Technology Foundation for a Real-Time Distributed Problem-Solving Environment Applied to Crisis and Disaster Scenarios

Technology foundation for a real-time distributed problem-solving
environment applied to crisis and disaster scenarios
Kas Osterbuhr
University of Denver University College
Capstone Project
for
Master of Applied Science ICT/GIS
August 17, 2012
_________________________
Holger Weinhardt, M.S.
Capstone Advisor
_________________________
Thomas Tierney, Ph.D.
Academic Director
Upon the Recommendation of the Department:
_________________________
Michael McGuire, Ph.D.
Interim Dean

Abstract
Disaster and crisis management is a global problem. Scenarios range from
short-term localized events to those with widespread impact persisting for
years or decades. From personal experience and research in the topic area,
there is clearly a need for a technology “platform” that can integrate
cross-disciplinary agencies, civilians, contractors, and any other
conceivable stakeholder. These stakeholders (including the environment
and the public) will benefit immensely from integration and
standardization in a problem-solving environment, especially in light of the
value of human life. This approach should lead to enhanced preservation of
life and safety, reduced environmental impact, and overall improvement in
disaster response and mitigation – irrespective of the disaster type or
scale.

These images of the Deepwater Horizon oil spill disaster serve as a
reminder of the impact an event of this scale can have upon our
lives. The DH oil spill is one of many disasters that have recently
affected the way we handle response and mitigation.

Contents
1.0 Understanding Disasters
1.1 Measurable Impact of a “Disaster”
1.2 Complexity and Scope
2.0 Gap Analysis: Status Quo and Observations
2.1 Commercial Off-the-shelf Solutions
2.2 Personal Experience with the 2012 Colorado Wildfires
3.0 Information Communication: Ethics, Issues, Solutions
3.1 Withholding Data Could be Dangerous
3.2 Innovative Solutions in Unlikely Places
4.0 Review of Literature and Survey of Technologies
4.1 Disaster Mitigation Depends on Information Technology
4.2 IT Architectures: Responding to Demand on a Large Scale
4.3 Communication Technology: Networks and Protocols
4.4 Data Technology: Sensing and Collecting Data
4.5 Unconventional Data and Methods: Crowdsourcing
4.6 Wisdom from Knowledge: Heuristics and Machine Learning
4.7 Human Gaming for Analysis and Simulation
4.8 Integration: Building Synthetic Environments
5.0 Characteristics of an Ideal Architecture Platform
5.1 General Characteristics Listed
5.2 Service Oriented Architecture in Detail
5.3 Stakeholder Unification and Open Sourcing

6.0 Pulling Together a Technology Foundation
6.1 Operating Model
6.2 Enterprise Architecture
6.3 IT Engagement Model
7.0 Funding Model
8.0 Further Research
9.0 Conclusion
10.0 References

Section 1
Understanding Disasters

1.0 Understanding Disasters
This research will primarily describe events of interest as disasters or
crises, but depending on the reader’s frame of reference, many other words
are suitable, e.g., emergency, situation, scenario, disturbance, extreme
event, catastrophe, or accident. Whatever the terminology, dealing with
disasters is difficult. My research will show that management of these
crises can benefit from a distributed tactic that leverages modern
technology paradigms such as social media, crowdsourcing, open source
development , 3D game engines, and cloud computing. More importantly, 1
the aggregation of existing platforms and technology building blocks
depends on methodologies designed for large-scale projects. No such
solutions currently exist, despite the presence of many viable building
blocks. It seems the primary obstacle is the lack of vision and execution in
pulling together necessary components into a unified platform. Therefore, I
will develop an extensible technology architecture plan capable of
improving distributed disaster response in a multidimensional synthetic
environment.
1.1 Measurable Impact of a “Disaster”
A disaster could be described in many ways, depending on the
viewpoint and context of the stakeholder. A citizen, for example, would
1 The term “open source” has been in widespread use for quite some time and describes projects that are
generally open to use free of commercial encumbrances and limitations. For more, see the Open Source
Initiative’s definition at: http://opensource.org/osd.html

have a different view of a hurricane (focused on the loss of a home or
property) than would the Government (concerned with widespread
economic loss and societal unrest). Developing a concrete working
definition of “disaster” is not necessarily required for this research, but one
description in particular does an excellent job of setting the context of this
term for the remainder of our discussion:
“A disaster is a situation which overwhelms local capacity,
necessitating a request to the national and international level for
external assistance, or is recognized by a multilateral agency or by
at least two sources, such as national, regional or international
assistance groups and the media.” (Senevirante, Baldry, and
Pathirage 2010, 378).
Furthermore, Dobel (2010) notes that “disasters unfold with little respect
for human boundaries [and] agreements across stakeholders will become
the cornerstone of modern approaches.” We may often recall familiar
disasters such as Hurricane Katrina or the Deepwater Horizon oil spill, but
it is imperative to consider that the entire planet is vulnerable to a diverse
spectrum of troubles – irrespective of race, religion, technological prowess,
wealth, or social status.
The United Nations Office for Disaster Risk Reduction (UNISDR)
issued a press release on June 13, 2012 explaining the massive cumulative
effects of natural disasters since 1992 (UNISDR excludes disasters that are

strictly technical such as a nuclear meltdown caused by an engineering
failure). Figure 1 documents the findings, showing 4.4 billion people
affected (1.3 million killed) and two trillion dollars in damage since 1992.
Figure 1 – Cumulative Impact of disasters since 1992. Adapted from the
UNISDR June 2012 press release.
1.2 Complexity and Scope
Disasters, crises, events, situations – whatever terminology is
preferred, when something goes wrong, the combination of variables is
almost limitless. In a virus outbreak, for example, the social behavior of
the population will be a critical factor. Do we need to keep the population
indoors and isolated, or is the real problem with our drinking water? In
Figure 2, the diversity of scope and complexity in natural disasters is
easily imagined in the context of the complex disaster landscape
(Guha-Sapir, Vos, Below, and Ponserre 2010). Technological disasters

comprise a category all their own and are not depicted in the table.
Examples of technological disasters are Chernobyl’s meltdown, Three Mile
Island, and the Deepwater Horizon oil spill. The recent tsunami in Japan is
perhaps a hybrid disaster – one component being the massive destruction
of the earthquake and tsunami and a secondary but equally important
aspect is the technological complexity of the Fukushima Daiichi reactor
meltdown. The lesson learned is that nothing happens as expected. Nearly
anything can go wrong and, worse yet, will likely happen at the most
inopportune time.
Figure 2 – Taken from Guha-Sapir (2010), the diversity of natural disaster
scenarios is clear when considering the numerous classification types.
Further complicating the notion of disaster management are “technological
disasters,” or those caused by humans (not depicted in table).

The financial impact of disasters is understandable, as is the loss of
property and, to some extent, the loss of human life. However, the
insidious aspect of disasters is their long-term impact in ways that most
people are unaware. For example, the work of Julca (2012, 507-508)
exposes often overlooked and unquantifiable effects such as the disruption
of public services, the adverse effects on human wellbeing, disease, or
disruptions to family life. Julca also notes many “cross-border” effects of
disasters and gives the example of water rights during severe drought,
stating that countries may be at odds with respect to usage and
distribution (2012, 509). Zahran and colleagues (2011) account for “poor
mental health” effects as the result of disasters, and are even able to
ascertain a mathematical relationship between total financial damage and
the number of poor mental health days following major disasters.

Section 2
Gap Analysis: Status Quo and Observations

2.0 Gap Analysis: Status Quo and Observations
Curiously, an article published in 1997 by Rhyne predicted that by
2002 (five years elapsed), “geographic information systems, digital
cartography, scientific visualization, computational modeling, and
videoconferencing will converge to support local and community decision
making.” Unfortunately, despite the passage of over 15 years since
publication, no such convergence is yet obvious. The most advanced
software programs in each of the disciplines mentioned by Rhyne have
certainly adopted elements of one another, but no fully combined
environment exists for use in disaster management. Even so, Rhyne’s
prediction excludes many components that a comprehensive
problem-solving environment would require such as scalable IT
infrastructure, data protocols, social network tools, or data mining
software. Many attempts at combining two or three of the components do
exist today and are in frequent use for disasters (e.g., Google Crisis
Response, InciWeb, or Ushahidi), but these platforms all have significant
disadvantages or shortcomings.
2.1 Commercial Off-the-shelf Solutions
A number of partial solutions exist for disaster management, but
they have lost funding or ceased progress, or do not address the full
spectrum of the problem area. For example, a project called

“Responsphere” purports to be an “IT infrastructure test-bed that 2
incorporates a multidisciplinary approach to emergency response.” Made
possible by funding from the National Science Foundation, Responsphere
seems to be a robust approach to information dissemination, but funding
stopped in 2008 and the website was last updated in 2009 (NSF 2004).
The website and research therein seem an exciting start but there is no
indication that developed concepts and technologies transferred to any
real-world solutions. Regardless, if the Responsphere project came to
fruition, it would still solve but one aspect of the problem space:
information flow and networking.
Google crisis response. Google constructed a crisis response
solution it shares with government and other authoritative publishers.
According to its website, Google created event-specific response interfaces
as early as 2005 (specifically for Hurricane Katrina). Google claims that to
initiate a new event-specific interface, it first “assesses the severity and
scope of the disaster” and subsequently choose what tools will be used for
the situation (Google 2012). Unfortunately, because Google chooses when 3
to launch the platform and what modules to include, the platform goes
unused by many who would otherwise be able to leverage its utility.
2 Responsphere is an effort by the RESCUE Project, the Center for Emergency Response Technologies and
California Institute for Telecommunications and Information Technology.
3 For quick reference, access the following links for the latest information: http://www.google.org/crisisresponse/
and http://www.google.org/crisisresponse/response.html

During the June 2012 Colorado wildfires, one of Google’s disaster
response engineers introduced the official Wildfire crisis portal with a blog
posting and paradoxically said “you can also see more authoritative,
local-level data from sources such as…” [emphasis added] (Geincke 2012).
Sadly, this public “disclaimer” suggests that stakeholders in the disaster
would be unable to rely on the Google response portal exclusively, and
they may in fact consider that the data perhaps lacks any authority.
Interestingly, one fire expert laments the lack of depth in the Google
interface and states that it does little to educate the public about why the
fires are occurring and what can be done to prevent them (Hopkins 2012).
The Google interface depicted in Figure 3 includes data layers highlighting
satellite imagery, burn perimeters, street maps, and a variety of other
specific resources such as webcams, shelters, and evacuation notices.

Figure 3 – Google’s crisis response interface created specifically for
disseminating information about wildfires burning across the country in
June 2012. In this screen capture, the map shows information specific to
the Waldo Canyon fire located near Colorado Springs.
Ushahidi. Another excellent tool is Ushahidi, an open source suite of
tools focused on gathering information from data feeds such as Twitter , 4
email, the web, and SMS (Figure 4 shows screen capture of the website). 5
Created in 2007 by David Kobia (then 29 years old), the tool was created
4 The reader is likely familiar with Twitter, the services is a public-facing Internet website that allows anyone to
post 140 character bits of information (including links to other websites).
5 Short Message Service is known to most people simply as “text messaging” and is the protocol of
communication used on cellular telephone devices for exchanging typed information.

as a means of soliciting eyewitness reports of violence and unrest during
the Kenyan presidential election (Greenwald 2010). Ushahidi includes a
suite of advanced information filtering tools it brands as SwiftRiver. With
SwiftRiver, it is possible to collect seemingly meaningless data from
various sources and subsequently add semantic meaning and context
(Ushahidi 2012). On the surface, one might suggest Ushahidi as a working
solution to the overall disaster problem, as it does an excellent job at
collecting and distilling information. Disaster problem solving, however, is
a wide-open arena full of challenges and Ushahidi is but one of many
information streams in the very large sea of information that must flow
through a disaster management portal. A properly designed technology
architecture will easily accommodate plug-in problem solving tools like
Ushahidi.

Figure 4 – A screen capture of Ushahidi’s homepage highlighting its central
theme of “information flow.” The Ushahidi platform is open source,
real-time collaborative mapping.
InciWeb. A highly active information portal during the Colorado
wildfires was InciWeb, an “interagency all-risk incident information
management system.” Typically deployed only for fire incidents, InciWeb
(Figure 5) is a great case study in the area of interagency collaboration,
considering that nine unique government agencies are responsible for its

design and deployment. The interface serves as an official distribution 6
portal for the agencies. Citizens, the media, and other interested parties
can go to InciWeb for a trustworthy current update on the status of an
incident. During the Colorado wildfires, Inciweb proved to be a common
denominator as the authoritative source for much of the information made
public. For disasters other than fire, there should obviously be efforts by
the appropriate agencies to collect information and alerts in an aggregated
portal similar to InciWeb. On its homepage, InciWeb’s primary logo reads
“Incident Information System,” suggesting it likely is suitable for a variety
of usage cases.
6 http://inciweb.org/about/ The nine agencies are: U.S. Forest Service, National Wildlife Coordinating Group,
Bureau of Land Management, Bureau of Indian Affairs, Fish and Wildlife Service, National Park Service, Office of
Aircraft Services, National Association of State Foresters, and the U.S. Fire Administration (a division of FEMA). A
Whois record shows that the inciweb.org domain was first registered in 2004 and is held by the USDA Forest
Service, but it is unclear from any source when the program was first conceived or successfully implemented.

