564 VOLUME 115 | NUMBER 4 | April 2007 • Environmental Health Perspectives

Research

The 2006–2011 National Institute ofEnvironmental Health Sciences’ (NIEHS)Strategic Plan New Frontiers in Environ-mental Sciences and Human Health (NIEHS2006c) seeks to “challeng[e] and energiz[e]the scientific community to use environmen-tal health sciences to understand the causes ofdisease and to improve human health”(National Institutes of Health 2006). Thestrategic plan emphasizes, among otherthings, integrative research, community-linked research, and prioritizing environmen-tal factors that are most likely contributing tohuman disease. Consistent with the goals ofthe strategic plan, new technologies in infor-mation systems, data federation, grid systems,and spatial analytics all hold great promise ofenabling integrated science teamwork whilehelping to disentangle genetic, environ-mental, and other factors that contribute tothe complex etiology of common human dis-eases. Such systems will provide effectivemeans of knowledge and data sharing amongscientists, as well as rapid and efficientapproaches for sorting through and analyzing

large data sets; thus, these systems will sup-port both the scientific and policy processeswhile accelerating the rate of knowledge pro-duction and discovery.

The need for such systems to addressenvironmental health concerns was madeespecially apparent when a series of devastat-ing hurricanes ripped across the Gulf Coast ofthe United States during 2005 (Schwartz2005). The effects of Hurricane Katrina wereespecially severe: The human and environ-mental health impacts on New Orleans,Louisiana, and other Gulf Coast communitieswill be felt for many decades to come[Daniels et al. 2006; Nates and Moyer 2005;Pastor et al. 2006; Pezzoli et al. 2006; U.S.Environmental Protection Agency (EPA)2006]. The Federal Emergency ManagementAgency (FEMA) estimates that Katrina’sdestruction (exacerbated by the chronic lossof wetlands, land subsidence, and inadequateinfrastructure) disrupted the lives of roughly650,000 Americans (U.S. Department ofCommerce 2006). Over 1,300 people died.The projected economic costs for recovery

and reconstruction are unprecedented—likelyto exceed $125 billion (Pastor et al. 2006).

The density of industrial operations inNew Orleans and the Mississippi Delta beforethe Katrina disaster combined with the releaseof contaminants into floodwaters and theirtransport throughout urbanized areas to giverise to many significant environmental healthconcerns. At the same time, the flooding ofthousands of homes and > 100 schools mayhave created indoor environments thatthreaten human respiratory and neurologichealth through the growth of mold and otherbioaeroallergens. All of these potential expo-sures took place within the context of highpsychosocial stress and, for many, low socio-economic status—conditions that often exacer-bate the effects of adverse environmentalexposures (Berube and Katz 2005). Health dis-parities that result from poverty and inade-quate infrastructure raise serious concernsabout environmental justice (Pellow and Brulle2005; Thomson et al. 2006). The scale and

Address correspondence to K. Pezzoli, Urban Studiesand Planning Program, 9500 Gilman Dr., #0517,La Jolla, CA 92093-0517 USA. Telephone: (858)534-3691. E-mail: kpezzoli@ucsd.edu

Key partners in development of the NIEHS Portalinclude Columbia University, the University ofKentucky, the Research Triangle Institute, SanDiego State University, the University of NewOrleans, Tulane University, Louisiana StateUniversity, the NIEHS, the U.S. EnvironmentalProtection Agency, the National Oceanic andAtmospheric Administration, the U.S. GeologicalSurvey, and multiple local and state agencies. Thedirector of the NIEHS, David A. Schwartz, providedthe leadership necessary to quickly bring theseresources together during a time of crisis.

The NIEHS Portal was made possible by a supple-mental grant (3 P42 ES010337-07 S1) to theUniversity of California, San Diego (UCSD)Superfund Basic Research Program (SBRP). UCSD’sSBRP is funded by NIEHS grant P42 ES010337. TheNIEHS Portal also benefited from a supplementalgrant to Duke University’s SBRP (5 P42-ES-010356-07 S1). The Telescience Portal infrastructure adaptedfor the NIEHS Portal was developed under supportfrom the National Center for Research Resources(NCRR) to M. Ellisman and P. Arzberger at UCSD(NCRR Award P41 RR004050 and P41 RR008605).

The authors declare they have no competingfinancial interests.

Received 9 October 2006; accepted 22 January2005.

The NIEHS Environmental Health Sciences Data Resource Portal: PlacingAdvanced Technologies in Service to Vulnerable Communities

Keith Pezzoli,1 Robert Tukey,2,3 Hiram Sarabia,1 Ilya Zaslavsky,4 Marie Lynn Miranda,5 William A. Suk,6

Abel Lin,7 and Mark Ellisman7

1Urban Studies and Planning Program, 2Department of Pharmacology, 3Department of Chemistry and Biochemistry, and 4San Diego Supercomputer Center, University of California San Diego, La Jolla, California, USA; 5Nicholas School of the Environmentand Earth Sciences, Duke University, Durham, North Carolina, USA; 6National Institute of Environmental Health Sciences, NationalInstitutes of Health, Department of Health and Human Services, Research Triangle Park, North Carolina, USA; 7National Center forMicroscopy and Imaging Research, Center for Research in Biological Systems and the Department of Neurosciences, University ofCalifornia San Diego, La Jolla, California, USA

BACKGROUND: Two devastating hurricanes ripped across the Gulf Coast of the United States during2005. The effects of Hurricane Katrina were especially severe: The human and environmental healthimpacts on New Orleans, Louisiana, and other Gulf Coast communities will be felt for decades tocome. The Federal Emergency Management Agency (FEMA) estimates that Katrina’s destructiondisrupted the lives of roughly 650,000 Americans. Over 1,300 people died. The projected economiccosts for recovery and reconstruction are likely to exceed $125 billion.

