indicates potentially deeper water, while the light blue areas indicatelower water levels. RADARSAT-2 ultra fine mode crisis image acquired on 7 May 2008.RADARSAT-2 Data and Products � MacDONALD, Dettwiler and Associates Ltd. 2008 –All Rights Reserved. RADARSAT is an official mark of the Canadian Space Agency. Basemap is a SPOT-5 image acquired on 2 March 2008, � CNES 2008. Map product courtesy ofSERTIT.

Geocarto International 95

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

Mission (SRTM) digital elevation models (DEMs) were used for the water columnheights reported on the 3 m pixel RADARSAT-2 imagery.

3.1.1. Tsunami

The massive Asian tsunami of December 2004 remains the biggest disaster coveredby the Charter in its entire history. It was extensively covered by means of threeseparate Charter activations and by a combined use of optical and radar satellites(IRS, ENVISAT, RADARSAT, SPOT and Landsat), including imagery procuredfrom high-resolution commercial satellites (IKONOS, QuickBird). Image productscomprised extracts of satellite imagery in real and in natural colours, space mapson various scales, ‘before-and-after’ images, damage maps and 3D overview videos.Some of these products can be found in a publication by Mahmood and Shokr(2008).

3.2. Volcanic eruption

The volcanic cloud often hampers observation of the crater and the surfacetopography during the eruption. Radar backscattering properties of the groundfeatures are used to delineate the different types of volcanic deposits. A goodexample of the use of SAR imagery for this disaster type comes from theCharter coverage of the Iceland volcanic eruption that halted air traffic over much ofEurope in April–May 2010. The map depicted in Figure 3 is based on comparingRADARSAT-2 multi-look Fine (approximately 6 m resolution) acquisition withthose of April 9 and 20 acquisitions to highlight phenomena induced by the volcaniceruption. The first feature to be noted is the shape of the crater and the lava/waterchannels to the north. The second element of the image map is the ash deposit withtwo rings on the southern and the eastern parts of the volcano as well as agriculturefields near the coast that are totally blurred. On the northwestern slope of thevolcano, the braided network is reactivated as seen on the 20 and 23 April images.The multi-look Fine imagery of RADARSAT-2 is a significant improvement overthe single-look Fine imagery of the predecessor RADARSAT-1 in terms of spectralnoise levels and, consequently, the ease of discerning the terrain units on the basis oftheir radar tones and textures.

3.3. Fire

The thermal and optical sensors onboard SPOT, MODIS and ALOS AVNIRalong with IRS satellites constitute the core Charter capability for monitoringthis disaster type. The coverage of forest fires in British Columbia (Canada) inAugust 2003, which affected some urban areas also, were adequately covered withSPOT, and products in various forms (drape-overs, fly-throughs, etc.) weredelivered to show both the burning foyers as well as the burnt surfaces. Large-scale burnt area mapping of the 2009 Greek fires was carried out with the abovesatellite sensors. The map in Figure 4 shows the burnt areas derived by means ofanalysis of SPOT-5 satellite data acquired on 25 August 2009, and the location ofactive hotspots on different days. The latter were automatically derived fromMODIS Fire Information System. The fire analysis underlines the fact thatdifferent sensors of the Charter constellation are tasked for particular disaster

96 A. Mahmood

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

targets, but the results are merged in single products for the facility of use by thefield staff.

3.4. Earthquake

The devastating Kashmir (Pakistan/India) earthquake of 8 October 2005 wascovered on two separate AU requests. The main data requirement was high-resolution imagery to detect damage to towns and villages, roads, communicationlines and other infrastructures. Given the hilly nature of the terrain, there was aparticular concern about the landslide damage. In fact, some new landslides weredetected on the satellite imagery; others were old landslides that were reactivated bythe earthquake. Fine mode RADARSAT-1 SAR and very high resolution (VHR)optical data from commercial satellite systems, such as IKONOS and Quickbird,were sought by the users. Topographic map updates using SRTM, Landsat andIKONOS data were prepared for logistical support and damage assessment

Figure 3. Interpretation of 20 April 2010 RADARSAT-2 Extended High beam image ofEyjafjallajokull volcano (interpretation based on comparison with reference image of 9 April2010). RADARSAT-2 Data and Products � MacDONALD, Dettwiler and Associates Ltd.2010 – All Rights Reserved. RADARSAT is an official mark of the Canadian Space Agency.Map product courtesy of SERTIT.