Figure 5 – Screen capture of the InciWeb homepage. InciWeb is used
almost exclusively by forest agencies for dissemination of data related to
wildfire incidents. Individual incidents have news, photos, maps, and
other associated data. http://www.inciweb.org
Capaware. Touted on its website as a “3D multilayer geographical
framework," the company’s showcase video ironically features a wildfire
demonstration scenario. The primary capabilities advertised are
communication, analysis, simulation, education, navigation, and
management. Capaware was “conceived as a system of 3D terrain
representation with multilayer representation of diverse types of resource

capabilities and application integration.” The first commercial deployment 7
of Capaware is Sademer3D, an emergency aid decision-making system. 8
Sademer3D states that its mission is to “show data in a usable and
intuitive way to eliminate noise and turn it into quality information.” The
tool seems to be a suitable product, especially for wildfire management,
and it would be worth studying the operation of this and other similar
software. The creators of Capaware published a paper in 2011 (Castrillon
et al.), but it was actually authored in 2009 and unfortunately it seems
work has ceased.
Changing the status quo. The quantity and type of existing products
is motivation for applying an architectural approach. The existing solution
components require aggregation into a suitable platform that is well known
to citizens and trusted by experts and emergency responders. Fortunately,
the surveyed technologies such as Ushahidi, Responsphere, InciWeb, and
others seem very well suited for aggregation into a technology
architecture. Creating the architecture so that it allows easy incorporation
of preexisting products implies resources will not be wasted solving
problems for which solutions already exist.
7 Capaware: http://www.capaware.org/index.php?option=com_content&view=article&id=55&Itemid=72
8 Sademer3D: http://www.sademer3d.com/

2.2 Personal Experience with the 2012 Colorado Wildfires
The High Park fire in Fort Collins ignited early June (2012) and
quickly grew to become the most destructive Colorado fire in history (259
homes destroyed) and second largest burned area in Colorado history
(87,284 acres), with a total cost of containment of around $40 million.
Within two weeks, the Waldo Canyon fire in Colorado Springs took over the
“most destructive” position by destroying 346 homes with approximately
20,000 acres burned. The containment cost alone was $17 Million with a
continuing daily expense of $25,000 as of July 17, 2012.
9
Becoming self-informed. Despite my experience formulating queries
with common search engines and my background in geospatial data,
finding maps and knowledge pertaining to the fires was difficult –
especially during the very early stages. Fire lines in the High Park incident
eventually moved to within three miles of my home, yet there was still no
authoritative single information portal and I found myself scavenging data
from various places to uncover facts that were undoubtedly already
examined in detail by experts. What roads are closed? What does the fire
progression look like over the past seven days? What is my risk of
evacuation?
Several data portals surfaced during the High Park fire incident,
mostly presenting basic map and evacuation data. Undoubtedly, many
9 Data from InciWeb pages: http://inciweb.org/incident/2904/ and http://inciweb.org/incident/2929/

were obfuscated by search engines and overall lack of popularity, but a
handful seemed to take hold on Twitter and on news websites. The more
prevalently shared maps during the first weeks were:
http://larimer.org/highparkfire/
http://inciweb.org/incident/2904/
http://www.cohighparkfiremap.org/
http://co.dtswildfire.com/home/Flex
http://activefiremaps.fs.fed.us/
http://www.fema.gov/news/event.fema?id=16493
http://www.geomac.gov/index.shtml
http://www.esri.com/services/disaster-response/wildlandfire/latest-n
ews-map.html
http://www.eriaconsultants.com/documents/high_park_fire.pdf
http://blackicegeospatial.com/
Ineffective solutions. In surveying these many different interfaces
and information sites, I observed several noteworthy characteristics and
specific deficiencies that a technology architecture approach would correct:
● The variety of ad hoc interfaces are confusing and not standardized,
making it difficult to interpret information.
● There is too much replication of data, and no indication of which
sources are authoritative.
● It is difficult for citizens to locate data during the event – with heavy
reliance on search engines, news broadcasts, or word of mouth.
Further, the search engines will not understand which sources are
authoritative until various sites get naturally cross-referenced.

● Lack of scalability and support infrastructure presents itself as slow
or unusable interfaces to the users.
● Interfaces were developed primarily to serve/publish basic maps with
no apparent effort made to actively collect data from users.
● The interfaces offer only basic measurement tools such as distance
and area. A more robust design would publish data with protocols
that open up the underlying datasets for analysis rather than
pre-rendered map tiles. For example, I should be able to easily mark
terrain data by slope or aspect – a trivial matter with a dedicated
GIS workstation but a potentially challenging task with a 10
distributed system that is accessed by numerous people.
● Most likely, none of the interfaces used advanced aggregation tools
to filter signals from the “noise” of input data. For example, simply
adding a “Twitter feed” really doesn’t add value, it pollutes the map
with too much information.
● New and important data are difficult for the authors of the interfaces
to discover; consequently, they are relying on tips from coworkers,
discussion boards, and social media to “discover” data.
Status quo frustration. While the efforts of government agencies,
corporations, and private citizens are noble, their variety of offerings
served to paralyze the average citizen with too many choices and lack of
10 Geographic Information Systems (GIS) is a term that broadly applies to mapping, cartography, and analysis of
spatial datasets.

hierarchy. Without a single authoritative distribution point, many users
had no means of finding answers to simple questions. A friend contacted
me in the early days of another wildfire incident in Southern Colorado. He
was frantically trying to ascertain the proximity of the fire to his family
property in the area and had asked me to locate a recent satellite image
of the burn. This was but one of many defining moments when I wondered
to myself why Colorado is spending millions of dollars each day to fight
fires but at the same time relying on good-natured private companies or
individuals to educate us about the facts of the event. The lack of a
disaster response infrastructure is not due to limitations of technology but
arises, rather, from flawed implementation philosophies.
One person can make a difference. Google gained increased
attention (popularity) when it publicized the web link to its wildfire map of
Colorado Springs. The Google crisis response team builds disaster portals
for selected crisis scenarios leveraging its technology infrastructure to
serve vast amounts of map data. For the Waldo Canyon fire in Colorado
Springs, Google added satellite images from DigitalGlobe to show post-fire
damage. I noticed a Twitter posting indicating images of the post-fire 11
damage were available with the Google interface, but wondered why the
excellent aerial photos taken days earlier had not yet been georectified 12
11 DigitalGlobe has provided overview images and a Google Earth file for download. Users can obtain the raw
image data after a qualification process at: http://www.digitalglobe.com/waldocanyonfire
12 Georectification is stretching and warping an image to fit known features (control points) on the earth’s
surface.

and released – especially considering their much higher resolution. It took
approximately four hours of effort for me to contact the photograph
copyright holder and subsequently publish a file to the Internet.
Surprisingly, even this limited investment yielded more accurate
information about the damage than previously available in this particular
data format, as depicted in Figure 6. The particular value in my work was
to put the aerial photographs into a format that anyone could use – a KMZ
file. This data file opens natively in Google Earth, where the contents will
overlay automatically on a composite map that includes street names and
other localized data. Homeowners could glance at a map, zoom into
streets they recognize (or even enter an address in a search bar), and
instantly view the status of their homes – all of which is possible on a
majority of the smartphones so prevalent today.

Technology Foundation for a Real-Time Distributed Problem-Solving Environment Applied to Crisis and Disaster Scenarios

Figure 6 – Top half shows a DigitalGlobe satellite image of a destroyed
area near Colorado Springs. The top image was the best-available
georectified data until this author rectified aerial photos previously taken
by Denver Post photographers (bottom). Note dramatic increase in
available information.

Section 3
Information Communication: Ethics, Issues, Solutions

3.0 Information Communication: Ethics, Issues, Solutions
During the Waldo Canyon fire, the Denver Post collected aerial
photographs of damaged homes in Colorado Springs. As a news agency, it
elected to release this information immediately for the benefit of the
public. Surprisingly, the city government was simultaneously withholding
data about damaged houses until there was a complete and accurate
survey of the neighborhoods (Handy 2012). Handy noted that the Colorado
Springs Mayor “disapproved” of these photos being circulated. Ironically,
this meant that homeowners came to rely on the Denver Post’s journalism
to discover the fate of their homes. Despite the city government’s
disapproval, the Denver Post reported (via Twitter) that it received emails
from 200 homeowners expressing thanks for releasing the images (only
approximately 350 homes were destroyed in total.) In this case, flow of
information was negatively affected by one stakeholder (the government),
and positively influenced by another stakeholder (a news agency). Ethical,
legal, and political threats are all important factors to concede and
mitigate when building a new disaster technology architecture.
13
3.1 Withholding Data Could be Dangerous
While my personal experience with information politics during the
High Park fire incident is edifying, a story from the Japan earthquake
disaster gives an even more emotional look into how citizens use
13 For more Twitter activity related to the damage photos: http://northlandfox.posterous.com/141984343

information. Noyes (2011) writes of his family’s predicament during the
early days of the reactor meltdown, noting in particular that he was
“forced” to trust the information distributed by Japanese officials regarding
the radiation levels. Noyes used extensive precaution with food and water
handling to avoid exposure and noted that news on the Internet was at
least several hours more current than available through traditional
channels (2011, 6). Alarmingly, he concluded that Japanese media,
motivated by financial reasons, would skew publicly released information
to favor the nuclear industry – effectively downplaying the actual levels of
radiation exposure.
While Noyes says that he knew the nuclear advocates were lying, he
“didn’t know who was right about radiation risks, and didn’t know where to
draw the line” (2011, 7). Noyes quoted one Japanese citizen as saying “we
don’t believe the government is telling the truth, but we have no choice
but to believe them.” [Emphasis added.] This is an important reminder
that the public will rely on published information to make decisions about
safety and wellbeing.
It also happens in the USA. In the United States, many would
believe that such an atrocity of misinformation (as in the Fukushima
meltdown) would simply not occur. But there were major documented
issues with the flow of information during the 2009 Deepwater Horizon oil
spill incident – issues that undoubtedly contributed to the inefficient

handling of the situation. Several stakeholders, including the White House,
British Petroleum, and government scientists, all endured criticism for
underestimating the amount of oil flowing from the breached underwater
wellhead (McNutt 2010, PEER 2011). While it is understandable for humans
(and corporations) to have a tendency toward self-preservation, we must
realize that distribution of unaltered factual data is in our collective best
interest when dealing with a disaster. One method to mitigate any
misconduct is through a disaster data portal where various stakeholders
can validate the flow and utilization of information. Any of the world’s
population can be involved in a disaster. In combination with a wide
spectrum of environmental ecosystems, this diversity of life greatly
increases the chance of unforeseen trickle-down effects when information
is mishandled.
3.2 Innovative Solutions in Unlikely Places
An extraordinary solution to the absence of radiation data was born
from the Japanese nuclear disaster. A private company called Safecast
started up by using a crowdsourced funding website called Kickstarter
(Figure 7). The founders of Safecast were “fed up with indefinite data” and
decided to take the opportunity to build a radiation sensor network that
would enable accurate radiation level sampling (Jamail 2011). The
company makes all collected data publicly available and summarizes the

results with a map interface (Figure 8). Safecast’s company director noted
during his discussion with Jamail, “Getting into this showed us there is a
lack of data everywhere.” The founders of Safecast and their unlikely
means of acquiring funding proves the potential that lays in public
resources.
Figure 7 – Safecast, founded through Kickstarter, again used the platform
this year to raise funds for building a new personal Geiger counter. Screen
capture shown is from http://www.kickstarter.com

Figure 8 – The map interface on Safecast’s website displays a “heat map”
of radiation levels near the Fukushima reactor in Japan. Citizens using
Safecast Geiger counters collected all data shown.
Social media. Another important type of “sensor” is the human.
Humans are capable of observing complex situations and processing those
observations into something we can share with others. Social media
websites such as Twitter are an excellent way to gather information over a
large network of human users. As Steiner (2012) writes, the majority of
the stakeholders (public included) in the Waldo Canyon fire near Colorado
Springs relied on Twitter for their information. I also found this to be true
in my own research during the event – I used Twitter almost exclusively to
uncover the most current news and other information by using the search

keyword: #WaldoCanyonFire. Steiner quoted the senior communication 14
specialist for the city of Colorado Springs as saying “Twitter was the route
that officials chose for immediate information, and [for] reining in any
errors flowing across social media streams.” Steiner also gives statistics
from a tracking site indicating the search term #WaldoCanyonFire had
reached 54.4 million people in 15 days with 119,000 total “tweets” (a
tweet is a single posted message).
Unfortunately, Twitter’s ease of use is not without consequences.
The information stream from 54 million people is sure to contain noise and
misinformation, but somewhere within lay the important facts that people
need for health and safety (Boulos et al. 2011, 2). Steiner (2012) talked
with an information officer from the Rocky Mountain Area Coordination
Center who indicated misinformation can take “a lot of effort” to correct.
Other social media sites such as Flickr (photo sharing) and Facebook also
played important roles in the information streams going into and out of
Colorado Springs. Such an assortment of data sources magnifies the
possibility for misinformation. Complex filtering algorithms can mitigate
this situation, but human intervention and supervision will likely remain a
necessary means of oversight.
14 The hash mark “#” is used on a Twitter posting when users want to add particular emphasis. As such, searching
for a word that is a recognized “hash tag” will yield excellent results.

Section 4
Review of Literature and Survey of Technologies

4.0 Review of Literature and Survey of Technologies
Constructing a problem-solving environment capable of improving
disaster response is predicated on a base layer of enabling infrastructure,
the primary aspects of which are processing, storage, and communication.
Built upon the infrastructure, various downstream software and sensor
technologies make it possible to send email, analyze scientific data, place
phone calls, or store information. Collectively, the entire pool of
technology is termed Information Communication Technology (ICT), often
shortened to just Information Technology, or even IT. My research goal is
to identify and characterize the nature of existing and near future IT and
to show how the various components can be woven together into an
architecture that allows for extensibility and for scalability. The
architectural approach will pull together solution “islands” into a unified
working system – as opposed to a disparate collection of competing
individual units, as seen in the status quo.
4.1 Disaster Mitigation Depends on Information Technology
A 1991 paper authored by Alexander effectively characterizes the role
of information technology in disasters and gives us insight into the “state
of the art” over 20 years ago. Surprisingly, even though written long ago,
the conceptual role of IT remains quite similar today. The author
proclaims, “The satellite and the microprocessor have revolutionized the

way in which natural disasters are monitored and managed” (1991, 238).
He goes on to remark that these technologies process a remarkable
amount of information. Anyone today would agree that both the satellites
and processors of 1991 would be woefully inferior compared to their
modern equivalents. This is a unique insight into how our view of
technology has evolved, with modern science exposing an almost limitless
hunger for more processing capability. Alexander cites work in the early
1980s, where computers aided responders during emergency situations,
thus establishing the long history and importance of information
technology in disasters.
Alexander also writes, “With respect to natural disasters, technology
designed to acquire and process large amounts of information has three
basic uses: data retrieval and display, monitoring and prediction of events,
and simulation for backcasting or forecasting” (1991, 239). Incredibly, it
seems that in the case of the Colorado wildfires of 2012 that we
completely neglected the third use of IT: simulation. At no point during
the wildfires were the vast amounts of geospatial data and numerical
modeling tools leveraged to model the fire (to my knowledge).
Surprisingly, commercially available wildfire simulation software is
available. Further, a handful of local Colorado researchers specialize in
predictive wildfire modeling.