OBJECTIVES: The NIEHS (National Institute of Environmental Health Sciences) Portal aims toprovide decision makers with the data, information, and the tools they need to a) monitor humanand environmental health impacts of disasters; b) assess and reduce human exposures to contami-nants; and c) develop science-based remediation, rebuilding, and repopulation strategies.

METHODS: The NIEHS Portal combines advances in geographic information systems (GIS), datamining/integration, and visualization technologies through new forms of grid-based (distributed,web-accessible) cyberinfrastructure.

RESULTS: The scale and complexity of the problems presented by Hurricane Katrina made it evi-dent that no stakeholder alone could tackle them and that there is a need for greater collaboration.The NIEHS Portal provides a collaboration-enabling, information-laden base necessary to respondto environmental health concerns in the Gulf Coast region while advancing integrative multi-disciplinary research.

CONCLUSIONS: The NIEHS Portal is poised to serve as a national resource to track environmentalhazards following natural and man-made disasters, focus medical and environmental response andrecovery resources in areas of greatest need, and function as a test bed for technologies that willhelp advance environmental health sciences research into the modern scientific and computing era.

KEY WORDS: community-linked research, cyberinfrastructure, disaster, environmental justice, GIS,grid, health disparities, integrative research, Katrina, telescience. Environ Health Perspect115:564–571 (2007). doi:10.1289/ehp.9817 available via http://dx.doi.org/ [Online 22 January 2007]

complexity of the ensuing environmentalhealth issues made tracking environmental haz-ards and focusing medical and environmentalresponse and recovery resources in areas ofgreatest need an obvious priority. A uniqueopportunity arose to deploy a system to meetthese needs, while also creating a test bed fortechnologies that will help advance environ-mental health sciences research into themodern scientific and computing era.

Objectives

As Hurricane Katrina made landfall, teams ofNIEHS staff mobilized various relief efforts.Within the Division of Extramural Researchand Training, a group of Superfund BasicResearch Program (SBRP) grantees pooledintellectual resources to create an internet por-tal that couples the power of geographic infor-mation systems (GIS) and grid technologies.The NIEHS Environmental Health SciencesData Resource Portal (NIEHS Portal) pro-vides common access and the ability to shareinformation, software, and computer processesacross organizational boundaries within a user-friendly, highly customizable environment(NIEHS 2006a).

The NIEHS Portal aims to provide scien-tists and decision makers with the data, infor-mation, and tools they need to a) monitor andevaluate human and environmental healthimpacts in the Gulf Coast Region; b) assessand reduce human exposures to contaminants;and c) develop science-based remediation,rebuilding, and repopulation strategies. In thisarticle we describe the NIEHS Portal’s aimsand prospects for its utilization from analyticaland applied perspectives. The NIEHS Portalallows users to access demographic, publichealth, infrastructure, and environmental data,all of which are geo-referenced. The spatialdata sets incorporated into the portal includebasic infrastructure data such as roads and elec-tric power plants, potential contaminant-release sources including Superfund and ToxicsRelease Inventory (TRI) sites, hurricane flood-ing data, Census data, physiographic data, andremote sensing imagery both pre- and post-Katrina. Creating a unified environmentalhealth GIS, integrating it in a secure web por-tal, and providing customized research envi-ronments for teams of collaborators are themain technical components of the project.

Context and Methods

The NIEHS Portal represents a proactiveresponse to four trends that, in addition tohelping explain the nature of the portal initia-tive itself, are key to understanding the newdirection and priorities spelled out in theNIEHS strategic plan (NIEHS 2006c). Thefour trends include a) the rising magnitude ofdisasters and vulnerability on a global scale;b) the emergence of new modes of knowledge

production, integration, and networking; c) anincreasing emphasis on integrative and com-munity-linked research in the culture andpractice of science; and d) new investments inexposure biology and gene–environmentinteraction studies.

Rising magnitude of disasters and vulnera-bility on a global scale. The continued func-tion and survival of any human society isdependent, to a significant degree, on itsadaptability, resilience, and vulnerability toenvironmental events (Young et al. 2006). Inspite of our increased awareness of environ-mental hazards and risks, the human, ecologic,and economic costs of disasters continue torise sharply worldwide. Over the last twodecades, disasters including floods, drought,hurricanes, earthquakes, and tsunamis haveaffected vast regions and claimed nearly a mil-lion human lives, indirectly impacted a billionpeople, and generated damages in the trillionsof dollars [Nates and Moyer 2005; U.K.Department for International Development(DFID) 2006]. The effects of these disasterson populations can be explained in part by agrowing human footprint in the environment,particularly in vulnerable areas (e.g., low-lyingcoastal areas, flood plains, steep hillsides),which makes encountering risk and the possi-bility of exposure to environmental hazardsmore likely (Nates and Moyer 2005). At thesame time, global climate change is expectedto increase the frequency and intensity offloods, hurricanes, and wildfires, and lead tosea level change (Barnett et al. 2001; Running2006; Westerling et al. 2006).

As globalization accelerates the rate andspatial scale of human–environment inter-actions, the distinction between natural andman-made disasters becomes blurred. So-callednatural disasters are often exacerbated byanthropogenic factors. In the case of Hurricane

Katrina, for instance, the development of andloss of low-lying wetlands (Figure 1) greatlydiminished a natural ecosystem buffer thatwould have otherwise provided a greater degreeof protection from the storm surge that killedmany people and did severe damage to indus-trial facilities. The later led to a number ofchemical spills that exacerbated the damagealready caused by the hurricane.