Geocarto International 97

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

purposes. In the case of an earlier Algerian earthquake that hit the northern part ofthe country in May 2003, the optical satellite imagery from SPOT helped in locatingthe evacuation and temporary refuge sites.

Since these earthquake events, the Charter has obtained regular access tocommercial VHR imagery, and earthquake damage assessment techniques haveimproved so that the very large amount of VHR data was analyzed applying somenew methods for recognising the degree of damage to buildings in the Port-au-Princearea as a result of the Haiti earthquake in January 2010. The map in Figure 5 isthe first of its kind in terms of urban damage assessment with satellite data. It showsthe location of buildings presumed to be badly damaged, some 4346 of them, and isbased on the observations made on the 50-cm resolution optical imagery of GeoEyesatellite over the city of Port-au-Prince. The building damage assessment follows thenumber of such buildings per km2. The methodology applied produces segmentationof urban blocks according to the landuse. Image analysis is carried out by means ofphoto-interpretation with two objectives for each segmentation block, i.e. the overallassessment of damage in terms of the three classes given in the caption of Figure 5,and the count of damaged buildings. The products are revised through an iterative

Figure 4. SPOT-5 crisis image of fires in the northern part of Attica, Greece, acquired on 25August 2009, � CNES 2009. Red colouring indicates burnt areas, while the yellow and orangedots indicate hot spots. USGS 2000; map produced by ZKI DLR.

98 A. Mahmood

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

process, and the damage assessment is coded as number of affected buildings withinthe area of each urban block.

3.5. Hurricane

For this disaster type, the Charter is activated as soon as an ocean storm attainshurricane strength and the landfall is imminent, which was the case with Katrina.Twenty-five space maps were produced covering the New Orleans flooded surfaces,Biloxi and the surrounding areas, and the coastal islands south of Mobile. Thesemaps were provided in addition to raw radar imagery. Some good quality value-added products showing different types of damage (flooding, oil spill due to damagedrigs, debris flow and other infrastructural damage) were produced and madeavailable to the AUs (see Figure 6).

Figure 5. GeoEye crisis image of Port-au-Prince acquired on 13 January 2010, � GeoEye2010. The map indicates damage to buildings as follows: red indicates obvious/widespreaddamage; yellow indicates evident but sporadic damage; green indicates scarce or non-visibledamage. Map product courtesy of SERTIT.

Geocarto International 99

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

4. Conclusion

The Charter operations are meant to assist the end users involved in disasterresponse with space-based information derived from the image products. Theincreasing number of satellites and sensors result in rapid response through shorterrevisit frequency and better recognition of the target with improved spectralresolution. A beneficial by-product of this entire process is that the data acquiredcan subsequently be made available for scientific research. This article hashighlighted some of the current processes of the Charter, some of the constantlyimproving interpretation techniques used, and the growing interest of disaster reliefand response organisations in this international initiative.

Acknowledgements

The author thanks the members of the Charter Executive Secretariat for their support and forhelping with the material used in this article.

References

Bessis, J.-L., Bequignon, J., and Mahmood, A., 2003. The International Charter ‘Space andMajor Disasters’ initiative. Acta Astronautica, 54, 183–190.

Bessis, J.-L., Bequignon, J., and Mahmood, A., 2004. Three examples of activation of theInternational Charter ‘Space andMajorDisasters’.Advances in SpaceResearch, 33, 244–248.

Mahmood, A., et al., 2002. An overview of the international charter ‘space and majordisasters’. In: IGARSS, 24–28 June, Toronto, Canada.

Mahmood, A., et al., 2004. Disaster response in Africa by the international charter. In: AfricanAssociation of remote sensing of the environment conference, 17–22 October, Nairobi,Kenya.

Figure 6. Multiple damage type assessment product following Hurricane Katrina.RADARSAT-1 Data � Canadian Space Agency 2005.

100 A. Mahmood

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014

Mahmood, A., et al., 2005. International charter ‘space and major disasters’ – status report.In: IGARSS, 25–29 July, Seoul, South Korea.

Mahmood, A., et al., 2008. South American perspective of the international charter ‘space andmajor disasters’. Advances in Geosciences, 14, 13–20.

Mahmood, A. and Shokr, M., 2008. Space measurements for disaster response: theInternational Charter. In: M. Gad-el-Hak, ed. In: Large-scale disasters: prediction, controland mitigation. Cambridge: Cambridge University Press, 453–541.

Geocarto International 101

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

04:

05 3

0 N

ovem

ber

2014