4.2 IT Architectures: Responding to Demand on a Large Scale
Digital information such as satellite imagery, social media data,
numerical models, or any other conceivable source must be maintained on
a base storage layer that is capable of rapid response and scalability to
accommodate millions of users. In addition to storage, processing power
must also be scalable – especially when approaching computationally
intensive tasks such as machine learning or scientific modeling. Of course,
to leverage processing and storage resources requires a suitable
communication network. A modern development known as “cloud
computing” is changing the IT hardware landscape by promising nearly
unlimited scalability and ubiquitous accessibility.
The “as-a-Service” paradigm: Cloud computing. The evolution of
modern computing environments has led to a service-oriented viewpoint of
IT functionality. The first and most recognizable “as-a-Service” archetype
is Software as a Service (SaaS). Software vendors traditionally sell licenses
for installation on computers or networks local to the customer. Software
as a service (SaaS) contrasts this approach by moving software and data
to a centralized computing system often far-removed from individual users.
This centralized system can be accessible over private networks, but when
moved to public networks (the Internet), the term used is cloud computing
(Brown et al. 2011, 390). Actually, many users may be using SaaS without
recognizing it as such. For example, many people use Yahoo’s webmail,

electronic commerce, and other services as a routine part of domain
hosting (Yahoo 2011, and Author). Even Facebook, which may not seem
like “software,” has become a commonplace SaaS application (Richard
2008).
A continuation of SaaS is Infrastructure as a Service (IaaS). Amazon
Web Services (AWS) is a widely popular IaaS provider. Anyone can get an
AWS account and subsequently pay an on-demand rate for running any
number of computers using remote access to AWS facilities over the
Internet. For example, rather than purchasing a second computer for home
or small office usage (perhaps for storage or added computing power), one
could instead pay Amazon a monthly usage fee for access to nearly any
type of computer system, with virtually unlimited storage capabilities. The
final “as-a-Service” buzzword is Platform as a Service (PaaS). PaaS is a
still undefined midpoint between pure software (webmail, navigation
maps) and infrastructure (which requires in-depth administrator level
knowledge). Regardless, having access to as much (or as little) hardware
as needed is a very powerful capability.
Recalling the work of Alexander (1991) mentioned earlier in this
section, we benefit now from understanding how IT has evolved and how it
is leveraged. Amazon Web Services (AWS) is one of many cloud service
providers offering IaaS. Due to its popularity, AWS has many success
stories. A few relevant case study metrics are as follows:

● Swisstopo, the office of topography in Switzerland, has eight
terabytes of data transfer per month, 250 million preprocessed
map tiles, and delivers 1,300 tiles per second on a typical day
to its user-base
(http://aws.amazon.com/solutions/case-studies/swisstopo/).
● NASA Jet Propulsion Laboratory is able to process 200,000
Cassini space images within a few hours at a cost of only $200
– without AWS JPL would spend 15 days completing the same
task
(http://aws.amazon.com/solutions/case-studies/nasa-jpl/).
● Foursquare, a popular social networking website, processes five
million daily “check-ins” for users and leverages the elastic
nature of AWS to grow and shrink its platform to match
demand
(http://aws.amazon.com/solutions/case-studies/foursquare/).
● Yelp (A local business feedback site) processes three terabytes
of data daily and has seen over 39 million unique visitors to its
site (http://aws.amazon.com/solutions/case-studies/yelp/).
The primary advantage of AWS and similar cloud computing platforms
is that of “elasticity.” A recent article by Jackson (2012) calls attention to
the rapid growth of a social networking service called Pinterest (hosted on
AWS). Launched in 2010, Pinterest now ranks 38th in global traffic and 16th

in the USA. Jackson cites the company as “one of the fastest growing 15
sites in the history of the Web.” Pinterest leverages the geographic
dispersion of AWS infrastructure to safely backup nearly 500 terabytes of
data. The company exercises the elasticity of AWS by taking servers offline
at night and adding more servers on weekends when traffic is high. During
disaster events, network traffic growth and resource demands are keenly
suited for this dynamic resource approach.
Finally, consider the story of Animoto, which developed an
application for Facebook and experienced such growth that it expanded
from 50 to 3,500 virtual machines in just three days (Economist 2008).
This would be essentially impossible to achieve in the physical sense.
Reliance on a preconfigured physical infrastructure such as AWS is the only
means to expand so rapidly. Figure 9 depicts a typical data center modular
container. A cloud provider such as AWS would likely have tens of
thousands of these available on its network and dispersed throughout the
globe at a dozen or more locations.
15 http://www.alexa.com/siteinfo/pinterest.com

Figure 9 – A modular container filled with server computers is typical of
the data centers powering Google, Amazon, and Facebook.
Cloud computing and the “as-a-service” paradigm is an effective
model for many aspects of the disaster problem solving environment. Each
stakeholder will have unique data and unique needs. The best means for
these stakeholders to interact is to publish information as a “service,” to
which anyone can subscribe. In the business vernacular, this is called a
“service oriented architecture.” One common example of a published
service is that of Google Maps. In addition to access via its portal
maps.google.com, the company built a web mapping service (WMS) that
anyone can use to pull live map data into a custom application. Twitter

data are also available as a service – an application developer can add a
small amount of software code to integrate this type of information feed.
In the disaster context, publishing data with service layers can be highly
effective in rapidly changing scenarios because transfer of static data files
is unnecessary. Services are more agile than discrete data transfer
because any subscriber will transparently receive updates in near
real-time.
4.3 Communication Technology: Networks and Protocols
Network transport. Information flow is essential to disaster
management and requires various communication technologies to convey
different types of data (e.g., voice, digital, analog, optical). In addition to
end-point hardware (modems, network cards, routers), the heart of
communication lies in the long distance travel medium and signal
technologies utilized. Examples (many cited by Alexander, 1991) are
telephone, fax/modem, cable TV, power line, fiber optic, laser optical,
two-way radios, HAM radio, AM/FM band radio stations, cellular or satellite
mobile phone, microwave radio towers, satellite data links, and WiFi (short
range wireless).
Diversity is a key aspect of communication within disaster stricken
areas. Data and voice traffic can easily overwhelm local infrastructure,
crippling the entire network. During the Colorado Springs fire, public

officials posted messages on Twitter reminding citizens to avoid further
use of overworked cell phone networks (Figure 10). In the figure, note that
user “Bettie” has posted to Twitter using her cell phone but at the same
time remarks that she is unable to make a phone call. This is due to
dual-mode availability as well as ingenious protocol usage. Twitter is
commonly used over wired Internet connections (e.g., cable TV or
dedicated phone line in the household), but the service is also available
over cellular phone networks using an efficient SMS protocol that can even
automatically append location data (Boulos 2011, 6&17). Communication
diversity is effective during disasters at times when limited infrastructure
is operating at full capacity – especially when core infrastructure such as
power and telephone landlines are crippled or destroyed. Luckily, the
widely known “Internet” is not the only way to network a large number of
systems.

Figure 10 – Four selected “Tweets” (of thousands at this moment in time).
Note the second and fourth entries both indicate issues with cell phone
networks. User “Bettie” reports she is unable to place a call, yet is
obviously able to send messages through Twitter.
The Internet is not the only network. Most readers of this work are
likely near a large metropolitan area with massive power grids, effective
telephone and cable networks, and even pervasive WiFi Internet. In parts
of the USA and certainly many parts of the world, such infrastructure is not
ubiquitous. When operating in these areas, responders would have

difficulty with cellular reception – let alone functional Internet
communication. Responsphere refers to these networks as “ground zero
communication” and depicts what it views as a prototype configuration to
include many network types as seen in Figure 11 below. The depiction
again emphasizes the concept of diversity as it applies to communication
modes.
Figure 11 – Responsphere’s view of the ideal “ground zero” network. The
premise of this system is to deploy rapidly, operate with diversity, and
maintain flexibility in deployment. (image source: www.responsphere.org)

Creating local ad hoc WiFi radio networks. The Internet is
obviously the most popular and prevalent network, but widespread outages
sometimes mandate the use of alternative technologies. An “ad hoc”
wireless network is one such possibility, whereby hardware components
(computers, routers, etc.) coalesce into a “local Internet.” Traffic control
and monitoring is an essential component of this type of system, as user
demand could quickly overcome network capacity. Arisoylu, Mishra, Rao,
and Lenert (2005) have built an ad hoc network capable of functioning in
locations where no functional networks exist. The novelty is in its
independence; it functions as a stand-alone, peer-to-peer network.
Because the work is slightly dated (now seven years old), it demonstrates
that the underlying technology building blocks are likely pervasive in
modern countries – as opposed to solutions premised on cutting edge
protocols and hardware that have no market infiltration. A majority of the
authors’ infrastructure operates on commercially available equipment,
giving rise to a very adaptive capability in disaster scenarios (no reliance
on proprietary hardware or access to specialized tools and infrastructure).
Adding diversity and distributing traffic. Most wireless
communications use frequencies in the 100 to 5000 megahertz range,
commonly referred to as “radio” waves. At much higher frequencies, those
same radio waves become visible light. This can ease frequency congestion
and creates unique opportunities for new methods of information

exchange. Miyair and Matsuda (2011) term this technology “visible light
communication,” or VLC. They suggest VLC is a reasonable augmentation
for cell phone connectivity and is possible using special LED light fixtures 16
installed in existing streetlights. They propose that a VLC-enabled
communication link can feed enough supplemental data to cell phones to
ease congestion on cellular network infrastructure. Their network
technology is currently only capable of delivering downstream data – thus
the public could easily consume published information much like watching
a television, but they could not upload their own data using the VLC
technology. This is an area they consider ripe for further development and
they believe VLC has many possible applications including linking nearly
any physical device to a network. Cell phone network saturation is
somewhat common during disaster events. Miyair and Matsuda noted that
cell phone networks were highly strained during the March 2011
earthquake, and they recognized the advantage of other communication
modes such as Twitter, which transmits data with much smaller bandwidth
requirements.
Data standards. Standardization of format facilitates communication
of data by increasing interoperability among systems. For example, when
sending a digital image by email, the most likely file format is JPG, a
standard created in the mid 1980s by the Joint Photographer’s Expert
16 LED: light emitting diode

Group. The PNNL gap analysis report noted in particular that the most 17
challenging aspect of integrating data from various sources is the lack of
standards. The report also notes, “Developing these standards can be a
complex and time-consuming effort” (2011, 3.14). With cross-disciplinary
stakeholders and global participation, the exchange of data is a
fundamental requirement. The ability of the disaster interface to eliminate
data format complications is a strategic advantage. Publishing data
“service layers” could allow data subscribers (e.g., the public, another
agency, or a responder at the scene) to access information without
compromising the security of data. With careful design, the source data
formats can remain proprietary and the service layer of the architecture
would expose information in a consumable format to any stakeholder with
sufficient security credentials. The service layer serves as an abstraction
element by translating and allowing the upstream and downstream
equipment to communicate easily.
XML. One popular data file format is the eXtensible Markup
Language (XML). The language is an evolution of two previous standards,
the first of which written in 1971 (Goldfarb 1996). Generally, the purpose
of a markup language is to separate the information content of documents
from their format. XML files are stored in human readable text (ASCII ), 18
17 For more on the JPG image standard, refer to: http://jpeg.org/jpeg/index.html
18 American Standard Code for Information Interchange – ASCII is a ubiquitous encoding scheme used to convert
digital bits to human readable characters, first published in 1963. http://en.wikipedia.org/wiki/ASCII

with the exception of embedded binary data files used in the more
advanced XML formats. A sample XML file might appear as follows:
<disaster_name>Waldo Canyon Fire</disaster_name>
<date>June 2012</date>
<resources>50 fire engines, 300 personnel</resources>
Each data entry is “tagged” with plaintext names that can be referenced in
a tag library. In the above example, the tag “resources” refers to the
deployed response units, in this case the 50 fire engines and 300
personnel. Despite the simplicity of the example, XML is capable of storing
very complex information sets. The advantage of the XML design is the
“extensibility.” Adding tags to the vocabulary and creating behaviors in the
applications is a matter of supplementing the programming code with a
new tag library (often referred to as a “schema”). Existing data would not
need to modification, as new tags generally add supplemental functionality
while maintaining backwards functionality.
CAP. An example XML filetype used in disaster response is the
Common Alerting Protocol (CAP). According to Jones and Westfall (2010),
CAP is a “simple but general format for exchanging all-hazard emergency
alerts and public warnings over all kinds of networks.” The authors go on
to say CAP “allows a consistent warning message to be disseminated
simultaneously over many different warning systems, thus increasing
warning effectiveness while simplifying the warning task.” Though Jones
and Westfall are speaking of CAP, the reasoning extrapolates easily to any

data exchange format. Publishing data in unknown or proprietary formats
forces downstream users to reverse engineer the format or purchase
expensive software to facilitate translation.
Open GeoSMS. Open GeoSMS is a standard that represents the
piggybacking of two components, the popular SMS (Short Message Service)
which we all use when sending text messages over cell phones, and a
geographic standard developed by the Open Geospatial Consortium (OGC).
The standard is used on the Ushahidi platform as well in Shahna 19
Disaster Management software and “has proven very useful in emergency
and crisis management systems” Boulos (2011, 3). Clearly, there are many
ways to build protocols and many specific applications for each. Open
GeoSMS, for example, is designed for mobile devices and specifically
facilitates adding spatial meaning to information.
Push/pull and publish/subscribe. In addition to data formats and
networks over which data flows, much thought goes into the many
methods of how authors and users become aware of services (Eugster,
Felber, Guerraoui, and Kermarrec 2003, 114-115). Push, pull, publish, and
subscribe are four words frequently used and well suited for grasping the
fundamental concepts (117). For example, consider the difference between
subscribing to a newspaper delivered to your door (“push” technology) and
finding the news yourself through Internet search (“pull” technology).
19 To read the OpenGeo SMS standard in detail, see: http://www.opengeospatial.org/standards/opengeosms