The negative impact of environmentalevents can be reduced by good planning. Forexample, this is evident in New Orleans whenwe compare the impact of flooding in newerneighborhoods built on risk-prone, low-lyingwetlands (i.e., New Orleans East) with oldersections of the city that the Spanish andFrench built at safer higher elevations.Similarly, the operation of industrial facilitiesnear densely populated areas can lead toincreased hazards. This is exemplified by theMurphy Oil spill in the St. Bernard Parish ofNew Orleans, where a storage tank at theMeraux Murphy Oil Refinery, adjacent to aheavily populated area, was damaged by floodwaters and spilled an estimated 1.05 milliongallons of crude oil affecting some 1,800 homes(Pezzoli et al. 2006).

Poverty-stricken communities experiencethe most severe social and economic impacts ofdisasters, often with long-lasting and devastat-ing effects (Daniels et al. 2006; Dilley et al.2005; U.K. DFID 2006). The Lower NinthWard, a poverty-stricken part of New Orleans,was utterly devastated (Figure 2). Disastershighlight the interdependence of communitiesat local, regional, and global scales as well asthe importance of building disaster responsesystems that are flexible, adaptable, andresilient (Daniels et al. 2006).

New modes of knowledge production andnetworking. In recent years, information andcommunications technology experts, in

NIEHS Portal

Environmental Health Perspectives • VOLUME 115 | NUMBER 4 | April 2007 565

Figure 1. Wetland destruction caused by Hurricane Katrina near New Orleans, April 2006 (photographedby K. Pezzoli).

partnership with researchers, have builtsophisticated virtual environments that areenabling multidisciplinary and geographicallydistributed teams to work together more effec-tively. This includes sharing resources andknowledge and responding to grand chal-lenges, such as those consequent to the hurri-canes and floods that devastated the GulfCoast in 2005. New modes of knowledge pro-duction and networking are exemplified bye-science and cyberinfrastructure that involvecollaborative use of distributed high-speednetworks, high-performance supercomputing,and massive storage to tackle the processingand storage-intensive questions of modern sci-entific research (Buetow 2005; Hey andTrefethen 2005; Lin et al. 2005a; Mattes et al.2004). In the same way that early bioinfor-matics played a crucial role in the completionof the Human Genome Project, the nextphase, the Human Proteonome FoldingProject, is now using grid-based technologiesto elucidate the structure and function of allthe proteins encoded by the Human Genome.This grid-based approach promises to providevaluable knowledge for understanding andcuring disease. Computing that takes advan-tage of grid technology can shorten computertime by many orders of magnitude.

The European Council for NuclearResearch (CERN) is one of the world’s lead-ers in grid development. CERN (2006)defined the grid in these terms:

Whereas the Web is a service for sharing informa-tion over the Internet, the Grid is a service forsharing computer power and data storage capacityover the Internet. The Grid goes well beyond sim-ple communication between computers and aimsultimately to turn the global network of comput-ers into one vast computational resource.

One of the best developed examples ofgrid-based cyberinfrastructure is the Tele-science model utilized by the NIEHS Portal.Through collaborative arrangements agreedupon by those developing and using theresource, the Telescience model integrates dis-tributed ultra–high speed networks, super-computing capacity, and access to massivestorage and databases with applications, virtualcollaboration spaces, advanced visualization,and remote control of laboratory instrumentsand field sensors with advanced authenticationand authorization features (Lathers et al. 2006;Lin et al. 2005b). This kind of collaborativecyberinfrastructure is essential to build capac-ity for data mining and data sharing in theenvironmental health sciences.

Another significant development involves“ontologies”—a kind of knowledge map tofacilitate data organization and mining for cer-tain predefined purposes (Noy and Musen2003; Rubin et al. 2006; Stoeckert et al.2006). Ontologies or knowledge maps thatinclude formal definitions of terms and their

relationship to one another can help to distin-guish between irrelevant information andinformation of interest, while also addressingimportant issues of communication amongmembers of multidisciplinary science teams(Martone et al. 2004). The NIEHS Portalaims to facilitate, among other things, knowl-edge networking within communities by cre-ating task- and domain-specific ontologies thatcan be used for applied purposes. The devel-opment of such ontologies must be participa-tory in so far as it requires a common languagethat cuts across disciplines and professionalboundaries (Spivey 2004).

Integrative and community-linked research.An important shift in the historical relation-ship between science and society has takenplace over the last few decades, particularly asit pertains to university–community partner-ships outside of the traditional university–government–industry triad (Chopyak andLevesque 2002). This has occurred in partbecause of changes in funding policies ofgovernmental and private agencies thatdemand greater social accountability fromuniversities and researchers. This is reflected inthe community outreach and/or research trans-lation requirements of many long-standing

and important sources of funding for univer-sity investigators (e.g., National ScienceFoundation, National Institutes of Health).The relationship between universities and thecommunities they serve has also continued toevolve. For example, after Hurricane Katrina,Tulane University saw a very significant por-tion of its research resources refocused toaddress the challenges facing New Orleansand its communities. The benefits of suchcommunity-linked research collaborations canbe far reaching for both the researchers and thecommunity because they provide a fertile set-ting with real-world feedback loops to put ideasand advances in science and technology to thetest while addressing community concerns(Anderson et al. 2002; Chopyak and Levesque2002). In the environmental health sciences,community-based research has an importantrole to play and is recognized by the NIEHS inits strategic plan (Schwartz et al. 2006).