Another common example is the emergency notification system that
interrupts our television and radio broadcasts when needed during
emergencies (or at predefined test times). This is a real-time “push” alert
that allows no option to remove it from our television (except, of course,
turning off the device). An alternative method might be a posted
evacuation message on a website or bulletin board.
In the subscribe/publish interaction pair, authors and users have a
mutual agreement of creating and consuming information. Eugster et al.
(2003, 128) examined the phenomenon in detail and concluded that it is
“well adapted to the deployment of scalable and loosely coupled systems.”
Most importantly, they have noted that the method of interaction can
potentially affect the underlying infrastructure, especially when that
infrastructure is not scalable (e.g., if communications consume more
bandwidth than another more optimum method).
4.4 Data Technology: Sensing and Collecting Data
Without knowledge about a disaster, it is impossible to react. Many
sources of information surface as events unfold; common examples are
news media coverage, satellite imagery of a devastated area, scientific
publications written specifically about the phenomenon of interest,
eyewitness testimony, and commissioned field studies. Gathering such
data is a careful science and when done improperly can undermine all

systems downstream of the data gathering activity. Presented below, a
sampling of the data types and collection methods hint at the complexity
of combinations and possibilities. A later section of this paper covers data
from social media sites in detail (known as “crowdsourcing”).
Satellite imagery. Extreme events are typically associated with a
particularly memorable image, and quite often the image is an overhead
photo taken from satellite or aircraft depicting the aftermath (Wang 2007,
1-2). Gathering imagery after a disaster is fundamental to assessing
damage and, though it might seem like just pictures, selecting a suitable
pixel resolution for such information is sometimes a tedious process.
Recording the highest resolution possible requires lengthy collection time
and yields good post-acquisition analysis results whereas a coarse
resolution may be unsuitable for interpretation (Wang 2007, 2). Designing
this thought process into a disaster management platform is crucial to
ensure data contribute to a healthy decision-making process. Stakeholders
should be able to publish a data needs document and the data gatherers
should be cognizant of post-acquisition requirements – thus the involved
parties can all strike a balance. This concept extrapolates for many types
of data collection – e.g., radiation levels recorded at certain spatial or
temporal intervals or soil samples distributed over an area of interest.
Mortality data. Zolala (2010) has researched the collection of
mortality data after earthquake incidents in Iran. The findings indicate that

the basic problems with status quo data collection methods relate mostly
to a lack of coordination among groups doing the collection. Zolala remarks
that measuring effects of natural disasters requires “ongoing and routine
data collection commencing immediately after a disaster” (2010, 542). The
author concludes that poor infrastructure (mostly in undeveloped countries)
can strongly inhibit effective data collection and that “multifaceted efforts
in the global and national level” are needed to strengthen health
information systems. Zolala’s work exposes yet another portion of the
extremely broad spectrum of issues faced when dealing with disasters
across the globe.
Collecting field data. In any disaster region, there is a need for
gathering information using humans equipped with sensors of various
types. For example, a scientist might travel to various locations and
measure air quality or observe animal species, or a citizen might discover
an important location within a debris field. Lwin and Murayama addressed
field data collection using GPS-enabled cell phones in moderate detail,
having begun by stating “accurate, timely and handy field data collection is
required for disaster management and quick response during emergencies”
(2011, 382). The authors utilized off-the-shelf software and with some
customization, successfully carried out field tests of a working system.
They found that substituting cell phones for more expensive equipment
such as portable computers was an effective approach, primarily because

cellular coverage networks have broader coverage than WiFi access
typically required for portable computer systems.
Machines that “tweet.” While many regard Twitter as a primarily
human-centric communication medium, Boulos discusses the important
potential Twitter has for connecting machines to a network (2011, 2).
Boulos believes Twitter would be an effective “publish-subscribe
infrastructure for non-human sensors,” allowing the sensors to
automatically push data such as temperature, time, location, or other
readings. The author also explains that similar machine communication
could occur over Bluetooth wireless (or similar wireless infrastructure)
using identity cards such as M2M (Machine-to-Machine) and SIM (Subscriber
Identity Module). The eventuality of our technological world is that all
objects will have embedded sensors, embedded connectivity, and likely a
communications protocol that can be activated manually or automatically
to facilitate the exchange of data among millions of devices.
4.5 Unconventional Data and Methods: Crowdsourcing
In our modern environment of nearly ubiquitous connectivity, our
society has become a collective conscience and woven network of human
sensors. The term “social media” is widely used to connote our digital
relationships with each other and our sharing of information. Twitter and
Facebook are two widely discussed social media web sites, but they are by

no means the only interfaces that use tight human integration to add
value. Social sites benefit from Metcalfe’s Law, which states that the value
of the network grows exponentially with the number of users. A new term,
“crowdsourcing,” has emerged to characterize the utility and possibilities
of social media. Euchner believes that with crowdsourcing, we are now
supplanting our trust in experts for the collective wisdom of the crowds
(2010, 7). He also thinks crowdsourcing is a “transformative capability with
wide application(s)” and cautions that use of the technology requires
careful constraint and that it is by no means a “panacea.” The most
poignant wisdom seems to be that crowdsourced solutions typically do not
proceed to completion because no one person can create a complete
solution and typically the crowdsourcing platforms have performed poorly
at allowing members to aggregate partial solutions. Euchner recommends
creating an open platform where individuals “have access to the collection
of partial ideas so that different parties can contribute unique expertise”
(2010, 8). He also cautions not to solicit complete solutions from any one
person, but rather to rely on individual expertise in areas of special
knowledge.
Not just crowdsourcing. Several “crowd” terms have emerged in
recent years. The most popular core term is crowdsourcing, with
permutations such as “crowdfunding” that suggest a monetary result or

“crowdsolving” which suggests the goal is to solve a problem. Consider the
following working examples of human networks in the spotlight recently:
1) “Sourcing” – After the Japan earthquake, the crowdsourced
network of radiation sensors launched by Safecast.com gathered
more accurate data than local governments (Jamail 2011).
2) “Investigating” – The website HelpMeInvestigate.com, conceived
by Paul Bradshaw, promises to change the flow of information
concerning issues of local interest such as health, education, and
welfare. The code for the website is open source (free to use by
anyone).
3) “Journalism” – The Ushahidi.com technology for citizen mapping
and journalism has proven the catalyst for change and awareness
in many global events. The founder, David Kobia, was awarded
“Humanitarian of the Year” in 2010 by Technology Review (MIT).
4) “Solving” – The Deepwater Horizon oil spill disaster response
included a strong campaign by BP and the US Government to
solicit solutions from the public, reportedly collecting over 20,000
ideas (Euchner 2010).
5) “Innovation” – InnoCentive.com has over 250,000 registered
solvers from 200 countries who have collectively amassed over
$35 million in award money, having solved more than 1400

challenges (Aron 2012). InnoCentive has a 300% to 600% higher
success rate than most established R&D facilities.
6) “Funding” – Kickstarter.com has successfully funded approximately
28,000 of 65,000 potential projects, doling out $300 million in
funding with seven of the projects each receiving over $1 million.
20
Volunteered information. The Colorado wildfires in 2012 discussed
earlier were not the first or only event to leverage the public for
information dissemination (Twitter). Geospatial Today’s article (2011) also
mentions the extensive role of Twitter in the floods in Queensland
Australia, emphasizing how the Queensland police used the website and a
keyword “#qldfloods” (known as a “hashtag”) to ensure the precise
exchange of data and communication among citizens – with over 15,000
tweets per hour during the peak of the event. The author explains that
crowdsourced volunteered geographic information (VGI) “plays a critical
role in sharing real-time information during crisis management, especially
when traditional sources of information are not accessible.” The same
statement is equally truthful when reduced to “volunteered information,” it
need not be “geographic” in nature. Many people carry a GPS enabled
phone, lending to the geographic aspect of information gathering, but the
reality is that many good ideas (information, perceptions, or even
20 Current stats are maintained at Kickstarter’s website for public review: http://www.kickstarter.com/help/stats

measured data) come from the public and have no location-relevant
aspect.
Trust but verify. The public may not be the best data warehouse –
for neither collection nor processing of information. As pointed out earlier,
the officials in the Colorado Springs Waldo Canyon fire had difficulty
correcting the spread of misinformation through social media sites. One
exemplary work investigating the use of social media in crisis management
is quick to point out that data from the public “often contains variable
amounts of ‘noise’, misinformation, and bias” (Boulos et al. 2011, 2).
Boulos and colleagues state also that the noise will undergo
“amplification” due to rapid rate of spread through social networks. They
believe it requires filtering and verification before it is useful. Despite this,
the authors believe human sensing is advantageous because of human
ability to “contextualize, discriminate, and filter” (2011, 6). They suggest
the combined use of expert analysis, crowdsourcing, and machine learning
to facilitate rapid analysis of large datasets (2011, 7). Furthermore, with
respect to crowdsourcing, the authors suggest that experts especially
should work in peer networks to best facilitate effectiveness.
4.6 Wisdom from Knowledge: Heuristics and Machine Learning
Images, measurements, or any other types of data are meaningless
until transformed into information; information is “some recognizable

pattern or meaning that influences a user” (Messerschmitt 2000, 39). The
most well known hierarchy (by Messerschmitt) is actually four levels of
progression: data, information, knowledge, wisdom. In the disaster
problem-solving architecture, each stage of the hierarchy needs
representation and treatment. To transform data all the way into wisdom
requires analysis, interpretation, and learning – all achieved with machines
(using software programs), humans, or some balance thereof.
Heuristics: Making decisions. Heuristics is the art of making
decisions. At times, we make calculated decisions, at other times we
estimate and use rules to allow us to make a decision that otherwise is
ambiguous. Defining the rules of decision making is difficult. Rolland,
Patterson, Ward, and Dodin (2010) have investigated decision rules
mathematically, specifically with regard to disaster response, and they
remark “decision support systems used in disaster management must cope
with complexity and uncertainty” (2010, 68). In their research, they note
that an investigation into the aftermath of the hurricane Katrina disaster
uncovered a lack of decisive action by all levels of government (federal,
state, and local). In light of this, their goal was to develop tools to
support the decision process in near real-time. Their findings indicate that
it is possible for decision makers to provide key input data to a
computational system and in return gain important knowledge about
best-case options. In particular, interagency operational effectiveness and

collaboration are both improved (2010, 74). The authors conclude that
further research in this area is merited, and they believe the value of their
approach is in “quickly assessing a disaster scenario and adapting to the
dynamic nature of the situation in a timely manner” (2010, 76).
Machine learning. Humans learn from one another through
interaction or training, and gain insight through individual experience.
Unfortunately though, many of the difficult patterns or perceptions in large
or complex datasets (such as during a disaster event) would require
unwieldy human resources to interpret. In fact, some problems may simply
be unsolvable without augmenting human knowledge with some degree of
artificial intelligence. From a purely mathematical standpoint, hundreds of
scholarly efforts attempt to address the task of building artificial “brains”
to make decisions very rapidly, very accurately, or in extremely demanding
situations. The work of Khouj, Lopez, Sarkaria, and Marti (2011) focuses on
the use of machines for simulating disaster evacuation scenarios using
software. They have constructed machine “agents” that individually
attempt to make decisions based on situational data awareness. By
exposing the agents to multiple scenarios, they hope to build artificial
“wisdom” that can be transferred to future real disaster events. Murthy
(2009) attempts a similar goal and cautions “human ingenuity can never
be replaced.” Murthy attempts to solve large-scale rescue procedures in
the hopes of “more predictable and planned results.” The author’s findings

discuss several algorithms which represent a typical “capability module”
that could be one of thousands of tools in the technology architecture
proposed in my paper.
4.7 Human Gaming for Analysis and Simulation
The power of the human mind is extraordinary; it is capable of
discerning patterns or moments of ingenuity, insight, or perception that no
machine could ever hope to achieve. The use of gaming and simulation is
relatively common in many industries and is a means to leverage human
capabilities. The spectrum between “simulation” and “gaming” deserves an
initial discussion. At one end, there lies pure simulation, in which
parameters are fed into a computer model and the calculated result
interpreted by users. At the other end lies gaming, where humans “play” in
an interactive environment and are free to make decisions (most typically
in a game involving many other human players). Somewhere between
simulation and gaming lies an ideal compromise. For example, if a
simulation has variable inputs, why not deploy the simulation to the public
and allow each individual to experiment? In a sense, this is similar to pure
“gaming” because individuals are competing for the best solution. In other
words, gaming in the familiar sense is typically interactive, with many
players, whereas we typically understand simulation to be a scientist or
engineer using predefined parameters to analyze problems. My contention

is that leveraging public input for mathematical simulations also qualifies
as “gaming” in the sense because individuals are competing to find the
best solution – perhaps the term “simulation gaming” best represents the
idea.
Many authors recognize the value of gaming as a means to iterate
potential scenarios in pursuit of an optimal outcome. Steward and Wan
believe gaming allows for “adapting strategy and tactics” and that it can
“support adaptive disaster response decision-making” (2007, 128). Taken
to maturity, the use of games and simulation can be “connected to
actuarial data for internal system and external context variables” and thus
“it is possible to monitor and mutate real-time system behavior while
facing disaster” (2007, 128) [emphasis added]. Steward and Wan also
expect that “more productive outcomes and more measurable readiness”
are possible when using matured game play in parallel with disaster
decision making (2007, 129). Games have a long history in war and conflict
training. Smith (2010) has a comprehensive treatment of the role “games”
have played since the Stone Age. In his look into the future of
“war-gaming”, Smith finds that “games, simulations, and virtual worlds”
are “becoming part of the global communication infrastructure …. They are
tools that can be used to think about all kinds of problems and to orient
dispersed audiences toward a shared problem in three dimensions” (2010,
18).

The IDSR (International Strategy for Disaster Reduction) created an
online disaster mitigation game designed to engage children but it is
equally effective at luring adults into play. Figure 12 is a screen capture
from the beginning of a session showing the play interface. The user must
build structures and modify the environment to best prepare for a disaster.
Each cell of a mapped area is configurable – for example, a hospital added,
the land raised as a levy, or even planting of trees is possible. The
simulation will run to completion and give the user feedback on the quality
of preparation for the disaster. With a sufficiently complex game, the
simulation results could reveal novel methods for mitigation and response.

Figure 12 – An interactive game originally designed to engage children in
simulated disaster scenarios. The user modifies every cell and will
attempt to build relief housing and support infrastructure before disaster
strikes. Game is available at http://www.stopdisastersgame.org
One online vendor, Linden Labs, has created a game so sophisticated
as to establish an entire user ecosystem complete with the exchange of
real money (users convert their currencies into “Linden” dollars),
interpersonal relationships, and even land ownership. Called Second Life,

users buy and sell objects to build and “live” in a world comprising
restaurants, transportation systems, recreation, and nearly any other
imaginable manifestation of real life (Figure 13). The Second Life project
could serve as a reference design for building large-scale scenarios to
model actual disasters. For example, it is conceivable to build a Second
Life world mimicking the Fukushima nuclear reactor meltdown. Players
would carry out their virtual lives in this world, perhaps divided initially
into groups to explore different situations. In a sense, this would be an
experiment involving the global community. Actually, the Second Life game
has achieved such complexity that one group utilized it as an experimental
laboratory for studies in human economic decision making. Duffy reviewed
the experiment and concludes that virtual worlds are “potentially rich
environments for experimental interventions” and that the subject pools
are “more easily accessible to researchers” (2011, 61). Duffy also notes
pitfalls in the use of virtual worlds for such testing, and cautions that
using such media for human study can be risky. This type of large-scale
simulation gaming seems a very promising utility for disaster scenario
experimentation.

Figure 13 – Second Life is an immersive 3D environment built entirely by
its user population. Users trade real dollars to buy objects, land, and
goods and services. The entire world exists in 3D with objects having
behaviors and utility – much like in real life.
Gaming has four primary purposes, as described by Walker, Giddings,
and Armstrong: teaching, training, operations, and experimentation (2011,
167-168). Teaching and training are self-explanatory. Operational games
investigate feasibility and review systems functionality. Experimental
games focus on “human decision-making behavior” and in the case of crisis
gaming can be used to “subject the players to a continuous process over

time in which they are both making decisions and living with prior
decisions” (2011, 168). Using humans to role-play in a reconstructed crisis
scenario seems suitable for determining the best course of action for
response and long-term remediation. Walker and colleagues created a
“roadmap” of gaming, reproduced here as Figure 14. They foresee a future
where it is possible to simulate a crisis in real-time and they are
“convinced that the boundary between gaming and reality will become ever
closer” (2011, 172). The authors also conclude that an open and
distributed architecture is best suited for gaming and, further, the
architecture should accommodate real world data feeds to facilitate
augmented reality user interfaces.

Figure 14 – A gaming technology roadmap with a particularly noteworthy
prediction that real-time scenario gaming and augmented reality are
integral to crisis management. From: Walker, Giddings, and Armstrong
2011.
4.8 Integration: Building Synthetic Environments
The culmination of disaster problem solving, and the key step lacking
in the status quo, is in the integration of all available components of the

technology landscape. With the architecture, stakeholders should have
access to (at a minimum) a spatial component (i.e., the relationship to the
earth or the walls of a building), a temporal component (timeline of
events), a data input-output module, analysis tools, and a communication
system capable of facilitating interaction among stakeholders. Integration
will occur in a natural workspace where the user interacts
two-dimensionally with flat data (news articles, for example) and
three-dimensionally when data relate to the earth’s surface (or in times
when data analysis necessitates a 3D environment). Collectively, these
concepts comprise what is commonly termed “situational awareness.”
Vision vs. reality. More than 20 years ago, experts in disaster
management predicted the aggregation of tools and techniques into a
single interface. Alexander (1991) suggested this type of integration would
“offer considerable potential for natural disaster management, especially if
real-time uses are developed by integrating these technologies.” Rhyne
(1997) also predicted a convergence of systems, having predicted that
geographic information systems (GIS), computational modeling, and video
conferencing would merge by the year 2002. Based on these early visions
of the future, it seems the industry remains in a limbo between
idealization and reality.
Early steps in the status quo. Work at York University in Canada has
recently demonstrated a real-time 3D environment premised on integration

within Google Earth (GE). The system, called 3D Town, uses sensors to
pass data to a standard GE interface where continuously updated visual
objects represent the motion of vehicles, pedestrians, and other tracked
objects. The research team based the system on a four-layer design: a 21
sensor layer for data acquisition, a protocol layer for homogenization of
data, a data management layer, and a responsive layer that renders the
environment (Boulos et al. 2011, 18-19). This tight integration and layered
design methodology is a step in the right direction for building immersive
problem-solving environments, but it is not yet comprehensive enough to
include all aspects of the disaster management architecture.
A 3D wildfire environment. The design team who engineered the
Capaware open source visualization tool mentioned in the earlier sections
of this paper also authored a paper specific to wildfire forecasting.
Castrillon et al. (2011) leveraged their Capaware software framework as a
base tool and integrated a simulation module called “FARSITE” for
modeling the growth of fires in three dimensions. The interface uses GPS
tracking to draw objects such as deployed helicopters or ground vehicles in
near real-time in the 3D environment. The diversity of resources involved
and the complexity of the problem dictate the need for a very powerful
support system (Castrillon et al. 2011, 392). They use an architectural
approach whereby the base layer is an operating system followed by a core
21 Video “New eye in the sky at York University” http://www.youtube.com/watch?v=_RcdaIdRmHs

toolset layer and finally a plug-in and library layer. The authors promote a
service oriented architecture approach, boasting the advantages of
interoperability, simplicity, up-to-date currency, reliability, and
functionality (2011, 392-393). Figure 15, taken from the authors’ paper,
portrays the high-level system design and service-oriented approach. Most
importantly, they have focused on integrating the simulation capability
into the situational graphic depiction, allowing stakeholders to predict and
estimate fire behavior in the same interface that feeds real-time data
about the actual scenario. It is surprising that this tool was not used (to
my knowledge) during the Colorado wildfires, as this is precisely the
system that would have allowed for improved situational awareness and
even prediction of the fire’s behavior.

Figure 15 – A Conceptual depiction of an integrated simulation and
monitoring system drawn by Castrillon et al. (2011). The system
integrates real-time data from field equipment with simulation modules
and a user interface.
Precision Information Environment. One extraordinary new
conceptual interface design promises to link a network of sensors and
systems into a user-centric Precision Information Environment (PIE),
depicted in Figure 16. Under development at the Pacific Northwest National
Laboratory (PNNL), the proposed PIE could revolutionize the way we
interact with data in our environment. A unique feature of the PIE is the 22
profiling of users that allows automatic filtering and adjustment of
22 PNNL’s video “Precision information environments envisioning the future of emergency management” is
hosted on YouTube at: http://www.youtube.com/watch?v=5tfnmhl-A54

incoming information streams, tailoring the data to each user.
Furthermore, the PIE uses a Natural User Interface (NUI) design
methodology. The NUI allows unfamiliar users to adapt to the system and
facilitates a “desirable and engaging user experience” (Boulos et al. 2011,
23-24). Collectively, the design features of PNNL’s PIE are an accurate
vision of an interface that immerses users in a multidimensional
environment where spatial context is a central theme and each stakeholder
has access to the appropriate information. PNNL has published an
excellent gap analysis report (PNNL 2011) detailing the shortcomings of
existing emergency management systems – an excellent resource for
narrowing future research efforts.

Figure 16 – A screen capture from PNNL’s video concept of a Precision
Information Environment depicting stakeholders collaborating on a graphic
table solving issues related to a wildfire. Video and more information
available at http://precisioninformation.org
Isn’t PIE good enough? Considering the very effective visuals in the
PIE concept video, it would seem the efforts of my research paper are
moot. In reality, this is precisely why I have assembled my research –
technology and creative research efforts already exist to accomplish what I
have envisioned. The missing component is the high-level architectural
view and broad vision that considers technology and how to treat the
disaster as a business so multiple agencies can collaborate free of

encumbrances. Another missing component is a means to ensure
engagement with the technology. In fact, many of the gaps discovered by
PNNL are solvable by the foundational “architecture” approach proposed in
this paper.
Where does this leave us? The concept of an integrated disaster
management platform with time and dimensional functionality is obviously
not unique. PNNL’s PIE concept is quite similar to what this paper
attempted to describe, but PNNL’s execution is lacking – at least for now.
Other efforts such as Capaware seem to have launched commercially but
inexplicably remain undiscovered or perhaps abandoned by their authors. 23
Aggregating technologies onto an “architectural platform” is one way of
allowing progress. Building an extensible foundation based on a service
layer model will ensure that new research discoveries easily integrate into
the system – something akin to a global “disasternet” that focuses on
information flow, early response, and long-term problem solving.
23 It seems Capaware, or its only known commercial deployment, Sademer3D, would have been a perfect tool for
use in the Colorado wildfires but I have seen no indication that any such tool was ever used. Sademer3D could
have predicted fire spread and fed information in real-time to responders and the public.

Section 5
Characteristics of the Ideal Disaster Solution Environment

5.0 Characteristics of an Ideal Architecture Platform
The optimum form of a distributed and extensible problem-solving
platform will continuously evolve with technology, but the core design
principles will hold true over time. Building on a foundation for execution
(FFE) that models the disaster management problem as a large business,
an assemblage of capability modules will integrate through a service
oriented architecture and an open source design philosophy. This approach
is capable of adapting to inevitable changes of technology over decades of
time while continually aligning stakeholders’ interests to ensure a
collaborative philosophy.
5.1 General Characteristics Listed
● Authoritative: The interface should be widely regarded as the source
of information. Stakeholders should trust and know the interface to
eliminate ambiguity of authority in times of crisis.
● Automatic: When disaster strikes, quickly aggregating and
disseminating information is important. If an event happens when
nobody is prepared, the interface should trigger creation of a new
problem workspace. To the extent possible, certain processes need
automation to reduce human workload. For example, deploying a
map, social media aggregator, and dedicated search engine module
is appropriate for almost all disasters.

● Adaptable: Each crisis will have new challenges and needs; the
interface should be pliable enough to accommodate variability. No
single design will be compatible with every disaster. Stakeholders
will have varying levels of design authority appropriate to their roles
and needs. The design of the human interface could evolve over time
by publishing slightly permuted configurations to a percentage of
users – the most effective version thus incrementally winning over
less desirable or less effective designs.
● Extensible and Modular: New capabilities emerge continually and
new advancements will need tight integration with existing systems.
Building services in an extensible manner ensures that new modules
of capability remain compatible with previously deployed systems.
Disasters can span decades, and so must the systems. A modular
approach means that solution building blocks are reused and
recombined in future scenarios.
● Elastic: The infrastructure and all systems built on the infrastructure
must be rapidly scalable so that 1) resources are conserved in times
of disuse, and 2) during rapid-onset crisis events, the system will
respond to exponential growth needs.
● Immersive: Response and mitigation require awareness. The
architecture will create a fully immersive environment capable of
putting any information from any point in time in the hands of any

stakeholder. Using role-based game play coupled with advanced
simulation techniques, the system could predict the outcome of
complex social or political scenarios or even the progression of
wildfires on an accelerated timeline – in time to react and alter the
real world outcome.
● Interactive: The premise of this project is that a user needs to do
more than just consume map data or public notices – the system
architecture must have interactivity as a tenet. Rather than just
glance at a map of a nuclear disaster, users should be able to click
on their locations and receive routing instructions through the safest
zones of radioactivity to the nearest shelters.
● Open: It is impossible for one stakeholder to address every aspect
of a crisis scenario. The design and maintenance of an architecture
must be open for all stakeholders’ involvement. Contrary to a
proprietary design maintained by a single entity, the open design
approach ensures there are no hidden agendas or commercial
motives.
● Secure: Different stakeholders must have different security levels
and the system must be secure from malicious activity. Each layer of
design will necessitate a certain level of protection, be it a password
login requirement or physical security such as locked doors and
tamper-proof equipment.

● Ubiquitous: Without access to this technology architecture, it is
useless. Achieving ubiquity will be very challenging, especially in
undeveloped countries or disaster stricken areas with completely
destroyed communication networks. Connectivity must be diverse,
leveraging analog and digital technologies as well as field deployable
units that can capture data for later upload through backup
infrastructure (even if that backup is as primitive as foot travel
carrying USB flash drives or optical discs).
5.2 Service Oriented Architecture in Detail
The heart of this project is to approach issues in the status quo with
a service oriented architecture (SOA). To understand the contextual
meaning of services, consider a restaurant as a “food preparation service.”
Waiters take orders (the instruction set), the kitchen gathers food and
prepares the meal (the service processes), and the waiter then delivers a
meal to the hungry customer (a stakeholder). Common online “services”
are web-mapping interfaces such as Google, Yahoo, or Mapquest – many
have used these websites at least once to find a local restaurant or retail
store. Rather than processing gigabytes of data and vector street
networks, users provide only a simple instruction set (the “from” and “to”
addresses), and the solution platform then returns a turnkey answer in the
form of driving directions or perhaps a phone number to a business. Even

just browsing such map interfaces is consuming a satellite imagery
service.
In the disaster architecture context, many such services make sense.
Satellite imagery is an obvious choice, but so is a current data service that
for example would continuously publish the best available wildfire
boundaries. In a public health disaster, the service could continuously
publish the locations of known infections. A more complicated service in a
disaster scenario might be an analytical solver module that processes user
data. By leveraging cross platform well-known file languages such as XML,
a researcher can build a plug-in module capable of processing data on a
“subscription” basis. Stakeholders (present and future) would have access
to the solving module by means of an advertised “solver service.”
What services are needed? Basic life support functions such as
food, water, and shelter will be necessities. Other basic services are
health (first aid, medicine, etc.) and transportation of goods. Beyond mere
survival, however, lies an array of technological needs, the first of which is
a computing infrastructure. As discussed previously, cloud service providers
such as Amazon Web Services frequently offer “infrastructure as a service,”
exposing users to a virtually unlimited resource pool of storage and
processing capability. The second need is communications networks.
Mobile phones are a critical tool, but as recent disasters have proven, they
are only one component to a successful communication network. In times

of overload, a secondary communication medium such as SMS texting,
Twitter, or other social outlets has proven a low-bandwidth alternative to
voice communication. When widespread power outages completely
preclude such communication, alternative means are necessary such as ad
hoc locally deployed network infrastructure. Comprehensively addressing
communication thus necessitates a broad spectrum of readiness. It is
conceivable, for example, that a creative software solution could patch
together hundreds of computers over short-range WiFi connections creating
a mesh network for a small city area.
Many technology service layers could exist and all have their
applications. Just a sampling of these might be:
● Data aggregation and filtering – Data feeds, especially during
unexpected events such as earthquakes, are notoriously noisy and
typically are high volume. It is difficult to filter signals from the
background chatter. Tools such as Ushahidi’s SwiftRiver are just one
example – including this type of processing through a simple query
to the overall architecture allows anyone to leverage the algorithms.
● Mapping – No disaster portal would be complete without a
comprehensive map interface. Within the map, many sub-services
would exist, some of which may be: measurement and analysis
interface, evacuation routing, weather data ingestion, user input and
modification, or location-based messaging.