The NIEHS Portal responds to environ-mental health issues and concerns facing com-munities of the Gulf Coast region whileproviding the cyberinfrastructure necessary toadvance integrative research and facilitatecommunity-linked research. In this regard,one important measure of performance for

Pezzoli et al.

566 VOLUME 115 | NUMBER 4 | April 2007 • Environmental Health Perspectives

Figure 2. Houses destroyed in Lower Ninth Ward of New Orleans, April 2006 (photographed by K. Pezzoli).

the portal will be the extent to which it con-tributes toward helping scientists and decisionmakers address community-based environ-mental health concerns and issues. Efforts ofthe NIEHS Portal development team toaddress community needs have been docu-mented (Pezzoli et al. 2006).

Exposure biology and gene–environmentinteraction studies. Progress in the study ofexposure biology and gene–environmentinteractions promises to lead to significantadvances in disease prediction, prevention,early detection, and treatment (Acharya et al.2004; Lee and Chang 2006; Shostak 2005;Weis et al. 2005). In at least two significantways, the NIEHS Portal provides a platformfor developing grid-based technologies thatdirectly address the needs of exposure biologyand gene–environment work. First, theNIEHS Portal creates capacity to manage theunprecedented amounts of data anticipatedfrom population-level studies involvinggenetic analyses, data collection by real-timepersonal exposure devices, and environmentalmonitoring. It also enables access to and inte-gration of external, distributed data sources.Second, by integrating distributed high-per-formance computing and GIS into one virtualonline workspace, the NIEHS Portal providessimultaneous capacity to integrate spatio-temporal queries and analyses (of demo-graphic, genetic, and environmental data sets)with modeling and other applications.

Assembling and implementing the portaltechnology. The portal’s infrastructure wasassembled using components developed withinthe University of California, San DiegoTelescience project (Lin et al. 2005b). Theportal-based GIS application is built over a

collection of geo-referenced data layers usingMicrosoft-based ESRI ArcIMS (ESRI,Redlands, CA), ERMapper Image Web Server(ERMapper, San Diego, CA), and open-sourcesolutions for managing spatial databases andlarge data grid storage space as the core soft-ware (Figure 3). In addition, street data areserved via a dedicated web mapping service,whereas navigation support is provided viaInxight’s Vizserver (Inxight Software, Inc,Sunnyvale, CA). The user–interface compo-nents, including an online map viewer and theStarTree hyperbolic tree visualization applica-tion (Inxight Sotware, Inc.), are integrated in aGridSphere-based portal (gridsphere.org),which also supports a variety of browse, query,and collaboration tools (Inxight Software,Inc.). A detailed discussion of the systemarchitecture, specifications, and function ispresented by Zaslavsky et al. (2006).

Within the main portal environment, aGIS application manages the large collectionof geo-referenced data. Standard componentsof the NIEHS Portal also enable local deci-sion makers and communities to contributedata, participate in discussions, and maintainresearch and collaboration workspaces. Theonline GIS, a core component of the NIEHSPortal, assembles spatially registered data forTexas, Louisiana, and Mississippi, with higherresolution layers available for areas affected bythe two most severe 2005 hurricanes, Katrinaand Rita (Figure 4). The information isextracted from publicly available sources andorganized in the following data categories:• Post-hurricane project-related data• Potential contaminant sources: U.S. EPA

TRI facilities (hazardous air pollutants,sources of carcinogens, metals, persistent

bioaccumulative toxic substances, and otherorganic compounds); multiple types ofindustrial and agricultural facilities, oil andgas facilities (including gas stations, oil andgas wells and platforms, refineries, storagefacilities, lube plants, and pipelines) andwastewater treatment plants, among others

• Electric facilities and drinking water intakes• Hurricane damage: Katrina and Rita damage

for all available dates (including impassablebridges and roads, and different degrees ofdamage), extent of oil spills and water conta-mination, extent and duration of flooding,debris, and estimated replacement costs

• Demographic data: racial composition andincome stratification by census tracts andblock groups

• Base map layers: states, counties, and cities;urban areas, streets and landmarks, hydro-logic layers, and elevation

• Imagery layers: 1-m aerial photography forNew Orleans taken in September 2005,15-m Terracolor imagery, 1-m Terraserveraerial photographs.

Each of the vector layers—that is, data inthe form of points, lines, and polygons—canbe downloaded from the portal along withmetadata (i.e., details about the data itself, suchas who collected it, over what time period, etc.).Although the NIEHS Portal has the potentialto bring forward crucial time-sensitive infor-mation that might not otherwise be readilyavailable (e.g., types and quantities of haz-ardous materials generated or stored at anindustrial facility), decision makers will stillhave to rely on trained and experienced profes-sionals to analyze and interpret those data andmake intelligent and informed managementdecisions. There are dangers associated withthe misuse of data—a reality facing manyinformation systems. The NIEHS Portal takesthis fact into account by including securitymeasures and authentication protocols. Otherfeatures of the portal, including clearly docu-mented metadata, are designed to help reduceand manage the risk of data misuse.

Users can navigate spatial layers andexplore layer metadata using StarTree hyper-bolic tree software. In addition to metadatamanagement, the current version of theGridSphere-based online mapping environ-ment provides a range of map navigation,address geocoding, and query and spatial selec-tion tools, including Web Portal, which pro-vides a means for local decision makers andcommunities to contribute data, participate indiscussions, and maintain research and collab-oration workspaces. As an example of online-mapping functionality available through thepublic section of the portal, Figure 5 shows ascreen capture of a workspace in which a userselected TRI facilities within 1 mile of drink-ing water intakes. Password-protected sectionsof the portal may provide access to additional

NIEHS Portal

Environmental Health Perspectives • VOLUME 115 | NUMBER 4 | April 2007 567

Public site Sedimentsproject

HEALProject Others...