● Citizen reporting – The public should have the technology to submit
photographs, text messages, and any other type of data or
observation through a convenient and user-friendly interface.
● Emergency messaging – Citizens need warnings about life saving
information by any means possible: television, radio, cell phone,
email, bulletin boards, police broadcast, social media, and any other
conceivable communication medium.
● Weather – It seems obvious, but integrating the most up-to-date
weather as a service layer is a fundamental requirement for many
analysis scenarios.
● Situational awareness – By sending latitude and longitude to a
situational awareness service, users can navigate to safety, toward a
predesignated meeting point with others, or learn about nearby
radiation levels or infection rates. The service could augment a user’s
view of reality by sending information to a “heads-up” display or
synthetic 3D environment.
● Global search – Not to be confused with a “Google search,” this
service would be a managed search portal that is situation-centric.
Public search engines (e.g., Google, Yahoo, Bing, Twitter, and
Facebook) will index almost everything, unfortunately including
“noise” and misinformation as a side effect. A customized search
service fine-tuned by stakeholders will keep information pathways

efficient. The service could leverage existing search engine
technologies, but perhaps include human supervision or a special
algorithm tailored to the event.
● Publish and subscribe – A service that aggregates data feeds and
understands stakeholder needs to facilitate accurate and efficient
exchange of information. For example, rather than choosing among
Facebook, Twitter, or mass email notification, this service allows one
posting to be delivered to any and all distribution portals depending
on the intended audience and the preferences of the subscribers. A
highly advanced version of this service might be predictive, in that it
would possibly evaluate a stakeholder’s interests in a given disaster
before it was even clear to decision makers that involvement was
necessary.
● Data validation and weighting – Providing a service that can
evaluate specific data for validity is useful. If a citizen submits a
photograph claiming damage in a certain area, the service could
immediately cross-check other photos in the area and grade the
validity of the claim. Likewise, another user who questions the
validity of certain information might call a phone number, issue a
spoken query concerning the validity, and in return discover if the
information is correct. Adding location awareness to all information
streams can strengthen validation algorithms. For example, a text

message coming from someone at the epicenter of a disaster might
have more validity than one that is 100 miles distant.
● Location-based push messaging – Rather than subscribing to an
emergency messaging service, users might receive notification on
their mobile phones based on proximity to local threats or safety
resources. When the user’s latitude and longitude come within a
certain area, a specific message would override the user interface –
perhaps with a warning about radiation levels. If the user did not
subscribe to a location-based service, it could still be possible to
trigger messages if built-in hardware capabilities override the user’s
preferences. Of course, this type of intrusive messaging may be
perceived as a privacy concern.
● Crowdsourcing – This service layer would facilitate requesting and
consuming crowd-based resources. One example is financial, when a
community is financially strained and needs to reach out for funding
from neighboring regions. Perhaps a research team is unable to
process a particular dataset effectively and wants to solicit public
insight on methods to manipulate the data. Another scenario would
be to solicit public interaction with a simulation (based on current
data for a particular disaster) – perhaps leading to a novel
understanding of the current event.

Service security. Individual stakeholders (companies, the public,
homeowners, data collection agencies, government entities, etc.) need a
robust method of publishing data that complies with a shared
specification. This specification facilitates information sharing and
aggregation and addresses usage rights, restrictions, and other factors.
The Open Web Application Security Project (OWASP) has invested
extensive effort into building security for web services; its work is an
excellent starting point for this project. It is clear from just a cursory
review OWASP’s work that security will comprise a significant part of the
final technology architecture. OWASP remarks on distributed system
security saying, “If web services system designers do not understand
those services and protocols … they cannot possibly design secure web
services systems.”
24
5.3 Stakeholder Unification and Open Sourcing
Open design and consistent sharing among entities is an excellent
way to encourage altruism in a community. My personal experience with
the Colorado wildfires this year solidified my understanding that unified
stakeholder involvement is in the collective best interest. During the fires,
stakeholders were not operating with the big picture in sight; they focused
on their own interests without a greater sense of how their information
plugs-in to other stakeholders’ viewpoints. Even though contributions were
24 For more, begin at https://www.owasp.org/index.php/Web_Services_Architecture_and_Security

excellent, the work of individuals, government, and private companies
seemed to be both repetitive and competitive … eventually leading to
confusion about the true authoritative source. Despite potential criticisms
and stakeholder rejection, an open design is the best solution with the
most potential for leveraging ingenuity in untapped resources.
Open sourcing the project may eliminate or reduce financial barriers
to implementation. Monetary backing will be necessary, but an open
design could bypass traditional contract design processes and lead to
vastly improved return on investment. For example, many of the tools
mentioned in this paper are open source. To contract the work 25
represented just by these examples might easily cost millions of dollars (if
not 10s or 100s of millions). As mentioned in an earlier section of the
paper, the crowdsourcing company Innocentive has achieved three to six
times better results than traditional research and development channels.
Unifying stakeholder interests will reduce parallel or duplicated
efforts and would help promote peer review and collaboration among
scientists in various disciplines. A disaster scenario tends to build a sense
of camaraderie itself, but having the collaborator philosophy in place
before and after will ensure that stakeholders work toward mutually
beneficial goals.
25 For more background on the meaning of “open source”, see the Open Source Initiative’s definition at:
http://opensource.org/osd.html or http://oreilly.com/catalog/opensources/book/perens.html

Section 6
Pulling Together a Technology Foundation

6.0 Pulling Together a Technology Foundation
The primary deficiency with disaster response portals in the present
environment is their lack of a “foundation for execution” (FFE). This term
originates in the business discipline, but extrapolates to the context of
disasters by treating the disaster management system as a company or
corporation (irrespective of whether or not the effort is commercial). FFE is
a means of leveraging a company’s technology to automate and optimize
its core capabilities (Ross, Weil, and Robertson 2006, 3-4). A core
capability is one that gives the company a specific advantage in the
marketplace – it is essential to functional efficiency. Figure 17 shows a
handful of core capabilities relating to disaster management. Because a
disaster portal has numerous core capabilities dependant on technology,
FFE is an appropriate viewpoint from which to begin overhauling status quo
paradigms.

Figure 17 – A grouping of core capabilities (processes) for disaster
management. Many auxiliary capabilities are available through an
extensible library. Plug-ins from the library may eventually become “core.”
Within this paper, many of the technologies discussed are core
capabilities (or nearly so, with future refinements). Figure 18 lists several
of these technologies grouped by capability descriptors. This listing is not
comprehensive, but serves to demonstrate the FFE approach to
“capability.” Often, technologies are considered “solutions”, but in the
high-level view, solutions are just temporary remedies. A capability is
developed extensively until it becomes integral to the system as basic
functionality – irrespective of the underlying technologies utilized.

Figure 18 – A representation of many of the technologies discussed in this
paper as they would be categorized into “core capabilities.”
Three building blocks comprise the FFE: the operating model, the
enterprise architecture (EA), and the information technology (IT)
engagement model (Ross, Weil, and Robertson 2006, 8-10). An operating
model is the high-level description that describes the degree to which the
business process is standardized and how well its systems are integrated.
The EA is the “organizing logic” and the overall map of how business

systems and processes function. Finally, the IT engagement model ensures
that individual projects align with overall business objectives.
6.1 Operating Model
A disaster management system’s primary goal is to reduce impact to
human life and the environment while balancing available resources. Much
like a functioning large company, disasters span cultural and physical
boundaries. Disaster stakeholders are like “business units” in an enterprise
– geographically dispersed and typically capable of some degree of
autonomous operation. Figure 19 depicts several of the more common
disaster stakeholders. The operating model pulls together these business
units and is a means of defining high-level system functionality. It gives
an overall sense of how closely the various units are integrated and
whether or not there is standardization of processes among them. The
operating model is more than just strategy; it actually provides a “stable
and actionable view of the company” (Ross, Weil, and Robertson 2006,
25-26).

Figure 19 – Part of a larger diagram, this graphic element identifies several
of the more notable stakeholders in a disaster. Many of the elements
would divide into sub-elements (e.g., “response agencies” would divide
into police, fire, and medical).
Without an operating model, disaster stakeholders (i.e., business
units) will function selfishly with reduced awareness of the benefits of
collaboration and information exchange. For example, during the 26
Deepwater Horizon (DH) oil spill off the coast of Louisiana, highly
advanced oil-skimming ships traveled from the Netherlands to assist the
United States with cleanup. This equipment was far more efficient and
capable than anything operated by the US – and the Dutch were eager to
help because they had a vested interest in the health of the environment.
Sadly, poor communication, a decades-old legal act, and a stubborn
Environmental Protection Agency (EPA) all contributed to major delays and
complications with deployment of the skimmers resulting in less than
26 I witnessed this first hand throughout the wildfire incidents in Colorado this summer though my involvement
with the GIS community, email discussion groups, and by extensive search and review of online news and data.

desirable oil recovery (Grundvig 2010). An operating model would have
fostered coordination among stakeholders, ensuring these stumbling
blocks would not exist and promoting an altruistic philosophy.
There are four operating models to consider, each with a different
degree of business process standardization and integration: 1) unified, 2)
diversified, 3) replicated, and 4) coordinated. Unification is the highest
level of standardization and integration; the other models sacrifice some
degree of process standardization or integration (Ross, Weil, and
Robertson 2006, 28-29). Figure 20 shows the four models as well as a
representative company for each. The choice and design of an operating
model has a “profound impact on [implementation of] business process
and IT infrastructure,” and businesses may transform to new models over
time (2006, 26, 40-41).

Figure 20 – The first component of a Foundation for Execution, the
Operating Model, can have varying degrees of business standardization
and integration. Standardization means individual business units are
identical (fast food restaurants) while integration implies tight data
sharing.
Choosing the operating model for a comprehensive global disaster
platform is difficult. Luckily, as Ross, Weil, and Robertson have explained,
a company need not adopt one model at all levels (2006, 39-40). The
disaster management platform should strive to attain unification at local
levels while functioning at a global level with a more diversified model.
Business process integration (the vertical scale in Figure 20) represents

sharing of data among business units. The service oriented architecture
(SOA) discussed previously is the enabling technology for data sharing – it
ensures that each business unit (stakeholder) has access to important
shared data. Standardization will likely remain a difficult characteristic to
define in a field as diverse as disaster management. During global events,
the level of standardization needs to be much higher than local events.
Consider an example of nuclear contamination in the context of
system scale. If global, the air sampling equipment and test procedures
used in America should be identical to those used in Germany to allow
comparison. But what is the probability that each country has identical
equipment beforehand? In the case of two isolated incidents (one in the
USA and another in Germany), each country can use its own equipment and
procedures without adversely affecting the other because sharing of
measurements has little meaning. No single operating model will work for
every level of the system, but an operating model of some type is
beneficial.
Geographic and cultural boundaries are a starting point for
aggregating stakeholders into operational groups. Over time, natural
groupings will likely evolve and may be dependent upon the nature of the
disaster (e.g., hurricane versus earthquake). Eventually, a library of
operational models will emerge, each particularly well-suited for a specific

type of event. Collectively, these operational models will play a role in
refining the top-level view of the disaster system.
6.2 Enterprise Architecture
The operating model alone will do little to implement the disaster
management system. But the next level of the FFE, Enterprise Architecture
(EA), leverages the model to define processes and integrating technologies
that carry out core processes. EA is a widely used problem-solving
approach that allows companies to integrate information systems,
technology, security, governance, and other business operations into a
unified and well-aligned whole (Ross, Weill, and Robertson 2006, 47-49).
For upper-level decision makers to understand the very complex and
detailed EA, architects create a simple one page “core diagram.” The
diagram becomes a “rallying point for managers responsible for building
out and exploiting the EA” (2006, 50-51). Businesses having the same
operating model often have similar core diagrams. Different operating
models will manifest as different core diagrams; but irrespective of the
model type, core diagrams typically emphasize:
a) Core business processes: those processes needed to execute
the operating model and respond to opportunities. In the
disaster portal context, one example would be emergency

notification, or resource coordination such as police,
ambulances, and other first responders.
b) Shared data driving the core processes: data shared across
locations unifies processes. The technology architecture in this
paper hopes to create a service oriented architecture (SOA) for
publishing data to all stakeholders; sharing data is a principle
of the design.
c) Key Linking and automation technologies: This is the
technological “guts” of the disaster portal. Elements are
networking infrastructure (wired and wireless), computing
platforms, sensor technologies, user interfaces, data
management, and sharing mechanisms.
d) Key stakeholders: In a disaster context, the stakeholders are
quite diverse – likely more so than in even the largest of
companies. Each core diagram at any particular system level
will identify stakeholders and their basic roles.
Figure 21 depicts a core diagram “pattern” applicable to disaster
management, comprising several elements of the four emphasis areas
given above: stakeholders, core capabilities, shared data, and linking
technologies. This diagram accentuates the link layer (networks, protocols,
etc.) and its strong connection to the central data repository. Stakeholders
interact with core processes through this link layer by means of a

credentialed access firewall. The data engine handles storage management
and additional security issues while maintaining transparency to the
appropriately credentialed user. The core processes listed are only a
sample of the entire process pool; the plug-in library facilitates the
extensibility of the core capabilities. As new modules are developed, they
will integrate either as new core functionality or as optional plug-in
modules. The diagram represents one possible embodiment but not
necessary the ideal form for all hierarchical levels of the overall system
(local, state, federal, global).