Telescience portal services (utilizing GridSphere framework): user management,application presentation: query, map generator, forums, etc.

Telescience ATOMIC services: authentication and authorization, resourcemanagement (data computation), workflows, etc.

ArcIMSservices

Image webserver WMS services

Spatial datamanagement

Distributedcomputing

management

Distributeddata

management

Figure 3. NIEHS Portal system architecture. Abbreviations: ATOMIC, Applications to MiddlewareInteraction Components; WMS, web mapping service.

resources and custom functions. These secureareas of the portal help users collaborate inspecific projects.

The NIEHS Portal is a work in progress;it aims to engage users and communities thatneed to share important information pertinentto an environmental disaster. Customizedresearch environments are identified anddeveloped according to two models: a) pro-ject-based (the portal team is enlisted byprospective end users to provide information

technology support for a particular project);and b) field of research [through conversationswith multiple teams of investigators, the portalteam provides data, knowledge, and tools for aselect field of research (e.g., basic and appliedstudies concerning contaminated sediments)].

In the first model, the portal team isexpected to a) create a separate space on theportal for the selected project; b) sculpt theparticular project’s user interface to includeonly the portal’s most useful existing data

layers (i.e., avoid clutter by eliminating layersdeemed nonessential); c) do some original datapreparation/packaging to meet the needs of theparticular project; and d) provide tools forsecure data archiving, sharing, integration,visualization, and analysis. In this model, theportal team becomes (among other things) aservice provider helping users meet particularneeds (e.g., the need to select a study site).

As an example of the first model of portalapplications, the portal team has developedresources to support and is communicatingwith the Head-Off Environmental Asthma inLouisiana (HEAL) research project. Thefloods caused by Hurricanes Katrina and Ritaleft many homes in the Gulf Region inun-dated, leading to prolific mold growth[according to the Centers for Disease Controland Prevention (CDC), almost 46% of homesinspected had mold growth]. The degradationof indoor/outdoor air quality and disruption ofthe health care delivery system had significanteffects on children living with asthma, which isalready the most common chronic diseaseamong children. The goal of the HEAL sectionof the portal is to support an epidemiologicstudy and the examination of genetic and envi-ronmental factors related to asthma, focusingon the implementation of a case managementprogram for children with asthma in theimpacted region. GIS support for the HEALinitiative can help investigators and otherstakeholders gauge levels of environmental riskassociated with schools that re-opened after thehurricane. GIS helps in this type of field evalu-ation, given the need to integrate multiplesources of information including data aboutthe populations and households served bythose schools, flooding depth and duration,potential environmental exposures, proximityto industrial facilities, proximity to roadways,and sociodemographic data (Figure 6).

Figure 6 illustrates a printable report for agiven point location that specifies the locationof each school vis-à-vis objects in several otherselected layers (e.g., potential contaminantsources, debris sites, duration of flooding).Available metadata at the “feature level”include many more parameters beyond thoseshown in Figure 6. It is an extensive collectionthat also includes facility identification infor-mation that is available in the original data-base. Metadata at the “layer level” contains alink to the original data source. In the exampleshown in Figure 6, there are nine feature-levelmetadata characteristics for debris sites arounda particular school. The debris site shapefile,downloadable from the portal, contains35 fields, including the purpose and opera-tions at the debris site, its size, land use andownership, and detailed information on whatspecific activities or items are permitted on thesite. The generation of metadata reports of thissort is not limited to schools. Metadata reports

Pezzoli et al.

568 VOLUME 115 | NUMBER 4 | April 2007 • Environmental Health Perspectives

Figure 4. NIEHS Portal GIS.

Figure 5. NIEHS Portal GIS showing workspace in which a user selected TRI facilities within 1 mile ofdrinking water surface intakes.

can be created on an as-needed basis by theuser for a wide range of particular locations.

In the second model, the portal team isexpected to: a) create a separate space on theportal for a strategically selected field ordomain of research—with a focus, initially, onone geographic area/hotspot of concern;b) make the data accessible (while respectingconfidentiality, security, intellectual propertyrights, etc.) to those who want to use it byputting it on the portal’s own servers and/or byenabling access to it via a participant’s externalserver through federated (collaborative grid-enabled) arrangements; and c) provide tools forsecure data archiving, sharing, integration,visualization, and analysis. The emphasis inthis second model is to enable a network of sci-entists to collectively develop a data grid in amutually defined area of concentration (i.e.,field or domain, as opposed to particular pro-ject) that would be beyond the capacity of anyone group to build and maintain on their own.This data grid would then become a collectiveresource for multiple and diverse projects.

As an example of the second model of por-tal applications, the portal team is developing acustomized environment to support basic andapplied studies of contaminated sediments.Since early September 2005, the U.S. EPA hascollected approximately 1,800 sediment andsoil samples from over 430 sites in Jefferson,Orleans, Plaquemines, and St. Bernard Parishes(parishes that were flooded with > 3 m of waterfrom Lake Pontchartrain and the MississippiRiver/Gulf of Mexico outlet). Most of the sam-ples, as noted in the U.S. EPA’s most recentfindings on this matter (U.S. EPA 2006), wereanalyzed for > 200 metals and organic chemi-cals. The U.S. EPA collaborated with manyagencies in this process, including the LouisianaDepartment of Environmental Quality, theCDC, the Agency for Toxic Substances andDisease Registry, the Louisiana Department ofHealth and Hospitals, and FEMA. Accordingto the U.S. EPA’s final report (U.S. EPA 2006),

The sample results indicate that, in general, thesediments left behind by the flooding from thehurricanes are not expected to cause adverse healthimpacts to individuals returning to New Orleans.