Figure 21 – Core diagram of the disaster response platform shows the
high-level view of what will become a sophisticated enterprise
architecture. Stakeholders (bottom) access core processes and plug-in
modules through a security layer. A data repository is integrated with the
linking technology.
A second form of the core diagram, depicted in Figure 22, shows the
base architecture expanded to allow regional customization (e.g., city,

state, federal). As discussed earlier, the operating model characterizes
integration and standardization of business processes. Integration is
accomplished with a horizontal “link layer” that is integral throughout all
regions including the data repository. Stakeholders and core processes will
differ based on scale – note the topology changes indicated among the
red, green, and blue regions. The data repository has grown in size and
includes additional elements to accommodate varying regional
requirements. All stakeholders, core processes, and the data repository
only connect through a security layer. Security measures give easy access
to properly credentialed stakeholders and curtail malicious attacks against
core processes and the data repository.

Figure 22 – Core diagram expanded with three similar groupings (shown as
red, green, blue) representing system levels at global, national, and local
scale. Multiple systems remain closely linked to a common data repository,
which has expanded to accommodate growth (compare to Figure 21).
Defining region groupings (Figure 22) is a difficult task.
Network-based infrastructure (cloud computing) is a key enabling
technology which places few, if any, limitations on how regions are
subdivided. The linking layer is best imagined as the ubiquitous “Internet,”
though as discussed previously, maintaining connectivity during disasters
requires extreme diversity in communication media. The data repository
appears as a single unit in the diagram, but is distributed geographically
to prevent a single point of failure. Distributed infrastructure is feasible

and is frequently used today for very large data-based companies (Google,
Facebook, Yahoo, and many others).
A discussion of EA in disaster management by Janssen, Lee,
Bharosa, and Cresswell (2009) concedes that many of the areas EA
attempts to tackle are specifically lacking in modern disaster management
efforts. They have concluded that individual agencies involved in disasters
are “isolated, overlapping in function and content, [and] highly
fragmented.” The authors conclude that the lack of multi-agency
collaboration remains a key handicap, despite major advances in
technology. Their work underscores the importance and relevance of the
business of disasters.
6.3 IT Engagement Model
The IT engagement model maps infiltration of information technology
within the business. Development of the full IT engagement model is an
enabling step to full EA adoption and focuses on pulling stakeholders
together to facilitate extraction of value from IT investments. More policy
and procedure than pictorial diagram, the best engagement models
typically contain three primary components: governance, project
management, and linking mechanisms. Governance ensures the
accountability of decisionmaking and cultivates “desirable behavior in the
use of IT.” Project management formalizes resource allocations as well as

the process of moving from beginning through completion. Finally, “linkage
mechanisms” are the keystones that pull together the goals of project
activities with governance (Ross, Weil, and Robertson 2006, 119).
The engagement model serves as framework from which to improve
operations efficiency, in particular with respect to leveraging technologies.
Without engagement, project leaders will execute in isolation, ignoring the
operational model’s goals of standardization and integration – typically by
deploying localized solutions to IT problems. In a disaster management
system, an earnest approach to IT engagement is advantageous especially
due to the very diverse stakeholder population. Given the current state of
disaster management and lack of alignment among stakeholders, IT
engagement is a promising step to completing the FFE successfully.
Ross, Weil, and Robertson present an effective summary of IT
engagement in their lengthy quote of Andre Spatz, who at the time was
Chief Information Officer at UNICEF. Spatz remarks:
We face high pressures for synergy across UNICEF and at the same
time, we have high pressures for local autonomy from the regional
and country offices. CIO leadership in a global IT organization is not
just command and execute. We need to continually empower people
with a vision and execution strategy, and position governance
elements within a global framework. Part of my role is to ensure that

we do not centralize too much and that our IT organization adapts
to the different cultural environments we work in (2006, 123-124).
The authors also ask and answer the question, “What is good IT
engagement?” They believe good governance gives clarity about decision
making and accountability, good project management limits risk and
improves the chance of meeting goals, and linking mechanisms ensure that
decision conversations can leverage the foundation for execution. In a case
study of eighteen companies, they identified several key principles of
successful engagement as follows (2006, 135-136):
● Engagement must identify clear strategic objectives that are
specific and actionable (i.e., it is possible to actually achieve
the objective).
● Incentives motivate unit leaders to achieve goals, especially at
the project manager and project architect levels.
● In complement to offering incentive, enforcing performance
builds credibility of a project. Enforcement allows change,
adaptation, cancellation, or exception when a project is off
track.
● IT engagement early can prevent bad solutions from evolving
into something larger. Technology architects must be involved
with business-level decisions as early as practical.

● Alignment and coordination between IT and business
objectives is not achieved, but rather is maintained. Consistent
two-way transparent communication is critical for leveraging
technology effectively, especially considering that many
high-level decision makers lack an in-depth knowledge of IT
limitations or capabilities.
It may seem like oversimplification, but achieving good engagement
seems to distill down to business common sense. How can IT
professionals expect to build a successful mapping interface when city
governments will not clearly explain their requirements? How can the
general public be expected to rely on an emergency messaging system
when a large portion do not have television or radio, but instead rely on
smartphones for the majority of their information? Is it reasonable to
expect a scientist to build an analysis module capable of predicting fire
propagation with no incentive to do so accurately or efficiently? Most
likely, using common sense is not the core problem. The real issue lies in
the interagency aspect of applying common sense; it is vastly more
complex to act when multiple stakeholders have difficulty collaborating
toward a common goal.

7.0 Funding Model
The defining characteristics of an “ideal interface” presented in an
earlier section are technically feasible, but many of the implementation
barriers to actually constructing such an interface could be large. Some of
the technologies are candidates for federal funding through research and
development grant programs. Other technologies may be better suited for
crowdsourcing or even private development. Once operational, continued
funding is required to support the IT equipment and personnel for system
operations. Using scalable cloud-based infrastructure, the system could
operate on standby using a very small team of people with relatively low
equipment costs (perhaps $1,000 to $5,000 per month). In the wake of a
large-scale disaster, however, IT operational costs may increase to
$100,000 per month or more with a commensurate increase in personnel
expenditures.
27
How many zeros? Statistics noted earlier in the paper show natural
disasters alone have affected 4.4 billion people, killed 1.3 million, and
caused $2 Trillion dollars in damage since 1992 (UNISDR 2011). This
averages approximately $270 Million dollars per day – a staggering number
that underscores the value of building better management architecture.
Factor in the cost of human life and other intangible costs, and the stakes
grow exponentially. The value of seemingly imperceptible improvements
27 The dynamic scaling of the disaster response portal closely matches the usage models advertised by cloud
computing providers in which the infrastructure behaves elastically and can respond to times of massive growth
while shrinking during periods of disuse.

could be in the billions of dollars. Consider the year 2011 in which the
Japan earthquake and Tsunami alone cost approximately $210 Billion, with
the total yearly cost of disasters nearing $400 Billion (CRED Crunch 2012).
Even a one-tenth of a percentage point (just 0.001 of the whole) could
fund a massive $600 million effort. Furthermore, considering the value of
human life in the USA approximated at $7 million, saving just one life is
significant – saving tens of thousands of lives would represent an
astronomical value.
28
Funding paradigms. Other means of revenue and funding are:
● From the standpoint of human life, we can all appreciate the value
of health or life insurance. Perhaps stakeholders in a certain
geographic region would consider funding a technology
architecture program as a type of “disaster insurance.”
● Similar to the Safecast revenue model, the project could be
funded by the public in advance or on-demand during events. With
over 4 billion people affected by disaster in the last 20 years,
even a small contribution from a portion of those individuals
would amount to significant revenue. A recent game console
crowdfunding project raised over $1 million in eight hours, $2
million by the end of the first day, $5 million after approximately
28 The Environmental Protection Agency has taken on the problem of “value of a statistical life” and has settled
on a number of $7.4 million in 2006 dollars.

five days, and leveled around the $6 million mark after two
weeks.
29
● Stakeholders could be “billed” after the technology architecture
has delivered results – perceived or actual. This would be difficult
in that investment capital would be required in advance, but it
could be a basis for future insurance models once baselines of
value are established.
Non-traditional financials. However funding is secured, the economic
model is likely to diverge from that of a typical corporation or enterprise.
Scope and complexity of each disaster will dictate who stakeholders are, a
proposition that could range from citywide to global. Crossing political and
social boundaries will be a trying scenario when dealing with economic
burdens. Capital investors must absorb the development costs or propose
a reasonable model for amortizing the initial expenses over the course of
future events. From this perspective, the extensibility of the platform is
fundamental – rather than spending billions up front to create the
“ultimate” technology architecture, clever design allows the system to
start small, then continue building additional value and functionality over
time.
Capital investments. Government funding for initial capital
requirements is likely, especially due to government’s almost ubiquitous
29
http://arstechnica.com/gaming/2012/07/oh-yeah-ouya-receives-over-1m-in-crowdfunding-in-less-than-8-hours/

involvement in disaster mitigation at city, county, state, and federal level.
For example, the National Science Foundation (NSF) offers grants to
disaster mitigation research. The NSF has a division called Infrastructure 30
Management and Extreme Events (IMME) that still actively disburses funds
for projects focusing “on the impact of large-scale hazards on civil
infrastructure and society and on related issues of preparedness, response,
mitigation, and recovery.” Some of the larger private companies may 31
express interest in capital investment as well. Companies like Google,
Amazon (AWS in particular), Comcast, Oracle, and other technology
“giants” may be interested in donating infrastructure or intellectual
capacity in exchange for tax benefits (donation write-offs) or even
improved public perception and involvement.
Global opportunity. Because disasters are a global issue with no
regard for social or political boundaries, funding opportunities are
abundant. One example is a donation-driven organization known as the
Global Facility for Disaster Reduction and Recovery (GFDRR). GFDRR
focuses on reducing susceptibility to natural hazards, not necessarily the
response and mitigation after disaster events. Regardless, its goals are
certainly in alignment with those of a disaster technology architecture, as
the architecture prescribed in this paper is not limited to a certain disaster
phase. On its homepage, GFDRR currently indicate the receipt of $332
30 The Responsphere interface mentioned earlier was awarded nearly $2 million over a four-year period. Review
the National Science Foundation at: http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=0403433
31 For more on the IMME division of NSF, see: http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=13353

million in pledged funds. The GFDRR has an innovative program of “risk 32
financing,” which has demonstrated excellent results in recent history.
While the program seems focused on helping individuals, the concept could
extrapolate to involve stakeholders at many levels.
32 The GFDRR homepage is located at http://www.gfdrr.org/gfdrr/

8.0 Further Research
The goal of my research was to approach high-level concepts rather
than deal with the details of technological processes. I have concluded
that it is both feasible and advantageous to approach disaster
management from a business viewpoint and to aggregate technologies
through an architectural paradigm – in particular, to implement service
oriented architecture where practical. Continuing research should primarily
concentrate on applying a “foundation for execution” (FFE) business
approach to disaster management with particular effort directed at
improving interagency collaboration and stakeholder participation.
The way forward: Funding. An important area to investigate further
is procurement of long-term funding. No matter how excellent the
proposed technology architecture, investor backing is required to
acquisition infrastructure and human resources. The simplistic funding
model proposed in the previous section of this paper is only the proverbial
“tip of the iceberg” and needs further development. Realistically, this
project in its current state is capable of securing just enough funding to
find more funding. With a small amount of private backing or even a small
government grant, the first area of focused research would be how to
secure this long-term financial backing. Perhaps local governments have a
budgetary allowance that would support further work, or maybe surveys of
private companies would uncover interest in backing the project.

Linking technology and business. Extensive research is already in
motion concerning the underlying technologies (e.g., broadband wireless,
self-assembling networks, wearable computing, simulation gaming, etc.).
Calling for additional effort in these areas is pointless because they are
already motivated independently of the disaster management problem. My
review of literature exposed hundreds of these burgeoning technologies,
all of which are candidate plug-in modules for a larger architecture. But I
have found little research concerning the integration of these existing and
emerging technologies into a disaster-centric architecture. Treating
disasters as a large enterprise opens up an entirely new research area.
Consequently, study areas are abundant on the topics of stakeholder
interactions, interagency collaboration, data sharing ethics, and
information ownership (and many others, of course). Combining
technology-centric research with business-centric is an important area of
future work related to this project.
Dissolving boundaries. Political, legal, and cultural issues present
major roadblocks by inhibiting open collaboration and information sharing.
This area of future research presents many issues in actual execution. The
primary goal is bridging boundaries that separate stakeholders, b y
ensuring adequate representation of all entities involved (citizens,
government, scientists, victims, local ecosystems, the earth, private
companies). Hands-on interviews and focus group discussions are a

recommended study area to assist in understanding this phenomenon.
Existing collaboration models seem to allow these stakeholders to operate
somewhat autonomously – without regard to their effects on others and
with no particular motivation to work jointly. This unfortunately creates
islands of knowledge, lack of central authority, and many other detrimental
side effects. Designing mutually beneficial working agreements could
potentially encourage stakeholders to behave altruistically – holding the
interests of others above their own in hopes of achieving an overall
improvement in outcome. Many agencies will be particularly sensitive
about releasing information, especially from a standpoint of legal
protection in instances where misinterpretation of data can lead to loss of
life or significant financial loss. Research and practical interaction with
these agencies is required to ensure that the operating model and
resulting technology architecture address key concerns.
Learning from missed opportunities. A pivotal lesson learned in my
review of literature was discovery of the Sademer3D tool based on the
Capaware platform. Touted as an emergency management decision-making
interface, its creators claim the tool shows data in a “usable and intuitive
way” while “eliminating noise.” Their wildfire case study seemed the
perfect proof of concept for application to the Colorado wildfires. But there
is no readily available evidence that anyone involved in the Colorado
incidents even knew of the tool’s existence. If Sademer3D had been

published as a capability module as part of a larger “technology
architecture,” Colorado stakeholders could have immediately put the tool
to use. Documenting other disasters and the technologies that could have
assisted in response and mitigation is an excellent way to secure the
future of this project. For example, retrospectively examining the
Fukushima disaster from the standpoint of missed opportunities could
demonstrate how technologies were available but were not used to their
potential (or not used at all). Such case studies might be a proving ground
for the proposed technology architecture and foundation for execution
approach to disaster response.