However, the same report indicates that hotspots of concern do exist. Some of the sitessampled showed elevated levels of arsenic, lead,benzo[a]pyrene, and diesel and oil rangeorganic petroleum chemicals. Besides the U.S.EPA, other groups including environmentalnon-governmental organizations, have also col-lected contamination data and have expressedstrong environmental health concerns. To sup-port integration of sediment sampling datafrom different sources with hurricane damageand potential exposure routes, the portal hoststhe “sediments” research space and an onlineGIS. The system is being designed to enable

analysis of spatial patterns in contaminatedsediments and help prioritize selection of addi-tional sampling sites, as well as to provide rele-vant information for risk assessment, riskcommunication, and remediation. Both of theexamples cited above exemplify the flexibilityof the NIEHS Portal.

ResultsWhether considering the public portion ofthe portal or applications developed undereither of the two models, the NIEHS Portalprovides the ability to• Create collaborative workspaces quickly and

easily (e.g., document and data set exchange,discussion groups, forums) and publish themonline for use by internal and external usercommunities

• Facilitate structured and secure informationexchange and knowledge sharing betweencollaborators, regardless of their physicallocation (users will be able to place theirdata sets in a secure managed environment,where, for example, these data sets can beshared with collaborators or discovered onsearch requests)

• Use advanced information management,analysis, and visualization applications, andconnect different applications in personalresearch workflows

• Enable search and query to multiple distrib-uted and separately maintained databases (toalleviate duplication and remove data main-tenance hassles, the different databases canbe accessed via a single log-on mechanism,thereby removing the hassle of logging into

each database separately and thus providingan easier platform for integrating disparateinformation)

• Synchronize user information from existingdirectories [i.e., MyProxy, LDAP (Light-weight Directory Access Protocol)]

• Within the same system, receive accessauthorization to computing resources thatare normally not available to regular webusers, such as supercomputers, clusters, orlarge data archives.

The NIEHS Portal is evolving as a systemthat enables multidisciplinary collaboration,including public–private sector partnerships,through two forms of integration: first, inte-gration of science to science through the join-ing of scientific perspectives, data, knowledge,and approaches across disciplines; and second,integration of science to society by creatingpathways to move research and technologyinto contexts where they are most needed fordecision making, problem solving, and inno-vation. Several goals spelled out in the2006–2011 NIEHS Strategic Plan (NIEHS2006c) emphasize the importance of thesetwo forms of integration: Goal III calls forenabling integrative and cross-cutting researchapproaches; Goal IV calls for improving andexpanding capacity for community-linkedresearch; and Goal VII underscores theimportance of supporting development ofpartnerships. Other types of integration arealso important. The portal can be used tointegrate data and information to allow viewsacross temporal and spatial scales. It also haspotential to integrate models used in societal

NIEHS Portal

Environmental Health Perspectives • VOLUME 115 | NUMBER 4 | April 2007 569

Figure 6. NIEHS Portal HEAL application (NIEHS 2006b).

Pezzoli et al.

570 VOLUME 115 | NUMBER 4 | April 2007 • Environmental Health Perspectives

and scientific communications—therebyenabling interaction across model-derived(rather than observational) data.

User-needs aassessment, system evaluation,and feedback. To identify and assess the needsof the NIEHS Portal user community, as wellas to test system performance with a diverserange of users, a series of meetings and publicworkshops were conducted in the Gulf CoastRegion in April and May of 2006. The find-ings of the NIEHS Portal User CommunityDevelopment effort were reported previously(Pezzoli et al. 2006). The NIEHS Portal usercommunity stakeholders identified as part ofthis process included local, regional, andfederal government agencies (LouisianaDepartment of Environmental Quality,planning agencies, U.S. EPA, FEMA, etc);researchers; non-governmental organizations,and the general public. The user communityexpressed an interest in seeing the portalpopulated with pre- and post-Katrina datasets on air quality, soil and sediment quality,surface and groundwater quality, health out-comes, demographics, National PollutantDischarge Elimination System permits, flooddata, rebuilding plans, building permits, andoverall data sets having greater spatial andtemporal representation. In addition, to theimmediate and long-term environmentalhealth research applications of the portal, theuser community expressed foreseeable appli-cations in evacuation studies, damage-assess-ment studies, landscape and wildlife studies,and inundation vulnerability and risk studies,among others. Comments from the user com-munity regarding the portal centered aroundmaking the system more “user friendly”,including developing front-end interfacesbased on applications such as Google Map(http://map.google.com/) and/or GoogleEarth (http://earth.google.com/). Also, othersuggestions included adding inventories ofexisting research projects in the region, link-ing technical reports to map locations, pro-viding interactive map generation, providingdynamic data entry tools, and providing clearand concise environmental health informa-tion designed for broad audiences in differentlanguages as well as a mechanism for userfeedback. Much of the feedback was used togenerate the first version of the portal, withother functionalities and features to be addedin the future. The present system is intendedto serve the Gulf Coast Region, but it is antic-ipated that similar systems could be built fordiverse applications involving larger or smallerspatial scales.

Conclusions

The NIEHS Portal combines advances inGIS, data integration, and visualization tech-nologies through new forms of grid-basedcyberinfrastructure. A core component of the

portal is an online interactive GIS that canseamlessly integrate disparate sources of com-parable environmental health data from dis-tributed databases, enabling visualization andanalyses of geographic relationships, as well asmodeling of contaminant transport, to helpidentify environmental hazards to humanhealth.

A key advantage of the NIEHS Portal isthat it enables the collaborative and multidis-ciplinary research needed to address complexenvironmental health issues by providing acommon data exchange platform and work-bench. The NIEHS Portal environment pro-vides members of the research communitywith a flexible framework in which to collabo-rate with research partners and conduct bothanalyses and visualization using a variety ofdata sets and web-based tools. Parts of theportal are open to the public, while others arereserved for individual and project researchspaces. The secure area of the portal can beused by researchers to upload and integratetheir own data into the existing data setsavailable within the portal, which can then beviewed or manipulated by other members oftheir defined research group. Customizedtools for a particular project can also be builtinto the portal, allowing researchers to utilizecommon applications among collaboratorswhile preserving security throughout. Thiscollaborative workspace for researchers isintended to be a tool that can help linkenvironmental health projects without eachresearch group having to set up and maintaina variety of data sets and information technol-ogy infrastructure.

The NIEHS Portal facilitates informationand knowledge sharing because it eliminatesthe need for each research group to design,build, and maintain its own collaborative sys-tem. For the NIEHS, this effort is in line withone of its priority funding and emphasis areas(i.e., infrastructure) and also supports thecritical challenge of enabling integrativeresearch. In addition to serving a critical needin disaster response and recovery, this effortserves as a prototype for environmental healthscience research in the modern scientific andcomputing era. At a global scale, many of theworld’s technology leaders (e.g., Europe, Asia)have begun to invest in cyberinfrastructuretechnology as part of a strategy to providetheir scientists with the tools to deal withincreasing amounts of data and informationbeing generated by large-scale projects whileaccelerating the production and sharing ofknowledge and its benefits to society.However, advances in technology and theavailability of powerful tools alone will notsuffice if both a culture and incentives formultidisciplinary collaboration are not inplace and systems are not built with userneeds in mind (Buetow 2005).

REFERENCES

Acharya T, Daar AS, Dowdeswell E, Singer PA, Thorsteinsdóttir H.2004. Genomics and Global Health: A Report of the GenomicsWorking Group of the Science and Technology Task Force ofthe United Nations Millennium Project. Toronto:CanadianProgram on Genomics and Global Health, University ofToronto Joint Centre for Bioethics.

Anderson B, Thompson C, Suk WA. 2002. The Superfund BasicResearch Program—making a difference: past, present,and future. Int J Hyg Environ Health 205:137–141.

Barnett T, Pierce D, Schnur R. 2001. Detection of anthro-pogenic climate change in the world’s oceans. Science292:270–273.

Berube A, Katz B. 2005. Katrina’s Window: ConfrontingConcentrated Poverty across America. Washington, DC:TheBrookings Institution, Metropolitan Policy Program.

Buetow KH. 2005. Cyberinfrastructure: empowering a “thirdway” in biomedical research. Science 308:821–824.

CERN (European Council for Nuclear Research). 2006. What Isthe Grid? Available: http://gridcafe.web.cern.ch/gridcafe/whatisgrid/whatis.html [accessed 1 August 2006].

Chopyak, J, Levesque PN. 2002. Community-based researchand changes in the research landscape. Bull Sci Tech Soc22:203–209.

Daniels RJ, Kettl DF, Kunreuther H. 2006. On Risk and Disaster:Lessons from Hurricane Katrina. Philadelphia:University ofPennsylvania Press.

Dilley M, Chen RS, Deichmann U, Lerner-Lam AL, Arnold M,Agwe J, et al. 2005. Natural Disaster Hotspots: A Global RiskAnalysis. Synthesis Report. Available: http://sedac.ciesin.columbia.edu/hazards/hotspots/synthesisreport.pdf[accessed 1 October 2006].

Hey T, Trefethen AE. 2005. Cyberinfrastructure for e-Science.Science 308:817–821.

Lathers A, Su M, Kulungowski A, Lin AW, Mehta G, Peltier ST,et al. 2006. Enabling parallel scientific applications withworkflow tools. In: Proceedings of the 6th InternationalWorkshop on Challenges of Large Applications inDistributed Environments, 19 June 2006, Paris, France.Available: http://birkdale.ucsd.edu/docs/alathers_clade.pdf[accessed 22 February 2007].

Lee WC, Chang CH. 2006. Assessing effects of disease genesand gene-environment interactions: the case-spousedesign and the counter-factual control analyses.J Epidemiol Community Health 60:683–685.

Lin AW, Dai L, Mock J, Peltier S, Ellisman MH. 2005a. The tele-science tools: version 2.0. In: Proceedings of the 1st IEEEInternational Conference on e-Science and GridComputing, 5–8 December 2005, Melbourne, Australia.Available ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1572209 [accessed 22 February 2007].

Lin AW, Dai L, Ung K, Peltier S, Ellisman MH. 2005b. The tele-science project: applications to middleware interactioncomponents. In: Proceedings of the 18th IEEE InternationalSymposium on Computer-Based Medical Systems, TrinityCollege Dublin, Ireland. Available ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1467748 [accessed 22 February 2007].

Martone ME, Gupta A, Ellisman MH. 2004. E-Neuroscience:challenges and triumphs in integrating distributed datafrom molecules to brains. Nat Neurosci 7:467–472.