9.0 Conclusion
The underlying philosophy that began this work was that
technologies exist that are capable of drastically improving disaster
response and long-term problem solving. My personal experience with a
local disaster (the June-July wildfires in Colorado) and the review of
literature both support the initial hypothesis and suggest that the core
problem is lack of high-level thinking capable of guiding development.
Another area of concern is open collaboration among stakeholders.
The technical challenges faced during disasters have similarities to
those faced by large companies – especially concerning collaboration and
management of IT resources. Enterprise Architecture (EA), or more broadly
the “foundation for execution” (FFE), is a business tool used to manage
large geographically dispersed companies. The companies that most
benefit from an FFE approach have many characteristics in common with
the stakeholder population typical of disaster events. Treating disasters as
a large enterprise is the logical means to realize an effective response,
mitigation, and problem-solving environment.
This project has exposed an excellent opportunity for applying
business philosophies to disaster response and mitigation. An initial round
of funding will allow a serious investigation into means of securing
long-term funding as well as resource allocations by large corporations or
government entities. If the next steps are successful, a team of experts

could begin working with stakeholders to identify requirements and build
working agreements among those stakeholders. With stakeholder buy-in,
work would progress methodically through typical design-build-test
iterative cycles. Overall guidance from existing frameworks (FFE, EA, and
others) will promote a successful implementation.
Collaborative philosophies will be crucial to the success of continued
work on this project. Selfless participation on behalf of all stakeholders is
necessary for handling disaster events. System designers must not only
find and deploy advanced technology systems, but also must help build
working relationships among stakeholders. The business-centric FFE view
of disaster management may seem an awkward application, but it is a
proven and highly effective means of ensuring stakeholders operate in the
best interest of the whole, not solely on their own behalf.
Fortunately, many of the problems documented in present-day
disaster systems are not intentional, or planned, but rather are an
unfortunate lack of awareness. Enabling technologies exist but are
encumbered by stakeholder isolation (intentional and unintentional). A
business view of disasters and a guided developmental process will
improve disaster preparation, response, and mitigation. Implementing a
technology architecture will leverage IT for the disaster stakeholders and
ultimately lead to reduced financial loss, lessened environmental impact,
and decreased loss of life.

10.0 References
Alexander, David. 1991. Information technology in real-time for monitoring
and managing natural disasters. Progress in Physical Geography 15,
238:260 (September).
Arisoylu, Mustafa, Rajesh Mishra, Ramesh Rao, and Leslie A. Lenert. 2005.
802.11 Wireless infrastructure to enhance medical response to
disasters. AMIA Symposium Proceedings.
Aron, Dave. 2012. InnoCentive positions itself as a challenge marketplace.
Gartner, Inc. Executive Programs. June 27.
https://www.innocentive.com/files/node/casestudy/gartner-report-inn
ocentive-positions-itself-challenge-marketplace.pdf (Accessed
July 25, 2012)
Boulos, Maged N Kamel, Bernd Resch, David Crowley, John Breslin, Gunho
Sohn, Russ Burtner, William Pike, Eduardo Jezierski, Kuo-Yu Slayer
Chuang. 2011. Crowdsourcing, citizen sensing and sensor web
technologies for public and environmental health surveillance and
crisis management: trends, OGC standards and application
examples. International Journal of Heath Geographics 10:67.
Brown, Carol V., Daniel W. DeHayes, Jeffrey A. Hoffer, E. Wainright Martin,
William C. Perkins. 2011. Managing information technology. 7th Ed.
Upper Saddle River, NJ: Pearson Education.

Castrillón, M., P.A. Jorge, I.J. López, A. Macías, D. Martín, R.J. Nebot, and
A. Trujillo, et al. 2011. Forecasting and visualization of wildfires in a
3D geographical information system. Computers & Geosciences 37,
no. 3: 390-396.
CRED Crunch. 2012. Disaster data: A balanced perspective. Centre for
Research on the Epidemiology of Disasters (CRED). Issue No. 27.
February. http://www.cred.be/sites/default/files/CredCrunch27.pdf
(Accessed June 21, 2012)
Deam, Jenny. 2010. Colorado scene: From aerial photos, a ray of hope
amid wildfires. Los Angeles Times. June 28.
http://www.latimes.com/news/nation/nationnow/la-na-nn-colorado-fir
es-photos-20120628,0,5017543.story (Accessed July 9, 2012)
Dobel, J. Patrick. 2010. Mission integrity in disaster management. Public
Administration Review. December (special Issue).
Duffy, John. 2011. Trust in Second Life. Southern Economic Journal
78:53-62.
Economist. 2008. A survey of corporate IT: Let it rise. The Economist
Newspaper. From the print edition, October 23.
http://www.economist.com/node/12411864/print (Accessed July 23,
2012)
Euchner, James A. 2010. The limits of crowds. Research Technology
Management 10: 7-8. December.

Eugster, Pathrick Th., Pascal A Felber, Rachid Guerraoui, and Anne-Marie
Kermarrec. 2003. The many faces of publish/subscribe. ACM
Computing Surveys Vol 35 iss. 2: 114-131. June.
Geincke, Pete. 2012. US Wildfires: New Google crisis map, Colorado
imagery and more. June 29.
http://google-latlong.blogspot.com/2012/06/us-wildfires-new-google-crisis-
map.html (Accessed July 1, 2012)
Geospatial Today. 2011. Disaster management: Social media to the
rescue. Geospatial Today (November 9).
Goldfarb, Charles F. 1996. The Roots of SGML – A Personal recollection.
http://www.sgmlsource.com/history/roots.htm (Accessed July 23,
2012)
Google. 2012. Google crisis response: Past response efforts. Webpage.
http://www.google.org/crisisresponse/response.html (Accessed July
8, 2012)
Greenwald, Ted. 2010. Software that helps populations cope with crises.
Technology Review. September/October 44-46.
Grundvig, James Ottar. 2010. The Odyssey of the Dutch oil skimmers.
Epoch Times. July 8.
http://www.theepochtimes.com/n2/united-states/gulf-oil-spill-netherl
ands-38810-print.html (Accessed August 8, 2012)

Guha-Sapir, Debby, Femke Vos, Regina Below, and Sylvain Ponserre. 2010.
Annual Disaster Statistical Review: The numbers and trends. Centre
for Research on the Epidemiology of Disasters (CRED). Universite
catholique de Louvain. Brussels, Belgium.
http://www.cred.be/sites/default/files/ADSR_2010.pdf (Accessed
June 21, 2012)
Handy, Ryan Maye. Waldo Canyon fire homeowners still in dark. Colorado
Springs Gazette. June 28.
http://www.gazette.com/articles/homes-140948-officials-fire.html
(Accessed July 9, 2012)
Hopkins, Curt. 2012. Google tracks hellish wildfire season in the American
West. Conde Nast. July 4.
http://arstechnica.com/security/2012/07/google-tracks-hellish-wildfir
e-season-in-the-american-west/ (Accessed July 6, 2012).
Jackson, Joab. 2012. Amazon cloud set stage for rapid Pinterest growth.
IDG News Service. April 19.
http://www.networkworld.com/news/2012/041912-amazon-cloud-set-stage-
for-258485.html (Accessed July 11, 2012)
Jamail, Dahr. 2011. Citizen group tracks down Japan’s radiation. Al
Jazeera. August 10.
http://www.aljazeera.com/news/asia-pacific/2011/08/2011810142915
166342.html (Accessed June 24, 2012).

Janssen, Marijn, JinKyu Lee, Nitesh Bharosa, and Anthony Cresswell. 2009.
Advances in multi-agency disaster management: Key elements in
disaster research. Information Systems Frontiers 12:1-7. May 9.
Jones, Elysa and Jacob Westfall. 2010. Common alerting protocol version
1.2 OASIS (Organization for the Advancement of Structured
Information Standards). July 1.
http://docs.oasis-open.org/emergency/cap/v1.2/CAP-v1.2-os.pdf
(Accessed July 8, 2012)
Julca, Alex. 2012. Natural disasters with un-natural effects: Why? Journal
of Economic Issues. 46-2: 499-510. June.
Khouj, Mohammed, Cesar Lopez, Sarbjit Sarkaria, and Jose Marti. 2011.
Disaster management in real time simulation using machine
learning. IEEE CCECE 001507. Niagara Falls, Canada.
Lwin, Ko Ko and Yuji Murayama. 2011. Web-based GIS system for
real-time field data collection using a personal mobile phone. Journal
of Geographic Information Systems 3: 382-389.
Messerschmitt, David G. 2000. Understanding Networked Applications: A
First Course. San Francisco: Morgan Kaufmann.
McNutt, Marcia. 2010. Email to “Plume Team.” Publicized email document
dated May 28th.
http://peer.org/docs/noaa/1_23_12_Email_WH_pressure.pdf
(Accessed June 25, 2012)

Miyair, Kantaro and Shunsuke Matsuda. 2011. Multi-use of urban
infrastructure: Wireless light curried digital communication based on
LED street lighting. 8th International Telecommunications Society
Conference. Taiwan. June. http://hdl.handle.net/10419/52333
(Accessed July 23, 2012).
Murthy, Sapna Guniguntla. 2009. Disaster recovery heuristic. Thesis.
Rochester Institute of Technology. UMI Publishing. August.
National Science Foundation (NSF). 2004. An IT infrastructure for
responding to the unexpected. NSF Award #0403433 granted
September 3, 2004.
http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=0403
433 (Accessed June 20, 2012)
Noyes, Matt. 2011. Tokoyo letter: After the disaster. Against the Current
153:5-7. July/August. Also available in online version:
http://www.solidarity-us.org/current/node/3311 (Accessed June 21,
2012).
PEER. 2011. Allegations of scientific and research misconduct. Public letter
of allegation dated January 23.
http://salsa.democracyinaction.org/dia/track.jsp?v=2&c=UbL6kFuMba
iiq5iphELTefZNDfJAZqwx (Accessed June 24, 2012)

PNNL. 2011. Gap assessment in the emergency response community.
Pacific Northwest National Laboratory. Batelle. February.
http://precisioninformation.org/?p=24 (Accessed April 15, 2012)
Rhyne, Theresa Marie. 1997. Going virtual with geographic information and
scientific visualization. Computers and Geosciences 23:4, 489-491.
Richard, Bryan. 2008. Facebook puts its SaaS where its mouth is.
QuinStreet Inc. June 2nd. http://www.linux-mag.com/id/6094/
(Accessed 23 Oct 2011).
Rolland, Erik, Raymond A. Patterson, Keith Ward, Bajis Dodin. 2010.
Decision support for disaster management. Operations Management
Research 3:68-79. March.
Ross, Jeanne W., Peter Weil and David C. Robertson. 2006. Enterprise
Architecture as strategy: Creating a foundation for business
execution. Boston: Harvard Business School Press.
Rushe, Dominic. 2012. BP sues Halliburton for Deepwater Horizon oil spill
clean-up costs. Guardian News. January 3.
http://www.guardian.co.uk/business/2012/jan/03/bp-sues-halliburton
-over-deepwater/print (Accessed June 21, 2012)
Senevirante, Krisanthi, David Baldry, and Chaminda Pathirage. 2010.
Disaster knowledge factors in managing disasters successfully.
International Journal of Strategic Property Management 14, 376-390.

Smith, Roger. 2010. The long history of gaming in military training.
Simulation Gaming 41: 6-19.
Steiner, Matt. 2012. Waldo Canyon fire: Social media dominates flow of
fire info. Colorado Springs Gazette. July 6.
http://www.gazette.com/common/printer/view.php?db=colgazette&id
=141309 (Accessed July 7, 2012)
Steward, Duane and Thomas T.H. Wan. 2007. The role of simulation and
modeling in disaster management. Journal of Medical Systems
31:125-130.
UNISDR. 2012. UNISDR Counts the cost of 20 years of inaction on climate
change and risk reduction. United Nations Office for Disaster Risk
Reduction. Press release. June 13.
http://www.unisdr.org/files/27162_2012no21.pdf (Accessed June 21,
2012)
Ushahidi SwiftRiver. 2012. Two-page information brochure. Ushahidi.
http://ushahidi.com/uploads/docs/SwiftRiver_1-Pager.pdf (Accessed
June 20, 2012)
Walker, Warren E., Jordan Giddings, and Stuart Armstrong. 2011. Training
and learning for crisis management using a virtual simulation/gaming
environment. Cognition, Technology, and Work 13:163-173.
Wang, Jiayu. 2007. A Cognitive assessment of post-disaster imagery
requirements. Thesis. University of South Carolina. UMI Publishing.

Yahoo Inc. 2011. Small business home page.
http://smallbusiness.yahoo.com/ (Accessed 23 Oct 2011).
Zahran, Sammy, Lori Peek, Jeffrey G. Snodgrass, Stephan Weiler, and Lynn
Hempel. 2011. Economics of disaster risk, social vulnerability and
mental health resilience. Risk Analysis 31-7: 1107-1119.
Zolala, Farzaneh. 2010. Data collection after massive natural disasters.
Disaster Prevention and Management 19:5, 541-547.
- - End - -

Technology Foundation for a Real-Time Distributed Problem-Solving Environment Applied to Crisis and Disaster Scenarios

More Related Content

What's hot (6)

Viewers also liked (7)

Similar to Technology Foundation for a Real-Time Distributed Problem-Solving Environment Applied to Crisis and Disaster Scenarios (20)

More from Enterprise Architecture Professional Journal (9)

Recently uploaded (20)

Technology Foundation for a Real-Time Distributed Problem-Solving Environment Applied to Crisis and Disaster Scenarios