Mattes WB, Pettit SD, Sansone SA, Bushel PR, Water MD.2004. Database development in toxicogenomics: issuesand efforts. Environ Health Perspect 112:495–505.

National Institutes of Health. 2006. NIEHS Director CelebratesOne Year Anniversary: Unveils New Strategic Plan forEnvironmental Health Sciences. Available: http://www.nih.gov/news/pr/may2006/niehs-01.htm [accessed 10 July2006].

NIEHS (National Institute of Environmental Health Sciences).2006a. NIEHS Hurricane Response. Available: http://www-apps.niehs.nih.gov/Katrina/ [accessed 1 August 2006].

NIEHS (National Institute of Environmental Health Sciences).2006b. NIEHS Portal HEAL Application. Available:http://balata.ucsd.edu:8080/gridsphere/gridsphere?cid-Map [accessed 22 February 2007].

NIEHS. 2006c. 2006–2011 Strategic Plan: New Frontiers inEnvironmental Sciences and Human Health. ResearchTriangle Park, NC:National Institute of EnvironmentalHealth Sciences. Available: http://www.niehs.nih.gov/external/plan2006/ [accessed 10 July 2006].

Nates JL, Moyer VA. 2005. Lessons from Hurricane Katrina,tsunamis, and other disasters. Lancet 366:1144–1146.

NIEHS Portal

Environmental Health Perspectives • VOLUME 115 | NUMBER 4 | April 2007 571

Noy NF, Musen MA. 2003. The PROMPT suite: interactive tools forontology merging and mapping. Int J Hum Comput Stud59:983–1024.

Pastor M, Bullard RD, Boyce JK, Fothergill A, Morello-Frosch R,Wright B. 2006. In the Wake of the Storm: Environment,Disaster and Race after Katrina. New York:Russell SageFoundation. Available: http://www.russellsage.org/news/060515.528528 [accessed 16 February 2007].

Pellow DN, Brulle RJ, eds. 2005. Power, Justice, and theEnvironment: A Critical Appraisal of the EnvironmentalJustice Movement. Cambridge, MA:MIT Press.

Pezzoli K, Sarabia H, Avakian M. 2006. Gulf Coast Trip Report.San Diego, CA:University of California, San Diego,Superfund Basic Research Program. Available: http://superfund.sdsc.edu/documents/2006_gulf_coast_trip_report.pdf [accessed 20 February 2007].

Rubin DL, Lewis SE, Mungall CJ, Misra S, Westerfield M,Ashburner M, et al. 2006. National Center for BiomedicalOntology: advancing biomedicine through structured orga-nization of scientific knowledge. OMICS 10:185–198.

Running SW. 2006. Is global warming causing more, largerwildfires? Science 313:927–928.

Schwartz DA. 2005. The NIEHS responds to Hurricane Katrina.Environ Health Perspect 113:A722.

Schwartz DA, Sassaman AP, Collman GW. 2006. Environmental

health sciences and the community. Environ HealthPerspect 114:A80.

Shostak S. 2005. The emergence of toxicogenomics: a casestudy of molecularization. Soc Stud Sci 35:367–403.

Spivey A. 2004. Systems biology: the big picture. Environ HealthPerspect 112:A938–943.

Stoeckert C, Ball C, Brazma A, Brinkman R, Causton H, Fan L,et al. 2006. Wrestling with SUMO and bio-ontologies. NatBiotechnol 24:21–22.

Thomson GE, Mitchell F, Williams M, eds. 2006. Examining theHealth Disparities Research Plan of the National Institutesof Health: Unfinished Business. Washington, DC:NationalAcademies Press.

U.S. Department of Commerce. 2006. Gulf Coast Recovery:7 Months After the Hurricanes. Available: http://www.economics.noaa.gov/library/documents/economics_for_coastal_marine_management/gulf-coast-recovery-esa.doc [accessed 21 August 2006].

U.S. EPA (Environmental Protection Agency). 2006. SummaryResults of Sediment Sampling Conducted by the EPA inResponse to Hurricanes Katrina and Rita. Available:http://www.epa.gov/katrina/testresults/sediments/summary.html [accessed 21 August 2006].

U.K. DFID. 2006. Reducing the Risk of Disasters—Helping toAchieve Sustainable Poverty Reduction in a Vulnerable

World: A DFID Policy Paper. London:Department forInternational Development. Available: http://www.unisdr.org/news/DFID-reducing-risk-of-disasters.pdf [accessed16 February 2007].

Weis BK, Balshaw D, Barr JR, Brown D, Ellisman M, Lioy P,et al. 2005. Personalized exposure assessment: promisingapproaches for human environmental health research.Environ Health Perspect 113:840–848.

Westerling AL, Hidalgo HG, Cayan DR, Swetman TW. 2006.Warming and earlier spring increase western U.S. forestwildfire activity. Science 313:940–943.

Young OR, Berkhout F, Gallopin G, Janssen M, Ostrom E,van der Leeuw S. 2006. The globalization of socio-ecologi-cal systems: an agenda for scientific research. GlobEnviron Change 16:304–316.

Zaslavsky I, Pezzoli K, Valentine D, Lin A, Sarabia H,Ellisman MH, et al. 2006. Integrating GIS and portal tech-nologies for assessing environmental health impacts ofHurricane Katrina. In: Proceedings from the SecondInternational Conference on Environmental Science andTechnology, 19–22 August 2006, Houston, TX, Vol 2 (StarrettSK, Hong J, Lyon WG, eds). Houston, TX:American SciencePress, 385–390.